Aligning Romance Grammar

Basque and Romance
Grammars and Language Sketches

of the World’s Languages
The titles published in this series are listed at brill.com/gswl

Basque and Romance
Aligning Grammars
Edited by
Ane Berro Urrizelki

Beatriz Fernández
Jon Ortiz de Urbina
LEIDEN | BOSTON
Library of Congress Cataloging-in-Publication Data
Names: Berro Urrizelki, Ane, editor. | Fernāandez, Beatriz, editor. | Ortiz de Urbina,
Jon, editor.
Title: Basque and Romance : aligning grammars / edited by Ane Berro Urrizelki,
Beatriz Fernandez, Jon Ortiz de Urbina.
Description: Leiden ; Boston : Brill, [2019] | Series: Grammars and language
sketches of the world's languages | Includes bibliographical references and
index.
Identifiers: LCCN 2019008329 (print) | LCCN 2019009310 (ebook) |
ISBN 9789004395398 (ebook) | ISBN 9789004395381 (hardback : alk. paper)
Subjects: LCSH: Basque language–Grammar, Comparative–Spanish. | Spanish
language–Grammar, Comparative–Basque. | Basque language–Grammar,
Comparative–Romance. | Romance languages–Grammar,
Comparative–Basque.
Classification: LCC PH5023 (ebook) | LCC PH5023 .B298 2019 (print) |
DDC 499/.925–dc23
LC record available at https://lccn.loc.gov/2019008329
Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill‑typeface.
ISSN 2352-9342
ISBN 978-90-04-39538-1 (hardback)
ISBN 978-90-04-39539-8 (e-book)
Copyright 2019 by Koninklijke Brill NV, Leiden, The Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi,
Brill Sense, Hotei Publishing, mentis Verlag, Verlag Ferdinand Schöningh and Wilhelm Fink Verlag.
All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system,
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
without prior written permission from the publisher.
Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided
that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive,
Suite 910, Danvers, MA 01923, USA. Fees are subject to change.
This book is printed on acid-free paper and produced in a sustainable manner.

Contents
1 Introduction 1
Ane Berro, Beatriz Fernández and Jon Ortiz de Urbina
2 Word Order 14
Jon Ortiz de Urbina
3 Tense, Aspect and Mood 59

José Ignacio Hualde and Céline Mounole
4 Non-verbal Participles in Basque and Spanish 82

Ane Berro
5 (In)transitive Verbs: Unergatives and Unaccusatives 139

Ane Berro
6 Light Verb Constructions in Basque and Romance 176

Víctor Acedo-Matellán and Anna Pineda
7 On Non-selected Datives: Ethical Datives in Basque and Spanish 221

Beatriz Fernández
8 Differential Object Marking in Basque and Spanish Dialects 243

Ane Odria
9 Complex Causative Verbs and Causes in Basque (and Romance) 276

Jon Ortiz de Urbina
10 Grammaticalization Processes in Causal Subordination 317

José Ignacio Hualde and Manuel Pérez Saldanya
Name Index 347

Subject Index 351
chapter 1
Introduction
Ane Berro, Beatriz Fernández and Jon Ortiz de Urbina
1 Introduction
The book Basque and Romance: Aligning Grammars is a collection of articles

describing and analyzing several of the most important morphosyntactic fea-
tures for which the formal comparison between Basque and its surrounding
Romance languages is relevant. In the context of a language virtually all of
whose speakers are bilingual in either Spanish or French, the theoretically
informed in-depth description offered in this volume focuses on the fine grain
of linguistic structures from languages typologically quite apart but coexisting
and probably interacting in the minds of speakers. The comparative approach
taken here directs the attention not only to broad macroparameters but also to
narrow empirical domains where microparametric differences show up in full
complexity. The coexistence of Basque with two different Romance languages
also allows us to contrast Basque varieties both from an internal and ‘external’
perspective. Some of the phenomena described in this book have a dialectal
distribution which can be linked to similar patterns in the Romance language
spoken in that area; this enables us to observe Basque internal phenomena
which may be connected with contact with different external contacting lan-
guages.
The aim of this book is therefore to produce a fine-grained comparison of
some areas of Basque and Romance grammar, delineating similarities and dif-
ferences, thus not only providing theoretical analyses of similar systems in
apparently dissimilar languages but also feeding studies in language contact.
We hope this grammar alignment, i.e., this close parallel examination of the
grammars of bilingual speakers, can shed some light on the types of interac-
tions between different systems and on the systems themselves. Moreover, we
strive to provide detailed and in-depth descriptions on which contact studies
could be reliably built as they address the analysis of the changes this type of
grammar coexistence may bring about, favor or, perhaps, also hinder.
© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004395398_002

2 berro, fernández and ortiz de urbina
2 Basque and Romance Grammars in Contact
Basque is a genetically isolated ergative language spoken by around 715,000

people in the Basque Country (5th Sociolinguistic Survey, Basque Government
2013), a territory divided into two different political and administrative regions,
Spain and France. It is in permanent contact with Romance languages, as all
Basque speakers are, nowadays, Spanish or French bilinguals. It used to be the
majority language three or four centuries ago (Hualde 2003), but today, it is
the native language of about 375,000 people, the 15 % of the total population
(5th Sociolinguistic Survey, Basque Government 2013). Thanks to the official and
semi-official status that Basque has in the Basque Autonomous Community
and Navarre, both in the Spanish side of the border, many young speakers have
learned the language through the educational system as a second language.
As a consequence, almost half of the speakers of Basque are L2 speakers. On
the other hand, Basque is also spoken in the French side of the border, in the
historical provinces of Lapurdi, Low Navarre and Zuberoa. In this region, the
language does not have official status and the number of speakers is decreas-
ing rapidly.
With virtual complete bilingualism, the contact between Basque and Ro-
mance has changed qualitatively in the 20th and 21st century, but, even if prob-
ably at a smaller scale, Basque has been in very close contact with Latin and
Romance languages in the last two millennia, with interactions that have been
identified in both directions. In the north, contact has been most intense with
Gascon, and French itself only came into generalized contact with northern
forms of Basque in modern times (Bayonne, the contemporary unofficial capi-
tal of the French Basque Country, was largely a Gascon speaking town). To the
south of the present border, Basque shows greatest similarities with Pyrenean
Aragonese dialects (Hualde 2016, Allières 1992) mostly extinct at present. Sim-
ilarities with these Romance forms, and with early Castilian Spanish are based
mostly on lexical, phonetic and phonological developments, and grammatical
contact phenomena are difficult to trace given the scarcity of Basque language
documents until the 16th century. But even so, they obviously existed. Imper-
fective markers derived from locative nominalizations are common enough not
to require an external explanation, but as for perfective participial markers, the
first isolated Basque glosses to Latin texts from the 10th or 11th century already
show a -tu marker from early Romance origin which would eventually oust
native -i in the course of the following centuries. Similarly, the verbal system
of the first Basque long texts in the 15th and 16th century displays clear signs of
an ongoing fairly thorough reorganization of the Tense, Aspect and Mood (TAM)
system (Lafon 1943, Mounole 2011) leading to a system which by the 18th cen-
introduction 3
tury was more similar to that of neighboring Romance. As Hualde and Perez
Saldanya (this volume) also show, grammaticalization patterns for causal sub-
ordinate clauses have followed similar paths. As a last instance, Basque shared
with neighboring languages (even with English) important developments such
as the deployment of second personal plural pronouns and verbal morphemes
as markers of respectful second singular forms. All this indicates that shared
developments and changes towards more Romance-like patterns have been
taking place in Basque for a long period now. In the context of this centuries-
long contact, the fact that we can still speak of Basque and Romance as being
typologically ‘different’ languages may also require some explanation. In the
last 150 years, however, factors such as the strengthening of French and Span-
ish state-level political organization, early Romance-exclusive schooling and
contemporary bilingual schooling, sociological changes in population make-up
and urbanization, the development of mass communication systems, among
many others, have led to the extended bilingualism which may underlie the
extensive changes Basque seems to be currently undergoing and which can
be seen in this book. Obviously, language coexistence in space compounded
with extensive language coexistence in the speakers’ brains triggers linguis-
tic changes which make of Basque and Romance contact a data-rich field for
linguistic studies from many different angles. The articles in this book aim at
striking a balance in presenting as wide an overview of the most important
phenomena as possible while providing at the same time detailed data and, in
many cases, theoretical analyses.
3 Topics Covered in This Book
Research on Basque and Romance has made it clear that despite their typo-
logical differences (ergative vs. accusative marking, head final vs. head initial
linearization, among others), Basque and Spanish/French show remarkable
similarities. Thus, even if Basque can be descriptively characterized as a fairly
typical OV language and both French and Spanish also display fairly standard
VO features, information packaging mechanisms, however they are construed,
produce in Basque clause-level orders which are (increasingly) similar to those
found in Spanish and French (Ortiz de Urbina, this volume). Similarities are
more evident in colloquial registers and in some dialectal contexts (schooling
is mostly done in the standard variety). Some important similarities are there-
fore formally related to Romance phenomena (whether dialectal or not). For
instance, Differential Object Marking (Bossong 1991, Lazard 2001, Aissen 2003)
is attested in some South-western varieties in the Spanish-speaking area of the
Basque territory and shows great similarities with Spanish general a-marking
(Torrego 2004), or even with the dialectal leísmo of the Basque Country (Fer-
nández & Rezac 2016, Odria 2017 this volume).
In the same vein, the use of the se clitic in French and Spanish is closely
related with the absolutive marking of the subject and izan ‘be’ auxiliary selec-
tion in Basque intransitive verbs, especially in Romance-loan verbs (Alberdi
2003). When the verb in the original Romance language occurs with the se
clitic, the borrowed predicate in Basque takes an absolutive subject and izan
‘be’ auxiliary, whereas when the verb in the original Romance language does
not take se, then the predicate is generally used with an ergative subject and
*edun ‘have’ auxiliary in Basque. As a matter of fact, when Spanish and French
differ in the use of se in certain verbs, the loan verb in Basque shows dialectal
variation, with a contrast between the Basque varieties in contact with Spanish
and those in contact with French. The analysis of loan verbs and the compari-
son of counterpart intransitive verbs in Basque and Romance is of particular
interest, bearing in mind the different case systems of the languages under
study—Basque being ergative (Levin 1983, Ortiz de Urbina 1989, Oyharçabal
1992 among others) and Spanish/French being accusative—and auxiliary alter-
nation. For instance, Spanish, unlike French, does not display auxiliary alterna-
tion, but, even so, the presence of the se clitic in Spanish goes hand in hand with
the selection of the izan ‘be’ auxiliary in Basque loan verbs (Berro, this volume).
Other similarities go beyond dialectally constrained structures, such as the
nature of datives in general and non-argumental datives in particular. Although
recent literature has focused on the high vs. low nature of applied datives espe-
cially after Pylkkänen 2008 [2002], even higher in the architecture of the sen-
tence seem to be the so-called ethical datives, as analyzed by Jaeggli (1982, 1986)
for Spanish and by Jouitteau & Rezac (2007) for French among others. The
sole existence of ethical datives in Basque is not uncontroversial, as affected
(experiencer) datives seem to be of an argumental type, and therefore an in-
depth analysis of affectedness in Basque and Spanish/Basque is important.
Actually, Bosse, Bruening & Yamada (2012) provide with a fine-grained typol-
ogy that includes Hebrew, French, Japanese among other languages but lacks
both Basque and Spanish materials. Thus, an incursion into affected datives
and even allocutives (Oyharçabal 1993) will give us the opportunity to review
classical syntactic tests (as in Borer & Grodzinsky 1986) and more novel criteria
proposed by Bosse et alia (2012) and Horn (2008, 2013) in order to see to what
extent affected datives are argumental or not in both Basque and Spanish (Fer-
nández, this volume). The interplay between goal, experiencer, ethical datives
on one hand and causee datives in causative structures on the other is also
described for Basque and Spanish in Ortiz de Urbina (this volume). The appar-
introduction 5
ent typological distance created by ergative marking and complex agreement

patterns in Basque largely dissolves to reveal both expected fundamental simi-
larities and also high degrees of convergence when dative clitics and a marking
in Romance is considered.
Another topic covered in this volume involves non-verbal participles in
Basque and Spanish. Resultative participles in many languages have been
reported to have hybrid category, as they are both verbs and adjectives at the
same time. As a consequence, they have been the subject of great discus-
sions, particularly regarding the study of the interface between syntax and
the lexicon (Wasow 1977) and syntax and morphology (Marantz 2001 2007,
Anagnostopoulou & Samioti 2014), as well as lexical categories and the pro-
jection of syntactic heads like v, Voice and Asp(ect). Spanish past participles
have been studied in a number of works (among others Luján 1981, Demonte
1983, Bosque 1999 2014, Marín 1997 2000 2004ab 2009, Gehrke & Marco 2014
among others), but there are not many studies about Basque resultative par-
ticiples (Ortiz de Urbina & Uribe-Etxebarria 1991, Krajewska 2012 2013, Berro
2017) and none comparing both Spanish and Basque. Both languages are sim-
ilar in building resultative participles by means of a locative copula (egon in
western and central Basque varieties and estar in Spanish) and an adjectival
(or adverbial) participle. However, Basque resultative participles are morpho-
logically more complex (showing an additional resultative morpheme attached
to the bare participle) and seem to be acceptable in more syntactic contexts,
since they allow event-related modification. Additionally, if the copula turns
into izan ‘be’ (substituting egon), Basque resultatives are compatible even with
non-restricted event-initiators, whereas in Spanish only a subset of initiators is
acceptable (Gehrke & Marco 2014). These contexts have been used as tests in
order to analyze the structural composition of resultative participles across lan-
guages (Gehrke 2011, Alexiadou, Gehrke & Anagnostopoulou 2014, Alexiadou,
Anagnostopoulou & Schäfer 2015) and are also useful in order to compare Span-
ish and Basque, and make a contribution to the current discussion of the topic
(Berro, this volume).
4 Overview of the Chapters
As mentioned in previous sections, Basque and the surrounding Romance lan-

guages display significant similarities as well as obvious differences both in
their structural characteristics and in their diachronic development. The chap-
ters in this book address and analyze in detail what may be considered to be the
most important areas where the comparison between Basque and Romance
grammar and the analysis of phenomena which may be due to their interaction
can have a wider linguistic interest. Throughout the chapters it contains, topics
such as word order, inflection, non-verbal participles, intransitive predicates,
light verb constructions, ethical datives, differential object marking, causative
constructions and causal clauses will be presented. In what follows, we briefly
outline the main contributions in each chapter.
The labels OV/VO or head-first/head-last, at least as general characterizations
of (some) central elements in (some) phrasal constituents, would seem to place
Spanish and French on one side of this typological divide and Basque on the
other. Jon Ortiz de Urbina’s chapter on ‘Word Order’ shows that while Basque
and Romance often differ in the location of their morphological and syntactic
heads, considerable similarities (sometimes even convergence) may be found
at the clausal level. Thus, SOV and SVO may correspond to grammatically neu-
tral, pragmatically bleached orders in both languages (even if, at least in the
case of Basque, certainly not to the most common ordering). However, there are
numerous information packaging mechanisms which account for the majority
of actual clause-level orders, and some of these, old and new, produce simi-
lar orderings in the two languages. Even though Basque does not have the right
peripheral information focus of Spanish, the latter does have as an option a pre-
verbal focus akin to the general-purpose focalization strategy of Basque, and
clause final corrective focus is also increasingly used in Basque. In the course
of the presentation of phrasal order, the chapter also points at ‘non-harmonic’
orders within lexical phrases, appositions and compounds. Their diachronic
development and the comparison with Romance suggests that, at least in some
cases, their presence is not just a function of well-known OV inconsistency.
With virtually universal bilingualism in either Spanish or French and, in the
western area, an important mass of new speakers that make up a good portion
of Basque speakers in heavily populated areas, it would be surprising if a cer-
tain degree of grammatical confluence did not take place. We seem to observe
increasing use of orders which exploit possibilities, traditional or not, which
are compatible with the Spanish/French grammar of the same users.
The chapter ‘Tense, Aspect and Mood’ coauthored by José Ignacio and Céline
Mounole explores the morphological shape of the Tense, Aspect and Mood
(TAM) system in finite verbal forms. As previously noticed in the literature
(Michelena 1981), Basque analytical verbal forms have Romance counterparts
that suggest similar diachronic development. Some similarities between
Basque and Romance are the distinction between indicative and subjunc-
tive, along with conditional and imperative, as far as the mood is concerned,
or the contrast between perfective and imperfective forms regarding aspect.
Nevertheless, there are also certain differences. For instance, Basque shows
introduction 7
a be/have auxiliary alternation whereas Spanish has only have auxiliary and
French selects mainly have and reserves be for a small class of intransitives.
These and other aspects show that Basque has both a convergent and a diver-
gent development with its Romance neighbors.
Ane Berro’s chapter ‘Non-verbal Participles in Basque and Spanish’ com-
pares and contrasts non-verbal participles in both languages, in light of the
literature on adjectival participles (among others, Wasow 1977, Marantz 2001
2007, Kratzer 1994 2000, Anagnostopoulou 2003, Embick 2004). The author
points out a number of differences, regarding the morphological shape of the
non-verbal participles and their aspectual interpretation. In fact, participles
take extra morphemes in Basque when they are adjectival or postpositional,
whereas in Spanish both verbal and non-verbal ones look similar. As for the
aspectual interpretation, Basque non-verbal participles can be used with both
resultative and experiential interpretation, but Spanish non-verbal participles
only trigger the resultative reading. Despite these differences, Spanish and
Basque participles share a number of properties. On the one hand, they show
a similar interaction between the lexical aspect of the verb embedded under
the participle and the aspectual interpretation of the whole structure. On the
other, non-verbal participles in both Basque and Spanish (at least, the Spanish
spoken in the Basque Country) (cf. Gehrke & Marco 2014), allow spatial and
temporal modification of the event underlying the participle.
The chapter ‘(In)transitive Verbs: Unergatives and Unaccusatives’, also
authored by Ane Berro, analyzes the unaccusative/unergative divide of intran-
sitive predicates (Perlmutter 1978 1989, Burzio 1981, 1986, Belletti 1988 1999,
Sorace 1993, Levin & Rappaport Hovav 1995 and subsequent works) in Basque
and several Romance languages. The author focuses on the verb classes that
are systematically aligned in the unergative way or, alternatively, in the unac-
cusative way, both in Basque and Romance, and also border-line verbs (like
Romance loan verbs and some new stative verbs). As proposed by the author,
intransitive stative verbs are the most variable verbs both cross-linguistically
and in Basque, precisely because the only argument of these predicates can be
introduced externally or internally without making a big semantic difference.
Víctor Acedo-Matellán and Anna Pineda’s chapter ‘Light Verb Constructions
in Basque and Romance’ analyzes the morphosyntactic and semantic prop-
erties of Light Verb Constructions (LVC) in Basque and compares them to
their counterparts in Romance—mainly Spanish and French. They argue that
the Light Verb (LV) is responsible for the event- and argument-properties of
the construction, whereas the Non-Verbal Element (NVE) provides conceptual
information of the event involved. As shown in Basque and also Spanish and
French, the choice of the LV is meaningful, given that only DO and GIVE give
rise to synonymous LVC s in these languages. Regarding the morphosyntactic

and semantic relation of the NVE and LV, the authors show that Basque LCV
allows NVE to be headed by postpositional or adverbial suffixes—with certain
aspectual interpretations—, contrary to Spanish and French, where the NVE
must be nominal. This nominal can be either bare—in French—or, in the case
of Spanish and Catalan, necessarily accompanied with a determiner. Finally,
this chapter also addresses the syntactic cohesion existing between the LV and
NVE and shows that there is variation among the different LVC s with respect to
the degree of incorporation.
Beatriz Fernández’s chapter entitled ‘On Non-selected Datives: Ethical
Datives in Basque and Spanish’ deals with non-selected datives. Basing on
Franco & Huidobro (2008), the chapter proposes a three-way typology of ethi-
cal datives: First of all, Class I, reminiscent of personal datives (Horn 2008 2013);
secondly, Class II, which correspond to affected experiencers in Bosse, Bruening
& Yamada (2012); and thirdly, Class III, ethical datives stricto sensu. All classes
are attested in Spanish, whereas as the author shows, only Class II—affected
experiencers—can be found in Basque. Both classical and new tests are used
in order to support this claim, such as the existence of person restrictions, dou-
bling, clitic clusters, and the presence of truth conditional and non-truth condi-
tional meaning. Although no particular analysis is provided, the author extends
Odria’s (2017) analysis of benefactives and external possessors to Class II eth-
ical datives, and claims that these datives are merged in the specifier position
of an Applicative head above v.
The chapter on Differential Object Marking (DOM) by Ane Odria compares
this phenomenon in both Basque and Spanish, particularly the variety of Span-
ish spoken in the Basque Country (Landa 1995). The chapter shows that DOM
in these languages (or language varieties) is conditioned by animacy and speci-
ficity, and has the same morphological shape of indirect objects, that is, dative
case and agreement. The author claims that DOM objects have the catego-
rial status of DP s, like causee, experiencer and possessor datives. By contrast,
datives of bivalent unergatives of the lagundu/ayudar ‘accompany’, ‘help’ and
ditransitive predicates are PP s. In fact, the author shows that DP datives, DOM
objects among others, can control depictive secondary predicates, whereas PP
datives cannot.
DOM-like phenomena are also discussed within the wider domain of causa-
tive formation in Jon Ortiz de Urbina’s chapter ‘Complex Causative Verbs and
Causees in Basque (and Romance)’. These constructions present another case
of a prima facie drastic typological difference between Basque morphological
and Romance analytical complex verb formation mechanisms in causatives
(Comrie 1989, Dixon 2000), which, when considered in detail, turn out to
introduction 9
present perhaps not-so-striking parallelisms. The Basque causative morpheme

is fairly transparently a verb with an archaic causative affix, while some
Romance fare/faire/hacer causatives with infinitival complement clauses have
been treated, at least since Zubizarreta (1985) as monoclausal structures. The
chapter also provides a fine-grained comparison of direct and indirect causa-
tion and impersonal causatives in Basque and Romance, resorting further to
Basque mediopassive causatives as a term of comparison with Romance pas-
sive causatives. Causatives involve the presence of an extra, causer argument
supplied by the causative verb, and case marking patterns for this argument are
also discussed in detail in this chapter. Dative marking for the causee is stan-
dard in Basque for some unergative causee subjects and for transitive subject
causees, while dialectal and substandard patterns increasingly resort to dative
marking even in these contexts with animate causees, a DOM pattern particu-
larly clear in dialects in contact with Spanish. This dative may in turn cooccur
with a dative argument of the verb in the caused event, whether a goal, an expe-
riencer or a non-argumental ethical dative, and the chapter presents a detailed
description of the availability of dative doubling in these contexts, comparing
it with dative clitic marking in Spanish.
José Ignacio Hualde and Manuel Pérez Saldanya’s chapter ‘Grammatical-
ization Processes in Causal Subordination’ focuses on Basque and Romance
causal clauses. As shown in the chapter, Basque exhibits more heterogeneity
than Romance regarding the subordinators involved in these clauses, given that
not only verbal suffixes but also prefixes and clause initial conjunctions are
attested in Basque. The chapter explores in-depth the development of Basque
and Romance causal structures in both their similarities and differences. From
a functional perspective, in both Basque and Romance, there is a causal subor-
dinator with a polyfunctional character and other subordinators appearing in
more restricted contexts. The authors show that the grammaticalization paths
of causal elements are similar in many respects, even though some Basque
strategies such as the grammaticalization of the enclitic eta ‘and’ as a clausal
connector lack a Romance counterpart.
References
Aissen, Judith. 2003. Differential object marking: iconicity vs. economy. Natural Lan-
guage and Linguistic Theory 21: 435–483.
Alberdi, Xabier. 2003. Euskal aditz mailegatuen erregimena: hurbilketa. In Jesus Mari
Makazaga & Bernard Oyharçabal (eds.), P. Lafitteren sortzearen mendemugako bilt-
zarra. Gramatika gaiak. Iker 14 (I). Bilbao: Euskaltzaindia. 37–60.
Alexiadou, Artemis, Elena Anagnostopoulou & Florian Schäfer. 2015. External Argu-
ments in Transitivity Altenations. A Layering Approach. Oxford: Oxford University
Press.
Allières, Jacques. 1992. Gascón y euskera: afinidades e interrelaciones lingüísticas. ASJU
26 (3): 801–812
Anagnostopoulou, Elena. 2003. Participles and Voice. In Artemis Alexiadou, Monika
Rathert and Armin von Stechow (eds.), Perfect explorations. Berlin: Mouton de
Gruyter. 1–36.
Anagnostopoulou, Elena & Yota Samioti. 2014. Domains within words and their mean-
ings: A case study. In Artemis Alexiadou, Hagit Borer and Florian Schäfer (eds.), The
Syntax of Roots and the Roots of Syntax. Oxford: Oxford University Press.
Belletti, Adriana. 1988. The case of unaccusatives, Linguistic Inquiry 19: 1–34.
Belletti, Adriana. 1999. “Inversion” as focalization and related questions, Catalan Work-
ing Papers in Linguistics 7: 9–45.
Berro, Ane. 2017. Resultative and non-resultative participles in Basque. Ms, University
of the Basque Country & SFL (UMR 7023, CNRS/Paris 8).
Borer, Hagit & Yosef Grodzinsky. 1986. Syntactic cliticization and lexical cliticization:
The Case of Hebrew Dative Clitics. In Hagit Borer (ed.), Syntax and Semantics 19.
New York: Academic Press. 175–215.
Bosse, Solveig, Benjamin Bruening & Masahiro Yamada. 2012. Affected experiencers.
Natural Language and Linguistic Theory (NLLT), 30(4): 1185–1230.
Bossong, Georg. 1991. Differential Object Marking in Romance and beyond. In Dieter
Wanner & Douglas A. Kibee (eds.), New Analyses in Romance Linguistics. Amster-
dam / Philadelphia: John Benjamins. 143–170.
Bosque, Ignacio. 1999. El sistema adjectival. Complementos y modificadores del adje-
tivo. Adjetivo y participio. In Ignacio Bosque & Violeta Demonte (eds.), Gramática
descriptiva de la lengua española. vol. 1. Madrid: Espasa Calpe. 217–310.
Bosque, Ignacio. 2014. On resultative past participles in Spanish. Catalan Journal of Lin-
guistics 13: 41–77.
Burzio, Luigi. 1981. Intransitive verbs and Italian auxiliaries. Ph.D. diss, MIT. Cambridge,
MA.
Burzio, Luigi. 1986. Italian Syntax. Dordrecht: Reidel.
Comrie, Bernard. 1989. Language Universals and Linguistic Typology. Chicago: Chicago
University Press.
Demonte, Violeta. 1983. Pasivas léxicas y pasivas sintácticas en español. Serta Philolog-
ica F. Lázaro Carreter. Madrid: Cátedra. 141–157.
Dixon, Robert M.W. 2000. A Typology of Causatives: Form, Meaning and Syntax. In
Robert M.W. Dixon & Alexandra Y. Aikhenwald, eds. Changing valency. Case stud-
ies in transitivity, Cambridge: Cambridge University Press, 30–83.
Embick, David. 2004. On the Structure of Resultative Participles in English. Linguistic
Inquiry 35-3: 355–392.
introduction 11
Fernández, Beatriz & Milan Rezac. 2016. Differential Object Marking in Basque vari-
eties. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
grammar of Basque. Amsterdam: John Benjamins. 93–139.
Franco, Jon & Susana Huidobro. 2008. Ethical Datives, Clitic doubling and the Theory
of pro. Joyce Brukes de Garavito & Elena Valenzuela (eds.), Selected Proceedings of
the 10th Hispanic linguistics Symposium. Somerville: Cascadilla Proceedings Project.
215–224.
Gehrke, Berit. 2011. Stative passives and event kinds. In I. Reich, E. Horch and D. Pauly
(eds.), Proceedings of Sinn und Bedeutung 15. Saarbrücken: Universaar—Saarland
University Press. 241–257.
Gehrke, Berit & Cristina Marco. 2014. The role of by-phrases in adjectival passives. Lin-
gua 149: 188–214.
Horn, Laurence R. (2008). I love me some him: The landscape of non-argument datives.
In Olivier Bonami & Patricia Cabredo Hofherr (eds.), Empirical Issues in Syntax
and Semantics 7, CSSP, Paris, 169–192. [On-line publication, http://www.cssp.cnrs.fr/
eiss7].
Horn, Laurence R. (2013). I love me some datives: Expresive meaning, free datives, and
F-implicature. In Daniel Gutzmann & Hans-Martin Gärtner (eds.), Beyond Expres-
sives: Explorations in Use-Conditional Meaning. Leiden: Brill. 153–201.
Hualde, José Ignacio. 2003. Introduction. In José Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A grammar of Basque. 1–14. Berlin: Mouton de Gruyter.
Hualde, José Ignacio. 2016. On Basque dialects. In Beatriz Fernández & Jon Ortiz de
Urbina (eds.), Microparameters in the Grammar of Basque. Amsterdam/Philadel-
phia: John Benjamins.15–35
Jaeggli, Osvaldo, A. 1982. Topics in Romance Syntax. Dordrecht: Foris.
Jaeggli, Osvaldo, A. 1986. Three issues in the theory of clitics: case, doubled NP s and
extraction. In Hagit Borer (ed.), Syntax and Semantics 19. New York: Academic Press.
15–42.
Jouitteau, Mélanie & Milan Rezac (2007), The French ethical dative, 13 syntactic tests.
Bucharest Working Papers in Linguistics, IX (1): 97–108.
Krajewska, Dorota. 2012. Resultatives in Basque: a diachronic study. Lingua Posnanien-
sis LIV(2): 56–67.
Krajewska, Dorota. 2013. The diachrony of resultative constructions in Basque. Master
Thesis, UPV/EHU.
Kratzer, Angelika. 1994. The Event Argument and the Semantics of Voice. Ms., Univer-
sity of Massachusetts at Amherst.
Kratzer, Angelika. 2000. Building Statives. Berkeley Linguistic Society 26. 385–399.
Lafon, R. 1980 [1943]. Le système du verbe basque au XVIe siècle. Zarautz: Elkar.
Landa, Alazne. 1995. Conditions on Null Objects in Basque Spanish and their relation to
Leísmo and Clitic Doubling. University of Southern California.
Lazard, Gilbert. 2001. Le marquage différential de l’object. In Martin Haspelmath, Ekke-

hard König, Wulf Oesterreicher & Wolfgang Raible (eds.), Language Typology and
Language Universals. An International Handbook. 2 vol. Berlin / New York: Mouton
de Gruyter. 873–885.
Levin, Beth. 1983. On the Nature of Ergativity. Doctoral Dissertation, MIT.
Levin, Beth & M. Rappaport Hovav. 1995. Unaccusativity. At the Syntax-Lexical Semantic
Interface. Cambridge, MA: the MIT Press.
Luján, Marta. 1981. The Spanish copulas as aspectual indicators. Lingua 54: 165–209.
Marantz, Alec. 2001. Words. WCCFL XX Handout, USC, February 2001.
Marantz, Alec. 2007. Phases and Words. In Sook-Hee Choe (ed.) Phases in the Theory of
Grammar. Seoul: Dong In. 191–222
Marín, Rafael. 1997. Participios con aspecto de adjetivos: entre la diacronía y la mor-
fología. Moenia 3: 365–376.
Marín, Rafael. 2000. El componente aspectual de la predicación. Doctoral dissertation.
Universitat Autònoma de Barcelona.
Marín, Rafael. 2004a. Entre ser y estar. Madrid: Arco/Libros.
Marín, Rafael. 2004b. Sobre pasivas adjetivales. Verba 31: 455–471.
Marín, Rafael. 2009. Del adjetivo al participio. In Elena de Miguel et al. (eds.), Fronteras
de un diccionario: las palabras en movimiento. San Millán de la Cogolla: Cilengua.
327–348.
Michelena, Luis. 1981. Lengua común y dialectos vascos. ASJU XV: 291–303. Reprinted
in: Luis Michelena. 2011. Obras completas [Supplements of ASJU VII]. ed. by Joseba
A. Lakarra & Iñ igo Ruiz Arzallus, Donostia-San Sebastián & Vitoria-Gasteiz: Gipuz-
koako Foru Aldundia & UPV/EHU. 517–544.
Mounole, Céline. 2011. Le verbe basque ancien: étude philologique et diachronique. Doc-
toral dissertation, University of the Basque Country (UPV/EHU) and Université
Michel de Montaigne-Bordeaux III.
Odria, Ane. 2017. Differential Object Marking and datives in Basque syntax. Doctoral dis-
sertation, UPV/EHU.
Ortiz de Urbina, Jon. 1989. Parameters in the Grammar of Basque. Studies in Generative
Grammar 33. Dordrecht: Foris Publications.
Ortiz de Urbina, Jon & Myriam Uribe-Etxebarria. 1991. Participial predication in Basque.
ASJU Gehigarriak 14-2. 993–1012.
Oyharçabal, Bernard. 1992. Structural and Inherent Case-Marking; Ergaccusativity in
Basque. In Joseba, A. Lakarra & Jon Ortiz de Urbina (eds), Syntactic Theory and
Basque Syntax. Supplements of ASJU 27. Donostia: UPV/EHU and Gipuzkoako Foru
Aldundia. 309–342.
Oyharçabal, Bernard. 1993. Verb agreement with non-arguments: On allocutive agree-
ment. In José Ignacio Hualde & Jon Ortiz de Urbina (eds.), Generative studies in
Basque linguistics. Amsterdam: John Benjamins. 89–114.
introduction 13
Perlmutter, David, M. 1978. Impersonal Passives and the Unaccusative Hypothesis. Pro-
ceedings of the 4th Annual Meeting of the Berkeley Linguistics Society. 157–190.
Perlmutter, David, M. 1989. ‘Multiattachment and the unaccusative hypothesis: the Per-
fect Auxiliary in Italian.’ Probus 1-1: 63–119.
Pylkkänen, Liina. 2008. Introducing arguments. Cambridge / London: MIT Press.
Sorace, Antonella. 1993. Unaccusativity and auxiliary choice in non-native grammars of
Italian and French: asymmetries and predictable indeterminacy. Journal of French
Language Studies 3: 71–93.
Wasow, Thomas. 1977. Transformations and the lexicon. In Peter Culicover, Thomas
Wasow & Adrian Akmajian (eds.), Formal Syntax. 327–360. New York: Academic
Press.
Zubizarreta, María Luisa. 1985. Romance accusatives. Linguistic Inquiry 16: 247–289.
chapter 2
Word Order
Jon Ortiz de Urbina
Basque and Spanish differ rather drastically in the organization of constituents

internal to major categories, the former being largely head-final while the lat-
ter is head-initial. On the other hand, word order at the clausal level is fairly
flexible in both languages. Thus, while Basque is usually classified as an SOV lan-
guage, this is largely a typological shorthand for word orders other than those
found for the combination of subject, object and verb with each other. While
SOV is often recognized as a ‘basic’ order, Basque is far from being a rigid verb-
final language, and heaviness as well as information packaging produce a wide
array of possible orders of major clausal constituents. It may not be misleading
to claim that Basque is a fairly typical OV/head-last language, while Spanish is a
fairly typical VO/head-first language. For the purposes of this chapter, however,
rather than discussing the validity of these macro-labels, it is more interest-
ing to examine in more detail the situation across categories and the possible
existence of changes which may increase convergence between the two lan-
guages.1
1 Morphology
Disregarding dvandva cocompounds, very productive in Basque but where

there is an additive, symmetric relation between their components (Hualde
1998), NN compounding is systematically right-headed: itsasgizona ‘seaman’
(itsas(o) ‘sea’ + gizona ‘man’) is structurally akin to the English translation.2 This
1 The abbreviations used in glosses are as follows: aux ‘auxiliary’, dat ‘dative’ det ‘determiner’,
comp ‘complementizer’, erg ‘ergative’, fut ‘future’, gen ‘genitive’, impf ‘imperfective’, nom
‘nominalizing affix’, part ‘partitive’, pl ‘plural’, q ‘question morpheme’. Glosses in a language
like Basque can be very obstrusive, so for readability purposes they have been kept to a min-
imum, providing in most cases only the grammatical information for which the example is
relevant. Thus, the determiner (and agreement) marker -a may be glossed as det, ‘the’ or
simply left out depending on the relevance to the discussion.
2 Apparent exceptions such as erpuru ‘thumb’ or hiriburu ‘capital’, (eri ‘finger’ and hiri ‘town’
+ buru ‘head’), designating types of fingers or towns may not be so when compared with the

word order 15
type of compounding is also extremely productive in Basque, while it is much

less so in Spanish; examples in the latter are systematically left-headed. Thus,
hombre rana contrasts structurally with both its English and its Basque coun-
terparts, frogman and igelgizona, respectively, and perro pastor ‘shepherd dog’
places the two elements in the opposite position with respect to English and
Basque artzain txakurra. Spanish often resorts to syntactic phrases where the
dependent is introduced by the preposition de or expressed as an adjective in
cases where Basque displays NN compounds: copo de nieve ‘snowflake’, Basque
elur maluta (elur ‘snow’, maluta ‘flake’); similarly, Spanish derechos humanos
‘human rights’ corresponds to Basque giza eskubideak (giza ‘human being’,
eskubide ‘right’).
NN compounds of the type briefly described in the previous lines are particu-
larly interesting in that they present mergers of elements which are not usually
combined in the (traditional view of) syntax. Other types of compounding of
categories have clearer parallels in the syntax, and often mirror the ordering
of their syntactic parallels. Thus, NV primary compounds are not very pro-
ductive in Basque, but existing ones are again right-headed from the earlier
known stages of the language. Thus Michelena (1977), Mounole (2009, 2011)
identify itxaron ‘wait’ as hitzedun (hitz ‘word’ + edun ‘have’), and iguriki ‘wait’
as eguneduki (egun ‘day’ + eduki ‘hold’); berdin ‘same’ is an old compound with
predicative complement ber- ‘same’ and copulative edin ‘be, become’. In a more
productive and contemporary vein, bahuvrihi compounds are very common
with what de Rijk (2008) identified as edun ‘have’ in the second element of
words like euskaldun ‘Basque (speaker)’ (euskara ‘Basque’), haurdun ‘pregnant’
(haur ‘child’), etc. Other such compounds with egin ‘make’ as second conjunct
are okin ‘baker’ (ogi ‘bread’) or iturgin ‘plumber’ (iturri ‘fountain’). Endocentric
compounds with a second verbal element are also productive, both in tradi-
tional words and neologisms: itxuraldatu ‘transform’ (itxura ‘shape’ + aldatu
‘change’), antzeman ‘guess’ (antza ‘appearance’ + eman ‘give’), lurrartu ‘land’
(lur ‘earth’ + (h)artu ‘take’), etc. Compounds with verbs in Spanish conform to
the basic VO pattern of the language, as in the very productive type forming
expected asteburu ‘weekend’ (aste ‘week’) or iturburu ‘spring’ (iturr(i) ‘fountain’). Notice that
while the latter are interpreted as the head of the week and perhaps as the head of a foun-
tain, the former are not the head of a finger or of a city. Rather, buru receives an interpretation
actually close to the one represented in capital, originally an adjective (from the Latin word
for head) in the collocation capital city, which the Basque word follows closely. Given that
the categorial distinction between nouns and adjectives was not clear in older stages of the
language, we may assume that buru in the first two examples is working as an adjective, and
that these are N Adj compounds of the types described below.
16 ortiz de urbina
nouns with a third person verbal form as first term: saltamontes ‘grasshopper’
(salta ‘jumps’ + montes ‘mountains’), sacacorchos ‘corkscrew’ (saca ‘takes out’ +
corchos ‘corks’). Basque NN secondary compounds (again just as productive in
Basque as they are scarce in Spanish) also follow systematically an OV pattern,
with the deverbal noun to the right as in ile apaingile ‘hairdresser’ (ile ‘hair’ +
apaingile ‘dresser’, from apain- ‘beautify’ and -gile ‘-er’) or suhiltzaile ‘firefighter’
(su ‘fire’ + hiltzaile ‘killer’, from hil ‘kill’ and -tzaile ‘-er’).3
We also find word orders similar to those of syntactic phrases in N Adj com-
pounds. These follow this order in both Spanish and Basque, as expected since
adjectives usually follow nouns in the syntax of both languages; cf. Bsq. bela-
rrimotz ‘small-eared, foreigner’ (belarri ‘ear’ + motz ‘short’), and Span. paticorto
‘short-legged’ (pata ‘leg’ + corto ‘short’). See 2.2.1.3 below for adjective ordering
within nominal phrases. There are few compounds containing adpositions in
either Basque or Spanish, and the few cases that do exist show the expected
order between the adposition head and its nominal complement. Spanish
sin ‘without’ is the most productive preposition found in compounds: sinsen-
tido ‘nonsense’, sinsabor ‘setback’, sinvergüenza ‘shameless, scoundrel’, sinsorgo
‘silly’;4 its Basque postpositional counterpart is gabe ‘without’, also very pro-
ductive, as in paregabe ‘matchless’, etengabe ‘continuous’, lit. ‘without break’,
lotsagabe ‘shameless’ or beldurgabe ‘fearless’. Other prepositions are only occa-
sionally found in Spanish: bajotecho ‘attic’, lit. ‘under roof’, sobrecama ‘bed-
spread’, lit. ‘over bed’, or sobrepeso ‘overweight’. There are even fewer examples
in Basque, partly because many Spanish prepositions correspond to Basque
‘case’ endings, bound inflectional morphemes. There are, however, indepen-
dent postpositions in Basque, of which gabe ‘without’ is a typical example.
Most postpositions, however, are relational nouns inflected for a locative case.
These combine with a noun in the genitive (or a verb in the participial form)
or may be compounded with it;5 in either case, the expected order is main-
3 Itself from (e)gi(n) ‘do’ and -le ‘-er’.

4 All are nouns/nominalized expressions, with only sinvergüenza having a possible adjectival
use. The predominance of negative meanings may be related to similar semantics in what
Bosque (1989) calls evaluative nominalization.
5 The absence of a genitive marking is not necessarily an indication of compounding. Thus,
etxe aurrean ‘in front of the house’ (lit. ‘at the house front’) is spelled as two independent
words, capturing a phonetic and perhaps morphological intuition. Notice that the postpo-
sition still requires a locative case, indicating the phrase has not become an independent
expression. There is a difference between ate aurrean ‘in front of the door’, with this type of
genitiveless phrase, and the bona fide compound ataurre ‘threshold’ (ate ‘door’ + aurre ‘front’).
The issue is irrelevant for ordering, however, so different types are used in the text as illustra-
tion.
word order 17
tained: etxe(aren) aurrean ‘in front of the house’, bazkalondo ‘chat after eating’
(cf. bazkaldu ondoren/ondoan ‘after having lunch’). Some relational nouns that
form the basis of postpositions, however, can be found in prenominal posi-
tion, only indirectly as a result of accommodation to Romance patterns. Thus,
azpi ‘below’ is used as the equivalent of the Romance prefix sub- in neologisms
like azpiegitura ‘substructure’ or azpimultzo ‘subset’. Basque has a wide array
of derivational suffixes but extremely few prefixes,6 and this type of seman-
tically transparent Latinate prefix is also transparently rendered by the noun
azpi. This is then a case of accommodation to the prefixation possibilities of
Romance, rather than to any kind of prepositional pattern. Notice that the pre-
fixed element is the basic noun without the locational marker, and that there is
no complement relationship between this noun and the one to its right: azpie-
gitura ‘substructure’ is not interpreted as ‘under the structure’ or ‘underside of
the structure’, but as a structure which is located under some other. The loca-
tional noun acts as an adjunct, not as a head.7 Similar considerations apply to
the use of aurre ‘before/front’ as equivalent to the prefix pre- in neologisms
like aurrizki ‘prefix’, aurreritzi ‘prejudice’, lit. ‘before opinion’. The latter is inter-
preted not as what is held before an opinion, but as an opinion formed previous
to contact with data and/or arguments.8
We have already briefly mentioned derivation while discussing relational
nouns. As indicated above, Basque is extremely limited in terms of derivational
prefixes. This in itself fits Greenberg’s findings on the relationship between
affixation and phrasal ordering, since he observed a tendency for SOV languages
to display suffixation exclusively, while SVO languages usually exhibit both. In
principle, if the affix is the head of a derived word, determining its category, one
6 As pointed out by de Rijk (2008), only three: ber- ‘re-’ and the loaned prefixes arra- ‘re’ and
des- ‘un-’.
7 This is also the use found in other words containing azpi- in the traditional vocabulary, such
as azpijoko ‘foul play’, lit. ‘under play’, or azpizapo ‘traitor’, lit. ‘under toad’.
8 Some morpheme ordering relationships within the number system were already claimed to
be linked to other phrasal relations in Lehman (1975). In particular, he claimed that in addi-
tive numbers (for instance, thirteen, three plus ten, as opposed to thirty, three times ten) the
smaller number added to the bigger one (the decimal in this type of system) behaves like
the verb in the O/V relation. Regardless of the overall validity of this correspondence, it does
capture the relation between the Basque and Spanish systems. Few numbers are additive in
these languages, 11 through 19 in Basque and 11 through 15 in Spanish, the remaining higher
numbers being either factorial (berrogei ‘40’, lit. 2 20 in Basque and cuarenta in Spanish, with
initial 4) or coordinations (berrogeitamar ‘50’, lit. 2 20 and 10 in Basque or dieciséis ‘16’, lit. 10
and 6 in Spanish). Basque additive numbers take the form ten + unit (hamairu ‘13’, lit. 10 3)
and, more opaquely, unit + ten in Spanish (doce ‘12’, 2?).
18 ortiz de urbina
would expect that head-first languages would favor prefixation, whereas head-
final languages like Basque would favor suffixation; however, this is clearly not
the case.9
2 Syntax
After examining morpheme ordering phenomena, let’s turn now to word order
phenomena in a more strict sense, devoting special attention to interlinguis-
tically or intralinguistically unexpected patterns, i.e., unexpected when com-
pared to patterns typical of OV languages and/or incongruent with other order-
ing internal to Basque. We begin with appositions in 2.1, devoting 2.2 to word
order in major lexical phrases and 2.3 to ordering at the clausal and sentence
level.
2.1 Appositions
The relative order between type and token in restrictive appositions correlates
with other VO/OV characteristics, so that the type expression corresponds to the
V position and the token phrase to the O position. In fact, this was the content
of Greenberg’s Universal 23, where he showed a correlation between proper
noun/common noun appositions and the order of genitive/noun pairs.10 Since
the genitive/noun order patterns with the V/O order, this type of apposition
would then show overall mirror image effects when comparing Basque and
Spanish. The differences between the two languages are clear in this respect:
(1) Agirre Jauna/ El señor Aguirre ‘Mr. Aguirre’
(2) Kantauri Itsasoa/Mar Cantábrico ‘Bay of Biscay’
(3) Ibaizabal ibaia/El río Ibaizabal ‘the river Ibaizabal’
(4) 2000 urtea/el año 2000 ‘the year 2000’
(5) Gernikako Arbola Plaza/ Plaza Gernikako Arbola ‘Gernikako Arbola

Square’
9 Kayne (2017) for an antisymmetric approach to this aspect of morphology.

10 The wording of this Universal was in fact mistaken, as noted in Bennet (1979); see Cinque
(2013: 110).
word order 19
(6) El País Egunkaria/Diario El País ‘the newspaper El País’
(7) purpura kolorea/el color púrpura ‘the color purple’
As usual, it is possible to find exceptions, more in Basque than in Spanish,11

although contact with Spanish may be the main driving force behind them.
Thus, the pressure of official designations, systematically in a Romance lan-
guage, accounts for the presence of VO-like apposition patterns in Basque
different from the general ones displayed above. It is perhaps not surprising
that the names of saints are identical in both languages, with a Saint+Name
order unexpected in Basque (San Juan). More interestingly, the same situation
obtained in earlier, pre-Trento times, when saint denominations were not sus-
pect of heresy and therefore were not subjected to uniformity constraints; the
same is the case for toponyms (which, themselves, often double with an offi-
cial designation). Earlier saint names usually contain the title Done (from Latin
domine)¸ often preceded by the same (preposed) Jaun in (1) above in the case
of male saints (Michelena & Yrigarai 1955): Jaun Done Petri ‘Saint Peter’, top.
Doneztebe ‘Saint Stephen’, Jaona domne Jakue ‘Saint Jacques’ (12th century).
Puristic tendencies at the end of the 19th century introduced the treatment
Deun(a) for the then standard Spanish San(to/ta), initially at the same prenom-
inal position but later in the expected OV position following the name: Agate
Deuna ‘Saint Agatha’ (at first Deun Agate), Mikel Deuna ‘Saint Michael’, etc. The
unexpectedly preposed title Jaun ‘Sir’ of traditional saint denominations can
also be found preceding the token in other official contexts where the pres-
sure of official Romance denominations would be easily felt. This was probably
the case of high status positions in a traditional society, such as mayor, priest,
doctor, governor, etc. Similarly, early texts also present prenominal Jaun with
proper names, probably following the Romance model too, and this position
has remained frequent in northern dialects.
Appositions whose type nominal corresponds to a family relation noun are
often built on a relation+name order. This is expected in the case of nouns like
aita ‘father’ and ama ‘mother’, only used in religious contexts akin to those
described in the previous paragraph, as in Aita Donostia ‘Father Donostia’ or
11 Cinque (2013) also notes counterexamples in both OV and VO languages, but the only
Romance language mentioned in this respect, Italian, conforms to the common/proper
order in all instances he checks. English displays more deviations (Church Street, Coney
Island, Potomac River), perhaps connected to the head last organization of NN com-
pounds.
20 ortiz de urbina
Ama Teresa Kalkutakoa ‘Mother Teresa of Calcutta’, or even Ama Lurra ‘Mother
Earth’, but less so in the much more productive cases involving relations like
izeko ‘aunt’, osaba ‘uncle’ or amona ‘grandmother’. Example (8) is a celebra-
tory radio message for the Bizkaia Irratia radio station indicating the persons
wishing their 6-year-old relative a happy birthday, while (9) displays two family
relation names, each in a different order:
(8) Bere aita Gaizkak, ama Isabelek eta neba Aritzek. Baita
her father G. mother I. and brother A. also
osaba Mikelek, aitona Jose Martin eta amona
uncle M. grandfather J.M. and grandmother
Mari Karmenek.
M.K.
‘Her father Gaizka, mother Isabel and brother Aritz. Also uncle Mikel,
grandfather Jose Martin and grandmother Mari Karmen.’
(9) Bere osaba Pedrok Guillerma amatxiren ganik ukan zuen etxe
his uncle Pedro Guillerma grandm.gen from have aux house
gotorrari.
strong.dat
‘To the strong house that his uncle Pedro had (received) from grand-
mother Guillerma.’ (Oncle, Ardoy SFran 84)
The Orotariko Euskal Hiztegia [Basque General Dictionary], from which the
latter is taken, lists examples with preposed and postposed osaba, but the ten-
dency seems to be towards preposing.
2.2 Order in Major Phrases

Although the distinction is epiphenomenal, we will discuss first ordering rela-
tionships in the major lexical phrases (and associated functional elements),
excluding verbal complements and modifiers, which will be considered in sec-
tion 2.3. This enables us to discuss in the latter more traditional ‘clause level’
ordering relationships.
2.2.1 NP s
Not unlike the situation described in 2.1, Basque behaves in general like an OV
language, with some ‘anomalies’, some but not all of which can be attributed
to contact phenomena with Spanish and Romance in general. We will describe
determiners and quantifiers in 2.2.1.1, adposition phrase modification of nouns
in 2.2.1.2, adjective modifiers in 2.2.1.3 and relative clauses in 2.2.1.4.
word order 21
2.1.1 Determiners and Quantifiers

Let’s briefly review first of all order relations between grammatical elements
such as determiner/quantifiers and nouns. One of Abney’s (1987) crucial argu-
ments in favor of the existence of Determiner Phrases headed by articles,
demonstratives and genitives in English was that if they are indeed heads tak-
ing NP complements, the expected order in an OV, head-last language would be
noun-determiner. This is indeed the difference between Spanish and Basque:
the three of them precede the noun in the former, while articles and demon-
strative follow the noun in Basque (possessives are expressed by postpositions,
as indicated above). In fact, while it is likely that the grammaticalization of a
demonstrative as an article was taking place as an areal phenomenon, affecting
Romance and Germanic languages as well as Basque at roughly the same early
medieval times, the grammaticalization in Basque retained the final position of
the original demonstrative, producing contemporary mirror image orders like
este chico/mutil hau ‘this guy’ (lit. ‘guy this’) and el chico/mutila ‘the guy’ (lit.
‘guy the’).
There is only one type of construction where the determiner apparently pre-
cedes the noun, Biscayan hori mutilori, where hori ‘that’ precedes the noun. But
notice that this is possible only if there is yet another demonstrative in the
expected position (the -ori form following the noun).12 There are in fact two
phrases involved here, and case marking must then be doubled. Thus, the erga-
tive form of this phrase would be horrek mutilorrek, with two -ek endings, akin
to Hungarian forms like abban a házban ‘in that house’, literally ‘in that in the
house’. The initial determiner is therefore not directly associated with the noun.
In the case of Spanish postnominal demonstratives like el chico ese ‘that guy’
we do not find demonstrative doubling, since the final demonstrative coexists
with the article el. The traditional analysis of ese as an adjective in this case,
rather than as a determiner, captures the distinction between the determiner
demonstrative and this one.13 It shares the postnominal position of adjectives,
although its positional possibilities are more varied than for normal adjectives:
el chico ese alto de gafas, el chico alto ese de gafas, el chico alto de gafas ese are
12 The sound [h] is not pronounced in this dialect, so there is no difference here between
the initial and final determiners.
13 The Biscayan pattern in the plural seems closer to the Spanish article+demonstrative pat-
tern: hónek mutilok ‘those guys’, with the proximal article -ok instead of a doubled demon-
strative. Since the proximal plural form -ok derives historically from the proximal plural
demonstrative *hauk (rather than the distal *hak, which is the source of the plural article),
the mismatch is actually smaller than seems to be the case at first sight. In any event, final
-ok is still the determiner, unlike the final demonstrative in Spanish.
22 ortiz de urbina
all ‘that tall guy with glasses’. In contrast, the final demonstrative in the Basque
examples is exclusively final, since it is a regular determiner after all.
Quantifiers make up a semantic class sometimes behaving like determiners,
and therefore incompatible with them. Thus, muchos in mis muchos amigos
‘my many friends’ is compatible with a determiner and, therefore, probably not
a determiner itself, while bastantes ‘quite a few, several’ cannot cooccur with
the genitive determiner, as shown in *mis bastantes amigos ‘my several friends’.
There is nothing semantically odd about the latter, since the same meaning can
be rendered in a fully grammatical way as bastantes amigos míos, with an adjec-
tival genitive míos, crucially not a determiner. Strong Basque quantifiers like
guzti ‘all’, den ‘all’, bakoitz ‘each’ or gehien ‘most’ must in fact cooccur with the
definite determiner, while weak quantifiers like asko ‘many’, batzu(e)k ‘some’,
gutxi ‘few’, zenbait ‘some’ and hainbat ‘some’ cannot coocur with them (see
Etxeberria 2005). Thus, we find lagun guzti*(ak) ‘all friends’ with obligatory plu-
ral determiner but lagun gutxi(*ak) ‘few friends’ with a bare quantifier. From
the point of view of word order, we find again the basic mirror image between
Basque and Spanish, so that the set expression to express thanks contains ini-
tial muchas ‘many’ in Spanish but final asko ‘many’ in Basque (muchas gracias
versus eskerrik asko ‘many thanks’). But, at the same time, we also find the occa-
sional disharmonic orders in OV languages, so that even if most quantifiers are
postnominal in Basque, there are also some prenominal ones. In particular,
zenbait ‘some’ and hainbat ‘some’ precede the noun they quantify over: zen-
bait lagun ‘some friends’.14 These prenominal orders may be syntactically more
complex than the postnominal ones, since the former contain operators which
typically induce displacement in clauses and/or noun phrases: zenbait appears
as zeinbeit in the earliest printed book in Basque, containing the wh-word zein-,
and hainbat contains hain ‘so’, which, as discussed in 2.2.2.2 below also appears
in a displaced position with respect to adjectives it combines with. If extra syn-
tactic positions and processes affect these prenominal quantifiers, then, the
overall mirror-image ordering of quantifiers and nouns in Spanish and Basque
would be more thorough at a more abstract level of analysis.15
14 Zenbait may (rarely) follow the noun Nahiko(a) ‘enough, quite a few’ is more complex,
since depending on the dialect it can precede or follow the noun it quantifies over; it can
also occur as an external quantifier, as if floating out of the NP, taking then the determiner:
a phrase like enough guys can be found across dialects as mutil-a nahiko-a, mutil nahiko-a,
nahiko-a mutil, nahiko-a mutil-a, nahiko mutil-a and nahiko mutil. The form is also mor-
phologically complex and may contain the genitive suffix -ko.
15 For an analysis of Basque NP s, see Artiagoitia (2002).
word order 23
2.2.1.2 PP Dependents of Nouns

Noun dependents and modifiers often take the form of PP s, which will then be
postnominal in Spanish but prenominal in Basque:
(10) etxearen jabea / el propietario de la casa

house.gen owner.det the owner of the house
‘the owner of the house’
(11) etxeko jauna / el señor de la casa

house.gen lord.det the lord of the house
‘the lord of the house’
(12) mugarik gabeko mundua / un mundo sin fronteras

border.part without.gen world.det a world without borders
‘a world without borders’
(13) bihotzeko mina / una pena en el corazón

heart.gen pain.det a pain in the heart
‘a pain in the heart’
(14) bihotz oneko pertsona / una persona de buen corazón

heart good.gen person.det a person of good heart
‘a person of good heart’
The preferred adnominal preposition in Spanish is de, for both argumental and
adjunct modifiers of the noun (as in (10) and (11)), as well as attributive mod-
ifiers (14). Other prepositions are also possible, as in (12) and (13). In all cases,
the PP follows the head noun. As famously put by de Rijk (1993), Basque nouns
are not very ‘hospitable’ to any type of phrasal dependents, which require the
intervention of a genitive marker to be able to attach to the head noun. Argu-
mental, thematic relations are typically marked with the genitive suffix -ren
(as in (10)), while more peripheral adnominals are marked with -ko, added to
an adverbial or a PP as in (12) or directly to the nominal in the case of loca-
tive/temporal relations like (11) or (13) and in attributive relations as in (14).16
Although the status of these ‘genitive’ markers is not uncontroversial (see de
Rijk 1993), positionally they are always placed at the end of the phrase, and
16 De Rijk (1993) claims all cases of apparent bare nominals to which -ko is attached actually
hide PP s, whereas genitive (r)en would attach to NP s.
24 ortiz de urbina
we can equate them with postpositions for word order considerations, forming
PP s just like those headed by de or other prepositions in Spanish. Prepositional
phrases are more ‘harmonic’ with head initial VP s and NP s, as in Spanish, while
postpositional phrases are more typically found with head final phrases, as in
the Basque examples above.
Let us examine now deviations from these patterns. As usual, it is OV lan-
guages like Basque that exhibit some degree of deviation from the expected
pattern: there are no prenominal PP s in Spanish, while postnominal PP s do
exist in Basque, if in a limited way. First of all, dative adnominals constitute
both a gap in the complementation pattern of nominal heads and a possible
anomaly in the expected ordering relationships between head noun and com-
plement. Thus, goal complements of nouns cannot be attached to the head by
using either of the genitive markers discussed above, neither as nominals with
-(r)en or as dative marked nominals with -ko:
(15) a. entzulegoari eskaini / eskertu

audience.dat offer / thank
‘to offer/thank the audience’
b. * entzulegoariko eskaintza / eskerrak

audience.dat.gen offer / thanks
‘an offer/thanks to the audience’
c. entzulegoaren eskaintza/eskerrak
audience.gen offer/thanks
‘the offer/thanks of the audience’
The dative verbal complement in (a) cannot be found as the complement to

the noun in (b), expected if -ko only attaches to adjuncts and datives are com-
plements. But if they are argumental complements, one would expect them
to occur as prenominal elements displaying the -(r)en genitive marker, as in
(c). However, as Kayne (1984) already noted, genitivized nominals cannot be
construed as goal arguments of the modified noun, so that entzulegoaren ‘of
the audience’ in (c) can only be interpreted as agent and/or theme. In order to
express a goal relationship, it is possible to join the two phrases, but then the
most common (although not exclusive) order is head-goal (see Fernández &
Sarasola 2010):
word order 25
(16) a. omenaldia Xalbadorri / Xalbadorri omenaldia

homage Xalbador.dat
‘homage to Xalbador’
b. abisua marinelei
warning sailors.dat
‘a warning to sailors’
Thus, a song like Gracias a la vida ‘Thanks to life’ is usually rendered as Eskerrak
bizitzari, literally ‘Thanks life to’, identical to the order in Spanish and English
in the relevant respect. It is not clear, however, whether the two phrases are
really integrated into a single expression or act with at least relative indepen-
dence.17 Certainly, in the rare cases where the goal can be expressed with a
dative, it occupies the pre-head position expected for an OV language. Com-
pare for instance the previous example eskerrak bizitzari ‘thanks to life’ with
an expression like berari esker ‘thanks to him’, where esker ‘thank’ has lost its
nominal status and is used as a postposition requiring a dative complement. In
any event, the possibility that the examples in (16) involve complements which
are not really fully integrated with the head means the variation in word order,
while still a piece of data that must be explained, is not a counterexample for
the right-headed nature of noun+adnominal pairs.
Real counterexamples would be expressions of the type N+XP-(r)en/-ko,
where the genitive bearing nominal follows the head noun, and these do occur;
however, most are old and found in lexicalized expressions of legal or religious
origin where Romance influence is expected, as in aita pontekoa ‘godfather’
(lit. ‘father of the [baptismal] font’), aita besoetakoa ‘godfather’ (lit. ‘father of
the arms’), aita zerukoa ‘father of heaven’, aingeru guardakoa ‘guardian angel’
(Span. ángel de la guarda), aita gurea ‘our father’ (as in the prayer, Span. Padre
nuestro), etc. Similarly, corresponding to Spanish last names with a toponym
preceded by de, earlier texts or forms fixed in songs often give sequences like
Peru Abendañoko, Span. Pedro de Abendaño, Matxalen Busturiko ‘Magdalen of
Busturia’; these, however, are not productive at present.
While complement and adjunct PP s are prenominal in Basque and post-
nominal in Spanish, their position relative to each other is more flexible. The
most neutral ordering is the expected one, namely adjunct-complement-head in
17 See Berro & Fernández (to appear). They show that these constructions may be found in
titles or as complements of a predicative head, but not in the normal nominal functions
such as subject, etc.
26 ortiz de urbina
Basque and its mirror image head-complement-adjunct in Spanish. Thus, in the

absence of any contextual information, a phrase with a picture noun head like
Spanish el retrato de Pedro de Miguel or Basque Mikelen Peruren irudia is inter-
preted as having Pedro/Peru as the person being portrayed and Miguel/Mikel as
owner or portrayer (or any of the contextually relevant interpretations of gen-
itive nouns). However, where function recoverability is contextually possible,
other orders are also acceptable, if perhaps more marked.
2.2.1.3 Attributive Adjectives

After examining PP (including genitive phrase) modification of nouns and
determiners, let us turn now to the other major noun modifiers, namely, adjec-
tives; relative clauses will be dealt with in the following subsection. As Dryer
(1992) showed, adjectives do not enter in correlation pairs with nouns, whereas
relative clauses do. Thus, the postnominal position of adjectives in Basque,
identical to the position of restrictive adjectives in Spanish, should not be
‘unexpected’ from a typological perspective. If in Roman Aquitanian inscrip-
tions like ummesahar ‘old child’, zahar ‘old’ is working as an adjective, as seems
likely, the postnominal position for adjectives would be quite old, certainly not
the result of recent convergence with Romance. Such order is found in N/Adj
compounds too, in words like abelgorri ‘bovine cattle’ (lit. red animals), zorion
‘congratulation’ (lit. good bird), etc.
Basque adjectives are much more robustly postnominal than Spanish ones,
which, depending on the type and with semantic differences, may occur in the
prenominal position: both especial interés and interés especial ‘special inter-
est’ are possible, while only interes berezia (lit. interest special) is in Basque.
In fact, according to Demonte (1999), only relational adjectives and adjecti-
val perfective participles are found exclusively in a postnominal position (la
distribución lingüística ‘the linguistic distribution’ but *la lingüística distribu-
ción, or la antena reparada ‘the repaired antenna’ but *la reparada antena).
In contrast, very few adjectives can occur preposed in Basque (see Trask 2003:
138–139 for a list); most prominently, adjectives of origin with the suffix -(t)ar
(bizkaitar abizena ‘Biscayne last name’, frantziar jendea ‘French people’, etc.) or
adjectives formed with the suffix -dun (akin to English -ed, as bizardun gizona
‘bearded man’, euskaldun jendea ‘Basque (language speaker) people’). Even with
these, the postposed position is far more common.18 The fact that preposabil-
18 Euskaltzaindia (1985: 243) points out that preposed adjectives can be found in literature,
giving examples like justu manamenduak ‘fair commands’, venial faltak ‘venial sins’ or
enganoso mundua ‘deceitful world’. The type of adjective and expression, hardly vernacu-
lar, may indicate this is/was a stylistic device for elevated register.
word order 27
ity hinges on these specific suffixes, rather than on a general semantic type,
for instance, indicates the positional properties may stem from properties of
the affixes themselves. Be that as it may, the restrictive versus non-restrictive
interpretations obtained in well-known postnominal and prenominal order-
ings of adjectives in Spanish phrases like nieve blanca/blanca nieve ‘white snow’
or incas valientes/valientes incas ‘bold incas’ does not have any structural cor-
relate in Basque, where both interpretations can be associated with the basic,
postnominal position of the adjective: lagun maiteak ‘dear friends’ receives
an unmarked non-restrictive interpretation (probably for pragmatic reasons),
while lagun minak ‘close friends’ receives a restrictive interpretation. We will
briefly dwell on the three types of adjectives which may be preposed in Basque
in the remainder of this subsection.
Beginning with -(t)ar derived items, de Rijk (2008: 127) explicitly treats them
as nouns: “Nouns derived with this suffix freely combine with other animate
nouns, acting as a first or as a second component of a bipartite compound”,
exemplifying with erromatar soldadua/soldadu erromatarra ‘the Roman sol-
dier’. Correspondingly, he translates mundutar ‘earthly’ as earthling, zerutar
‘celestial, heavenly’ as ‘celestial being’ or ezkertiar as ‘leftist, left-winger’. But
at least some -(t)ar derivatives seem to be used as bona fide adjectives, shar-
ing nonetheless the positional properties of better-behaved items: soñu eztitsu
ta zerutarra ‘a sweet and celestial tune’ (Bilbao, Ipuin Barreka 158), zerutar ta
jainkotiar gauzetaz ‘on celestial and devout things’ (Gazt MusIx 166).19 Even
if most -(t)ar derivatives were actually nouns and we were dealing with com-
pounds, however, we would probably be transferring the problem to morphol-
ogy: erromatar soldadua ‘Roman soldier’ may be equivalent to Erroma soldadua
‘Rome soldier’, a right-headed compound or an apposition, but it is more diffi-
cult to analyze soldadu erromatarra, the alternative order, as headed by a noun
erromatarra: this expression corresponds to Spanish el soldado romano, rather
than something like appositive el romano soldado. It is of course possible that
some of these derived words have become adjectives, but that only nouns can
be found in prenominal position, so the basic line of de Rijk’s analysis may per-
haps still be maintained.
Basque grammars (Euskaltzaindia 1985: 244; Trask 2003: 138) often mention
that loaned adjectives of origin like frantses ‘French’ share with the native
19 These are literary uses, where the derived item has a qualifying function; the preposed
order here may be taking a free ride on the preposing capacity of -(t)ar forms, produc-
ing a pattern similar to preposed qualifying adjectives in Spanish. See below for a short
discussion on qualifying and classifying uses of preposed adjectives.
28 ortiz de urbina
adjectives built on -(t)ar like frantziar the ability to appear preposed, so that
both frantses/frantziar mutilak ‘French youths’ are acceptable, as well as post-
posed mutil frantsesak/frantziarrak. But, leaving aside the preference for post-
posed orders, this is far from a general fact: many adjectives do not sound
felicitous in prenominal position (??aleman/*suizo/*espainol mutilak ‘German/
Swiss/Spanish youths’), and frantses is by far the most common adjective of
this group. Like nafar ‘Navarrese’ (not a loanword), these are found in some
lexicalized expressions like frantses belar ‘alfalfa’ (lit. French grass), nafar hiz-
kuntza ‘Navarrese language’ (i.e., Basque), or nafar-ilar ‘Navarrese pea, gar-
banzo bean’,20 where it is not clear the first element is always an adjective,
rather than a noun.21 Certainly, euskal ‘Basque’ as in euskal dantzak ‘Basque
dances’ is a nominal bound form of euskara ‘Basque language’, and as such can-
not even be found following the noun.22
Turning now to the case of words formed with the suffix -dun, the standard
analysis identifies it as a reduced form of the finite relative verb duen ‘that has
…’, so that a form like euskaldun jendea ‘Basque (speaking) people’ would actu-
ally be something like ‘people who have the Basque language’. The prenominal
order would follow directly from the fact that relative clauses are prenomi-
nal in Basque. Postnominal orders would then have to be explained, probably
as analogy with normal adjectives. As de Rijk (2008:503) points out, however,
there is little evidence for a finite form doubling as a suffix, at least in Basque,
while there is solid evidence for verbs as second elements of exocentric com-
pounds being grammaticalized as derivational suffixes. This would be the case
of -gin ‘-er’ (<egin ‘make, do’), -garri ‘-able’ (<ekarri ‘bring’) and, in this case
-dun ‘-ed’ (<*edun ‘have’). If this is so, then, the ability for words bearing this
suffix to appear prenominally must now be explained, since -garri adjectives,
for instance, do not share it.23
20 The -ar sequence in nafar ‘Navarrese’ belongs to the stem, although its similarity with the
suffix -ar may also be connected to its preposability.
21 Perez Gaztelu et al. (2004: 158) explicitly analyze these as exocentric compounds, although
the categorial identification of the first element is not questioned.
22 That is, something like *dantza euskarak or *kirol euskarak ‘Basque sports’ does not seem
to be possible, although Trask (2003: 139) points out that Eastern varieties admit expres-
sions like liburu euskara bat ‘a Basque [language] book’, specifically referring to the lan-
guage.
23 If the widespread analysis of this suffix as akin to the relative finite form reflects not
just an analysis by linguists but a reanalysis by speakers, we would expect prenominal
forms to be less common in earlier texts. As far as I know, this has not been researched so
far.
word order 29
Notice that it is (some) mostly classifying ‘adjectives’, as opposed to quali-

fying adjectives, that can be preposed in Basque, while classifying adjectives
are obligatorily postnominal in Spanish. Qualifying adjectives like zentzudun
‘sensible’ or bihoztun ‘(kind)hearted’ and, perhaps, zerutar ‘celestial’ above
can be modified by degree words like oso ‘very’ or nahiko ‘rather’ in predica-
tive position, but in attributive positions they must be postnominal: oso jende
zentzuduna/??oso zentzudun jendea ‘very sensible people’, nahiko jende bihoz-
tuna/*nahiko bihoztun jendea ‘rather kindhearted people’. This defectiveness
is shared by the occasional qualifying adjectives which may occur prenom-
inally, like gazte ‘young’ in gazte jendea ‘young people’; modification is not
possible (*oso gazte jendea/oso jende gaztea). These facts square well with the
modified version of de Rijk’s (2008) analysis mentioned above, namely, that
the prenominal and postnominal positions may in fact be occupied by words
with different categories: prenominal nouns and postnominal adjectives. If
that is the case, a sequence like *oso gazte jendea would be ungrammatical
because there would be no adjective that the degree word could combine
with.
2.2.1.4 Relative Clauses

Turning now to relative clauses, as indicated above, their position does corre-
late with the OV/VO order: prenominal relatives are typically found in OV lan-
guages like Basque, whereas a VO language like Spanish displays postnominal
relative clauses. This does not mean that prenominal relatives are predominant
in OV languages, since in Dryer’s (2013) language sample more OV languages
display postnominal than prenominal relatives, but, rather, that RelN orders
are almost exclusively found in OV languages.24 Cinque (2013: 130) proposes
to link OV rigidity with relative clause placement, claiming that OV languages
that allow for postverbal subordinate clauses, i.e., non-rigid OV languages like
Basque, also allow for postnominal relatives.25 This is in fact borne out by the
Basque data: while prenominal relatives are without doubt the major type,
postnominal types are also occasionally encountered, albeit far less frequently.
According to Oyharçabal (2003: 765), they were more common in earlier texts,
24 While roughly similar numbers of OV languages are NRel or RelN (with varying propor-
tions depending on the sample used), VO languages with RelN are exceedingly rare: 1 out
of 61 in Dryer’s 1992 sample and 5 out of 421 in Dryer (2013). Mandarin Chinese (a language
with other OV features such as prenominal genitives) is the clearest example.
25 Cinque actually classifies Basque as a rigid OV language; the fact that it is not turns out to
be actually helpful for his analysis, in view of the data discussed in the text.
30 ortiz de urbina
and at present are still occasionally found in eastern dialects. The following
examples illustrate the more standard prenominal relative (a) and the post-
nominal substandard one (b):
(17) a. [Zerbait dakite-n] eskualdun zaharr-ak alde dituzte.

something know-comp Basque old-det.pl favor have
‘They have the support of old Basques who know something.’
b. Eskualdun zahar [zerbait dakite-n-ak] alde dituzte.

Basque old something know-comp-det.pl favor have
‘They have the support of old Basques who know something.’
More frequent in older texts, but following more closely Romance relative
structure and rarely crossing the barrier into the spoken language, is a postnom-
inal non-restrictive relativization strategy with overt relative pronouns (i.e.,
special uses of otherwise interrogative pronouns) as in the following example
from Leizarraga’s 1571 translation of the New Testament:
(18) Iaincoa, ceinec iustificaturen baitu circoncisionea fedetic

God who.erg justify.fut comp.aux circumcision faith.from
‘God, who will justify the circumcision by faith’
The relative clause is introduced by the pronoun zein ‘which’, identical to the
interrogative, and is marked with the complementizer bait-. In contrast, (17b)
displays the typical gap strategy of Basque relatives, varying only in the left
position of the antecedent; since case marking is attached to the end of the
constituent, the absolutive plural determiner -ak that is attached to the postce-
dent in a standard Basque relative like (17a) appears in (17b) at the end of the
relative, attached to the complementizer -n. The Romance type of postnomi-
nal relative is, however, less interesting from the point of view of ordering, since
its position may be accounted for by the fact that it is a non-restrictive paren-
thetical element, rather than by the type of phrase internal relations we are
examining here. More interesting is the type of relative clause which appears
to the right of and in an appositive relation with the antecedent, as in (19), from
de Rijk (2006: 488):
(19) Zuk ez al duzu lehengusu bat [filarmonika jotzen du-en-a]?

you not Q have cousin one accordion play aux-comp-det
‘Don’t you have a cousin who plays the accordion?’
word order 31
The appositive relative (Oyharçabal 2003: 802–807, de Rijk 2006: 487–490)

is a headless relative clause occurring in apposition to the right of the ante-
cedent. Since there are two nominal phrases in the previous example, two
determiners are attached, each at the end of the phrase, the indefinite bat ‘one’
and the matching ‘definite’ -a. This contrasts with (17), where the antecedent
is not case marked and a single determiner is attached at the end of the sin-
gle noun+relative constituent. Most of these postnominal structures are non-
restrictive appositions, and, as indicated above, perhaps not directly relevant
to the current discussion. They are intonationally and (often) graphically sep-
arated from the first nominal although linked with it by usually replicating
its case (absolutive in (19) above). However, there are also some examples of
restrictive appositions, like the previous one or the following (Oyharçabal 2003:
806):26
(20) Eta ez dut uste baden emazterik ezetz ihardetsiko

and not aux think is.comp wife.part no reply.fut
didanik.
aux.comp.part
‘And I do not think there is any wife who will say ‘no’.’
Oyharçabal (2003: 804) states appositive relatives are necessarily non-restric-

tive. The example in the text, actually taken from his article, is then analyzed
there as an instance of an extraposed relative, perhaps because even if nothing
intervenes between the antecedent and the relative clause, the relative clause
occupies a position at the end of the clause. Restricting our attention to the
few examples of restrictive appositives of this type, the extraposition analy-
sis seems more appropriate than a restrictive apposition analysis: in principle,
one would expect the latter type of structure to conform to the specific+general
order found in OV appositives, while here the specification introduced by the
relative clause follows the phrase introducing the specified noun in Basque.
If these were indeed appositions, they would be quite atypical. One possible
objection to the extraposition approach is that if the relative clause is indeed
26 A very similar sentence is given in de Rijk (2008: 489) as an example of restrictive appos-
itive, although in fact a negated existential verb intervenes here between the antecedent
and the relative:
(i) Harginik ez da edaten ez duenik
stonecutter.part not is drink not aux.comp.part
‘There is no stonecutter who does not drink’
As discussed in the following text, these could also be analyzed as extraposed relatives.
32 ortiz de urbina
extraposed, we would expect restrictive relative clauses in general to undergo

the process, so the scarcity of the ‘appositive’ restrictive relatives under discus-
sion would be surprising. In favor of the approach, however, is the related fact
that relative extraposition is generally possible from indefinite antecedents,27
as are the examples offered in the literature. Relative clauses and adjectives are
similar in this respect, since extraposition of an attributive adjective in Basque
is possible only over nouns with the indefinite determiner bat:
(21) liburu bat ederra / *liburua ederra

book one beautiful book.det beautiful
‘one beautiful book/the beautiful book’
It is difficult to analyze these as appositions, since they involve different types

of phrases. The functional parallelism between the adjective and the relative
clause points at shared semantic constraints in the two constructions.
2.2.2 Adjective Phrases

Continuing with an examination of word order relations within the major lexi-
cal phrases, we turn now to AdjP. We will first describe their structure in 2.2.2.1,
turning then in 2.2.2.2 to the attributive uses of complex adjective phrases
(i.e., those containing not only the adjective, already discussed in the previ-
ous section, but the adjective accompanied by a degree word and/or depen-
dents), looking for differences between attributive and predicative Adjective
Phrases.
2.2.2.1 Degree Words, Complements and Adjuncts

Spanish adjectives take PP complements and adjuncts in the expected order:
orgulloso de su obra sin motivo ‘proud of his/her work without reason’ is more
neutral than orgulloso sin motivo de su obra ‘proud without reason of his/her
work’. If the default P for Spanish noun complements is de, adjective comple-
ments display more variability, so we also find fiel a ‘loyal to’, acorde con ‘in
agreement with’ or apto para ‘apt for’. The smaller degree of grammaticalization
of adjective complements has some consequence for word order in Basque.
Noun complements and adjuncts cannot be directly joined to the head noun,
and a genitive marker is required; this genitive marker is a postposition and
the genitive PP must be prenominal, as discussed in 2.2.1.2. On the other hand,
27 Compare with English: the??/a man came to see me who I had met the day before.
word order 33
PP complements and adjuncts to adjectives are far less common28 and can be
attached directly to the Adjective head without any other intervening gram-
matical element which may rule its position relative to the Adjective. In con-
structing these complex AdjPs, while the position of adjective complements
and adjuncts with respect to their head is quite flexible, there is a preference in
more neutral contexts for a head final order:
(22) a. hizkuntzari leial (??leial hizkuntzari) irautea

language.dat loyal remain
‘to remain loyal to the language’
b. Zu behar zaitut nitaz seguru. (?seguru nitaz)

you need aux I.about sure
‘It is you that I need sure about me.’
Similarly, in metalinguistic contexts such as citations or listings, the neutral

order is complement-head, as in esaten erraza ‘easy to say’, horretan trebea
‘skillful at that’ or edozertarako prest ‘ready for anything’. Therefore, the neu-
tral positioning of the core lexical elements of AdjPs correlates with the OV/VO
or head-last/head-first basic nature of Basque and Spanish.
Turning to more grammatical components of the adjective phrase, degree
words are assumed to take the adjective as complement, and, hence, would be
expected to pattern differently in head first and head last languages. A look at
Dryer and Haspelmath’s WALS online data combining degree word/adjective
and object/verb ordering, however, shows no significant correlation, with, in
fact, a slight tendency for the opposite. Excluding languages without dominant
order for either pair, their sample gives 107 VO languages with Adj-Deg order,
and 88 with Deg-Adj order, while, for OV languages, Deg-Adj orders outnumber
Adj-Deg by 128 to 72. It is then perhaps not surprising that almost all degree
words precede the adjective in Basque: oso/ nahiko/ hain/ arras/ txit/ sobera/
erabat handia ‘very/ rather/ so/ completely/ very/ exceedingly/ totally big’; sim-
ilarly, the polarity degree work batere ‘at all’ precedes the adjective. Only samar
‘rather’ systematically follows the adjective, as in handi samarra ‘rather big’.29
Some of the degree words listed can occasionally be found following the adjec-
28 As Goenaga (in progress) points out, there are relatively few complement-taking adjec-
tives in Basque, especially, as will be seen below, in attributive contexts.
29 Comparatives and superlatives are realized as inflectional suffixes on the adjective (han-
diago ‘bigger’, handien ‘biggest’), so their ordering is not considered here.
34 ortiz de urbina
tive, but, unlike samar ‘rather’, they also follow the determiner, so that they
seem to be extraposed: compare common handi samarra and marked handia
erabat.
Manner adverbs (biziki, zeharo, izugarri) and PP s (benetan ‘really, truly’)
which modify adjectives receive degree interpretations; they form a relatively
open class, and occupy a preadjectival position; in similar Spanish examples,
on the other hand, degree adverbs precede (just like regular degree words)
while PP s follow, just as regular PP s without degree interpretation: tremen-
damente aburrido ‘terribly boring’, but aburrido en exceso ‘boring in excess’.
As Bosque (1999: 220–223) points out, however, both of them function gram-
matically like degree expressions and are therefore incompatible with regular
degree words: *tremendamente muy aburrido ‘terribly very boring’, *muy abu-
rrido en exceso ‘very boring in excess’. Similarly in Basque *biziki oso ederra
‘extremely very beautiful’, *oso biziki ederra ‘very extremely beautiful’ (with
very as degree for beautiful), *benetan oso ederra ‘truly very beautiful’,30 *oso
benetan ederra ‘very truly beautiful’ (again, very connected with the adjective).
The quantifier asko ‘many’ (hitz asko ‘many words’) is occasionally used as a
degree word with adverbs and adjectives. In that case, it still occurs predomi-
nantly in final position, as seen with the adverbial examples maiz/laster/pozik
asko ‘very often/soon/happy’. The use of asko as a degree word with adjec-
tives, sparsely attested in earlier texts in spite of the fairly common expression
segur asko ‘most likely, lit. very certain’, is currently found in central dialects
(according to Trask 2003: 141 and the OEH entry for asko),31 as in eder askoa
da ‘it is very beautiful’, where askoa bears the agreement marker as a pred-
icative element. We will return to its attributive use in the following sec-
tion.
The extension of asko from its original function as a noun quantifier to a
degree word for gradable adjectives can be related to the ‘across the board’
nature of some Basque quantifiers/degree words in terms of their categorial
selectional properties. As de Rijk (2008: 242) points out, especially in north-
ern dialects words like oso, anitz, biziki, sobera ‘very’ can modify across nouns,
verbs, adjectives and adverbs: oso adiskide ‘very (much a) friend’, oso poztu ‘to
rejoice very much’, oso erraza ‘very easy’, oso errazki ‘very easily’. Some Span-
ish modifiers are also quite flexible in terms of the category of the lexical
element they modify: bastante amigo ‘rather friend’, bastante rápida(mente)
30 Benetan oso ederra is acceptable with a non-degree interpretation of benetan more similar
to ‘in truth’ than to ‘truly’, the interpretation relevant in this paragraph.
31 The use of askoa described here is specifically classified as Gipuzkoan in the OEH.
word order 35
‘rather quick(ly)’, fumar bastante ‘smoke quite a lot’. Muy also extends to nouns,
targeting the prototypical qualities associated with the entity denoted by the
noun (muy amigo, muy gallo) but not to verbs, where mucho is used instead
( fumar mucho).
The Basque quantifier asko ‘many’ has developed in different directions in
terms of the distribution with categories other than nouns with respect to the
equivalent Spanish quantifier mucho/a(s) ‘many’. On one hand, the extension
to adjectives described above (eder askoa ‘very beautiful’) is not paralleled in
Spanish, where mucho cannot occur with adjectives; on the other hand, in
southern dialects, asko ‘many, much’ has virtually displaced oso ‘very’ as a ver-
bal modifier, presumably helped by the fact that its Spanish equivalent muy
‘very’ does not modify verbs, while mucho ‘many, much’ does.
2.2.2.2 Complex Adjective Phrases in Attributive Position

The constraint against attributive adjective phrases where the adjective is sep-
arated from the noun by any complement or adjunct (a proud person vs. *a
proud of his achievements person, with *[Adj PP] N sequencing), does not apply
in Spanish, trivially so since adjective phrases follow the noun and their head
precedes any complements or adjuncts they may have (N [Adj PP], una persona
orgullosa de sus logros). Basque presents an intermediate situation in terms of
ordering, since adjective complements/adjuncts may precede their head, pro-
ducing in principle orders such as N [PP Adj]. As the following examples show,
the order is unacceptable:
(23) a. ??egitasmo egiten zaila

project make hard
‘a project hard to carry out’
b. * gizon horretan trebea

man that.in skilled
‘a man skilled at that’
There is a slight improvement in acceptability when the complex adjective is

combined in restrictive apposition with the noun. The improvement is more
drastic when the noun bears the indefinite determiner bat ‘a, one’:32
32 Non-restrictive apposition is of course acceptable regardless of the type of determiner the

head noun bears.
36 ortiz de urbina
(24) a. ??egitasmoa(k) egiten zaila(k)

project.det do hard.det
‘project(s) hard to carry out’
b. ??gizona(k) horretan trebea(k)

man(pl) that.in skillful(pl)
‘a man (men) skillful at that’
(25) a. egitasmo bat egiten zaila

project a do hard
‘a project hard to carry out’
b. gizon bat horretan trebea

man a that.in skillful
‘a man skillful at that’
The increased acceptability in forming restrictive appositions of complex

adjectival phrases with the indefinite determiner, then, parallels the possibil-
ity of simple adjective postposing with that determiner discussed at the end of
section 2.2.1.4.33
We have described attributive AdjP complexity arising from the presence of
complements and adjuncts. Let’s now focus on the type of complexity added
by the presence of degree words like those described for predicative adjec-
tives in the previous section. Since degree words usually precede adjectives,
and adjectives follow nouns, the expected configuration is [N [Deg Adj] Det],
with a postnominal adjective phrase and the determiner occupying the final
position in the NP. Leaving aside the placement of determiners, the previous
order is standard in Spanish: una casa muy grande ‘a very big house’. On the
other hand, the expected Basque order, etxe oso handi bat, while not ungram-
matical, is uncommon and, according to de Rijk (2008: 243), only found under
heavy emphasis. To a certain extent, then, the presence of intervening material
between noun and adjective prevents the latter from appearing with other ele-
ments which would be perfectly normal and unmarked in a predicative phrase.
In the case of degree words, however, there exists an alternative location for the
latter to the left of the noun phrase. The following phrases provide examples of
such ‘extracted’ degree words:
33 The similarity mentioned there with respect to relative clause extraposition with indefi-
nite head nouns in English is intriguing in that there is no apparent extraposition involved
in the Basque examples with simple or complex appositive adjectives.
word order 37
(26) a. oso / nahiko / hain etxe handi bat

very / rather / so house big a
‘a very/rather/so big house’
b. oso herriko festa nazkagarriak

very village.of festival disgusting
‘very disgusting village festivals’
c. (ez dira) batere festa interesgarriak

not are at.all festival interesting
‘they are not festivals interesting at all’
The second example shows that the position is not only to the left of the noun
but, more specifically, at the left periphery of the nominal complex. The last
example shows the similar behavior of the negative polarity degree word batere
‘at all’.34 Moreover, at least oso ‘very’ can be extraposed to the right, as in etxe
handia oso ‘a very big house’ (but not *etxe handia hain/nahiko ‘so big a house/a
rather big house’).
There are also restrictions on the cooccurrence of degree words on the adjec-
tive on one hand and the type of noun determiner on the other (in particular,
its definiteness/specificity). As noted in Saltarelli (1988), Euskaltzaindia (1985)
and de Rijk (2008), degree words like oso/nahiko are often at odds with definite
determiners and demonstratives. This is shared with Spanish:
(27) a. oso etxe handi-a / ??la casa muy grande

very house big-the the house very big
b. oso etxe handi bat / una casa muy grande

very house big a a house very big
c. ??oso etxe handi hori / ??esa casa muy grande

hat that
d. hain etxe handi hori / esa casa tan grande

so that that so
34 These preposed degree words resemble similar constructions in English like so/how/too/
that beautiful a house, where the degree element seems to pied-pipe the adjective phrase
to a pre-determiner position (in fact, always indefinite a).
38 ortiz de urbina
The difference between the two languages in (a) is smaller than it appears
to be when one bears in mind that the Basque phrase is interpreted as indefi-
nite (but specific in appropriate contexts) rather than definite. As Pastor (2011)
shows, there are also interactions between some degree words and the indefi-
nite determiner in Spanish, since the non-specific interpretation of indefinite
determiners one can find in a phrase like una casa grande ‘a big house’ becomes
less difficult to obtain in (b).
We mentioned in 2.2.2.1 that some degree words follow the adjective, so
they do not intervene between the noun and the adjective if the latter is not
accompanied by any other complement or adjunct. The expected order is then
actually unmarked (N [Adj Deg] Det) and no extraction or preposing can be
found, so we only have etxe handi samar bat ‘a rather big house’, but no *samar
etxe handi bat. One of the degree words that follow the adjective is asko ‘rather,
fairly’ in Gipuzkoan. When the adjective+asko phrase is used attributively, the
degree word ends up in final position of the nominal phrase and it will then be
followed by the case or agreement markers for the latter. Thus, in these dialects,
a phrase like telebisio merke asko ‘many cheap TV s’ with the standard quanti-
fier asko contrasts with telebisio merke askoa ‘a very cheap TV’. The indefinite
quantifier asko in the first phrase occurs without a determiner, as is general
when a noun is quantified by an indefinite quantifier. However, askoa in the
second phrase acts as a degree word to the adjective. This means that the noun
in that phrase is not quantified, so a determiner -a is attached at the end, cor-
responding to the function of the whole nominal phrase in the clause. In the
following examples, this determiner stands for an agreement marker on the
predicate NP in (a) below and for an indefinite determiner in an argumental
position (b,c):
(28) a. Emakume fin askoa zan.

women smart very.det was
‘She was a very smart woman.’ (Albeniz 25)
b. Sotana utzi ta traje dotor askoa jarri zuan.

cassock leave and suit elegant very.det put.on aux
‘He left the cassock and put on a very elegant suit.’ (BBarand 103)
c. Gauza pollit askoak esan ditut nik.

thing nice many.det.pl say aux I
‘I have said many nice things.’ (Iztueta)
word order 39
Notice that in the latter the definite determiner actually has an indefinite
interpretation. In neither case is it possible to use the indefinite determiner
bat even though the same phrase with standard degree word oso ‘very’ would
admit it (oso traje dotore bat/*traje dotore asko bat ‘a very elegant suit’).
Degree words can actually be extracted from the whole nominal expression
and placed in emphatic position before the verb. As de Rijk (2008: 243) indi-
cates, this may be found only with copulative verbs like izan, egon ‘be’ or their
transitive counterparts:
(29) a. Oso da etxe handia. (etxe handia da oso?)

very is house big.det
‘It is a very big house.’
b. hain da etxe handia ze …

such is house big.det that
‘it is such a big house that …’
Definiteness effects are reinforced under focus extraction; compare the ex-
tracted and basic phrasal patterns:
(30) a. Oso da etxe handia. vs. oso etxe handia

very is house big.det
b. ??Oso da etxe handi bat. vs. ?oso etxe handi bat

very is house big a
c. * oso da etxe handi hori vs. ??oso etxe handi hori

very is house big that
‘It is that very big house’
2.3 Order of Major Clausal Constituents

In both Basque and Spanish, word order possibilities are far more numerous
among clausal constituents than inside the types of phrases discussed so far, so
that, as indicated above, characterizations of Basque as SOV or Spanish as SVO
often reflect the former types of facts. Nonetheless, it is customary to abstract
away from word order ‘freedom’ and recognize a neutral, ‘basic’ pattern, along
with some information packaging strategies for different types of emphasis,
topicalization and heaviness. Section 2.3.1 discusses ‘basic’ word order patterns,
40 ortiz de urbina
while 2.3.2 identifies some of the information packaging strategies which can
be seen to be operative in the two languages. The difference between main and
embedded orders is briefly mentioned in section 2.3.3, and weight and heavi-
ness are dealt with in 2.3.4.
2.3.1 Neutral Orders

S-Adjunct-IO-O-V sequences are considered ‘neutral’ or ‘purely grammatical’
for Navarrese-Labourdin in Lafitte (1944: 46), and de Rijk (1969) and Villasante
(1980) call the SOV order ‘neutral’ and ‘aseptic’. None of these designations are
meant to indicate that SOV orders are statistically more common in Basque
transitive clauses with lexical arguments; in fact they are not. Hidalgo (1995a)
presents detailed counts which, leaving the works of grammarians aside, sys-
tematically show SVO orders to outnumber SOV ones. Thus, the analysis of
folk tale transcriptions from western Biscay yields 37,5 % SVO and 27,9% SOV
orders.35 This is in clear contrast with the SVO order of corresponding sen-
tences in Spanish, which, although not as pervasive as in English, still accounts
for some 85% of these clauses, depending on the counts (Siewierska 1998).
Nonetheless, the basic order is helpful to understand why, in Basque, when
there are constraints on the relative order between subject and object, it is
always this sequence that is required, or why when the position of the verb
is restricted (outside focalization structures), it is a verb final order that is
required (Ortiz de Urbina 2003: 448–451). Thus, for instance, multiple wh-word
questions display nork zer ‘who what’ sequencing (a), and relative clauses, as
mentioned in section 2.2.1.4, are verb final (b):
(31) a. Eta zure prestutasunaz … nork zer erranen du?

and your virtue.on who what say.fut aux
‘And about your virtue … who will say what?’
b. trenari harriak botatzen zizkioten mutilak

train.to stones throw aux.comp boys
‘the boys who threw stones at [lit. to] the train’
35 The other orders are far less common: 11,5% (OSV), 10,6 % (OVS), 8,6 % (VSO) and 3,8 %
(VOS). Notice however that these are raw counts which do not take into account the pres-
ence of topicalizations or focalizations in the samples. SVO orders are also shown to be
more common in the dialect of Ondarroa, as described in Rotaetxe (1978).
word order 41
These preferences have as a consequence that Basque does not display the
types of SV/VS asymmetries in neutral orders which are found in Spanish and
other Romance languages, mostly with unaccusative verbs. The importance of
verb initial basic orders led Contreras (1976) to postulate a basic VSO pattern
in Spanish and has produced an important body of literature which assumes
subjects need not be in Spec of IP in this language or actually correspond to
topics in Clitic Left Dislocation structures when preceding the subject (Alexi-
adou & Anagnostopoulou 1998; Ordoñez & Treviño 1999 among many others).
Thus, indefinite subjects of unaccusative verbs like caer ‘fall’ or morir ‘die’ are
introduced into the conversation more neutrally in the postverbal position (a).
In preverbal position (b), the indefinite phrase seems to receive a marked, con-
trastive topic type of interpretation. With definite subjects, however, the most
neutral position is preverbal (d), rather than postverbal (c):
(32) a. Cayó una roca sobre el pueblo.

fell a stone over the village
‘A stone fell over the village.’
b. Una roca cayó sobre el pueblo.
c. Cayó la roca sobre el pueblo.

the stone
‘The stone fell over the village.’
d. La roca cayó sobre el pueblo.
(33) a. Murió mucha gente en aquella guerra.

died many people in that war
‘Many people died in that war.’
b. Mucha gente murió en aquella guerra.
c. Murió mi abuelo en aquella guerra.

my grandfather
‘My grandfather died in that war.’
d. Mi abuelo murió en aquella guerra.
If preverbal subjects are topicalized given information, it is not surprising that

indefinites prefer a postverbal position. Although, as we will see, postverbal
42 ortiz de urbina
corrective focus strategies can be found in colloquial contemporary Basque,

verb initial patterns are not very prominent in the standard language, perhaps
due to the more ‘basic’ status of verb final patterns and to the fact that preverbal
focalization strategies will often place some material before the verb.36,37
Leaving VS orders behind, the major difference between SVO and SOV orders
is that objects (and adjuncts) follow the verb in the former and precede it in the
latter. Postverbal objects and adjuncts, as indicated above, have been and are
in fact more common than preverbal ones (Michelena 1978; Hidalgo 1995a,b).
At least since Altube (1929), the ‘leftward’ placement of the verb away from the
right periphery of the clause in neutral contexts in Basque has been attributed
to influence from Spanish,38 and it seems that in colloquial registers, especially
among younger (let alone ‘new’) speakers, postverbal constituents are increas-
ing in frequency so that major constituent order is not as different between the
two languages as might be expected. Since more detailed statistic studies have
not been conducted, especially comparisons between word order in the two
languages for the same speakers, we will leave this question aside. Instead, the
discussion will concentrate on the types of focalization and topicalization pro-
cesses which produce many of the observed constituent sequences in the two
languages (see Elordieta 2013 for Basque).
3.3.2 Information Packaging

Basque shows fairly standard ‘left periphery’ V2-like orders in questions (34a)
and focalization (34b) structures (see for example Etxepare & Ortiz de Urbina
36 The postverbal position is obligatory for bare subjects (and objects) in Spanish, another
type of situation that does not arise in Basque, where bare nominals are virtually absent,
mostly restricted to N egin verbs like lan egin ‘work, lit. do work’, amets egin ‘dream, lit. do
dream’, etc.:
i. Cayeron rocas sobre el pueblo vs. *Rocas cayeron sobre el pueblo
fell rocks on the village rocks fell on the village
ii. Murieron soldados jóvenes en esa guerra vs. *Soldados jóvenes murieron en esa
died soldiers young in that war soldiers young died in that
guerra
war
iii. *Cigüeñas anidan aquí vs. Aquí anidan cigüeñas ‘Storks nest here’
iv. *Refugiados trabajan aquí vs. Aquí trabajan refugiados ‘Refugees work here’
37 In literary contexts such as novels, direct speech quotations are followed by V-Subject
orders (esan zuen Jonek ‘said John’), even if relatively ‘neutral’ in that neither the quo-
tation nor the subject (or the verb, for that matter) are emphasized. The presence of the
same unexpected order in English may indicate this is a literary stylistic cliché.
38 Michelena (1978) says that, at least in narrative texts, the (X)VO pattern is traditional rather
than innovative.
word order 43
2003 or Elordieta 2013). Further left periphery phenomena affect tensed verbs
in negative clauses, which are attracted to the left. This is more salient where
the verb is complex, since the auxiliary, which usually follows the verb as
expected in an OV language, appears to the left of the main verb, as well as to
the left of the clause (34c), as described in Laka (1994):
(34) a. Noiz erosi zuen Jonek bere kotxea?

when buy aux Jon his car
‘When did Jon buy his car?’
b. Iaz erosi zuen Jonek bere kotxea.

last.year
‘Jon bought his car last year.’
c. Jonek ez zuen bere kotxea iaz erosi.

Jon not aux his car last.year buy
‘Jon did not buy his car last year.’
The immediate preverbal, left peripheral position is general for wh-words, and
quite general in western dialects as well as the standard language in the case
of foci, whether new information, contrastive or corrective. Thus, contrary to
Spanish, the answer to a wh-question will have narrow focus in the preverbal
position of the answer, as shown in (b) above, whereas this new information
will be postverbal, and probably final in Spanish:
(35) a. ¿Quién ha comprado un coche nuevo?

who has bought a car new
‘Who bought a new car?’
b. (Ha comprado un coche nuevo) Juan.

has bought a car new Juan
‘Juan (has bought a new car.)’
(36) a. ¿Quién ha salido?

who has left
b. Ha salido Juan.
has left Juan
‘Juan has left.’
44 ortiz de urbina
Given appropriate structural assumptions, the position of sentence stress

correlates well with the predictions of the Nuclear Stress Rule (in the general
reformulation by Cinque 1993) in both Basque and Spanish, so that sentence
focus correlates fairly well with prosody. In neutral orders, the most deeply
embedded element would be the object in both languages (Zubizarreta 1998,
as well Elordieta 2001 & Arregi 2002 for different dialects of Basque). Elements
displaced to the right and left in Basque would be attached higher in the struc-
ture, so the rightmost element to the left of the verb would be deepest and,
hence stressed, accounting in essence for the preverbal position of stressed
elements but not necessarily for other properties of non-neutral focalized con-
stituents in Basque. In the Lekeitio dialect described by Hualde, Elordieta &
Elordieta (1994), a special focal stress must be present in all sentences, always
to the left of the verb. However, only a sentence with SOV order, with a focally
accented object, can be interpreted as ‘neutral’, i.e., with prosodic prominence
of an element but without any foregrounding interpretation for that element.
This follows if SOV is the basic word order in Basque.39
Spanish also has focalization structures similar to the V2-like one described
for Basque, and also analyzed as involving V/T movement to some left periph-
eral focus or C head. Focus fronting to the specifier of that same head pro-
duces the type of V2 adjacency described for Basque (although not in Southern
Peninsular Spanish according to Jiménez-Fernández 2015, who reviews these
patterns). The Spanish structures, however, are more restrictive in terms of
the pragmatic contexts in which they can be used, so they remain marked,
while the Basque structure is pervasive. Examples (37) through (40) illustrate
some of them, namely, Focus Fronting (Jiménez Fernández 2015; Ortega 2016),
Verum focus-inducing Fronting (Escandell-Vidal & Leonetti 2014), Mirative
Focus (Cruschina 2012; Bianchi et al. 2014; Jiménez-Fernández 2015) and Quan-
tifier Fronting (Quer 2002):
(37) a. El coche quiero.

the car I.want
‘I want the car.’
b. El coche ha vendido Juan, no la moto.

the car has sold Juan not the motorcycle
‘Juan has sold the car, not the motorcycle.’
39 This assumes SOV is a basic pattern in a ‘symmetric’ syntax.

word order 45
(38) a. ¡ Menudo jaleo montó Pedro!

small riot set.up Pedro
‘What a riot Pedro made!’
b. ¡ Bonita faena me has hecho!

nice trick me have done
‘What a nasty trick you’ve played on me!’
(39) ¡ No me lo puedo creer! ¡Tres trozos de tarta se ha comido

not me it can believe three pieces of cake her has eaten
Ángela!
Angela
‘I can’t believe it. Angela ate three pieces of cake!’
(40) [Digo yo que] algo le habrá hecho Pedro a Juan.

say I that something him must.have done Pedro to Juan
‘I guess that Pedro must have done something [wrong] to Juan.’
The fronted elements cannot be interpreted as left dislocated topics, and no

clitic doubling is then found even though an object has been fronted. The Focus
Fronting example in (37) is contrastive in the context discussed in Ortega (2016:
26), where a child is given the choice between two toys. Similar constructions
have been reported in Spanish in contact with Basque (Sainz-Maza 2017). The
Verum-Focus Fronting in (38) would correspond to a positive polarity focal-
ization, according to Escandell-Vidal & Leonetti (2014), with pragmatic ironic
reading. Unlike the type of fronting in (37), there is no contrastive or informa-
tion focus interpretation associated with the focalized elements in (38). Neither
are they information foci, since, as Escandell-Vidal & Leonetti show, a sentence
like (38b) cannot be uttered as an answer to a question requesting information
like ¿Qué he hecho yo? ‘What did I do?’. These sentences seem similar to excla-
matives, with quantificational readings of the key words menudo and bonita.
Similar examples with noun modifiers like Kristoren ‘of Christ’ in Basque also
involve V2 structures:
(41) Kristo(re)n zarata egin zuten atzo auzokoek!

Christ.gen noise make aux yesterday neighbors
‘Yesterday our neighbors made a terrible noise.’
The mirative type of focus in a sentence like (39) would provide unexpected
and surprising new information, in this example perhaps because Angela is
46 ortiz de urbina
known not to like cakes at all or to eat extremely little, or any other contex-
tually defined presupposed knowledge which would make her eating three
slices unexpected. Finally, Quantifier Fronting in (40) does not convey any con-
trastive interpretation for the fronted element. Leonetti & Escandell (2014)
analyze it as an instance of Verum Focus Fronting, with polarity emphasis.
Jiménez-Fernández (2015) reviews the different types of pragmatic situations
where these constructions can be used, as well as some of the syntactic dif-
ferences such as adjacency with the verb, availability in embedded contexts,
incompatibility among different types of fronting etc. As indicated above, these
seem to be relatively marked structures in Spanish, often with overlapping
and unclear features. All of them could (and probably would) be expressed
through the focalization strategy described above for Basque, with verbal adja-
cency, availability in embedded contexts, incompatibility with further focal-
ized material, etc., even though much remains to be done on the fine details in
central and eastern dialects.40 Those that are well-described from the point of
view of prosody, such as the dialect of Lekeitio (Hualde, Elordieta & Elordieta
1994) indicate that, far from being marked sentence types, all sentences must
include a focally accented element in the preverbal position. From this per-
spective, and abstracting away from the actual focalization structures, Basque
can be qualified as a discourse-configurational language in the sense of Kiss
(1998).
There are further focalization strategies in Basque present in the colloquial
language and often branded as resulting from interference with Spanish. Thus,
Etxepare (1998) describes some minor, marked and not fully grammaticalized
constructions which he considers contrastive, but which Irurtzun (2016: 259)
takes as instances of mirative focus. An example is given in (42):
(42) Jonek ardoa ekarri du.

Jon.erg wine.det brought aux
‘John brought the wine.’
Observe that the emphasized constituent is not preverbal, and, according to

Etxepare (1998), would occupy a low A-position. Whereas the eventuality
denoted by the proposition that a statement like (42) could answer would be
conversationally implicated in the standard construction with verb adjacency,
it would be conventionally implicated in this construction.
40 Irurtzun (2016: 244), however, considers the standard, preverbal strategy to be available
across all dialects of the language.
word order 47
A corrective, right peripheral focalization is described in Ortiz de Urbina

(2002), Irurtzun (2016) for Basque, and Ortega (2016) for Spanish, both ana-
lyzed as involving focalization to the left periphery followed by remnant move-
ment of the TP to its left, resulting in the apparently right peripheral position
of the focalized constituent:
(43) a. Ha comprado la revista Pedro (no María).

has bought the magazine Pedro not María
‘Pedro bought the magazine, not María.’
b. Aldizkaria erosi du Peruk (ez Mirenek).

magazine bought aux Peru not Miren
‘Peru bought the magazine, not Miren.’
Irurtzun (2016: 255) considers the Basque corrective structure as more con-
trastive and presuppositional than the standard preverbal focalization strategy,
with an import akin to that of cleft constructions. As in Spanish, a previous
proposition is assumed to be negated, that somebody other than Peter bought
the magazine.
Corrective focalization is found in southern dialects, in very similar ways
for the two languages among the same speakers. In contrast, there is another
‘reinforced’ focalization strategy, found in northern dialects and probably not
connected to Romance patterns, which has been considered ‘standard’ (though
dialectal) since the first descriptions. It differs from the regular preverbal strat-
egy in that the focalized constituent precedes not the whole verbal complex,
but only the auxiliary, which has been separated from the lexical verb:
(44) Jonek du ardoa ekarri.

Jon.erg aux wine.det brought
‘John brought (the) wine.’
In this split-verb strategy, the lexical verb need not be final but may follow the
auxiliary or other constituents to its right, probably resulting from whichever
processes are responsible for the apparent dislocation of constituents to the
right, always possible in Basque, as will be discussed below. The split verb strat-
egy, as Irurtzun (2016) points out, contains a focalized element with reinforced
focal features, and Lafitte (1944: 48) claimed it must be rendered in French
with ce que cleft constructions. We can also interpret Altube’s (1929: 15) remark
that the focus preceding the auxiliary acquires ‘a particularly marked relief’ as
pointing in the same direction.
48 ortiz de urbina
Where movement of the verbal material to some left peripheral focus-

related head, that is, the basic focalization strategy, occurs without an overt
element being placed to its left, the result is a V1-like order. This is found in
various ‘verb focalization’ structures across dialects, with a range of emphatic
interpretations including new information focus, contrastive focus or positive
polarity focus (see Euskaltzaindia 1985 and Elordieta & Haddican 2016 for a
review and analyses). Similar orders can be found in Spanish in jussive con-
texts as in (46):
(45) Eta bukatu zuten gizonek beren lana.

and finish aux men their work
‘And the men finished their work.’ (Atx. Ob.: 39)
(46) a. ¡ Prestad todos atención!

loan all attention
‘Pay attention all!’
b. ¡ Que diga él lo que quiera!

that say he the that want
‘Let him say what he wants!’
Turning now to non-focalized material, non-focal constituents can be topi-

calized in both languages, often through left clitic dislocation in the case of
Spanish objects. Basque allows for scrambling of that material. For core argu-
ments, this could be interpreted as left dislocation with covert pro clitics, given
extensive agreement in this language. However, there seem to be no differ-
ences between argument and adjunct scrambling possibilities, even though
only the first ones would be associated with pro (see Duguine 2008 for an ellip-
sis alternative to covert clitics). We can try to see non-focal constituent order
possibilities by using a sentence like (47) or (48) whose focus is not the neutral,
most deeply embedded element:
(47) a. [San Juan egunean / Jonek / sua] egunkari batez piztu zuen.
Saint John day.on Jon fire newspaper a.with light aux
‘John lit the fire on Saint John’s day with a newspaper.’
b. egunkari batez piztu zuen [Jonek/sutea/San Juan egunean]

word order 49
(48) (El día de San Juan) Jon encendió la hoguera (el día de San Juan)
the day of S. Juan Jon lit the bonfire
con un periódico (el día de San Juan).
with a newspaper
‘John lit the fire on Saint John’s day with a newspaper.’
Pre-focal constituents receive a listing intonation in Basque, ending in rising

pitch, and may occur in any order, indicated by the slash in (47a). The same
constituents may occur to the right (47b), with roughly the same sequencing
possibilities as in the case of left scrambling. This may indicate that, on top of
‘basic’ orders, right scrambling is also possible in Basque. A listing intonation is
again found in these cases, without the final rising pitch separating left periph-
eral topics from the focalized constituent.
2.3.3 Heavy Constituents

Although elusive in terms of precise formalization (see for instance Shih &
Grafmiller 2011), the principle of ‘end weight’ is commonly held to play a role
in the relative position of ‘heavy’ and ‘light’ constituents: heavy constituents
tend to be located towards the end of the major constituent they belong to.
Heaviness considerations may play a role in the explanation of some of the
unexpected patterns found in the attributive use of complex adjective phrases
discussed in 2.2.2.2, but in this section we will only briefly touch upon end
weight effects at the clausal level. Moreover, we will concentrate on clauses as
prototypically heavier elements than nominals which could play the same syn-
tactic functions. Given that, as shown by Shih & Grafmiller (2011), sheer word
number can actually be used as an appropriate and sufficient proxy for com-
plexity and weight, we will use short nominals and relatively longer clauses
so that length co-varies with structural complexity in the examples. As will
be shown, heaviness accounts for ordering preferences for some subordinate
clauses, ranging from slightly (dis)preferred options to virtually unacceptable
ones.
In clear contrast to end weight analyses, however, Altube (1929: 18–21) explic-
itly claims that, contrary to Spanish, long, ‘heavy’ elements are placed to the
left in Basque, following linking topics, rather than to the right. The evidence
adduced, however, is not particularly convincing since the long elements in his
examples happen to be either focalized (as in (49), one of Altube’s examples)
or conditional clauses. The first ones would have to be preverbal regardless of
weight, and topic-like protheses are usually placed before apodoses across lan-
guages:
50 ortiz de urbina
(49) Exerita zeuden lagunakanantz joan zan, geldi-geldi, umea.

sitting were friends.towards go aux slow-slow child
‘The child went, slowly, towards the friends that were sitting.’
It seems then that there is no basic asymmetry in this respect between Span-
ish and Basque, contrary to what Altube claimed. If we turn to typically heavy
elements such as clauses (as opposed to normal nominals), it is fairly clear that
they prefer a rightward position in both languages. Given basic SOV and SVO
orders as terms of the comparison, this preference will in fact be also observ-
able in the postverbal position of object and, quite generally, verbal dependent
clauses in Basque. Thus, in the following examples, the (non-focal) sentential
subject of the transitive verb and of the copulative predicate is less common
in the preverbal position (a) than in the postverbal one (b) in both languages,
just as in the English translation:
(50) a. ? Que Juan no te dijera nada me sorprende / es sorprendente.

that John not tell nothing surprises / is surprising
‘That John did not tell you anything surprises me/is surprising.’
b. Me sorprende/es sorprendente que Juan no te dijera nada.
(51) a. [Jonek zuri errua botatzen dizula] nabaritzen/nabaria da.

Jon you.dat fault throw aux.that appear/apparent is
‘It is evident that Jon blames you.’
b. Nabaritzen/nabaria da [Jonek zuri errua botatzen dizula].
Similarly, while ‘inversion’ yes/no questions are possible in Spanish, their

acceptability decreases with subject length, so they are maximally acceptable
with a short subject like eso in (a), and, just as in English, minimally so with a
clausal subject as in (b):
(52) a. ¿Es eso cierto?

‘Is that true?’
b. */??¿Es que ha llegado tarde cierto?
*/??‘Is that he has arrived late true?’
As indicated above, clausal objects also exhibit end-weight behavior in Basque,

where they would be expected to precede the verb in a neutral order. Thus,
Euskaltzaindia (1999: 39) explicitly attributes to the ‘complexity’ of the clausal
word order 51
object the fact that, unlike a nominal object, they can be postverbal even if
neutral. There is, moreover, some evidence that such postverbal clauses in non-
canonical position ‘count’ as preverbal. Thus, in the Lekeitio dialect described
in detail in Hualde at al. (1994), the preverbal element must bear focal stress,
and in the basic SOV pattern, objects can be interpreted as neutral or focal-
ized, always bearing focal stress. Unlike neutral nominal objects, however,
non-emphatic object clauses are not preverbal but postverbal. In the same
way, whereas sentence initial root tensed verbal forms are usually ruled out in
Basque, first person tensed forms like nago ‘I am [of the opinion], I think’, or
daukat ‘I have [the opinion], I think’ can be found in some dialects with clausal
complements to their right and occurring at the beginning of the sentence:
(53) Nago […] adiskideek ez dutela erratearen beharrik.

am friends not have.that say.of need.part
‘I think that friends do not have any need to say [that].’ (HU Zez 17)
(54) Daukat ezen horretara erraztu egiten dela euskararen

have that that.way simplify do.impf aux.that Basque.of
ikastea eta comprenitzea.
learn.nom.det and understand.nom.det
‘I think that in that way learning and understanding Basque becomes eas-
ier.’ (Roman Garmendia, Nire Filosofiaren muina)41
In any event, subject and object clauses in Basque pattern with Spanish subject
clauses in their ability to be located in the right periphery of the clause with a
neutral interpretation.
There are other contexts where end-weight effects on clauses can be checked
in the two languages independently of the verb. Thus, in ditransitive construc-
tions heaviness can be checked relative to each of the two verbal complements,
while a small clause complement can in turn contain a clausal subject heavier
than normal small clause predicates. We will only exemplify with the first type
of situation. In the following Spanish ditransitive examples, the clausal con-
stituent will occupy the final position regardless of its function (DO in (a) and
IO in (b), which also corresponds to the canonic order):
41 This example is in fact consciously built to sound like an Axular-like sentence, i.e, a sen-
tence in 17th century Basque. On top of the spelling, notice the use of ezen as an intro-
ductory completive marker, added to the final complementizer -ela; only the latter is
harmonic with OV languages, another expected difference between Spanish and Basque.
52 ortiz de urbina
(55) a. atribuir al azar [que no le haya tocado la lotería]

attribute to luck that not him/her has touched the lottery
‘attribute to luck that s/he did not win the lottery’
b. atribuir su suerte [a que haya sabido jugar bien sus bazas]

attribute his luck to that has known play well his cards
‘attribute his/her luck to having known how to play his/her hand well’
The order in (55b) corresponds to the canonic DO-IO, and only a very heavily
emphatic DO would be able to appear to the right. There is therefore a differ-
ence between the IO-OD orders with clausal and with nominal IO s, as in the
following examples:
(56) a. atribuir [a que haya sabido jugar bien sus bazas] su suerte
b. atribuir al azar su suerte
While the deviation from the canonical order with the short IO in (56b) pro-
duces at most only a mild, presentational type of emphasis, the same devia-
tion over a clausal IO in (56a) would most likely require a heavy, corrective-
like emphasis, since it involves a deviation from both the unmarked DO-IO
order and the expected light-heavy sequencing. Similar heaviness facts can
be found for Basque, modulo the ‘neutral’ final position of the verb and pre-
ferred focalization patterns. Since tensed clauses in this language can only
be found in absolutive positions, the dative constituent will be exemplified
with a nominalized clause (57a). An independent focal element is used to
prevent the DO/IO from being focalized; this also fronts the verb, produc-
ing an S-V left peripheral pattern superficially similar to the Spanish neutral
order:
(57) a. Jonek leporatzen dio krisialdia [bere garaian gauzak argi

Jon attribute.impf aux crisis its time.at things clear
ez uzteari].
not leave.nom.dat
‘Jon attributes the crisis to not having left things clear at the right
time.’
b. ?Jonek leporatzen dio [bere garaian gauzak argi ez uzteari] krisialdia.

word order 53
(58) a. Jonek leporatzen dio Peruri [urte batean aurrezki guztiak

Jon.erg attribute aux Peru.dat year one.in savings all
xahutu dituela].
use.up aux.that
‘Jon blames Peru for having used up all the savings in one year.’
b. ?Jonek leporatzen dio [… aurrezki guztiak xahutu dituela] Peruri.
Again, the heavy clausal constituent is construed as most neutral when it fol-
lows the lighter nominal one, regardless of grammatical function. In the heavy-
light sequences in (57b) and (58b) above, the presence of a preverbal focus bars
the final, light element from receiving a corrective focus interpretation, and
these sentences are therefore less natural than the corresponding ones in (57a)
and (58a).
3 Final Remarks
As indicated at the outset of this article, the OV and VO labels, taken descrip-
tively as sets of expected orderings of certain (mostly) pairs of constituents,
fit relatively well as general characterization of the major word order profile
of Basque and Spanish, respectively. It is at the phrasal level (NP s, PP s, AdjPs,
etc.) that differences are more marked, but even there confluence is occasion-
ally found. At the clausal level, SOV and SVO stand as grammatically neutral,
pragmatically bleached orders, and, at least in the case of Basque, certainly not
as the most common sequence types. Over these, there are numerous informa-
tion packaging mechanisms which account for the majority of actual orders,
and some of these, old and new, also produce similar orderings of clausal con-
stituents in the two languages. Even though Basque does not have the right
peripheral information focus of Spanish, the latter does have as an option a
preverbal focus akin to the general purpose focalization strategy of Basque,
and clause final corrective focus is also increasingly used in Basque. With vir-
tually universal bilingualism with either Spanish or French and, in the western
area, an important mass of new speakers that make up a good portion of Basque
speakers in heavily populated areas, it would be surprising if a certain degree
of grammatical confluence did not take place. The current grammar of Basque
presents therefore an interesting case for language contact studies. In this con-
text, even though this article only aims at presenting the ground description of
the main data which may be relevant for those studies, it may be worth point-
ing out that in the case of word order, this confluence is found in the apparently
54 ortiz de urbina
increasing use of orders which exploit possibilities, traditional or not, which are
compatible with the Spanish grammar of the same users. As a simple example
of the type of word order developments one can find in contemporary Basque
and which will have to be systematically researched, we can consider example
(59), uttered by a young native speaker of a Biscay variety. A variety of factors,
standard some and substandard others, result in a sentential order of the ele-
ments in this sentence closely aligned with Spanish ¿Podemos repetirlos todos?:
(59) Ahal doguz repetidu danak?

can aux repeat all
‘Can we repeat all?’
The substandard use of semiauxiliary ahal ‘can’ (cf. standard repetidu ahal
doguz), which produces a verbal sequence linearly identical to Spanish and
English poder repetir/can repeat, and the final informational focus as opposed
to a preverbal focalized position of the quantifier combine to produce a more
Spanish-like order of clausal constituents in this example. Probably, the declar-
ative counterpart would still differ greatly from the Spanish structure even for
this type of speaker, so that, in general, the conditions, extent and specific
mechanisms of these accommodation phenomena will have to be examined.
On top of the interest of this line of research for the analysis of Basque, both
from a synchronic and a diachronic perspective, we believe the specific socio-
linguistic situation of Basque vis-à-vis its Romance neighbors can make of this
a potentially fruitful research area for grammar contact studies.
Acknowledgements
This work has been supported by the Spanish Ministry of Economy and Com-
petitiveness (FFI2014–51878-P). I would like to thank Ane Berro and Beatriz
Fernández for multiple comments on this and related issues.
References
Abney, Stephen. 1987. The English Noun Phrase in its Sentential Aspect. Ph.D. disserta-
tion, MIT.
Alexiadou, Artemis & Elena Anagnostopoulou. 1998. “Parametrizing AGR: Word order,
V-movement and EPP-checking”. Natural Language and Linguistic Theory 16: 491–
539.
word order 55
Altube, Seber. 1929. “Erderismos”. Euzkerea 10. [Repr. 1975, Bilbao: Indauchu]
Arregi, Karlos. 2002. Focus on Basque Movements. Ph.D. dissertation, MIT.
Artiagoitia, Xabier. 2002. “The Functional Structure of the Basque Noun Phrase.” In
Xabier Artiagoitia, Patxi Goenaga & Joseba Lakarra (eds.), Erramu Boneta: Festschrift
for Rudolf P.G. de Rijk. Vitoria-Gasteiz: ASJU 73–90.
Bennett, Paul A. 1979. “On Universal 23”. Linguistic Inquiry 10: 510–511.
Bianchi Valentina, Giulio Bocci & Silvio Cruschina. 2014. Focus fronting, unexpected-
ness, and the evaluative dimension. Ms. [Available at http://ling.auf.net/lingbuzz/
002166]
Bosque, Ignacio. 1989. Las categorías gramaticales. Madrid: Síntesis.
Bosque, Ignacio. 1999. “El sintagma adjetival. Modificadores y complementos del adje-
tivo. Adjetivo y participio”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática
descriptiva de la lengua española. Madrid: Espasa. 217–310.
Cinque, Guglielmo. 1993. “A null theory of phrase and compound stress”. Linguistic
Inquiry 24: 239–297.
Cinque, Guglielmo. 2013. Typological Studies. Word Order and Relative Clauses. New
York: Routledge.
Contreras, Heles. 1976. A theory of word order with special reference to Spanish (Vol. 29).
Amsterdam: North-Holland Publishing Company.
Cruschina Silvio. 2012. Discourse-related features and functional projections. New York/
Oxford: Oxford University Press.
De Rijk, Rudolf P.G. 1969. Is Basque an SOV language? Fontes Linguae Vasconum 3: 319–
351.
De Rijk, Rudolf P.G. 1993. Basque Hospitality and the Suffix -ko. In Jose Ignacio Hualde
& Jon Ortiz de Urbina (eds.), Generative Studies in Basque Linguistics. Amsterdam/
Philadelphia. 145–162.
De Rijk, Rudolf P.G. 2008. Standard Basque: A Progressive Grammar. Cambridge, MA:
The MIT Press.
Demonte, Violeta. 1999. “El adjetivo: clases y usos. La posición del adjetivo en el sin-
tagma nominal”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva
de la lengua española. Madrid: Espasa. 129–215.
Dryer, Matthew. 1992. “The Greenbergian Word Order Correlations”. Language 68: 81–
138.
Dryer, Matthew. 2013. “Relationship between the Order of Object and Verb and the
Order of Relative Clause and Noun”. In Matthew Dryer & Martin Haspelmath (eds.),
The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evo-
lutionary Anthropology. [Available online at http://wals.info/chapter/96. Accessed
on 2016-03-02.]
Duguine, Maia. 2008. “Silent Arguments without pro”. In Theresa Biberauer (ed.), The
Limits of Syntactic Variation. Amsterdam: John Benjamins. 311–320.
56 ortiz de urbina
Elordieta, Arantzazu & Bill Haddican. 2016. “Strategies of verb and verb phrase focus
across Basque dialects”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Micropa-
rameters in the Grammar of Basque. Amsterdam/Philadelphia: John Benjamins. 221–
242.
Elordieta, Arantzazu. 2001. Verb movement and constituent permutation in Basque. Ph.D.
dissertation, Holland Institute of Generative Linguistics/Leiden University.
Elordieta, Arantzazu. 2013. “On the Relevance of the Head Parameter in a Mixed OV
Language”. In Theresa Biberauer & Michelle Sheehan (eds.), Theoretical Approaches
to Disharmonic Word Order. Oxford: Oxford University Press. 306–329.
Escandell-Vidal, Victoria & Manuel Leonetti. 2014. “Fronting and irony in Spanish”.
In Andreas Dufter & Alvaro S. Octavio de Toledo (eds.), Left Sentence Peripheries
in Spanish: Diachronic, Variationist and Comparative Perspectives. Amsterdam: John
Benjamins. 309–342. [doi: 10.1075/la.214.16esc]
Etxeberria, Urtzi. 2005. “Contextual Restriction and Quantification in Basque”. In Svet-
lana Vogeleer (ed.), Bare Plurals, Indefinites and Weak-Strong Distinction. Belgian
Journal of Linguistics vol. 19. Amsterdam/Philadelphia: John Benjamins.
Etxepare, Ricardo & Myriam Uribe-Etxebarria. 2008. “On negation and focus in Spanish
and Basque”. In Xabier Artiagoitia Beaskoetxea & J. Lakarra Andrinua (eds.), Gra-
matika jaietan: Patxi Goenagaren omenez. Bilbo/Donostia: UPV/Gipuzkoako Foru
Aldundia. [ASJU supplement LI]. 287–310.
Etxepare, Ricardo & Jon Ortiz de Urbina. 2003. “Focalization”. In Jose Ignacio Hualde
& Jon Ortiz de Urbina (eds.), A Grammar of Basque. Berlin/New York: Mouton. 459–
516.
Etxepare, Ricardo. 1998. “A case for two types of Focus in Basque”. In Elena Benedicto,
Maribel Romero & Satoshi Tomioka (eds.), UMOP 21: Proceedings of the workshop of
Focus. Amherst: GLSA, 65–81.
Euskaltzaindia [Academy of the Basque Language]. 1985. Euskal Gramatika. Lehen
Urratsak I. Iruñea: Euskaltzaindia.
Euskaltzaindia [Academy of the Basque Language]. 1999. Euskal Gramatika. Lehen
Urratsak V. Bilbo: Euskaltzaindia.
Fernández, Beatriz & Ibon Sarasola. 2010. “Marinelei abisua: izen ondoko datibo sin-
tagmak izenburuen sintaxian” [A warning to sailors: postnominal dative phrases in
the sintax of nouns]. Lapurdum, 14:55–75.
Goenaga, Patxi. In progress. Euskal gramatika: egiturak eta osagaiak [Basque Grammar:
Structures and Constituents]. Euskaltzaindia.
Hidalgo, Bittor. 1995a. “Ohar estatistiko garrantzitsuak euskararen hitz ordenaren
inguru. Euskara S.O.V.?”. Fontes Linguae Vasconum 70:401–420.
Hidalgo, Bittor. 1995b. Hitzen ordena euskaraz [Order of Words in Basque]. Ph.D. disser-
tation, University of the Basque Country.
Hualde, Jose Ignacio, Gorka Elordieta & Arantzazu Elordieta. 1994. The Basque Dialect
of Lekeitio. Bilbo: Universidad del País Vasco. [ASJU Supplement XXXIV]
word order 57
Irurtzun, Aritz. 2016. “Strategies for argument and adjunct focalization in Basque”. In
Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the Grammar of
Basque. Amsterdam/Philadelphia: John Benjamins. 243–263.
Jiménez-Fernández, Angel L. 2015. “Towards a typology of focus: subject position and
microvariation at the discourse-syntax interface”. Ampersand 2:48–60.
Kayne, Richard. 1984. “Unambiguous Paths”. In Connectedness and Binary Branching.
Dordrecht: Foris Publications. 129–164.
Kayne, Richard. 2017. “Antisymmetry and Morphology. Prefixes vs. Suffixes”. Ms.
Kiss, Katalin É. (ed.). 1998. Discourse Configurational Languages. New York/Oxford:
Oxford University Press.
Lafitte, Pierres. 1979 [1944]. Grammaire Basque. Navarro-Labourdin Litteraire. Donostia:
Elkar.
Laka, Itziar. 1994. On the Syntax of Negation. New York/London: Garland Publishing Co.
Lehmann, Winfred P. 1975. Proto-Indo-European Syntax. Austin, TX: Univ of Texas Press.
Leonetti Manuel & Victoria Escandell. 2014. “Fronting and verum-focus in Spanish”. In
Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages.
Amsterdam/Philadelphia: John Benjamins. 155–204.
Michelena, Luis & Angel Yrigarai. 1955. “Nombres vascos de persona”. ASJU II: 107–
127.
Michelena, Luis. 1977. “Notas sobre compuestos verbales vascos”. Revista de dialectolo-
gía y tradiciones populares XXXIII: 245–271 [Repr. in Joseba Lakarra & Iñigo Ruiz
Arzalluz. 2011. Luis Michelena. Obras Completas VII. Donostia: Universidad del País
Vasco. [ASJU Supplement LX]
Michelena, Luis. 1978. “Miscelánea filológica vasca”. Fontes Linguae Vasconum. 205–228.
Ordoñez, Francisco & Esthela Treviño. 1999. “Left dislocated subjects and the pro-drop
parameter: A case study of Spanish”. Lingua 107:39–68.
Ortega, Ivan. 2016. Focus Related Operations at the Right Edge in Spanish: Focus and Ellip-
sis. Amsterdam/Philadelphia: John Benjamins.
Ortiz de Urbina, Jon. 2002. “Focus of correction and remnant movement in Basque”.
In Xabier Artiagoitia Beaskoetxea, Patxi Goenaga Mendizabal & Joseba A. Lakarra
Andrinua (eds.), Erramu Boneta. Festschrift for Rudolf P.G. de Rijk. Donostia: Univer-
sidad del País Vasco. 511–524. [ASJU Supplement XLIV]
Ortiz de Urbina, Jon. 2003. “Word Order”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A Grammar of Basque. Berlin/New York: Mouton. 448–459.
Oyharçabal, Bernard. 2003. “Relative Clauses”. In Jose Ignacio Hualde & Jon Ortiz de
Urbina (eds.), A Grammar of Basque. Berlin/New York: Mouton. 762–823.
Pastor, Alberto. 2011. “Degree Modifiers, Definiteness, and Specificity”. In Luis A. Ortiz-
López (ed.), Selected Proceedings of the 13th Hispanic Linguistics Symposium. Somer-
ville, MA: Cascadilla Proceedings Project.
Pérez Gaztelu, Elixabete, Lluisa Gràcia & Igone Zabala (eds.), 2004. Estructuras morfo-
58 ortiz de urbina
lógicas y estructuras sintácticas: la frontera de la composición en lenguas románicas

y en vasco. Donostia: Universidad de Deusto.
Quer, Josep. 2002. “Edging quantifiers. On QP-fronting in Western Romance”. In Claire
Beyssade, Reineke Bok-Bennema, Frank Drijkoningen & Paola Monachesi (eds.),
Romance languages and linguistic theory 2000. Amsterdam/Philadelphia: John Ben-
jamins. 253–270.
Rotaetxe, Karmele. 1978. Estudio estructural del euskera de Ondárroa. Durango: Zugaza.
Sainz-Maza, Lorena. 2017. Interactions among focus, exhaustivity, and constituent order
in Spanish and Basque. Ph.D. dissertation, Ohio State University.
Saltarelli, Mario et al. 1988. Basque [Croom Helm Descriptive Series]. London: Croom
Helm.
Shih, Stephanie & Jason Grafmiller. 2011. “Weighing in on End Weight”. Presented at the
85 Annual LSA Meeting. January 9, Pittsburgh, Pennsylvania.
Siewierska, Anna (ed.). 1998. Constituent order in the languages of Europe. Berlin: Mou-
ton de Gruyter.
Trask, Larry. 2003. “The Noun Phrase”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A Grammar of Basque. Berlin/New York: Mouton. 113–170.
Villasante, Luis. 1980. Sintaxis de la oración simple. Oñate: Editorial Franciscana.
Zubizarreta, María Luisa. 1998. Prosody, focus, and word order. Cambridge, MA: MIT
Press.
chapter 3
Tense, Aspect and mood

José Ignacio Hualde and Céline Mounole
1 Introduction
Whereas the morphology of finite forms in Basque differs greatly from that of
the Romance languages, in the organization of the Tense/Aspect/Mood (TAM)
system, on the other hand, we find interesting similarities, some of which
are likely due to language contact. These similarities are particularly appar-
ent when we consider diachronic processes of grammaticalization. Thus, it has
long been noticed that some of the analytical or periphrastic constructions of
Basque have a direct corresponding structure in Romance, suggesting parallel
developments in Basque and the neighboring Romance languages (Michelena
1981; Trask 1997: 237; Gómez & Sainz 1995). There are, however, important dif-
ferences as well.
When comparing the present-day TAM system in Basque and Romance lan-
guages like Spanish and French, we notice the following common features:
(a) Regarding mood, there is an indicative/subjunctive contrast, in addition
to conditional and imperative forms. (In Basque there are some addi-
tional forms, see below).
(b) Regarding tense distinctions, in the indicative, there is a future/present/
past contrast, whereas in the subjunctive only present and past are dis-
tinguished. (In Spanish there is an obsolescent future subjunctive, which
did not exist in Latin. In French, the past subjunctive is literary).
(c) Regarding aspect, there is a morphological contrast between perfective
and imperfective forms, but only in the past indicative.
Differences between Romance and Basque are most apparent in the morphol-
ogy of verbal forms:
(a) In the Romance languages, all verbs have both simple and compound
forms. In Basque most verbs only possess compound forms.
(b) In Basque all transitive verbs and a class of intransitive verbs (unerga-
tive verbs) take the auxiliary ‘have’ and all other intransitive vebs (unac-
cusative) take the auxiliary ‘be’. In French, most verbs take ‘have’ and
only a small class of intransitives are conjugated with ‘be’ in their com-
pound forms. Finally, in modern Spanish ‘have’ has been generalized as
an auxiliary for all verbs (and has been lost as a main verb). The choice

60 hualde and mounole
of auxiliaries in Basque is similar to the situation in Italian. However, in

Basque this choice has a much more profound effect in conjugation as it
runs throughout the TAM system, affecting both perfective and imperfec-
tive forms.
(c) Regarding mood, Basque possesses distinct irrealis and potential forms.
2 Simple and Compound Indicative Forms
In comparing the structure of the TAM system in Basque and Romance, it is

useful to take a diachronic perspective. We will start with Romance, where the
diachronic facts are more easily established, by comparison with Latin.
2.1 Romance Tenses Originating from Latin Periphrases

An important development in the verb system from Latin to Romance was the
creation of compound forms involving a conjugated form of the auxiliary and
the past participle; e.g. Sp he comprado ‘I have bought’, había comprado ‘I had
bought’, habré comprado ‘I will have bought’, habría comprado ‘I would have
bought’, haya comprado ‘I have bought, subjunctive’, hubiera comprado ‘I had
bought, subjunctive’, etc.
These Romance compound forms appear to have arisen from the grammat-
icalization of Latin resultative constructions, e.g. Lat habeō litteras scriptas ~
habeō scriptas litteras ‘I have letters (which are) written’ > It ho scritto le lettere,
Fr j’ai écrit les lettres, Sp he escrito las cartas ‘I have written the letters’ (see
Vincent 1982; Fleischman 1983; Ledgeway 2011: 452–458, among others). Notice
the loss of agreement between participle and direct object. For the intermedi-
ate stage in Old Spanish, showing agreement in gender and number with the
object (as well as greater flexibility in the order of the two verbal elements), cf.,
e.g., from El Cid: yo xxx he ganados ‘I have earned 30’ (Mod Sp yo he ganado 30);
los .vj. dias de plazo passados los an ‘The 6 days of the grace period they have
passed’ (Mod Sp los 6 días de plazo los han pasado); La oraçion fecha, la missa
acabada la an ‘The prayer ended, they have finished the Mass’ (Mod Sp la ora-
ción hecha, la misa la han acabado). The same grammaticalization processes
took place in English and other Germanic languages (see, e.g. Trask 1996: 137–
139; Bybee et al. 1994: 68–69).
Initially, the resultative construction with ‘have’ would have been possible
only with transitive verbs. With intransitive verbs, a parallel construction with
‘be’ arose. In present-day Spanish (like in English) all verbs take the auxiliary
‘have’ (which, unlike in English, has been lost as a main verb), but in Old Span-
ish we find remnants of the use of ‘be’ with intransitive verbs; e.g. El dia es exido
tense, aspect and mood 61
‘the day is gone’ (El Cid). In Modern Italian, a relatively large class of intransi-
tive verbs—known as unaccusative verbs—still take ‘be’, cf. It ho comprato un
libro ‘I have bought a book’ (transitive), ho lavorato ‘I have worked’ (unergative)
vs. sono arrivato/-a ‘I have arrived’ (lit. ‘I am arrived, masc/fem.’, unaccusative).
Notice that when the ‘be’ auxiliary is used, the participle shows gender (and
number) agreement with the subject.
The development of the present perfect created a contrast that did not exist
in Latin, where the perfect had both values of simple past and present per-
fect; e.g.: Lat amāvī ‘I loved; I have loved’ vs Sp amé ‘I loved’, he amado ‘I have
loved’. In French the contrast in meaning was later lost again, with the disap-
pearance of the simple past from the spoken language. In present-day French,
the form deriving from the Latin perfect (i.e. the passé simple, e.g. Fr j’ aimai)
has been relegated to the written language and nowadays only the compound
form (passé composé, e.g. Fr j’ai aimé) is used both as a present perfect and as
a perfective past in the spoken language.
Another major development in the verb system from Latin to Romance,
which is relevant for the comparison for Basque that we want to make here,
was the loss of the synthetic future (e.g. Lat cantābō ‘I will sing’) and its replace-
ment with a periphrasis of obligation formed with the present of ‘have’ and
an infinitive. This periphrasis was later morphologized as a single word; e.g.:
Lat cantāre habeō ‘I have to sing’ > OSp cantar (h)e > Sp cantaré ‘I will sing’
(see, e.g., Fleischman 1982). The transitional phase can be observed in the fol-
lowing examples from El Cid: doblarvos he la soldada ‘I will double your salary’
(ModSp os doblaré el sueldo); El Campeador dexar las ha en vuestra mano ‘The
Campeador will leave them in your hand’ (ModSp las dejará en vuestra mano)
The same construction, but with the past imperfect of ‘have’ gave rise to the
Romance conditional; e.g.: Lat cantāre habēbam ‘I had to sing’ > OSp cantar ia
> Sp cantaría ‘I would sing’.
With these Romance diachronic facts in mind, let us now consider the struc-
ture of verbal forms in Basque.
2.2 Basic Analytical Forms in Basque

Leaving aside a handful of verbs that possess synthetic forms in a few tenses,
conjugating a verb in Basque always involves combining a nonfinite form of
the verb with an auxiliary. A respect in which Basque differs from Romance is
the rarity of synthetic forms. Whereas, in Romance, like in English, tenses may
be simple (e.g. Sp compro ‘I buy’, compré ‘I bought’) or compound (e.g. he com-
prado ‘I have bought’, había comprado ‘I had bought’), all tenses in Basque are
compound tenses, again with the exception of a very small number of verbs,
which have some synthetic forms in present-day Basque (see section 4).
table 3.1 Non-finite forms
Perfective part. erosi erori hartu eman

Imperf. part. erosten erortzen hartzen ematen
Prospective part. erosiko eroriko hartuko emango, emanen
Radical eros eror har eman
The nonfinite forms used in the conjugation of verbs include the perfective,
imperfective and prospective participles, in addition to the radical. The per-
fective participle is the most basic form of the verb in the sense that, knowing
the perfective participle, it is possible to derive all other forms. The nonfinite
forms of erosi ‘to buy’, erori ‘to fall’, eman ‘to give’ and hartu ‘to take’ are shown
in Table 3.1.
The prospective participle bears a suffix -ko that can be identified, at least
diachronically, with the relational suffix, which, among other functions, serves
to form the locative genitive as in mendi-ko ‘of the mountain’. In the standard
language, this suffix -ko competes with -en for verbs whose perfective partici-
ple ends in -n. In northeastern dialects, however, the suffix -(r)en is also used
with other verbs; e.g. harturen, erosiren instead of hartuko, erosiko. From a
diachronic perspective, the latter is the same suffix that forms the regular gen-
itive when attached to noun phrases, e.g. mutil-a-ren ‘of the boy’, mutil-en ‘of
(the) boys’. Both suffixes -ko and -en may have benefactive or destinative value.
Although these usages of these suffixes are restricted in the modern language
(e.g. bihar-ko ‘for tomorrow’), they were prevalent in earlier centuries, which
may explain their use in the prospective participle (Mounole 2011: 312–315).
Consider the examples in (1), from B. Etxepare (1545):1
(1) a. amoretan plazer baten mila dira dolore

love.loc pleasure on.gen 1000 are pain
‘In love, for one pleasure there are a thousand pains’
b. Guiçonaren prouechuco emaztea bethi da

man.gen profit.rel wife always is
‘A wife is always for the benefit of man’ (Etxepare II, 10)
1 Abbreviations used in the glosses in this chapter: loc = locative, gen = genitive, rel = rela-
tional.
table 3.2 Basic indicative tenses (transitive con-

jugation)
erosi dut ‘I have bought’

erosi nuen ‘I bought’
erosten dut ‘I buy’
erosten nuen ‘I was buying; I used to buy’
erosiko dut ‘I will buy’
erosiko nuen ‘I would buy’
The imperfective participle is diachronically an inflected form of the verbal

noun, with the addition of the locative suffix -n (as has long been noticed, see
Aldai 2002: 109–110). The verbal noun is created by adding -tze or -te to the radi-
cal; e.g. hartze ‘act of taking’. In the standard language, the allomorph -te is used
with verbs whose radical ends in a sibilant or a nasal; in some dialects, this allo-
morph, which is diachronically prior (see Urgell 2007), has a wider distribution.
In its use, the verbal noun corresponds, roughly, to the Romance infinitive; e.g.
asko hartze-a ‘to take a lot’, cf. Sp ‘el tomar mucho’.
A development in Western and Central dialects has been the replacement of
the radical with the perfective participle in all contexts, so that in these dialects
the radical is now only found in a few fixed expressions (see Mounole 2007 for
a detailed diachronic study).
The basic indicative forms of a verb not possessing synthetic forms arise
from the combination of one of the participles in Table 3.1 with the present
or past tense of an auxiliary. In Table 3.2 we illustrate with the first person sin-
gular indicative forms of erosi ‘to buy’, which include all possible combinations
of the three participles with the auxiliaries dut ‘I have’ and nuen ‘I had’.
As mentioned, unlike in present-day Spanish or English, but like in earlier
stages of these languages and in modern Italian, French and German, there
are two auxiliaries, ‘have’ and ‘be’. All transitive verbs take the ‘have’ auxiliary.
On the other hand, some intransitive verbs take the ‘be’ auxiliary and others
the ‘have’ auxiliary, a matter to which we will return. Table 3.3 shows the same
forms as in Table 3.2, but for erori ‘to fall’. The forms of the auxiliary that are
shown, when used alone, are the copulas naiz ‘I am’ and nintzen ‘I was’.
This system of indicative forms is the result of relatively recent processes
of grammaticalization. As Mounole (2011) argues, the earliest Basque texts,
from the 16th century, appear to show the transition from a very different
older system where synthetic forms were more prevalent, and these were
opposed to analytical perfective forms based on the radical (which have now
table 3.3 Basic indicative tenses (intransitive con-

jugation)
erori naiz ‘I have fallen’

erori nintzen ‘I fell’
erortzen naiz ‘I fall’
erortzen nintzen ‘I was falling; I used to fall’
eroriko naiz ‘I will fall’
eroriko nintzen ‘I would fall’
been confined to the subjunctive and other non-indicative moods) and other
periphrases, now obsolete, which we will consider below (see section 5.4).
Given this, it seems sensible to consider the possible role of language contact
in the development of the system of forms illustrated in Tables 3.2 and 3.3.
The present perfect form erosi dut ‘I have bought’ has the same structure as
in the Romance and Germanic languages, as it combines a perfective participle
with the present tense of the auxiliary.
In both Peninsular Spanish and Basque, the present perfect, in addition to
its perfect meaning, functions as a hodiernal or recent past; e.g. Esta mañana
he visto a Mikel/ Gaur goizean Mikel ikusi dut ‘I saw Mikel this morning’ (vs Ayer
vi a Mikel/Atzo Mikel ikusi nuen ‘Yesterday I saw Mikel’). In dialects spoken in
France, this form can also be used as a non-hodiernal past. Although the latter
development mirrors that in French, it appears that it is due to an indepen-
dent evolution, as Mounole (2011) argues, since the widening of meaning took
place in those Basque varieties in the 17th century, before it did in French, and,
furthermore, the contact Romance language at the time was not French but
Gascon, which did not undergo this change.
On the other hand, notice that erosi nuen, which combines the perfective
participle with the past of ‘have’ does not have a pluperfect meaning, ‘I had
bought’ (Sp había comprado), but, rather it is a simple perfective nonhodiernal
or remote past ‘I bought’.
Both tenses with the perfective participle must have developed from resul-
tative structures through the same grammaticalization path as we find in Ro-
mance and Germanic (see Aldai 2001; Mounole 2011). Nevertheless, some struc-
tural differences should be noted: in Basque there is no singular/plural agree-
ment between participle and direct object, not even in the earliest texts (from
the 15th–16th centuries), when the structure was clearly biclausal and had a
resultative reading (Obviously, there cannot be gender agreement, as Basque
lacks this morphological category). The explanation would be that, when the
structure arose, the use of the definite article had not yet been extended to
predicative constructions (morphological plurality developed from the gram-
maticalization as articles of singular and plural demonstratives, Manterola
2015).
In fact, in 16th century texts, the form we are exemplifying with erosi nuen
could function both as a perfective past and as a pluperfect (Mounole 2011:
55). The change in meaning from pluperfect (‘I had bought’) to perfective past
(‘I bought’) is thus a recent innovation, without a direct parallel in Romance,
although along a well-attested grammaticalization path (Aldai 2002: 2017–218):
pluperfect > pluperfect & perfective nonhodiernal past > perfective nonhodier-
nal past; cf. also the evolution of the Latin pluperfect indicative to past subjunc-
tive in Spanish, e.g. Lat amāveram ‘I had loved’ > Sp amara ‘I loved, subjunctive’.
In Old Basque we also find a structure erosi duket ‘I will have bought’, where
the auxiliary bears the suffix -ke. This form has a future perfect interpretation
and may also express probability (see section 3.4).
Nowadays, to express a pluperfect meaning (i.e. a past of the past), as op-
posed to a remote past, Basque makes use of a “double-compound past perfect”
(Aldai 2002), as in erosi izan nuen ‘I had bought’, where izan is the radical and
past participle of ‘be’ (there is also a parallel double-compound present perfect,
erosi izan dut ‘I have bought’, which usually expresses the experiential perfect).
Double-compound forms have a formal parallel in the formes surcomposées of
French and Occitan (Mounole 2011: 47). A meaning close to that of the pluper-
fect can also be expressed by means of a resultative construction, erosia nuen
~ erosita neukan ‘I had it bought’ vs erosia dut ~ erosita daukat ‘I have it (in the
state of being) bought’, which also has a parallel in Spanish (e.g. las tengo com-
pradas ‘I have bought them, fem.’).
The forms with the prospective participle, erosiko dut ‘I will buy’, erosiko nuen
‘I would buy’, also appear to share a diachronic origin with the corresponding
tenses in Spanish and most other Romance languages (Michelena 1981). The
dialectal alternation between suffixes in Basque, e.g. erosiko dut ~ erosiren dut
‘I will buy’ supports this identification between the Romance and Basque struc-
tures to a certain extent. There is a more exact parallel with a construction
of obligation in Spanish with the preposition de ‘of’, e.g. Sp he de comprar ‘I
have to buy’ (although this Spanish construction retains a meaning of obli-
gation and has not been grammaticalized as a future). Nevertheless, notice
that the Basque prospective participle is built on the perfective participle. A
literal Spanish translation of erosiko dut would be *he de comprado. The par-
allel, thus, is not perfect. It is more likely that the grammaticalization of the
suffixes -ko and -en in the prospective participle actually arose from their use
in phrases with benefactive or destinative value (Mounole 2011: 312–315). Struc-
table 3.4 Modern Spanish simple and compound TAM forms and their Basque counterparts
Present compro erosten dut ‘I buy’

Past imperf. compraba erosten nuen ‘I bought’
Past perfective compré erosi nuen ‘I bought’
Future compraré erosiko dut ‘I will buy’
Conditional compraría erosiko nuen ‘I would buy’
Present perfect he comprado erosi dut ‘I (have) bought’
Pluperfect había comprado erosi izan nuen ‘I had bought’
(hube comprado)
Future perfect habré comprado erosi izango dut ‘I will have bought’
Conditional perf habría comprado erosi izango nuen ‘I would have bought’
tures like, e.g., bihar-ko dut ‘I have (to do) it for tomorrow’ may have provided
the original model for erosi-ko dut.
The Basque forms with the imperfective participle do not have a Spanish
parallel. The fact that the imperfective participle incorporates a locative suf-
fix suggests a similarity with, e.g., Eng I am on buying > I’m a-buying. In this
respect, we may note that Gascon has a similar progressive construction with
‘be’ and the infinitive preceded by the preposition a/en, e.g. Gasc èste a fer quau-
còm ‘s/he is doing something’ (Haase 1992: 93), (cf. also Port estou a comprar
‘I am buying’). If the Basque present indicative form is the result of a simi-
lar process of grammaticalization starting from a progressive construction, it
may originally have taken the ‘be’ auxiliary with all verbs, with a later shift
in auxiliary for transitive and unergative verbs, following the choice of aux-
iliary in other tenses (erortzen naiz ‘I am falling, I fall’ and also *erosten naiz
> erosten dut ‘I am buying, I buy’) (see Mounole 2008, 2011). Western Basque
also has a periphrasis with egon ‘be, stay’ for all verbs (e.g. erosten nago ‘I am
buying’) that more closely resembles the Spanish progressive (e.g. estoy com-
prando).
In Table 3.4, a comparison of modern Spanish and Basque indicative forms is
offered, arranged according to Spanish morphology, with simple forms before
compound forms. As can be observed, in spite of the diachronic parallels noted
above, in the present-day languages the only forms that show agreement in
their morphology are those of the present perfect.
Table 3.5 offers the same comparison between modern Basque and Spanish
forms, but using the basic forms of the indicative in Basque (excluding double-
compound forms) as point of comparison and pointing out parallel evolutions.
Forms that are identical in structure and meaning are indicated with an equal
table 3.5 Modern Basque TAM forms and their Spanish formal counterparts
erosi dut ‘I (have) bought’ = he comprado

erosi nuen ‘I bought’ (perfective past) ≈ había comprado ‘I had bought’ (pluperfect)
erosiko dut ‘I will buy’ ≈ compraré ‘I will buy’ (< comparāre habeō)
he de comprar ‘I have to buy’
erosiko nuen ‘I would buy’ ≈ compraría ‘I would buy’ (< comparāre
habēbam)
erosten dut ‘I buy’ No formal equivalent
erosten nuen ‘I bought’ (imperfect) No formal equivalent
sign, whereas the “approximately” sign (≈) is used for less perfect equivalencies
(with the caveats noted in the text above).
To conclude this section, the most striking difference between Basque and
Romance is the complete lack of synthetic forms in Basque for most verbs, so
that even for the present tense a compound form must necessarily be used. Also
strikingly different is the consistent choice in Basque between the ‘be’ and the
‘have’ auxiliary, depending on the semantics of the verb, across all compound
tenses.
When the contemporary languages are considered, a comparison between
Basque and Romance indicative forms reveals an exact match only in the struc-
ture of the present perfect. When the diachronic evolution of the compound
forms of Basque is also considered, other similarities with Romance develop-
ments become apparent, including in the form of the perfective past, whose
structure mirrors the Romance pluperfect, and, to some extent, the future and
the future of the past or conditional.
Regarding the distinctions that are expressed morphologically, we notice the
existence of present, past and future forms, with a further contrast in both
Romance and Basque between perfective and imperfective in the past. This
restriction of the contrast between perfective and imperfective to the past
tense is cross-linguistically common (Dahl 1985; Bybee et al. 1994:83).
3 Other Moods
3.1 Subjunctive
In both Basque and Romance, we find a contrast in mood between indicative
and subjunctive in subordinate clauses. In Basque, the subjunctive is formed
with the radical and a present or past form of a distinct auxiliary verb. As in the
table 3.6 Subjunctive forms in Basque
Present subjunctive eros dezadan compre ‘(that) I buy’

eror nadin caiga ‘(that) I fall’
Past subjunctive eros nezan comprara ‘(that) I bought’
Past subjunctive eror nendin cayera ‘(that) I fell’
indicative forms considered in the previous section, the auxiliary is different

for transitive and unergative verbs, on the one hand, and unaccusative verbs,
on the other. The auxiliaries used in the subjunctive do not have participial
forms and are reconstructed as *edin (intransitive) and *ezan (transitive) (in
Bizkaian or Western Basque, the non-indicative transitive auxiliary is egin ‘do’,
instead). Examples are given in Table 3.6, together with their corresponding
forms in Spanish for comparison.
Whereas in the Romance languages, the existence of a separate subjunctive
mood is inherited from Latin, in Basque, the subjunctive has arisen as an inde-
pendent mood only in very recent historical times—between the 15th and 18th
centuries—as a consequence of the restriction of forms containing the radical
and *edin/*ezan to non-assertive contexts (Mounole 2011: 335). In 16th century
texts, forms like eros nezan ‘I bought’, which are nowadays used only in certain
types of subordinate clauses, appear also in main clauses with the value of a
simple perfective past; e.g. jaigi zidi nagia, erra zizan uria (Refranes y Senten-
cias 1596: 294) ‘the loafer got up and burned the town’. Conversely, forms that
nowadays only have an indicative value could be used in contexts from which
they are now excluded, such as to express wishes, etc.; e.g. Jainkoak hala nahi
duela ‘May God will it thus’ (Axular, 1643), cf. modern Bq Jainkoak hala nahi
dezala.
In the past subjunctive there is no perfective/imperfective contrast in either
Romance or Basque. In Spanish, for instance, the contrast between María cantó
muy bien ‘M. sang very well (perfective)’ and María cantaba muy bien ‘M. sang
(imperfective)/was singing/used to sing very well’ is neutralized if the clause
is embedded under a matrix verb requiring the subjunctive mood in the sub-
ordinate clause: No creo que María cantara muy bien ‘I don’t think that María
sang (perfective or imperfective) very well’. In Basque, the fact that the subjunc-
tive makes use of the radical precludes the possibility of an aspectual contrast.
(Although, as mentioned, in Western and Central dialects the radical has been
replaced by the perfective participle in all contexts).
In Basque, the subjunctive is used less extensively than in Romance, as unin-
flected clauses with the verbal noun can be used as an alternative to the sub-
junctive in contexts where that is not an option in Romance, such as indirect

commands and expressions of desire; e.g. Sp quiero que lo compres, Bq zuk eros
dezazun nahi dut ~ zuk erostea nahi dut ‘I want you to buy it’. In addition, the
subjunctive is not used to express doubt, unlike in Romance; e.g. Sp no creo que
lo compre, Bq ez dut uste erosiko duenik ‘I don’t think s/he will buy it (present
subjunctive in Sp vs future indicative in Bq)’.
French and Spanish do not use the subjunctive in exactly the same contexts.
For instance, in temporal subordinate clauses with future reference, Spanish
uses the present subjunctive and French the future indicative, e.g. Sp cuando
lo veas, Fr. quand tu le verras ‘when you see it’. In Basque the present indicative
is the preferred option in this context, although the subjunctive is sometimes
used in writing, Bq ikusten duzunean (~ ikus dezazunean) ‘when you see it’.
In sum, the subjunctive is a very recently grammaticalized mood in Basque,
with a more restricted usage than in Spanish or French.
3.2 Imperative, Exhortative and Jussive

In exhortatives or first person plural commands, subjunctive forms are used in
both Basque and Romance: Bq eros dezagun, Sp comprémoslo ‘let’s buy it!’; Bq
ez gaitezen eror, Sp no caigamos ‘Let’s not fall!’
Third person commands or jussives are expressed in Basque either by means
of subjunctive forms with the complementizer suffix -la ‘that’ or with forms
obtained from the same auxiliary as in the subjunctive, but with the prefix b-
: eros dezala ~ eros beza ‘let him/her buy it’, eror dadila ~ eror bedi ‘let him/her
fall’. In Spanish and French subjunctive forms are used for this purpose.
True (second person) commands also show some affinity with the subjunc-
tive in both Basque and Romance, but the details are not identical. In Basque
there are distinct imperative forms, which are used in both positive and neg-
ative commands; e.g. eros ezazu! ‘buy it!’, ez ezazu eros! ‘do not buy it’, except
that for verbs that have synthetic imperative forms these can only be used in
positive commands. In Eastern and Central varieties, subjunctive forms are
nevertheless preferred in negative commands, ez dezazula eros ‘do not buy
it!’. In Spanish and Occitan (including Gascon), distinct imperative forms are
only used in positive (informal) commands, whereas negative commands are
expressed by the subjunctive: Sp compra ‘buy!’, no compres ‘do not buy!’, Gasc
non tornes pas tà casa ‘Do not go back home!’ (Birabent & Salles-Loustau 1989:
86).
3.3 Conditional
In Spanish and French the morphological conditional is used in the apo-
dosis of hypothetical conditional sentences, whereas past subjunctive or past
table 3.7 Basque potential forms
Present potential eros dezaket puedo comprar I can buy

eror naiteke puedo caer I can fall
Past potential eros nezakeen podía comprar I could buy
eror nintekeen podía caer I could fall
Hypothetical potential eros nezake podría comprar I could buy
eror ninteke podría caer I could fall
imperfective indicative forms are used in the protasis: Sp si lo encontrara,

lo compraría / Fr si je le trouvais, je l’acheterai ‘If I found it, I would buy it’.
The morphological conditional is also used as a future of the past, in indi-
rect speech: Sp dije que lo compraría ‘I said I would buy it’. There is a corre-
sponding compound form used in counterfactuals or failed conditions: Sp si lo
hubiera encontrado, lo habría comprado ‘If I had found it, I would have bought
it’.
Basque, on the other hand, possesses special conditional forms in both the
protasis and the apodosis of conditional clauses: ikusiko banu, erosiko nuke ‘If I
saw it, I would buy it’ (or, in Northeastern varieties, ikusten banu, erosiko nuke).
In principle this allows for a distinction with the future of the past: esan nuen
erosiko nuela ‘I said I would buy it’ (non-conditional, reporting ‘I said: I will buy
it’) vs esan nuen erosiko nukeela ‘I said I would buy it’ (conditional, if some con-
dition were to be met). In practice, this difference is not always observed and
the future of the past may appear in conditional sentences.
Morphologically, the Basque conditional auxiliary forms are related to past
indicative forms cf, e.g., nu-en ‘I had’, ba-nu ‘if I had’, nu-ke ‘I would have’; nintz-
en ‘I was’, ba-nintz ‘If I were’, nintza-teke ‘I would be’.
3.4 Potential
An aspect in which the Basque TAM system differs markedly from Romance is
in the existence of a set of potential forms. Potential forms bear a suffix -ke or
-teke and are based on the same auxiliary as in the subjunctive. In Romance
potential forms are periphrastic constructions with ‘be able’ and the infini-
tive.
The suffix -(te)ke of potential forms is clearly the same suffix that is found
in the conditional apodosis. In older texts and (rarely) in present-day literary
usage, this suffix may also attach to the indicative auxiliary to indicate either
future time or probability, e.g. ikusten dugu ‘we see’, ikusten dukegu ‘we may see,
we will see’ (also ikusi dukegu).
Basque can also express the potential in periphrases with ahal ‘to be able’
and ezin ‘not to be able’; e.g. erosi ahal dut ~ erosten ahal dut ‘I can buy it’ (with
the perfective or the imperfective participle depending on the dialect, see sec-
tion 6).
4 Synthetic Forms in Basque
As has been mentioned, a handful of Basque verbs deviate from the patterns
illustrated above in that, besides compound or analytical forms, they also pos-
sess synthetic forms. In present-day Basque, in addition to the two auxiliaries
of the indicative, which also function as main verbs with the meanings of ‘have’
and ‘be’, the following verbs have synthetic forms for most speakers:
a) With intransitive morphology: joan ‘go’, etorri ‘come’, ibili ‘walk’
b) With transitive morphology: eduki ‘have’, jakin ‘know’, ekarri ‘bring’, era-
man ‘take’, *io(n) ‘give’
A few other verbs have a more restricted use as synthetic, limited sometimes to
a few forms, e.g. erion ‘flow’, jardun ‘be engaged in’, iraun ‘last, endure’, etzan ‘lie’.
These verbs possess a synthetic present and past indicative and, in some
cases, also conditional, potential and imperative forms; e.g. etorri ‘come’: dator
‘s/he comes, is coming’, zetorren ‘s/he came, was coming’, baletor ‘if s/he came’,
letorke ‘s/he would come’, betor ‘may s/he come’, zetorkeen ‘s/he could come’,
zatoz ‘come!’; jakin ‘know’: dakit ‘I know’, nekien ‘I knew, used to know’, baneki
‘if I knew’, nekike ‘I would know’.
Referring to the contrast between synthetic and compound forms, as in
dator ‘s/he is coming (continuous aspect)’ vs etortzen da ‘s/he comes (gener-
ally or habitually’), Trask (1997:212) writes: “What is striking is that, in contrast
to nearly all other European languages which have developed such an aspec-
tual contrast, in Basque it is the synthetic form which is marked for continuous
aspect […] and this in spite of the fact that the periphrastic form is etymologi-
cally transparent as involving the locative of a verbal noun”.
The online Basque grammar Sareko Euskal Gramatika lists 26 verbs with syn-
thetic forms, but most of the forms listed are not commonly used by a majority
of speakers. In the oldest texts, from the 16th and 17th centuries, we find a much
larger number of verbs with synthetic forms and also additional synthetic TAM
forms (Lafon 1944; Mounole 2011). The progressive loss of synthetic forms does
not seem to be directly attributable to contact with Romance. Influence from
Romance should have resulted instead in a generalization of synthetic forms
for all verbs. What we see in this development is, instead, the application of a
universal tendency to replace synthetic forms with periphrases. This tendency
operated in Romance, for instance, in the loss of the Latin synthetic passive
and the synthetic future, as well as the recent loss of the passé simple in spoken
French.
As other researchers have noticed, however, all verbs that are or have been
synthetically conjugated in historical Basque belong to the class whose par-
ticipial forms and radical take the prefix e- (Lakarra 2017:77), so that the exten-
sion of the synthetic conjugation to other classes of verbs could not easily
follow an established pattern.
5 Less Grammaticalized Constructions (Periphrases)
5.1 Progressive Constructions

In Spanish the simple present and imperfective past may be used to convey
progressive aspect. In addition, there is a progressive structure with estar ‘be,
stay’ and the present participle, very much like in English; e.g. estoy comprando
‘I am buying’, estaba comprando ‘I was buying’. French, on the other hand, lacks
progressive tenses and uses instead a periphrasis with être en train ‘to be in the
process of’; e.g. je suis/étais en train d’acheter ‘I am/was buying’.
Western Basque (Bizkaian) has a progressive construction with egon ‘be,
stay’ and the perfect participle that mirrors the Spanish construction. In this
construction, the auxiliary is always egon. That is, we do not find a transitivity-
conditioned alternation in auxiliaries; e.g. erosten nago ‘I am buying’.
Other dialects, instead, make use of a periphrasis with the perfective partici-
ple, the semi-auxiliary ari ‘to be engaged in’ and the ‘be’ auxiliary; e.g. erosten
ari naiz ‘I am buying’.
Aldai (2002: 135) suggests that the lack of auxiliary alternation in both of
these Basque progressive constructions may be due to Romance influence.
However, this may also be explained as a sign of the less grammaticalized
character of these constructions. As mentioned above (and as suggested by
Mounole 2011), the present indicative as well (erosten dut ‘I buy’, erortzen naiz
‘I fall’) may have started as a construction with ‘be’ for all verbs, with a later
introduction of the auxiliary alternation that developed in the present perfect
because of the originally resultative source of this tense. In fact, the possibility
of having auxiliary alternation with ari ‘to be engaged in’ appears to have arisen
in some dialects, starting in the second half of the 19th century, indicating a fur-
ther step in the grammaticalization process (Mounole 2008, 2010).
5.2 Resultative
Besides the compound present perfect, Spanish has developed a new resulta-
tive construction with the new ‘have’ verb tener (as mentioned, haber has been
lost in Spanish as a main verb indicating possession and only functions as an
auxiliary and in the ‘there be’ construction). In this resultative construction,
the participle agrees with the direct object in gender and number, cf he escrito
las cartas ‘I have written the letters’ vs tengo escritas las cartas ‘I have the letters
written’ (see Harre 1991). Basque also has a parallel construction, which differs
from the present perfect in that the participle takes the article -a(k), cf gutunak
idatzi ditut ‘I have written the letter’ vs gutunak idatziak ditut ‘I have the letters
written’.
In addition, Basque possesses an experiential construction of identical
structure, with an inflected perfective participle, but with the ‘be’ auxiliary.
This experiential construction can be used with both intransitive and transitive
verbs, e.g. Parisen egona naiz ‘I have been in Paris’, ‘I have had the experience
of being in Paris’ (vs Parisen egon naiz ‘I have been in Paris’), gutun asko idatzia
naiz ‘I have written many letters’ (e.g. ‘I have the experience of having written
many letters’), gorriak ikusia naiz ‘I’ve had a hard time’, lit. ‘I am (having) seen
them red’ (Villasante 1980: 152; Zabala 2003: 431).
Sentences with an inflected participle are perhaps to be analyzed as bi-
clausal constructions, with a larger domain of applicability, beyond their resul-
tative and experiential functions (Ortiz de Urbina & Uribe-Etxebarria 1991).
Basque surpasses Romance in the flexibility of participial predicates.
With the same meaning and functions, Basque also has other structures
where the past participle bears the suffix -ta (probably from the conjunction
eta ‘and’, Krajewska 2013) or -rik (partitive, probably form an earlier ablative
usage of this suffix, Mounole 2008: 291–292) and inflected forms of egon ‘be,
stay’ and eduki ‘have’, e.g. nekatuta/nekaturik nago ‘I am tired’, ikusita/ikusirik
daukat ‘I have seen it’.
5.3 Must, Need, Want

To express the notions of ‘must’ or ‘need’, Basque has developed a construction
involving the noun behar ‘need, obligation’ in its uninflected form and ‘have’;
e.g. telefono bat behar dut ‘I need a telephone’. This construction can combine
with a participial clause; e.g. telefono bat erosi behar dut ‘I need to/must buy
a telephone’; etxera joan behar dut ‘I must go home’. As it is crosslinguistically
common for obligation expressions, the construction with behar has also devel-
oped the meaning of an immediate future (Bybee et al. 1994: 258–259), e.g.
Tokatzen baldin bazait aurten loteria, erosi beharko det Citroën berria (song by
X. Lete) ‘If I win the lottery this year, I am going to buy a new Citroën’.
The closest Romance parallel is offered by the avoir besoin de ‘need, must,
have need of’ construction in French: j’ai besoin d’un téléphone ‘I need a tele-
phone’, j’ai besoin d’acheter un téléphone ‘I need to/must buy a telephone’ (or
It avere bisogno). Notice, however, that in Basque the object of behar appears in
the absolutive case, not in the genitive, showing a higher degree of grammati-
calization as a verbal construction than in French. In fact, behar ‘need’ can also
be inflected and be used as a regular noun, in which case it takes a genitive
complement: telefono baten (erosteko) beharra dut ‘I have the need of (buying)
a telephone’. In present-day Basque, then, we must distinguish between the ver-
bal construction behar + ‘have’ and the noun behar ‘obligation, need, work’.
Another indication of a high degree of grammaticalization of the verbal con-
struction behar + ‘have’ is that in this construction the word behar can take the
ending -ko of the prospective participle directly. As an alternative, however, the
prospective participle of izan may also be added: telefonoa erosi beharko nuke
~ telefonoa erosi behar izango nuke ‘I should buy the telephone’. On the other
hand, in this construction, behar cannot bear the imperfective suffix (irrele-
vantly, the denominal verb behartu ‘oblige, force’ conjugates like a regular verb).
Both Spanish and French make use of deber/devoir as a modal verb with
deontic (obligation) and epistemic (probability) value, like English must,
should: Sp debes comprarlo, Fr tu dois l’acheter ‘you must buy’, Sp debe haber
caído, Fr il doit être tombé ‘It must have fallen’. In Spanish, to express epistemic
meaning this construction competes with the future (possibility in the present)
and conditional (possibility in the past). As already mentioned, Basque also has
this usage.
To express obligation or need, French and Spanish make use of some addi-
tional constructions. French possesses a construction with falloir ‘be neces-
sary’ used as an impersonal verb with a subjunctive complement, e.g. il faut
que je l’achète ‘I must buy it’. This construction has no parallel in present-day
Spanish or in Basque. On its part, Spanish has constructions with haber de +
infinitive and tener que + infinitive (e.g. he de comprarlo ~ tengo que comprarlo
‘I must buy it’). Above we mentioned the similarity between the haber de con-
struction of Spanish and the Basque future. An older Spanish construction with
caler ‘need’ (lit. ‘to be hot’) is now obsolete in this language (but alive in Judeo-
Spanish and Catalan).
With a parallel structure to the behar ‘need’ + ‘have’ construction, Basque
possesses another construction with nahi expressing the notion of ‘want’. This
construction has resulted from the grammaticalization of the uninflected noun
nahi ‘desire’ with ‘have’. In some Western varieties gura (< Lat gula) is used
instead of nahi. Whereas in standard Basque behar and nahi always take the
‘have’ auxiliary, a further step in the grammaticalization process, found in some
local varieties, is an alternation between ‘have’ and ‘be’ depending on the main
verb; e.g. joan nahi ba duzu > joan nahi bazara ‘If you want to go’, etorri behar du
> etorri behar da ‘s/he must come’ (see Mounole 2010). Here as well we see that
in its tendency to expand the lexically-determined alternation between ‘have’
and ‘be’ across verbal forms, Basque has been moving in the opposite direction
from Romance.
5.4 Other Periphrases

Basque and Spanish share several other periphrases showing parallel grammat-
icalization patterns. One of them is the frequentative periphrasis with ‘walk’, Sp
andar, Bq ibili: Sp Mikel anda aprendiendo inglés/ Bq Mikel ingelesa ikasten dabil
‘Mikel is engaged in learning English’. Another common, but more restricted,
shared periphrasis employs ‘carry’ (Sp llevar, Bq eraman) with a time expres-
sion to indicate the duration of an activity: Sp Mikel lleva dos años aprendiendo
inglés / Bq Mikelek bi urte daramatza ingelesa ikasten ‘Mikel has been learning
English for two years’.
In some Western Basque varieties, we find a habitual construction with jakin
‘to know’, e.g. Sanantolinétan nire lagúnak bádakídxe etórten Lekittora ‘By St
Antolin’s day my friends usually come to Lekeitio’ (lit. ‘they know how to come’)
(Hualde, Elordieta & Elordieta 1994: 145). A parallel construction is found in
the Spanish of Ecuador and some other South American countries, but it is no
longer found in Peninsular Spanish, cf., e.g. No sabe venir el gozo sin pensiones de
pesares ‘Pleasure doesn’t usually come without a price in pain’ (Tirso de Molina
1629).2
In older Basque texts, we find another habitual construction with the per-
fective participle and joan ‘go’ or eraman ~ eroan ‘carry’, depending on the
transitivity of the main verb (Lakarra 1996: 179–181, Mounole 2011), e.g. Oski
gitxi urratu doa (Refranes y Sentencias 1596: 275) ‘few shoes tend to break apart’;
Mundu honek anhitz jende enganatu darama ‘This world tends to deceive many
people’. This construction, which in the Bizkaian area survived until the 19th
century but seems to be now obsolete in Basque, has a possible Old Spanish
parallel in expressions with ir ‘go’ and a past participle, although this Spanish
construction has a resultative, rather than habitual, meaning: Sin Dios todo va
perdido (González de Eslava 1569) ‘without God, everything is lost’; Sardina que
gato lleva, galduda va (Marqués de Santillana 1454) ‘A sardine that is carried by
a cat, is lost’; de manera que va vendido por junto el trigo a excesivos precios ‘So
that wheat is sold together at excessive prices’ (Pedro de Valencia 1605); parte
2 Historical examples in Spanish are taken from the Royal Spanish Academy’s CORDE database.
del dicho pasadizo va cerrado con celosías (Luis Cabrera de Córdoba 1599–1644)
‘Part of that passageway is closed by lattices’. The less frequent corresponding
Spanish expression with llevar ‘carry’ and a past participle also has a resultative
meaning: como llevo declarado ‘As I have already declared’.
The ‘go’ +infinitive construction, which, like in English, has given rise to a
periphrasis with future value in Spanish, French and Portuguese (Sp voy a can-
tar = Fr je vais chanter = Port vou cantar ‘I am going to sing’) has also a parallel in
Basque: kantatzera noa, where the verbal noun takes the allative suffix (but in
some Basque varieties, kantatzen noa). Incidentally, this construction has been
grammaticalized with perfective past value in Catalan, Cat vaig cantar ‘I sang’,
and in some varieties of Gascon, Gasc son mari va revenir à l’ ostel ‘her husband
went back to the hotel’ (Rohlfs 1977: 217–218).
6 Preverbal Particles in Basque and Romance Parallels
Basque uses a set of particles for modality and aspectual purposes. These par-
ticles are usually placed before the finite verb, i.e. before a synthetic verb or
between main verb and auxiliary in analytical constructions.
There is only one aspectual particle, ohi, which expresses habitual aspect
and has been documented in all dialects since the earliest texts. In Western
dialects it is used in the ‘perfective participle + izan/*edun construction’ and
in Eastern dialects in the ‘imperfective participle + izan/*edun construction’,
instead. The construction with ohi does not have the same status in both dialec-
tal areas, due to different paths of grammaticalization. Whereas in Eastern
dialects ohi is clearly an aspectual marker, in the Western area it behaves as
a semiauxiliary (Ortiz de Urbina 2003: 321–323).3
The use of ohi is not necessary to express habituality. As mentioned above,
Basque, like the Romance languages, can use imperfective forms to refer to
habitual actions, e.g.: goizeko bederatzietan etortzen da/zen ~ goizeko bederatzi-
etan etorri ohi da/zen ‘s-he usually comes/s-he used to come at nine in the
morning’.
The other preverbal particles of Basque belong to the sphere of modality.
Among them, we find ahal ‘desiderative, epistemic modality, yes/no questions’,
bide ‘apparently’, omen ‘hearsay information’ and ei ‘hearsay information’.
3 There are, for instance, syntactic differences in negative sentences; e.g. Eastern etortzen ohi
da ‘s-he usually comes’→ ez ohi da etortzen ‘s-he usually does not come’, where ohi is fronted
with the auxiliary and negative ez ‘not’, vs Western etorri ohi da → ez da etorri ohi.
Ahal is a desiderative and epistemic modality particle, e.g.: ez ahal da

berandu etorriko ‘I hope s-he does not come late!’, gizon ederra egin ahal da ‘he
must have become a handsome man’. As mentioned above, it is also used as a
modal particle is potential forms, e.g.: etorri ahal da = etortzen ahal da = etorri
(ahal) daiteke ‘s-he may/can come’. In addition it is used as an interrogative
particle in Central dialects (spelled al), e.g.: etorriko al da ‘will s-he come?’. This
last use is very recent, attested only since the 19th century, although it is easy
to explain as a reinterpretation of the epistemic value.
Bide expresses probability: etxean gelditu bide da ‘s-he may have stayed at
home’, ez bide da etorriko ‘s-he may not come’. As was mentioned above, in the
Romance languages epistemic modality is mainly expressed by a periphrasis of
obligation (Fr devoir + infinitive: Il a dû bien grandir ‘he must have grown a lot’,
Sp deber (de) + infinitive: Debe de haber crecido mucho) or by the future tense
(Sp Estará enfermo ‘he may be sick’, Old Fr Il sera malade). In Basque varieties
in contact with Spanish the future tense can be also be used to express prob-
ability in the present (and the future of the past to express probability in the
past): gaixorik egongo/egonen da ‘s-he may be sick’. In varieties in contact with
French, on the other hand, the periphrasis with behar izan (see section 5.3) has
recently started to be used with this meaning: eri izan behar da ‘s-he may be
sick’.
Basque also displays particles for reporting hearsay information, omen in
Central and Eastern dialects and ei in the Western area: eri omen da ‘s-he is
apparently sick’, etorriko omen dira ‘it seems that they will come’. The Romance
languages in contact do not have any grammatical elements with this function,
although dizque (< dicen que ‘they say’) is used in some varieties of Latin Ameri-
can Spanish with a similar meaning, e.g.: dizque vendrán hoy ‘they are supposed
to be coming today’, dizque lloverá mañana ‘apparently it will rain tomorrow’.
In some varieties of French d’après may also have this function.
As for the origin of the Basque preverbal particles, most of them clearly
derive from nouns and are still used as such: bide ‘road’, ahal ‘power, ability’,
omen ‘fame, reputation; rumor’. The origin of ei remains obscure.
Gascon is the Romance language that employs the most preverbal particles
and some authors have sought a historical connection between the Gascon and
Basque particles (Allières 1992). However, the Gascon particles are far from hav-
ing a modal function, as they are mainly “enunciative”. The best known if these
particles is que, which is frequently employed preceding the verb in the main
clause of affirmative sentences: que i a ua hont qui canta ‘there is a spring that
sings’, qu’arriba ‘s/he is coming’ (Birabent & Salles-Loustau 1989: 72). This has
undoubtedly originated from the conjunction que ‘that’ heading a subordinate
clause introduced by a clause or an expression indicating certainty (‘I’m sure
that, the point is that; certainly, of course’). The main clause or the adverb must
gradually have fallen out of use, and then this use of que was extended to all
affirmative sentences (Rohlfs 1977: 206). In Spanish and Romanian as well, the
particle que is very often employed for emphasizing the clause (Sp ¡que voy!),
and for marking negation (Sp ¡que no quiero!, Rum ca no pot ‘I can’t’). Historical
Basque has no particle of this kind. Nevertheless, Gómez & Sainz (1995) estab-
lished a comparison between the 3rd person past prefix z- and the conjunction
eze, proposing that originally it must have appeared as an enuntiative particle
in narration, which was later reinterpreted as a past morpheme.
Gascon displays a few more particles: be (exclamative particle, be parlas plan
lo gascon! ‘you speak Gascon very well’), e (essentially interrogative particle and
conjunction e sabes parlar gascon? ‘Do you speak Gascon?’) and ja (exclama-
tive and affirmative particle, J’ac sabi ‘I know it’). The latter—also present in
Spanish and Catalan—can be compared to the Basque prefix ba-. It is indeed
the case that the Basque particle originates from bai ‘yes’. In Basque the use
of this preverbal particle was extended and has become obligatory when the
finite verb is located at the beginning of the clause (badakit nik, *dakit nik ‘I
know (it)’, badator ‘s/he’s coming’) (Birabent & Salles-Loustau 1989: 73; Rohlfs
1977: 208–211).
7 Summary
In this paper we have undertaken a detailed comparison, from a diachronic

perspective, of the Basque TAM system with that of the Romance languages
with which it is in contact. We have pointed out that, whereas a number of rela-
tively recent developments in Basque, including the development of a subjunc-
tive mood and several periphrases, show convergence between Basque and its
Romance neighbors, not every important development has been convergent. In
particular, in Basque the diachronic tendency may have been towards strength-
ening the choice between transitive and intransitive auxiliary, depending on
the semantics of the verb, while in Spanish an original ‘be’/ ‘have’ alternation
in auxiliaries was later lost. Some apparent morphosyntactic similarities, such
as the presence of preverbal particles in Basque and Gascon, and the morpho-
logical structure of the future tense, turn out to be less obvious when examined
in some detail.
Acknowledgements
The first author acknowledges the research funding for this project received
from the Spanish Ministry of Economy and Competitiveness (FFI2014–51878-
P).
References
Aldai, Gontzal. 2002. The grammaticalization of present and past in Basque. Ph.D. thesis,
University of Southern California, Los Angeles, CA.
Allières, Jacques. 1992. “Gascón y euskera: afinidades e interrelaciones lingüísticas”.
ASJU 26(3). 801–912.
Axular, Pedro. 1643. Gero. Bordeaux: G. Milanges. Edition and Spanish translation by
L. Villasante, 1964, Barcelona: J. Flors; Critical edition by B. Urgell, 2015, Bilbao:
Euskaltzaindia & Nafarroako Gobernua.
Bec, Pierre. 1968. Les interférences linguistiques entre gascon et languedocien dans les
parles du Comminges et du Couserans. Paris.
Birabent Jean-Pierre, & Salles-Loustau Jean. 1989. Memento grammatical du gascon,
Pau: Escòla Gaston Febus & Nosauts de Bigòrra.
Bybee, Joan, Revere Parkins & William Pagliuca. 1994. The evolution of grammar: Tense,
aspect and modality in the languages of the world. Chicago: University of Chicago
Press.
Creissels, Denis & Mounole, Céline. 2012. “Non-canonical valency patterns in Basque:
variation and evolution”. 45th Annual Meeting of the Societas Linguistica Europaea
29 August–1 September 2012, Workshop Contrastive studies of verbal valency in Euro-
pean languages.
Dahl, Östen. 1985. Tense and aspect systems. Oxford: Blackwell.
Etxepare, Bernard. 1545. Lingue Vasconum Primitiae. Bordeaux. Facsimile, critical edi-
tion and translation by P. Altuna, 1995. Bilbao: Euskaltzaindia.
Fleischman, Suzanne. 1982. The future in thought and language: Diachronic evidence
from Romance. Cambridge: Cambridge Univ. Press.
Felischman, Suzanne. 1983. “From pragmatics to grammar: Diachronic reflections on
complex pasts and futures in Romance”. Lingua 60(2–3). 183–214
Gómez, Ricardo & Sainz Koldo. 1996. “On the origin of the Finite Forms of the Basque
Verb”. In Hualde Jose Ignacio, Joseba A. Lakarra & Trask Larry (eds.), Towards a His-
tory of Basque. Amsterdam/Philadelphia: John Benjamins. 235–254.
Haase, Martin. 1992. Sprachkontakt und Sprachwandel in Baskenland: Die Einflüsse der
Gaskgonischen und Französischen auf das Baskische. Hamburg: Helmut Buske Ver-
lag.
Harre, Catherine. 1991. Tener + Past Participle: A case study in linguistic description. Lon-
don: Routledge.
Hualde, José Ignacio, Gorka Elordieta & Arantzazu Elordieta. 1994. The Basque dialect
of Lekeitio. [Supplements of ASJU XXXIV]. Bilbao/Donostia-San Sebastián: Univ. of
the Basque Country Press/ Diputación Foral de Gipuzkoa.
Krajewska, Dorota. 2013. The diachrony of resultative constructions in Basque. Master’s
Thesis, UPV/EHU.
Lakarra, Joseba A. 1996. Refranes y sentencias (1596): Ikerketak eta edizioa. Bilbo: Real
Academia de la Lengua Vasca/Euskaltzaindia.
Lakarra, Joseba A. 2017. “Basque and the reconstruction of isolated languages”. In Lyle
Campbell (ed.), Language isolates. London: Routledge. 59–99.
Ledgeway, Adam. 2011. “Syntactic and morphosyntactic typology and change”. In Mar-
tin Maiden, John Charless Smith & Adam Ledgeway (eds.), The Romance languages,
vol. 1: Structures. Cambridge: Cambridge Univ. Press. 382–471.
Manterola, Julen. 2015. Euskararen morfologiaren historiarako: artikuluak eta erakus-
leak/ Towards a history of Basque morphology: Articles and demonstratives. Doctoral
dissertation, UPV/EHU, Vitoria-Gasteiz.
Michelena, Luis. 1981. “Lengua común y dialectos vascos”. ASJU XV: 291–303. Reprinted
in: Luis Michelena. 2011. Obras completas [Supplements of ASJU VII]. ed. by Joseba
A. Lakarra & Iñigo Ruiz Arzallus, Donostia-San Sebastián & Vitoria-Gasteiz:
Gipuzkoako Foru Aldundia & UPV/EHU. 517–544.
Mounole, Céline. 2007. “Perifrasi zaharra mendebalde eta erdialdeko euskara zahar-
rean: Azterketa kuantifikatiboa eta proposamen berria”. ASJU 41(1): 67–138.
Mounole, Céline. 2008. “Sintaxi diakronikoa eta aditz multzuaren garapena: inper-
fektibozko perifrasiaren sorreraz”. In Xabier Artiagoitia & Joseba A. Lakarra (eds.)
Gramatika jaietan Patxi Goenagaren omenez [Supplements of ASJU LI]. 548–604.
Mounole, Céline. 2010. “Alineazio aldaketak euskararen diakronian”, In Beatriz Fernán-
dez, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: aldakortasun
sintaktikoa aztergai [Supplements of ASJU LII]. 151–170.
Mounole, Céline. 2011. Le verbe basque ancien: étude philologique et diachronique. Ph.D.
thesis, UPV/EHU. ASJU 48 (2014 [2018]). Bilbao: UPV/EHU.
Ortiz de Urbina, Jon. 2003. “Semiauxiliary verbs”. In José Ignacio Hualde & Jon Ortiz
de Urbina, (eds.), A Grammar of Basque, 300–328. Berlin/New York: Mouton de
Gruyter.
Ortiz de Urbina, Jon & Myriam Uribe-Etxebarria. 1991. “Participial predication in
Basque”. In Joseba A. Lakarra (ed.), Memoriae L. Mitxelena Magistri Sacrum, vol. 2,
[Supplements of ASJU XIV]. Donostia-San Sebastián: Diputación de Gipuzkoa. 993–
1012.
Real Academia Española. Corpus diacrónico del español (CORDE). [Available on the
internet: http://www.rae.es]
Rohlfs, Gerhard. 1977. Le gascon: Etudes de philologie pyrénéenne. Tübingen: Max Nie-
meyer Verlag.
Trask, Robert L. 1997. The history of Basque. London: Routledge.
Urgell, Blanca. 2007. “Para la historia del sustantivo verbal en vasco”. In Joseba A. La-
karra & José Ignacio Hualde (eds.), Studies in Basque and historical linguistics in
memory of R.L. Trask/R.L. Trasken oroitzapenetan ikerketak euskalaritzaz eta hiz-
kuntzalaritza historikoaz [ASJU XL]. 921–948.
Villasante, Luis. 1980. Sintaxis de la oración simple. Oñati: Editorial Franciscana Arán-
zazu.
Vincent, Nigel. 1982. “The development of the auxiliaries habere and esse in Romance”.
In Nigel Vincent & Martin Harris (eds.), Studies in the Romance verb. London: Croom
Helm. 71–96.
Zabala, Igone. 2003. “Nominal predication: copulative sentences and secondary predi-
cation”. In José Ignacio Hualde & Jon Ortiz de Urbina, (eds.), A grammar of Basque,
Berlin/New York: Mouton de Gruyter. 426–448.
chapter 4
Non-verbal Participles in Basque and Spanish

Ane Berro
1 Introduction
Adjectival participles have been argued to be mixed categories, since they show,
at the same time, properties belonging to verbs and to adjectives. On the one
hand, they are superficially adjectives: they denote a state predicated about an
entity, and can occur in attributive (1) or predicative position (2). On the other
hand, they can also be internally verbal, because, in some cases, participles also
imply a previous event that has as a result the state denoted by the participle.
(1) The broken tree.
(2) The tree is broken.
The different types of participles constitute an area of substantive study for the
analysis of the interface between syntax and the lexicon (Wasow 1977) and syn-
tax and morphology (Marantz 2001, 2007; Anagnostopoulou & Samioti 2014),
as well as for analysis of lexical categories and the projection of syntactic heads
like v, Voice and Aspect. Depending on whether an event is implied or not, and
on the acceptability of different event-related and subject-oriented modifiers,
adjectival participles have been argued to be of different types, with differ-
ent structural compositions, particularly involving different verbal layers (e.g.
Anagnostopoulou 2003; Embick 2004; Alexiadou & Anagnostopoulou 2008;
Bruening 2014; Doron 2014). Those without an event implication have been
considered to be built on top of an acategorial Root (Embick 2004). In con-
trast, others may involve, at different degrees, verbal projections like vP, VoiceP
and even a perfect AspP (Alexiadou et al. 2014; Alexiadou et al. 2015, see also
Sleeman 2011, 2014).
In this chapter, Basque and Spanish adjectival (and adpositional) partici-
ples will be addressed, in the light of the literature on adjectival participles.
As in most of the literature following Kratzer (1994, 2000), I will assume that
there is a differentiation between adjectival and verbal participles: adjectival
participles denote a state, whereas verbal participles describe an event. Addi-
tionally, I will also consider that adjectival participles can be phrasal and that

non-verbal participles in basque and spanish 83
they can also have event implications. With this picture in mind, in this chapter,
I will explore the properties of phrasal and non-phrasal non-verbal partici-
ples in Basque and Spanish, comparing them with verbal ones, and analyzing
whether the presence/absence of these properties is subject to a classification
of different types of non-verbal participles, such as resultant state / target state
participles (Kratzer 2000) and resultative/stative participles (Embick 2004).
Spanish past participles have been studied in a number of works (among
many others, Luján 1981; Demonte 1983; Bosque 1999, 2014; Marín 1997, 2000,
2004ab, 2009; Gehrke & Marco 2014), but there are not many studies about
Basque non-verbal participles (Ortiz de Urbina & Uribe-Etxebarria 1991; Kra-
jewska 2012, 2013; Berro 2017) and none comparing both Spanish and Basque.
Both languages, although being typologically distant, build and use non-verbal
participles in a strikingly similar way: they are alike in building resultative par-
ticiples by means of a stage-level copula (egon in western and central Basque
varieties and estar in Spanish) and an adjectival (or adpositional) partici-
ple. Additionally, as will be shown, both languages show a similar interaction
between lexical aspect and grammatical aspect, with adjectival (and adposi-
tional) participles interpreted in a non-resultative way when the embedded
event is non-dynamic.
On the other hand, they also differ in a number of aspects. For example,
Basque non-verbal participles are morphologically more complex (showing an
additional resultative morpheme attached to the bare participle) and, depend-
ing on the suffix, they can be adjectival and adpositional. Moreover, they are, in
principle, acceptable in more syntactic contexts, given that Basque non-verbal
participles allow temporal and spatial modification of the event underlying the
participle and, if the copula turns into izan ‘be’ (substituting egon), are com-
patible with all kinds of agent arguments. In Spanish, it has been argued that
the insertion of por-phrases (by-phrases) in adjectival participles is restricted
to those that modify an event kind (Gehrke & Marco 2014). However, as will be
shown, at least, in the Spanish spoken in the Basque Country, temporal and spa-
tial modification of the event is acceptable, and por-phrases with wide scope,
strong determiners and discourse referents are also judged grammatical. These
features have been taken as indicators for the presence of verbal projections
in adjectival participles cross-linguistically (Gehrke 2011, 2015; Alexiadou et al.
2014; Alexiadou et al. 2015) and will be considered in this chapter in order to
approach the structural composition of non-verbal participles in Basque and
Spanish.
Apart from the different degrees of eventiveness implied by the participles,
in this chapter, I will also pay attention to the configuration in which the par-
ticiple occurs, particularly, to whether the participle occurs in the attributive or
84 berro
predicative position, and regarding the latter, to whether the copula is intransi-
tive or transitive. As will be seen, these aspects are sometimes related with the
interpretation of the participle and also with the acceptability of event-related
modifiers.
The structure of the chapter is the following: in section 2, the form and
meaning of non-verbal participles in Spanish and Basque will be analyzed,
paying attention to their morphological and clausal composition, their com-
bination with the copula, their category and their position in the clause. In
section 3, the interaction between the lexical aspect and grammatical aspect
is studied, following particularly Bosque (2014), and also addressing the expe-
riential interpretation that Basque adjectival and adpositional participles can
have (Krajewska 2012, 2013ab). In section 4, the acceptability of non-verbal par-
ticiples with different types of event-oriented modifiers is explored, in the light
of the literature on adjectival participles in other languages such as German
and Greek (Kratzer 1994, 2000; Anagnostopoulou 2003; Alexiadou & Anagnos-
topoulou 2008; Gehrke 2011, 2015; Alexiadou et al. 2014; Bruening 2014; Alexi-
adou et al. 2015). In section 5, the restrictions that hold on event-modification
are addressed, particularly following the event token and event kind distinction
introduced in Gehrke (2011 and subsequent work) and discussing Spanish data
that suggest that adjectival participles in Spanish involve event kinds, and not
events that are instantiated (Gehrke & Marco 2014). I will argue that, in Basque
adjectival and adpositional participles, there is an event token, and that that
may also be the case in Spanish varieties spoken in the Basque Country.
2 Non-verbal Participles in Basque and Spanish
Like in English and German, adjectival participles in Spanish have the same
form as verbal participles used in perfect clauses (4), with the participial ending
attached to the verbal Root [+theme vowel]. Particularly, they look very simi-
lar to verbal passive participles, since adjectival participles show gender and
number agreement with the entity they are predicated about (4), just like ver-
bal passive participles (5).1
1 These are the abbreviations used in the glosses: abl = ablative; abs = absolutive; cl = clitic;
dat = dative; erg = ergative; fem = feminine; imperf = imperfect; ine = inessive adposi-
tion; instr = instrumental; loc = locative (used in the locative copula); masc = masculine;
part = partitive; pl = plural; prt = participle; pst = past; rel = relational adposition; res =
resultative suffix; sg = singular.
(3) He comprado un coche.

have.1sg buy.prt a car
‘I have bought a car.’
(4) a. El coche está comprado.

the car be.loc.3sg buy.prt.masc
‘The car is bought.’
b. Las manzanas están compradas.

the apple.pl be.loc.3pl buy.prt.fem.pl
‘The apples are bought.’
(5) Las manzanas han sido compradas por mí.

the apple.pl have.3pl be.prt buy.prt.fem.pl by me
‘The apples have been bought by me.’
In Basque, the participle used in perfect clauses is headed by the suffix -tu/-
i/-n/-ø, and, like in Spanish, it does not agree with the arguments of the verb.
Differing from Spanish, adjectival participles in Basque are formed attaching
an additional suffix to the participle used in perfect clauses (glossed here as
Res(ultative), see section 2.1): -a.
(6) Auto bat eros-i dut.

car a buy-prt have.1sg.erg
‘I have bought a car.’
(7) a. Auto-a eros-i-a dago.

car-the buy-prt-res be.loc.3sg.abs
‘The car is bought.’
b. Sagarr-a-k erosi-a-k daude.

apple-the-pl buy-prt-res-pl be.loc.3pl.abs
‘The apples are bought.’
When the participle is headed by -a, it agrees in number with the entity is pred-
icated about. As will be explained in the following subsection, Basque has other
suffixes to form resultative participles, but unlike -a ending ones, those do not
agree in number with the entity they are predicated about.
86 berro
2.1 Different Participial Suffixes in Basque

Apart from the -a ending, resultative participles can also be built with the suf-
fixes -ta [-da after a nasal or lateral] or -rik [-ik after a consonant].
(8) Mahai-a apur-tu-ta dago.

table-the buy-prt-res be.loc.3sg.abs
‘The table is broken.’
(9) Gizon-a Amazonia-n gal-du-rik dago.

man-the Amazonia-ine lose-prt-res be.loc.3sg.abs
‘The man is lost in Amazonia.’
-Ta and -rik are dialectally distributed: in western varieties, we find -ta headed
participles, whereas in eastern dialects -rik participles are more prominent. In
central varieties, both types of participles are found (Krajewska 2013). -A par-
ticiples, on the other hand, are used in all dialects.
Participles headed by -a are considered in the Basque linguistic tradition
adjectival, because they can occur in attributive and predicative configurations
and because they agree in number with the absolutive noun they are predicated
about (7), just like individual level adjectives (10b).
(10) a. Mendi altu-a-k

mountain high-the-pl
‘(The) high mountains’
b. Mendi-a-k altu-a-k dira.

Mountain-the-pl high-the-pl be.3pl.abs
‘Mountains are high.’
The suffix -a is homophonous with the determiner meaning ‘a’ or ‘the’, and has
been analyzed as a resultative aspect marker (Oyharçabal 1992) or a predicative
marker (Ortiz de Urbina & Uribe-Etxebarria 1991; Eguren 2012) in the context
of participles. As shown in example (10), -a also occurs heading individual level
adjectives like altu-a-k [tall-a-pl]. Eguren (2012), in this respect, claims that -a,
in the latter context, is the exponent of a Pred(icative) head (Baker 2003).
On the other hand, -ta and -rik headed participles are usually referred to as
adverbial (Rebuschi 1984), given that they cannot appear in attributive posi-
tion and they do not show number agreement with the absolutive argument.
Their historical origin is adpositional. According to Krajewska (2013), -rik has
evolved from the ablative adposition: it started to be used to mark converbs
(a diachronic tendency that is attested cross-linguistically), then secondary

participial predicates, and finally turned into a resultative suffix. As for -ta,
it originated from the copulative conjunction eta ‘and’ (de Rijk 2008: 692),
obtained the temporal meaning of after, and ultimately was reanalyzed as a
suffix attached to the verb.
Synchronically, -ta and -rik can be argued to be adpositional, and this way, we
would be able to account for their lack of number agreement. Moreover, these
suffixes behave like adpositions in not being acceptable in attributive position,
unless they are additionally headed by the relational suffix -ko (Goenaga 1978;
de Rijk 1998 [1993]; Elordieta 2001; Oyharçabal 2010). Adpositions cannot be
adnominal modifiers unless they are headed by -ko, a distributional feature that
is not shared by adjectives.
(11) a. *Laser bide-z depilazio-a

laser way-instr hair removal-the
Intended: ‘hair removal through laser’, ‘laser hair removal’
b. Laser bide-z-ko depilazio-a

laser way-instr-rel depilation-the
‘Laser hair removal’
(12) a. *gal-du-ta / *gal-du-rik gizon-a

lose-prt-res / lose-prt-res man-the
Intended: ‘The lost man’
b. galdu-ta-ko / galdu-ri-ko gizon-a

lose-res-rel / lose-res-rel man-the
‘The lost man’
In this respect, -ta and -rik participles replicate the behavior of adpositions.
They are good adnominal modifiers only in the presence of the relational mor-
pheme. This adnominal use of adpositional resultatives is usually referred to as
a relative form (Hualde et al. 1994).
All in all, Basque seems to have adjectival (ending in -a) and adpositional
(ending in -ta and -rik) participles, apart from the verbal participial form (end-
ing in -tu/-i/-n) used in perfect clauses.
2.2 Stative and Resultative Participles in Basque and Spanish

The suffixes -a/-ta/-rik in Basque can be used to form stative participles. Embick
(2004) proposed that adjectival participles in English can be stative or resulta-
88 berro
tives. Stative participles denote a characteristic state, like a non-derived adjec-

tive, whereas the resultative denotes a state that is the result of a previous event,
where this event is grammatically represented. In the former, the participle is
syntactically structured with an Aspectual head (Asp) projected directly on top
of an acategorial Root, whereas in the latter, the Root is verbalized (in a verbal
phrase, vP) before being selected by Asp.
There are different diagnostics to tell apart stative participles from resulta-
tives ones. One involves the use of manner adverbs that modify the manner in
which the event has taken place (Kratzer 1994). Stative participles (called lexi-
cal adjectival participles in Kratzer 1994) do not allow this kind of modification,
whereas resultatives do (13). Another one is the ability to occur as complement
of verbs of creation like build, create or make (Embick 2004) (14). Resultative
participles are not acceptable in this environment, since they imply a previous
event and this is incompatible with the presence of the verb of creation. Stative
adjectival participles, in contrast, are perfectly grammatical.
(13) a. The package remained carefully opened.

b. *The package remained carefully open.
(14) a. The door was built open.

b. *The door was built opened.
Thus, even if the same root (√open) is used in (13ab) and (14ab), the contrasts
found suggest that it is being used in different configurations, and this is sup-
ported by the participial morphology (-ed) only found in the resultative and
not in the stative adjective. In other cases, however, the same participial form
is used in both (e.g. closed).
(15) a. The package was carefully closed.

b. The door was built closed.
In Basque, the participial suffixes -a/-ta/-rik can be used to form resultative and
stative participles. In the case of the Root zabal ‘open, wide’, the stative and the
resultative participles are morphologically differentiated. Zabal-ik, consisting
of the Root and the suffix -rik, is used as a stative adjective, whereas zabal-
du-a/-ta/-rik is used in the resultative participle. In this case, the resultative
involves the Root (zabal), the participial suffix -tu (-du after -n/-l) (the one used
in perfect clauses) and the adjectival/adpositional suffixes -a/-ta/-rik.
(16) a. Ate-a zabal-ik eraiki dute.

door-the open-res build.prt have.3sg.abs.3pl.erg
‘They have built the door open’
b. ??Ate-a zabal-du-a/ta/rik eraiki dute.

door-the open-prt-res build.prt have.3sg.abs.3pl.erg
‘*They have built the door opened’
(17) a. *Ate-a kontu handi-z zabal-ik dago.

door-the care big-instr open-res be.loc.3sg.abs
‘*The door is open carefully’
b. Ate-a kontu handi-z zabal-du-a/ta/rik dago.

door-the care big-instr open-prt-res be.loc.3sg.abs
‘The door is opened carefully’
Contrast like this one are attested for a number of other participles such as huts-
ik ‘empty’ vs. hus-tu-rik ‘emptied’, biluz-ik ‘naked’ vs. biluz-tu-rik ‘undressed’,
poz-ik ‘happy’ vs. poz-tu-rik ‘delighted’ and bakarr-ik ‘alone’ vs. bakar-tu-rik ‘iso-
lated’. The first member of the pairs is not usually considered participial, since
it does not involve the participial suffix -tu that is used in perfect clauses. These
stative elements are referred to as stage-level adjectives (Zabala 1993) or stative
adverbs (de Rijk 2008).
In the case of the Roots ireki ‘open’ and itxi ‘close’, in contrast, the same form
is used as the stative and the resultative participles. It must be noted that, in
these cases, the bare participle has the same form as the bare Root.
(18) a. Ate-a ireki-ta eraiki dute.

door-the open.prt-res build.prt have.3sg.abs.3pl.erg
‘They have built the door open.’
b. Ate-a kontu handi-z ireki-ta dago.

door-the care big-instr open.prt-res be.loc.3sg.abs
‘*The door is open carefully.’
(19) a. Ate-a itxi-ta eraiki dute.

door-the close.prt-res build.prt have.3sg.abs.3pl.erg
‘They have built the door open.’
90 berro
b. Ate-a kontu handi-z itxi-ta dago.

door-the care big-instr close.prt-res be.loc.3sg.abs
‘*The door is open carefully.’
In Spanish, we find a similar scenario. Like in English and Basque, Spanish has
also stative adjectives, called perfective adjectives (Bosque 1999, Fábregas &
Marín 2015): lleno (vs. the participial llenado), seco (vs. the participial secado),
limpio (vs. the participial limpiado), sucio (vs. the participial ensuciado), vacío
(vs. vaciado) etc. They are defined as follows (Bosque 1999: 185; translation
taken from Fábregas & Marín 2015).
(20) [Perfective adjectives] denote states reached by the entities they are pred-
icated of.
Bosque (2014) argues that when participial forms have counterpart perfective
adjectives of this sort, the post-nominal participial form can only have eventive
(passive) meaning and not resultative, in Embick’s terms.
(21) a. una piscina llenada

a.fem swimming.pool filled.fem
‘A swimming pool filled up’
vs.
b. Una piscina llena

a.fem swimming.pool full.fem
‘A full swimming pool’
(22) a. la ropa secada

the.fem clothes dried.fem
‘The clothes dried’
vs.
b. La ropa seca
the.fem clothes dry.fem
‘The dry clothes’
According to Bosque (2014), (21a) and (22a) have a passive eventive meaning,
corresponding to their active counterpart (and do not denote the state reached
by these actions). In his opinion, the resultative interpretation is blocked when

the lexicon contains a lexical item for the adjective in the lower state of the
causative structure. In relation to this, the adjectives in (21b) and (22b) seem
to belong to the stative class proposed in Embick’s (2004), given that they are
incompatible with a verb of creation.
(23) a. Esta piscina se construyó llena.

this.fem swimming.pool cl build.3sg.pst full.fem
‘This swimming pool was built full.’
b. *Esta piscina se construyó llenada.

this.fem swimming.pool cl build.3sg.pst filled.fem
‘*This swimming pool was built filled up.’
Both the participle and the perfective adjective denote states. The difference
is that in the perfective participle, an event has not necessarily taken place.
Thus, in (23a), there is not an event of filling up, just like in stage-level adjec-
tives headed by -rik in Basque. In relation to this, Fábregas & Marín (2015)
defended the view that perfective adjectives are the exponents of a Result
subevent (syntactically represented in a First Phase Syntax, Ramchand 2008)
and an adjectival phrase (AP). Thus, according to these authors too, perfective
adjectives correspond merely to a stative layer, with no dynamic event. They
do not involve the Process subevent (responsible for providing dynamicity), so
that the Result is simply interpreted as a stative subevent. They must be then
paralleled to Embick’s stative participles, rather than to resultative ones. Fáb-
regas & Marín (2015) support this analysis by means of several criteria, such as
the following:
(24) a. Juan dejó la casa limpia / *limpiada.

John left the.fem house clean.fem / *cleaned.fem
‘John left the house clean/*cleaned.’
b. La casa quedo limpia / *limpiada.

the.fem house stayed clean.fem / *cleaned.fem
‘The house stayed clean/*cleaned.’
The verbs dejar ‘to leave’ and its anticausative counterpart quedar ‘to stay’ focal-
ize the result state and are only possible with telic verbs (García Fernández et
al. 2006). According to Fábregas & Marín (2015), dejar and quedar lexicalize,
subsequently, the Initiation + Process subevents and the Process subevent. The
92 berro
Result subevent is, in turn, spelled out by the perfective adjective. Adjectival
participles like limpiada ‘cleaned’ are not acceptable in this context because
they are blocked by the perfective participle. In the cases where the adjectival
participle entails a change of state, they claim that they necessarily lexicalize
Process.
2.3 Copular Verbs and the Biclausal Analysis

In Spanish, adjectival and verbal participles (resultative and eventive par-
ticiples in the terminology proposed by Embick 2004) are differentiated on
the basis of the inflectional element accompanying the participle. Unlike in
English, and more similar to German, the inflectional element used with verbal
passive participles and adjectival resultative participles is different. In adjecti-
val passives, the inflectional element is a copular verb, estar ‘(the stage level)
be’, whereas in verbal passive participles, it is an auxiliary, namely ser ‘be’.
(25) Las sillas están tiradas.

the.fem.pl chair.pl be.loc.3pl throw.prt.fem.pl
‘The chairs are thrown.’
(26) Las sillas son tiradas (todas las mañanas).

the.fem.pl chair.pl be.3pl throw.prt.fem.pl (every the mornings)
‘The chairs are thrown every morning.’
The configuration involving the verbal passive, with the auxiliary ser ‘be’ in
the present tense, has a habitual interpretation, with the event of throwing
the chairs repeated every morning. The structure with the adjectival partici-
ple, in contrast, only asserts that the state denoted by the participle holds at
the moment of the reference time.
Verbal passive participles occur in a monoclausal configuration, where the
main predicate is the participle and the inflected element stands as an aux-
iliary, whereas resultative participles are bi-clausal: the main predicate is the
copula (estar in Spanish), which takes the adjectival participle as its comple-
ment. Regarding Basque, the bi-clausal analysis of adjectival and adpositional
participles is proposed by Hualde et al. (1994) and de Rijk (2008), and is par-
ticularly analyzed and defended in Ortiz de Urbina & Uribe-Etxebarria (1991).
The monoclausal vs. bi-clausal analysis in Basque has been more controversial,
though, precisely because adjectival participles in Basque are acceptable with
ergative marked agent arguments, just like in monoclausal clauses, and because
they can occur with the copular izan ‘be’, instead of with egon ‘(the stage level)
be’.
(27) Gutun-a Miren-ek idatz-i-a da.

letter-the Mary-erg write-prt-res be.3sg.abs
‘The letter is written by Mary.’
Ortiz de Urbina & Uribe-Etxebaria (1991) argue in favor of a bi-clausal analysis

of configurations like (27), which involve an adjectival participle, particularly
on the basis of the absence of ergative agreement markers on the auxiliary and
word-order restrictions that differentiate clearly configurations like (27) and
perfect monoclausal clauses like (28). See section 8.3.3 for an overview of these
facts.
(28) (Miren-ek) gutun-a (Miren-ek) idatz-i du.

(Mary-erg) letter-the (Mary-erg) write-prt have.3sg.abs.3sg.erg
‘The letter is written by Mary.’
In both Spanish and Basque, apart from using the intransitive copulas (izan ‘be’
and egon ‘(the stage level) be’ in Basque and estar ‘(the stage level) be’ in Span-
ish), adjectival (an adpositional) participles can occur with transitive copulas
(edun and eduki ‘have’ in Basque and tener ‘have’ in Spanish). This configu-
ration has been called possessive resultative (Nedjalkov & Jaxontov 1988) and
it has been attested that, in the diachronic evolution of many languages, the
perfect has evolved from earlier possessive constructions of this sort (among
others, Detges 2000; Pinkster 1987). Actually, the perfect analytic configuration
of modern Basque, with the participle ending in -tu/-i/-n, used to be employed
as a resultative construction in previous stages of the language (Mounole 2011).
In possessive resultatives, the state denoted by the participle is predicated
about the object of the transitive copula, and the subject may or may not be
interpreted as the agent of the event underlying the participle. For instance, in
(29) and (30), the subject of ‘have’ can be interpreted merely as the possessor
or holder of the state denoted by the object, but not as the agent initiating the
event of cleaning.
(29) a. Ni-k sagarr-a-k eros-i-ta dauzkat.

I-erg apple-the-pl buy-prt-res have.3pl.abs.3sg.erg
‘I have the apples bought.’
b. Ni-k zapata-k garbi-tu-ta dauzkat.

I-erg shoes.the-pl buy-prt-res have.3pl.abs.3sg.erg
‘I have the shoes cleaned.’
94 berro
(30) a. Tengo las manzanas compradas.

have.1sg the.fem.pl apple.pl bought.fem.pl
‘I have the apples bought’
b. Tengo los zapatos limpiados.

have.1sg the.masc.pl shoes.pl cleaned.masc.pl
‘I have the apples cleaned.’
Similar constructions can be found in many other languages, like English and
French. In these cases, like in Spanish, the non-pronominal object of ‘have’
takes a pre-participial position, instead of its usual post-participial position,
and the participle agrees in number and gender with the object in the case
of French. In these configurations, the subject is interpreted as the posses-
sor/holder but not necessarily as the agent of the event underlying the par-
ticiple.
(31) I have my shoes cleaned.
(32) J’ai une lettre écrite.

I have.1sg a.fem letter written.fem
‘I have a letter written.’
Iatridou (1996) analyzes similar configurations in Modern Greek, and proposes

that, in these configurations, the object forms a passive small clause with the
participle, and the small clause is the complement of ‘have’.
(33) echo grameno ta grama.

have.1sg written.neut.sg.acc the letter
‘I have the letter written.’
According to Iatridou (1996), in (33), the entity ‘the letter’ is not the object
of ‘have’ since (33) does not assert a possessive relation between the subject
(1st person singular) and ‘the letter’. This fact is explained considering that the
whole passive small clause is the complement of ‘have’, and consequently, that
there is no thematic relation between ‘have’ and the entity ‘the letter’, as argued
in Iatridou (1996). Additionally, she argues that the agreement relation between
‘the letter’ and the participle in Modern Greek, and in similar configurations
in French (32), must not be regarded as a case of object agreement, but that
it should be analyzed as subject-predicate agreement inside a passive small
clause.
2.4 Categorial Status: Tests to Tell Verbal and Adjectival (and

Adpositional) Participles Apart
In the literature, several diagnostics have been identified to differentiate verbal
and adjectival participles (Wasow 1977; Levin & Rappaport Hovav 1986; Embick
2004; Sleeman 2011; McIntyre 2013 etc.): prenominal position in attributive par-
ticiples, un- prefixation, semantic presence/absence of an external argument,
selection by adjective phrase taking verbs, incompatibility with double objects
etc. As pointed out by Bosque (2014), though, the majority of these tests cannot
be applied to Romance, and, actually, neither to Basque.
As mentioned in section 3.2.3, adjectival participles in Spanish occur with
the copular verbs estar ‘(the stage level) be’ or tener ‘have’, whereas verbal pas-
sive participles are combined with the auxiliary ser ‘be’ or haber ‘have’. This
way, adjectival participles occurring in predicative position can be differenti-
ated from verbal passive participles.
(34) Esta mesa está rota.

this.fem table be.loc.3sg broken.fem
‘This table is broken.’
(35) Esta mesa fue rota.

this.fem table be.3sg.pst broken.fem
‘This table was broken.’
Number and gender agreement between the participle and the predicated
theme cannot be taken as a diagnostic to differentiate adjectival participles
from verbal ones. For instance, as shown in section 3.2, verbal passive partici-
ples also agree with the derived subject.
Bosque (2014) considers that both participles depicted in (34) and (35) are
verbal,2 and not adjectival, and points out that the ability of resultative par-
ticiples to occur with secondary predicates (36) and infinitive complements in
causative sentences (37a) or verbal set phrases (37b) support their verbal status.
(36) a. Un acusado considerado culpable

a.masc defendant considered.masc guilty
‘A defendant found guilty’
2 According to Bosque (2014), the difference between (34) and (35) is that, in (34), the participle
is resultative, whereas in (35), it is eventive. Estar is only compatible with resultative partici-
ples.
96 berro
b. Un concejal elegido alcalde

a.masc councilman elected.masc mayor
‘A councilman elected mayor’
(37) a. El lío de las pruebas hechas desaparecer

the.masc mess of the proofs made.fem.pl disappear
‘The mess of the proofs that were made disappear’ (El País, 30/01/2012)
b. Garbanzos puestos a remojo

chickpeas put.masc.pl to soak
‘Chickpeas left to soak’
Nevertheless, their grammaticality in these contexts may be explained if we

consider that they syntactically involve a vP layer, embedded under the adjec-
tival layer.
In Basque, adjectival participles in predicative position can occur either with
the copula egon ‘(stage level) be’ or with the copula izan ‘be’. The latter can also
be used as an auxiliary in intransitive analytic verbal configurations (perfect,
perfective, imperfective and prospective).
Another diagnostic that has been identified in the literature to differenti-
ate adjectival and verbal participles is the ability to occur in the complement
position of verbs like act, appear, become, look, remain, seem, smell and sound
in English (Wasow 1997), which is restricted to adjectives. Both Spanish and
Basque adjectival participles are licit in this context.
(38) María parece cansada / triste.

Mary looked tired.fem / sad
‘Mary looked tired/sad.’
(39) neka-tu-a / triste dirudizu.

exhaust-prt-res / sad seem.2sg.erg
‘you look exhausted/sad’
Even Basque adpositional participles (headed by -ta and -rik) can occur as the
complements of verbs like eman ‘to seem’ and of irudi ‘to seem, to look like’
(40). In contrast, bare participles in Basque—those headed by the participial
ending -tu/-u/-n—are ungrammatical in all these contexts. They cannot be the
complement of eman ‘to seem’ and of irudi ‘to seem, to look like’ and, moreover,
they are incompatible with the copulas egon ‘(the stage level) be’ and eduki
‘have’.
(40) neka-tu-ta/-rik dirudizu.

exhaust-prt-res seem.2sg.erg
‘you look exhausted’
(41) *neka-tu dirudizu

exhaust-prt seem.2sg.erg
Intended: ‘you look exhausted’
(42) *Liburu-a-k apur-tu daude /

book-the-pl break-prt be.loc.3pl.abs /
dauzkat
have(eduki).3pl.abs.1sg.erg
Intended: ‘(The) books are broken / I have (the) books broken’
Another test that is usually referred to is the compatibility of adjectival partici-

ples with degree modifiers. As shown by Kratzer (2000), degree modifiers are
only compatible with target state participles and ungrammatical with verbal
resultant state participles (see also Anagnostopoulou 2003). The modifier very,
in particular, selects for gradable adjectives that have a context-sensitive stan-
dard, in terms of Kennedy and McNally (1999). This means that they usually
select for adjectives that denote open scales. In Spanish, adjectival participles
denoting open scales, both in predicative position (43) and in attributive posi-
tion (44), can be modified by muy ‘very’, unlike verbal passive participles (45).
(43) a. Alex Mumbrú, que estuvo muy vigilado en el

Alex Mumbrú, that be.loc.3sg.pst very guarded.masc in the
poste bajo, …
post low
‘Alex Mumbrú, who was very guarded in the low post, …’ (Deia, 12/05/
2017)
b. Los globos están muy inflados

the.masc.pl ballon.pl be.loc.3pl very inflated.masc.pl
‘The ballons are very inflated’
c. Las casas están muy

the.fem.pl house.pl be.loc.3pl very
encarecidas
become.expensive.prt.fem.pl
‘The houses have become very expensive’ Lit. ‘The houses are very
become expensive’
98 berro
(44) a. Un edificio muy vigilado (por la policía)

a.masc building very guarded.masc (by the police)
b. Un político corrupto muy protegido por sus

a.masc politician corrupt.masc very protected.masc by his
colegas en los últimos años
colleagues in the last years
‘A corrupt politician very much protected by his colleagues in the last
years.’ (Bosque 2014: 54, footnote 12).
(45) a. *Las casas fueron muy

the.fem.pl house.pl be.3pl.pst very
encarecidas
become.expensive.prt.fem.pl
b. *Los globos fueron muy inflados

the.masc.pl ballon.pl be.3pl.pst very inflated.masc.pl
Degree modifiers are also compatible with Basque adjectival and adpositional
participles (46) (47), which, once again, contrast with bare participles used in
perfect simple clauses (48).
(46) a. Oso leku zain-du-a da.

very place protect-prt-res be.3sg.abs
‘It is a very protected place.’
b. Estilo oso puz-tu-a du.

style very boast-prt-res have.3sg.abs.3sg.erg
‘He/she has a very boasted style.’
(47) a. Mahai-a oso apur-tu-ta dago.

table-the very break-prt-res be.loc.3sg.abs
‘The table is very broken.’
b. Puxika-k oso puz-tu-a-k / puz-tu-ta

balloon.the-pl very inflate-prt-res / inflate-prt-res
daude.
be.loc.3pl.abs
‘The balloons are very inflated.’
(48) a. *Mahai-a oso apur-tu da

table-the very break-prt be.3sg.abs
b. *Puxika-k oso puz-tu dira /

balloon.the-pl very inflate-prt be.3pl.abs /
ditut
have.1sg.erg.3pl.abs
As can be seen, in Spanish, adjectival participles that are modified by muy ‘very’
can also occur with certain por-phrases (44). Interestingly, the por-phrases
occurring in this context modifies the state denoted by the participle, rather
than the previous event (see Alexiadou et al. 2014 and Alexiadou et al. 2015 for
similar cases in Greek).
2.5 Location in the Clause: Attributive vs. Predicative Participles

The standard assumption in the literature is that verbal participles (eventive
passives in Embick’s 2004 terminology) cannot occur in attributive position
because only adjectives, and not verbs, can be prenominal modifiers (Embick
2004). Sleeman (2011 2014) proposes that attributive participles can be either
adjectival (resultative) or verbal (eventive), showing that, in Dutch and English,
prenominal passive participles are adjectival while post-nominal ones are ver-
bal.
In Basque, adjectival participles in attributive position are only possible in
post-nominal position, but there exists a contrast that seems to be related to
the prenominal vs. post-nominal position of (participial) adjectives in other
languages like English. In Basque, adjectival participles can modify a bare noun
or a DP, and these two options present a number of contrasting distributional
features.
As shown in section 2.3, adjectival participles in predicative position can
license an agentive argument bearing ergative case marking, which unlike in
monoclausal simple clauses, does not agree with the inflectional element and
has a fixed position: it must remain between the absolutive theme and the
adjectival participle (Ortiz de Urbina and Uribe-Etxebarria 1992, see also sec-
tion 4.2.2 in this chapter). In attributive position, the introduction of an ergative
argument is more restricted. It can only be introduced in adjectival participles
that modify a DP nominal, and not in those that modify a bare noun.
(49) Gutun-a guraso-ek sina-tu-a da. Predicative

letter-the parents-erg firm-prt-res be.3sg.abs
‘The letter signed by the parents’
100 berro
(50) a. Gutun-a sina-tu-a (ekarri behar dute.) Attributive with DP

letter-the firm-prt-res (they have to bring)
‘(They have to bring) the letter firmed’
b. Gutun-a guraso-ek sina-tu-a (ekarri behar dute.)

letter-the parents-erg firm-prt-res (they have to bring)
‘(They have to bring) the letter firmed by the parents’
(51) a. Gutun sina-tu-a (ekarri behar dute.) Attributive with

letter firm-prt-res (they have to bring) bare noun
‘(They have to bring) the firmed letter’
b. *Gutun guraso-ek sina-tu-a (ekarri behar dute.)

letter parents-erg firm-prt-res (they have to bring)
‘*(They have to bring) the firmed letter by the parents’
The adjectival participle without the ergative argument can modify either a DP
(50a) or a bare noun (51a). In the presence of the ergative argument, the partici-
ple can only modify a DP, cf. (50b) (51b). The same contrast arises with indirect
dative arguments. A dative indirect object, participant of the event underlying
the participle, can only be introduced when the adjectival participle modifies
a DP.
(52) a. Gutun-a bidal-i-a (ikusi nahi dut.) Attributive with DP

letter-the send-prt-res (I want to see)
‘(I want to see) the letter sent.’
b. Gutun-a Miren-i bidal-i-a (ikusi nahi dut.)

letter-the Mary-dat send-prt-res (I want to see)
‘(I want to see) the letter sent to Mary.’
(53) a. Gutun bidal-i-a (ikusi nahi dut.) Attributive with bare noun
letter send-prt-res (I want to see)
‘(I want to see) the sent letter.’
b. *Gutun Miren-i bidal-i-a (ikusi nahi dut.)

letter Mary-dat send-prt-res (I want to see)
‘*(I want to see) the sent letter to Mary.’
The neutral relative order of the adjective, noun and the determiner in
Basque is depicted in (54). As can be seen, adjectives occur between the bare
noun and the determiner.
(54) Noun Adjective Det

Auto gorri -a
‘The red car’
But there are also instances where the adjective can modify a DP (55b). Let us
compare the example in (55a), with the adjective modifying the bare noun, and
the example in (55b). There is a subtle meaning contrast between them, which
seems to be similar to that obtained between prenominal and post-nominal
adjectives in English.
(55) a. Auto gorri-a nahi dut

car red-the want have.1sg.abs.1sg.erg
‘I want the red car’
b. Auto-a gorri-a nahi dut

car-the red-the want have.1sg.abs.1sg.erg
‘I want the car red’
In the example in (55a), the subject wants a specific car; it picks up an individ-
ual, which is distinguished from its red color. In the example (55b), in contrast,
the subject expresses which property the car he/she wants must have.
The inability of prenominal adjectival participles to take further modi-
fiers finds a correlate in non-derived adjectives. Actually, as shown by Sadler
and Arnold (1994), prenominal and post-nominal non-derived adjectives show
a clear contrast with respect to their ability to take modifiers and comple-
ments.
(56) a. a child grateful for the present

b. *a grateful for the present child
Sadler and Arnold (1994) point out several interpretative contrasts between
prenominal and post-nominal adjectives: in prenominal position, the reading
of the adjective is characteristic, timeless or defining property of the noun,
whereas in post-nominal position, the interpretation is of a temporary quality
or property. This is particularly observed in the following example:
102 berro
(57) a. We need to find a responsible person

b. We need to find the person responsible
As suggested by Sadler & Arnold, a responsible person is someone who has

generally the property of being responsible, whereas the person responsible is
the individual who has performed a particular action. This contrast can be
easily related to the individual level vs. stage level distinction (Carlson 1977).
In relation to this contrast, Svenonius (1993) argued that post-nominal posi-
tion is restricted to stage-level adjectives, but as pointed out in Sadler and
Arnold (1994), individual levels may also occur post-nominally. In any case, as
claimed by Sadler and Arnold (1994: 8), “prenominal adjectives are better when
they provide a ‘natural’ classification in combination with the noun” or when
they “denote some plausibly characteristic property of the noun they mod-
ify”.
A possible analysis differentiating prenominal and post-nominal adjectives
is to consider that post-nominal adjectives are reduced relative clauses, so that
they would be really predicative, involving some kind of silent copular verb.
Actually, Sadler and Arnold (1994) list several aspects in which post-nominal
adjectives and predicative adjectives behave alike in English. For instance, (i)
adjectives that cannot occur in predicative position cannot occur in attribu-
tive post-nominal position (58), (ii) when an adjective in the prenominal and
the post-nominal position has different senses, the post-nominal sense is the
same as that obtained in predicative position (59), (iii) prenominal adjectives
allow idiomatic interpretations in combination with the noun, whereas post-
nominal adjectives and predicative adjectives do not (60).
(58) a. *Those policemen are former

b. *policemen former
c. former policemen
(59) a. The present editors

b. The editors present
c. The editors are present
(60) a. a white lie

b. ??a lie white and without malice
c. ??a lie which is white
Nevertheless, Sandler and Arnold do not consider that the analysis of post-
positional adjectives can be reduced to that of predicative ones, since, as they
argue, the semantics of post-nominal adjectives is not identical to that of pred-

icative adjectives. As a matter of fact, this analysis would not be able to explain
the individual level vs. stage level dimension attributed to prenominal and
post-nominal adjectives, given that predicative adjectives can be either indi-
vidual level or stage level (Carlson 1977).3
3 Interaction between Lexical Aspect and Grammatical Aspect
In this section, I will analyze the lexical aspect of the verbs involved in resul-
tative participles and how the lexical aspect interacts with the interpretation
of the adjectival/adpositional participle. I will show that, even though verbs
involving an internal argument are the most frequent verbs giving rise to
adjectival/adpositional participles in Basque and Spanish (and generally, cross-
linguistically), unergative verbs and also transitive verbs where the theme of
predication is co-referent with the subject can also be found in this configu-
ration in Basque (and sometimes also in Spanish). As will be explained, the
occurrence of these verbs and interpretations is closely correlated with the
aspectual meaning of the participle.
3.1 Classes of Verbs Involved in Non-verbal Participles

3.1.1 Unaccusative/Transitive Verbs vs. Unergative Verbs
The standard analysis of resultatives assumes that adjectival participles are
derived from verbs that have an internal argument, that is, unaccusative (61a)
(62a) and transitive verbs (61b) (62b). The theme of predication, agreeing with
the copula, corresponds to the internal argument of the verb underlying in the
participle.
(61) a. Aulki-a jaus-i-ta dago.

chair-the fall-prt-res be.loc.3sg.abs
‘The chair is fallen.’
3 It must be highlighted that, although there are some semantic correspondences between
attributive participles with bare NP s and attributive participles with DP s in Basque, on the
one hand, and pre-nominal and post-nominal adjectives in English on the other, I do not
want to state that they are syntactically parallel, given that attributive participles with DP s
in Basque can be regarded as being really predicative, but post-nominal adjectives in English
do not need to be so.
104 berro
b. Mahai-a apur-tu-ta dago.

table-the break-prt-res be.loc.3sg.abs
(62) a. La silla está caída.

the.fem chair be.loc.3sg fallen.fem
b. La mesa está rota.

the.fem table be.loc.3sg broken.fem
Cross-linguistically, unergative verbs do not occur in this configuration,

because unergative verbs do not have an internal argument. Nevertheless, in
Basque, adjectival and adpositional participles built on unergative verbs are
sometimes acceptable. As will be shown in section 3.2.2, in the majority of
cases, the use of an unergative verb in this configuration goes hand in hand
with the experiential interpretation of the participle. In some particular cases,
however, adjectival and adpositional participles built on unergatives have sta-
tive interpretation, like in the following examples.
(63) a. Bazkal-du dut.

have.lunch-prt have.1sg.erg
‘I have eaten lunch’
b. Bazkal-du-ta nago
have.lunch-prt-res be.loc.1sg.abs
‘I am had lunch’
(64) a. Hori sines-ten dut.

that believe-imprf have.1sg.erg
b. Hala sines-tu-ta nago.

that.way believe-prt-res be.loc.1sg.abs
‘Like this I am believed’
In these examples, the theme of predication corresponds to the subject of the

event embedded under the participle; a subject that is marked with ergative
case in simple clauses (63a) (64a), like subjects of unergative and transitive
verbs. There are also some cases in Spanish where the theme of predication
is correferent with the subject of the unergative verb embedded in the adjecti-
val participle. These uses involve consuming verbs similar to that in (63): beber
‘to drink’ and comer ‘to eat’.
(65) Los niños están bebidos y comidos.

the.masc.pl children be.loc.3pl drunk.masc.pl and eaten.masc.pl
‘The children are drunk and eaten.’
The verbs beber ‘to drink’ and comer ‘to eat’ underlying the participle in the
example (65) do not accept direct objects in this configuration. For instance,
the example in (66), which includes direct objects, is not acceptable.
(66) *Los niños están bebidos agua y

the.masc.pl children be.loc.3pl drunk.masc.pl water and
comidos papilla
eaten.masc.pl porridge
‘The children are drunk water and eaten porridge.’
Like in the Basque examples, in the Spanish example of (65), the theme of pred-
ication (los niños ‘children’) corresponds to the unergative subject of the verbs
underlying the participles. Subjects of unergative verbs, as well as subjects of
transitive verbs, are considered to belong not the verbal projection, but intro-
duced by a Voice head projected on top of vP/VP (Kratzer 1996 and following
works). This is why, in an analysis where adjectival participles do not involve
external arguments (Levin & Rapapport 1986, Kratzer 1994 2000 among others),
the fact that predicates of this type are found in resultative participle configu-
rations is unexpected (see section 4.2 for a discussion on this aspect). On the
other hand, the ability of these verbs to occur in this configuration could also be
related to the fact that their subjects, apart from being interpreted as agentive
or initiators, can also be argued to undergo a change (see actually Ramchand
2008, where it is argued that the subject of consuming verbs is both an initiator
and an undergoer), and can, therefore, hold the consequent state denoted by
the participle.
3.1.2 Different Aspectual Classes of Verbs

Bosque (2014) observes for Spanish resultative participles that the participle
can have different interpretations depending on the aspectual class of the
underlying verb. He shows that, if the verb belongs to the class of change-of-
state verbs, manner verbs or incremental verbs, the resulting participle has a
bounded interpretation, meaning that the participle denotes a state obtained
106 berro
as a consequence of a past event. This is the case of the participles derived from
the verbs castigar ‘to punish’ and atrapar ‘to catch’.
(67) a. Un niño castigado

a.masc child punished.masc
‘A punished child’
b. Ladrones atrapados por la policía

thieves caught.masc.pl by the police
‘Thieves caught by the police’
In contrast, when the participle is derived from certain activity verbs (e.g. vigi-
lar ‘to guard’, buscar ‘to seek’, perseguir ‘to chase’), extent verbs (e.g. ocupar ‘to
occupy’) or verbs expressing constituency and other similar physical proper-
ties (e.g. formar ‘to constitute’), the meaning of the participle is unbounded:
the participle denotes a state that holds at the relevant evaluation time, but
does not imply a past event.
(68) a. Un edificio vigilado

a.masc building guarded.masc
‘A guarded building’
b. Una ciudad rodeada de montañas

a.fem city surrounded.fem of mountains
‘A city surrounded by mountains’
The state can have started sometime in the past, e.g. in (68a), but this is irrele-
vant for the meaning of the participle, particularly in (68b). The activity verbs
giving rise to the so-called unbounded past participles in Bosque (2014) also
include the so-called D(avidsonian)-states (Maienborn 2005 2007, Rothmayr
2009, Fábregas & Marín 2017, Berro 2015) (e.g. esperar ‘to wait’, gobernar ‘to gov-
ern’): verbs that share properties with both eventive and stative verbs. They are
eventive and, thus, behave like eventive verbs in some tests (they can occur in
the progressive, they can be infinitive complements of perception verbs, they
can be located in time and space etc.), but non-dynamic, and therefore act like
stative verbs in some other aspects (they give rise to an on-going, non-habitual
interpretation in the simple present, they are bad with dynamic adverbs etc.).
Extent verbs (Gawron 2009; Koontz-Garboden 2010) and verbs express-
ing constituency and other physical property are predicates that denote non-
dynamic eventualities, and which alternate with dynamic interpretations (e.g.
obstruct in English, Kratzer 2000). For example, the verb ocupar ‘to occupy’ is
ambiguous between a dynamic and a non-dynamic reading:
(69) Los estudiantes ocuparon / ocupaban la

the.masc.pl students occupy.3pl.pst / occupy.3pl.imperf the
sala
room
‘The students occupied the room.’
(70) La cama ocupaba toda la habitación

the.fem bed occupy.3sg.imperf all the room
‘The bed occupied the entire room.’
The unbounded reading of participles is derived from the non-dynamic inter-

pretation of the verbs, and these two interpretations can be thought to be
interrelated. For instance, the events denoted by the verb vigilar ‘to guard’ or
the non-dynamic interpretation of rodear ‘surround’ do not cause any change
on the internal argument. Consequently, the state they hold after the event of
guarding or surrounding has taken place is basically interpreted as a state that
may have started sometime in the past, but not as a consequent state reached
after an event has taken place.
The fact that adjectival participles built on certain verbs can have un-
bounded interpretation has been addressed in Anagnostopoulou & Samioti
(2014: 14). As explained in section 2.2, Embick (2004) proposes that adjecti-
val participles can be either stative or resultative. Stative adjectival participles
denote a simple state that is not the result of a previous event. In this sense,
stative participles seem to correspond roughly to Bosque’s unbounded partici-
ples, because in both types, a past event that gives rise to the state denoted by
the participle is not implied.4 Embick (2004) claims that resultative adjecti-
val participles are derived from a verb (a vP), whereas stative ones are built on
an acategorial Root. The presence of the semi-functional head v in the former
type of participles is responsible for the eventive interpretation of resultatives.
Anagnostopoulou & Samioti (2014) oppose this view, showing that, in Greek,
adjectival participles denoting characteristic states (without event implica-
tions) can also be built on verbs. In Greek, -tos ending participles have been
4 Note, however, that Bosque (2014) argues that unbounded participles are a subtype of resul-
tative participles (opposed to bounded ones). Embick (2004), in contrast, makes a distinction
between resultative participles and stative participles.
108 berro
claimed to be built at the Root level and not to entail the existence of a prior
event (Alexiadou & Anagnostopoulou 2008). However, certain -tos participles
involve the morphological exponents of verbalizers (e.g. -iz, -on, -a etc.), show-
ing that -tos participles can also be derived from verbs.
(71) Verbs Corresponding participle
a. axn-iz ‘steam’ axn-iz-tos ‘steaming hot’

b. koudoun-iz-o ‘ring (a bell)’ koudoun-is-tos ‘ringing’
c. vathoul-on-o ‘hollow out’ vathoul-o-tos ‘hollow’
d. if-en-o ‘weave’ if-an-tos ‘woven’
e. fit-ev-o ‘plant’ fit-ef-tos ‘planted’
Despite the presence of verbalizers, -tos participles do not have eventive mean-
ing and cannot be modified by adverbs (72), contrasting with -menos ending
participles (73).
(72) *Ta fita ine fitefta me diaforetika regalia

The plants are planted with different instruments
(73) Ta malia ine atsala htenismena

The hair are sloppily combed
Spanish unbounded participles (Bosque 2014) derived from activity verbs can
be modified by manner adverbs, a fact which shows that even though un-
bounded participles of this type do not imply a past event, they involve the
relevant functional projection that licenses manner adverbs.
(74) Un edificio vigilado cuidadosamente

a.masc building guarded.masc carefully
‘A carefully guarded building’
In contrast, participles derived from extent verbs and verbs expressing con-
stituency and other physical property behave differently. Bosque (2014) argues,
building on Gawron (2009), that these classes of verbs should be considered
eventive. Nevertheless, it seems that the participial forms derived from these
predicates do not accept easily manner adverbs, unlike participles derived from
non-dynamic activity verbs.
(75) a. Una ciudad rodeada (*cuidadosamente) de montañas

a.fem city surrounded.fem (*carefully) of mountains
b. Un apartamento habitado (??silenciosamente) por

an.masc apartment inhabited.masc (??silently) by
recien-casados
newly-weds
c. Un equipo formado (*detalladamente) por tres

a.masc team formed.masc (*fine.grained) by three
jugadores
players
If these predicates involved an event, it is not clear why their participial forms
do not license manner modifiers. Moreover, if they were eventive in their non-
dynamic interpretation, they would be grammatical in the progressive, con-
trary to fact:
(76) a. Las montañas rodean la ciudad

the.fem.pl mountains surround the.fem city
b. #Las montañas están rodeando la ciudad

#the.fem.pl mountains are surrounding the.fem city
(77) a. Los reciencasados habitan el apartamento

the.masc.pl newlyweds inhabit the.masc apartment
b. *Los recién-casados están habitando el

the.masc.pl newly-weds are inhabiting the.masc
apartamento
apartment
(78) a. Tres jugadores forman el equipo

three players form the.masc team
b. #Tres jugadores están formando el equipo

#three players are forming the.masc team
110 berro
3.1.3 Morphological Evidence of “Verbalness” in Stative Adjectival

Participles in Basque
In Basque, we also find morphological evidence of verbs in adjectival and adpo-
sitional participles that are interpreted in a non-resultative way. This happens
in participles derived from extent verbs and verbs expressing constituency and
physical properties, like in Spanish.
(79) Hiri bat mendi-ek ingura-tu-a

city a mountain-erg.pl surround-prt-res
‘A city surrounded by mountains’
(80) Talde bat lau jokalari-k osa-tu-a

team a four players-erg form-prt-res
‘A team formed by four players’
As explained at the beginning of section 2, Basque resultative participles are

formed attaching the suffixes -a/-ta or -rik to the bare participle used in sim-
ple perfect clauses. Bare participles are generally headed by -tu or -i, being -tu
the only productive suffix in contemporary Basque. Berro (2015) argues that
-tu is the exponent of an eventivizer, since it can be used very productively to
form new predicates (see de Rijk 2008). These are some examples of -tu headed
verbs derived from an adjectival root (81a), an adverbial root (81b), a nominal
root (81c) and an adposition (81d).
(81) a. gorri ‘red’ gorri-tu ‘to redden’

b. hurbil ‘near’ hurbil-du ‘to approach, to get closer’
c. ama ‘mother’ ama-tu ‘to become a mother’ ‘to turn sb into a
mother’
d. etxe-ra ‘to home’ etxe-ra-tu ‘to go home’ ‘to carry sb home’
The -tu suffix only attaches to the Root when the Root is being used as a verb. In
some cases, the Root changes its final vowel depending on its categorial envi-
ronment. Artiagoitia (2004) addresses this phonological change affecting the
final vowel of certain Roots in verbal environments (where those ending in /e/,
/o/ or /u/ turn to /a/) and proposes that the /a/ ending variants have verbal
category, that is to say, involve the projection of a verbal layer.5
5 This phonological change is similar to a wider process of derivative morphology mentioned

in Hualde (1991), and termed Minor Apocope (MA) in De Rijk (2008). It is a lowering process,
(82) -tu derived verbs (de Rijk 2008: 150–152; Artiagoitia 2004: 151)
neke ‘effort, fatigue’ nekatu ‘to get tired’
aipu ‘citation’ aipatu ‘to cite’
zoro ‘crazy’ zoratu ‘to go crazy’
gorroto ‘hatred’ gorrotatu ‘to hate’
errespetu ‘respect’ errespetatu ‘to respect’
When they are headed by -tu, the Roots end in /a/, showing that they are being
used in a verbal environment, namely, in complement position of a verbalizer.
The two non-resultative (or unbounded, in terms of Bosque 2014) adjectival
participles illustrated in (79) and (80) are also examples of this phonological
change. The verb inguratu ‘to surround’ and osatu ‘to constitute’ are derived
respectively from the nominal Root inguru ‘surrounding’ and the adjectival
Root oso ‘whole, complete’. Both Roots change their final vowel to /a/ when
they are used as verbs.
(83) Root Verb Non-verbal participle
a. inguru ‘surrounding’ ingura-tu ‘to surround’ ingura-tu-a/ta/rik

b. oso ‘whole, complete’ osa-tu ‘to form’ osa-tu-a/ta/rik
The adjectival participles in (79) and (80) involve the /a/ ending Roots, as
well as the -tu suffix, showing that both participles are built on verbs, and not
directly on acategorial Roots. In other words, these participles are derived from
verbs, and not from Roots. Thus, these pieces of data show that stative partici-
ples (those without event implications) can be derived from verbs, supporting
the claim made in Anagnostopoulou & Samioti (2014), and against the oppo-
sition between stative and resultative participles in terms of the absence or
presence of v, as presented in Embick (2004).
3.2 Aspectual Interpretations of the Participle

Adjectival (and adpositional) participles typically have the meaning of resulta-
tive perfects (Kratzer 2000): they are derived from telic verbs and they have a
affecting non-high vowels of bisyllabic Roots. In the context of verbalizing derivational mor-
phemes (80), however, it can also affect trisyllabic Roots and Roots ending in the high vowel
/u/, as can be seen in the examples.
112 berro
meaning where the result of the event holds at the relevant evaluation time.
However, in some cases, adjectival (and adpositional) participles in Basque
(and sometimes also in Spanish) can have a broader use, and cover meanings
pertaining to the experiential perfect, as observed for Basque by Krajewska
(2012 2013ab). The experiential perfect can be built on any type of verb. It
asserts that the subject has a certain experience related to the fact that the pred-
icate has held at one or more intervals in the relevant evaluation time (Comrie
1976; Iatridou, Anagnostopoulou & Izvorski 2001/2003 among others). Finally,
as already advanced in section 3.1.2 and 3.1.3, adjectival (and adpositional) par-
ticiples can also have a continuous interpretation (Bosque 2014), also called
persistent or universal (Comrie 1976; Iatridou et al. 2001/2003; Pancheva 2003),
where the predicate holds from a certain point in the past till the relevant eval-
uation time. In the literature, it has been observed that the universal perfect is
obtained from unbounded predicates such as states and progressives. In con-
figurations involving adjectival (adpositional) participles, I have shown that
it arises when the participle denotes a non-resultative state—unbounded in
terms of Bosque (2014) or stative in terms of Embick (2004)
3.2.1 Resultative Interpretation vs. Continuous Interpretation

According to Kratzer (2000), resultant state participles involve an aspectual
operator that maps properties of eventualities into properties of times. More
specifically, the aspectual operator present in resultant state participles asserts
that a property holds at a time interval that is preceded by the running time
of a completed event. Thus, the meaning of resultant state participles is sim-
ilar to the perfect of result. Gehrke (2011) (and also Bosque 2014) notes that
the state asserted by the participle in resultative participles does not need to
be the result of a change of state, but should be rather regarded as the con-
sequent or subsequent state of a previous event. They are compatible also
with activity verbs such as push or pet as long as they get a ‘job is done’ or
‘that’s over’ interpretation (Kratzer 2000). Thus, all telic (or telicized) verbs like
accomplishments or verbs allowing an inchoative interpretation are also fine
in resultative participles.
(84) a. Gutun-a idatz-i-a / irakurr-i-a dago

letter-the write-prt-res / read-prt-res be.loc.3sg.abs
‘The letter is written/read.’
b. Mutiko-a haserre-tu-a dago

boy-the become.angry-prt-res be.loc.3sg.abs
‘The boy is angered’
Kratzer (2000) identifies another type of adjectival participles, different

from resultant state participles: target state participles. She borrows both terms
from Parsons (1990) and uses them to differentiate between states that are non-
reversible (resultant states) from those that are reversible (target states). As
shown in Kratzer (2000), target state participles differ from resultant state ones
in that they are compatible with the German adverb immer noch ‘still’. Accord-
ing to Kratzer (1998, 2000), target state participles are only compatible with
verbs that are lexically specified with an eventive and a stative component. As
can be seen in the examples, the adverb oraindik in Basque is only compatible
with participles derived from change of state verbs like puztu ‘to inflate’ or ireki
‘to open’, but is non-acceptable with participles built on incremental verbs like
irakurri ‘to read’. The same contrasts are observed in Spanish with the adverb
todavía ‘still’.
(85) a. Puxik-a-k oraindik puz-tu-a-k daude.

balloon-the-pl still inflate-prt-res-pl be.loc.3pl.abs
‘The balloons are still inflated.’
b. Ate-a oraindik ireki-ta dago.

balloon-the still open.prt-res be.loc.3pl.abs
‘The door is still open.’
(86) *Gutun-a oraindik irakurr-i-ta dago.

letter-the still read-prt-res be.loc.3sg.abs
‘The letter is still read.’
In Kratzer’s account (see also Alexiadou et al. 2015), the participial morpheme
is meaningless in target state participles; its only function is to categorize the
syntactic object as an adjective and to existentially quantify the Davidsonian
argument of a predicate that has additionally a target state.
Embick (2004) subsumes Kratzer’s target state participles into stative par-
ticiples—those that denote a characteristic state, without event implications.
Nevertheless, Alexiadou et al. (2014) argue, following previous work like Anag-
nostopoulou (2003) and Alexiadou & Anagnostopoulou (2008), that target
state participles can have event implications, and consequently, should be con-
sidered to be derived from vP.
With respect to unbounded participles understood as in Bosque (2014)—
those that are interpreted as underived states, without a causing previous
event, see section 3.1.2—, their aspectual interpretation is not similar to the
perfect of result, but rather seems to be closer to the continuous (Bosque
114 berro
2014), also called persistent (Comrie 1976) or universal perfect (Iatridou et al.
2001/2003; Pancheva 2003). The interpretation of the universal perfect arises
when participles are derived from unbounded predicates like states and pro-
gressives, and especially when a durative temporal adverbial like since-or for-
adverbs are used.
(87) I have lived here since 2016.
The example in (87) asserts that the property expressed by the predicate (living
here) holds since 2016. Thus, the property does not hold at a time following the
completion of the event denoted by the predicate, as in the perfect of result,
but holds at the same interval as the predicate.
The interpretation obtained in adjectival participles derived from non-
dynamic eventualities like vigilar ‘to guard’ or perseguir ‘to chase’ in Spanish,
or extent verbs and verbs expressing physical properties is similar.
(88) Un criminal perseguido por la policía

a.masc criminal chased.masc by the police
‘A criminal chased by the police’ (Bosque 2014: 51)
As pointed out in Bosque (2014), the aspectual interpretation of (88) is continu-

ous: the state predicated of the theme does not hold after the event denoted by
the verb is completed, but the interval in which the state holds is simultaneous
with the running time of the event.
3.2.2 Experiential Interpretation

As noted at the beginning of section 3.2, Basque and Spanish adjectival par-
ticiples can also have experiential interpretation. This interpretation is partic-
ularly frequent in Basque (Hualde et al. 1994; Krajewska 2012 2013ab), when
the theme of predication is correferent with the subject initiating the event
embedded in the participle. Nonetheless, Basque is not alone in this aspect
since Spanish adjectival participles can also license experiential interpretation
in particular circumstances, as pointed out in Bosque (2014).
In Basque, in the majority of configurations where an adjectival or adposi-
tional participle is used, the state denoted by the participle is predicated about
the direct object or the intransitive subject of the verb underlying the partici-
ple.
(89) Mahai-a apur-tu-ta dago.

table-the break-prt-res be.loc.3sg.abs
(90) Aulki-a jaus-i-ta dago.

chair-the fall-prt-res be.loc.3sg.abs
This interpretation is highly predominant when the verb embedded in the par-
ticiple is a transitive or intransitive change of state verbs, like in (89) and (90).
Actually, a participle derived from a change of state verb like hondatu ‘break’ is
quite hard with a frequency adverb like askotan ‘many times’.
(91) ??Auto-a askotan honda-tu-ta dago.

car-the many.times break-prt-res be.loc.3sg.abs
‘The car is broken many times.’
In contrast, when the participle is predicated about the subject initiating the
event, the experiential or perfect readings of the participle arises naturally. The
experiential interpretation contrasts with the resultative one in that the out-
come of the event does not necessarily hold at the relevant evaluation time.
Rather than the consequent state, it is the event itself that is focused. These are
some examples taken from Krajewska (2013):
(92) Meza-ra? —ez, entzun-ik nago.

mass.the-all? no, listen.prt-res be.loc.3sg.abs
‘To mass?—no, I have already listened’, lit. ‘I am already listened.’ (Iri-
garay)
(93) —Ezagu-tzen zenuen hiri? —ez; ez

—know-imprf have.2sg.erg.1sg.abs.pst city? No; no
nintzen inoiz izan-a.
be.1sg.abs.pst never be.prt-res
‘—Did you know the city?—no, I had never been there.’, lit. ‘I was never
been.’ (Satrustegi)
(94) Ni askotxo ikus-i-a nago.

I.abs a.lot see-prt-res be.loc.1sg.abs
‘I have seen many things’, lit. ‘I am a lot seen.’ (Anabitarte)
116 berro
(95) Asko-tan ene baitan erran-ik nago: “Zenbat itz

A.lot-ine my ine say.prt-res be.loc.1sg.abs: “how.many words
deusik ez errateko”.
nothing no to say”
‘I have said many times to myself: “How many words to say nothing”.’, lit.
‘I am a lot said many times to myself: …’ (Irigaray)
In examples (92), (94) and (95), the theme of predication (the absolutive argu-
ment) corresponds to the subject of the previous event (of ‘listening’ in (92), of
‘seeing’ in (94) and of ‘saying’ in (95)), so that the action affects mainly the state
held by the subject (Krajewska 2013). As a consequence of being the subject of
the event (once or more times), the theme of predication accumulates an expe-
rience. The experiential interpretation is particularly salient when the copula is
in the past tense, and moreover when the copula is *edun or eduki ‘have’ instead
of izan ‘be’. The latter configuration corresponds to the possessive resultative
commented in section 2.3. As mentioned, this configuration has been attested
diachronically in many languages as the source of the perfect. For instance, in
the past tense, the configuration has the meaning of a past perfect (Hualde et
al. 1994: 145–146); a past event completed before some other past event.
(96) Etxe-tik joan-a nintzen (zu iritsi

home-abl go.prt-res be.1sg.abs.pst (you.abs arrive.prt
zinenerako).
be.2sg.abs.pst-by.the.time)
‘I was gone by the time you arrived.’
(97) Pelikula ikus-i-ta daukat.

film.the see-prt-res have(eduki).3sg.abs.1sg.erg
‘I have the film seen.’
The fact that these participles in Basque can have experiential interpretation
can suggest that these are really perfect constructions, where the participles
would be verbal rather than adjectival or adpositional. Krajewska (2012 2013ab)
actually claims that Basque -a/-ta and -rik ending participles are in the mid-
dle of the grammaticalization path towards a new perfect. In her diachronic
study, Krajewska mentions that, in the historical evolution of the Basque lan-
guage in the last centuries, the majority of verbs used in the periphrases involv-
ing -a/-ta and -rik participles are telic change of state verbs where the state
is predicated about the object of the event or the subject of an intransitive
event. Nevertheless, in the cases where other type of predicates are used—
such as mental, perception and communication verbs, as well as stative and

atelic verbs—perfect and experiential interpretations have become more com-
mon in the last centuries. According to her, these configurations have not yet
become full-fledged perfects, as the frequency of the latter type of verb classes
has not increased in comparison to change of state verbs, which are still highly
predominant. Krajewska then concludes that the configuration involving -a/-ta
and -rik participles is still not a perfect, but that it is in the middle of a gram-
maticalization path, and that it is acquiring properties usually attributed to the
perfect.
This issue raises a number of questions regarding the structural composi-
tion of -a/-ta and -rik participles. Are -a/-ta and -rik participles with resulta-
tive meaning structurally different from those with experiential meaning? In
relation to this, is there a categorial difference between resultative and expe-
riential participles? Do the former ones occur in bi-clausal configurations and
the latter ones in monoclausal ones? As for the first question, my suggestion
is that there has to be some structural and/or semantic difference between
them, as they have different interpretations. Regarding the categorial or bi-
clausal/monoclausal configuration, it seems that even if -a/-ta and -rik partici-
ples get experiential interpretation, they are still in a bi-clausal structure. The
bi-clausal analysis of experiential participles is supported by two facts: (i) they
can occur with the copulas egon ‘(the stage-level) be’ and eduki ‘have’, which
are never used as auxiliaries, and (ii) they combine with intransitive copulas
(izan ‘be’ and egon ‘(the stage-level) be’) also when the underlying verb of the
participle is transitive or unergative (92) (94) (95), a fact that is not attested in
perfect monoclausal configurations.
Be it as it may, Basque -a/-ta and -rik participles are not alone in the ability
to get experiential interpretation. As a matter of fact, Bosque (2014) shows that
accompanied by frequency adverbs, attributive participles can have experien-
tial interpretations.
(98) Un coche averiado en muchas ocasiones no es una buena

a.masc car broken.masc in many occasions no is a good
inversión.
investment
‘A car many times broken down in not a good investment.’ (Bosque 2014:
51)
Participles occurring in predicative position with the copula estar ‘(the stage-
level) be’ can also have this interpretation in certain cases, although it is not a
generalized tendency.
118 berro
(99) Este ordenador está arreglado varias veces.

this.masc computer be.loc.3g fixed.masc several times
‘This computer is fixed several times.’
(100) ?? Este coche está averiado varias veces.

this.masc car be.loc.3g broken.masc several times
‘This computer is fixed several times.’
In the examples (98) and (99), the participle has experiential interpretation,
rather than resultative, since a car cannot be in a state of being broken many
times and a computer cannot be in a state of being fixed many times. Still, the
experiential interpretation of participles in this context is not always available
(100). It is clear that the experiential interpretation is not as easily obtained as
in perfect clauses. For instance, in a perfect clause in Spanish, we can get an
experiential interpretation when the participle is derived from a stative verb
(e.g. estar enfermo ‘be sick’)
(101) He estado enfermo antes.

have.1sg been sick.masc before
‘I have been sick before.’
Participles occurring in predicative position accompanied by the copula estar

cannot be built on stative verbs (unlike in Basque (93)), and thus, they cannot
have an experiential interpretation.
(102) *María está estada enferma

Mary be.loc.3sg been.fem sick.fem
(103) *Este asunto está conocido por todos

this issue be.loc.3sg known.masc by all
‘This issue is known by everybody.’
As in Basque, if the participle combines with the transitive copula tener ‘have’,
an experiential interpretation is much easier to obtain.
(104) Tengo la película vista

have.1sg the.fem film seen.fem
‘I have the film seen.’
(105) Tengo el libro leído

have.1sg the.masc book read.masc
‘I have the book read.’
McFadden & Alexiadou (2010) analyze the structural differences between ex-
perientials and resultatives, particularly in the context of German and older
English have and be auxiliary alternation (see also Alexiadou 2015). They pro-
pose that, in resultative adjectival participles, the verb is headed by a resultative
Asp(ectual) head that introduces a state that is the result of a prior event. AspP
is, in turn, the complement of the copular verb -be-. In experientials, on the
other hand, the inflected element -have- is an auxiliary located in an aspec-
tual Perf(ect) head, which is below T, and which selects an Asp head. In their
account, this configuration has the meaning of a perfect.
4 Event-Related Modification
In this section, I will deal with the modification of the event underlying the
participle. The ability of adjectival participles to accept different types of event-
related modifiers is used in the literature as a diagnostic to find out the struc-
tural composition of adjectival participles; e.g. to test the presence/absence of
a verb phrase (vP), a Voice phrase and also an aspectual (Asp) phrase (see also
section 5.2).
The fact that there are different classes of adjectival participles has been
noted from early on. Starting from Wasow (1977), deverbal -ed forms in English
have been claimed to be formed at different levels of the language architec-
ture. Wasow (1977) proposed that adjectival passives are created in the lexicon,
whereas verbal passives are created in the syntax. Adding a further distinc-
tion to this two-way classification, Kratzer (1994) claimed that adjectival pas-
sives could be classified into two groups: phrasal passives and lexical passives.
According to her, adjectival participles, which in German are distinguished by
the use of the auxiliary sein ‘be’, are headed by an adjectival passive morpheme
that may be merged into the structure at different positions. In phrasal passives,
the adjectivizing head merges to the verbal phrase (VP in her account), whereas
in lexical passives, it merges to the verb (V). As a consequence, the resulting
participle has different syntactic properties, and in this way, she accounts for
the contrasting behavior between the two, for example in allowing manner
adverbials and the ability to undergo -un prefixation. On the one hand, phrasal
adjectival participles allow modifying adverbs (106a)—unlike underived adjec-
tives (106b)–.
120 berro
(106) a. Das Haar war ziemlich schlampig gekämmt.

The hair was rather sloppily combed
‘The hair was rather sloppily combed.’
b. *Das Haar war ziemlich schlampig fettig

The hair was rather sloppily greasy
‘*The hair was rather sloppily greasy’
On the other hand, lexical participles can be negated by the un- suffix and can-
not be combined with manner adverbs. Thus, lexical participles pattern with
underived adjectives and contrast with phrasal participles.
(107) *Das Haar war hässlich ungekämmt

The hair was ugly uncombed
‘*The hair was ugly uncombed’
Another test that is used to explore the structural composition of adjectival par-
ticiples concerns the temporal and spatial location of the event embedded in
the participle. If participles are compatible with temporal and spatial adverbs
that locate the event at a time and place that are different from those of the
state, that means that the event is actually instantiated (Gehrke 2011, 2015). A
modifier like recently or a time-frame adverb like three days ago can be used
with this purpose. Gehrke (2011: 246) shows that in German adjectival partici-
ples, recently can only modify the state rather than the event (108), and that the
adverb three days ago is non-acceptable (109) (von Stechow 1998; Rapp 1996,
1997).
(108) Die Tür war kürzlich geöffnet.

the door was recently opened
‘The door was in the opened state recently, but probably no longer.’
(NOT: the door is in the opened state, the opening took place recently)
(109) a. *Der Computer ist vor drei Tagen repariert

the computer is before three days repaired
Intended: ‘The computer is repaired three days ago’
b. Der Computer ist seit drei Tagen repariert.

the computer is since three days repaired
As can be seen, only a temporal adverb that modifies the state, like seit drei
Tagen ‘since three days ago’, is compatible with German adjectival participles.
In the next sections, I will argue that Basque and Spanish are more flexible than
German adjectival participles in this aspect, as they are compatible with this
kind of modification.
4.1 Event-Related Modification in Basque and Spanish

Like in German, adjectival (and adpositional) participles in Basque are com-
patible with manner adverbs. As can be seen in the examples below, Basque
resultative participles are compatible with modifiers like arretarik gabe ‘slop-
pily’, kontu handiz ‘carefully’ an eskuz ‘by hand’.
(110) a. Auto-a kontu handi-z konpon-du-a/-ta/-rik

car-the.abs care big-instr fix-prt-res
dago.
be.loc.3sg.abs
‘The car is fixed carefully’
b. Auto-a arreta-rik gabe konpon-du-ta

car-the.abs attention-part without fix-prt-res
dago.
be.loc.3sg.abs
‘The car is fixed sloppily’
c. Horma hau edozein modu-tan margo-tu-ta dago.

wall this.abs any way-ine fix-prt-res be.loc.3sg.abs
‘This wall is painted sloppily’
d. Estatua hori esku-z egin-a/da/ik dago.

statu that.abs hand-instr do.prt-res be.loc.3sg.abs
‘This statu is done by hand’
The compatibility with manner modifiers show that adjectival and adposi-
tional participles in Basque are actually phrasal, derived from a vP which
denotes an event. Spanish behaves similarly, as can be observed in the examples
below, where the adjectival participle combines with a manner adverb ending
in -mente.
(111) Un edificio concienzudamente vigilado

a.masc building thoughfully guarded.masc
(Bosque 2014: 54, ex. 21b)
122 berro
(112) (Se nota que) este árbol está cortado

(it isnoted that) this.masc tree be.loc.3sg cut.masc
rápidamente.
rapidly
‘It is noted that this tree has been cut rapidly.’, lit., ‘this tree is cut
rapidly’
(113) El pelo está peinado descuidadamente.

the.masc hair be.loc.3sg combed.masc sloppily
‘The hair is combed sloppily,’
The manner modifiers used in the above examples can be argued to be describ-
ing some aspect of the event that is relevant in the state denoted by the par-
ticiple. Actually, in example (112) the initial se nota que ‘it is noted that’ has
been introduced in order to make the sentence more natural. However, in other
cases, the adverb seems to be modifying the event itself, rather than the state.
(114) Un niño injustamente castigado

a.masc kid unfairly punished.masc
(adapted from Bosque 2014: 54, ex. 21a)
Contrasting with German, both Basque and Spanish allow temporal and spatial
modification of the event embedded in the participle.
(115) Aspaldi geldi-tu-ta gaude.

long.ago meet-prt-res be.loc.1pl.abs
‘We did an appointment long ago.’, lit. ‘we are done an appointment
long ago.’
(116) a. Pastel hori duela hiru egun egin-a/-da/-ik dago.

cake this.abs ago three days do.prt-res be.loc.3sg.abs
‘That cake is done three days ago.’
b. Pastel hori atzo goiz-ean egin-a/-da-/ik

cake this.abs yesterday morning-ine do.prt-res
dago.
be.loc.3sg.abs
‘That cake is done yesterday morning.’
(117) a. Pastel hori Lourdes-en etxe-an egin-a/-da/-ik

cake this.abs Lourdes-gen house-ine do.prt-res
dago.
be.loc.3sg.abs
‘This cake is done in Lourdes’ house.’
b. Makina hau Danobat-en egin-a/-da/-ik dago.

machine this.abs Danobat-ine do.prt-res be.loc.3sg.abs
‘This machine is done in Danobat’
Temporal and spatial modification of the event is also possible in Spanish

adjectival participles, both in predicative and attributive position. As can be
seen in examples (119), (120) and (121), the adjectival participle occurs with the
copula estar and it is modified by temporal and spatial adverbs that specify
when and where the event (bringing the state asserted by the participle) has
taken place.
(118) El ordenador está arreglado recientemente.

the.masc computer be.loc.3sg fixed.masc recently
‘The computer is fixed recently.’
(119) El ordenador está arreglado hace poco en la tienda de

the computer be.loc.3sg fixed.masc recently in the shop of
un amigo mío.
a friend mine
‘The computer is fixed recently in the shop of a friend of mine.’
(120) Este edificio está construido en 1800.

this.masc building be.loc.3sg built.masc in 1800
‘This building is built in 1800.’
(121) Un artículo publicado ayer en la prensa de la

an.masc article published.masc yesterday in the press of the
tarde.
evening
‘An article published yesterday in the evening press.’ (Bosque 2014: 48,
ex. 10b)
Thus, both Basque and Spanish allow temporal and spatial modification of
the event. In this aspect, they patter with Greek, which, unlike German, also
accepts temporal adverbs like three days ago and spatial modification.
124 berro
(122) a. To pc itan diorthomeno prin tris meres. Greek

the pc was repaired three days ago
‘The pc was repaired three days ago.’
b. To pedi itan htenismeno sto banio.

the child was combed in.the bathroom
‘The child was combed in the bathroom.’
(123) a. *Der Computer ist vor drei Tagen repariert. German

the computer is before three days repaired
‘The computer is repaired three days ago.’
b. ??? DasKind war im Badezimmer gekämmt.

the child was in.the bathroom combed
‘The child was combed in the bathroom.’
4.2 Presence of Voice

In the study of adjectival participles, another aspect that has generated much
discussion is related to whether adjectival participles do actually involve the
projection of Voice, the head introducing the external argument (Kratzer 1994,
1996). Kratzer (1994) claimed that adjectival participles are different from ver-
bal ones precisely in this aspect: unlike verbal passive participles, adjectival
ones are not VoiceP-derived, and thus, an implicit external argument, e.g. an
agent, is totally absent from them.
Later works have, nevertheless, questioned this claim. Anagnostopoulou
(2003) and Alexiadou & Anagnostopoulou (2008) showed that VoiceP can be
projected in Greek adjectival participles. For instance, in Greek resultant state
participles, an agent can be explicitly introduced in an apo-phrase, similar to a
by-phrase in English.
(124) To psari itan tiganismeno apo tin Maria Greek

the fish was fried by the Mary
‘The fish was fried by Mary’
Recent studies on English and German have argued that VoiceP can be involved
also in English and German adjectival participles (e.g. Alexiadou et al. 2014;
Alexiadou et al. 2015 for German; McIntyre 2013; Bruening 2014 and Alexi-
adou et al. 2014 for English). In the literature, the absence of the external
argument in adjectival participles has been analyzed using several diagnos-
tics, such as the inability to control into purpose clauses, and the absence of
the disjoint reference effect. Nevertheless, the lack of the disjoint reference
effect as a test to prove the absence of an implicit external argument in Ger-
man (Kratzer 1994) has been called into question in McIntyre (2013), Alexiadou
et al. (2014) and Alexiadou et al. (2015). Particularly, it seems that when non-
reflexive verbs are used, adjectival participles give rise to the disjoint effect,
just like verbal passive participles, so that this diagnostic does not prove that
VoiceP is not projected in German adjectival participles. On the other hand, the
inability of adjectival passives to control into purpose clauses in German and
English has also been counter-evidenced in McIntyre (2013) and Alexiadou et
al. (2014).
In the following two subsections, I will provide several pieces of data that
indicate than an (implicit) external argument is also present in adjectival (and
adpositional) participles in Basque and Spanish. Particularly, I will show that
adjectival (and adpositional) participles are compatible with modifiers that
rely on the presence of an agent.
4.2.1 Subject-Oriented Modification and Control into Purpose Clauses

In both Basque and Spanish, adjectival (and adpositional) participles are com-
patible with controlled purpose clauses, where the implicit external argument
controls the verb in the purpose clause (125ab). Additionally, agent-related
modifiers nahita ‘on purpose’ in Basque (126a) and a propósito ‘on purpose’ in
Spanish (126b) are also naturally admitted.
(125) Control into purpose clauses

a. Horma-k beltz-ez margo-tu-ta daude
wall.the.abs-pl black-instr paint-prt-res be.loc.3pl.abs
gela ilunago bihurtze-ko.
room.the.abs darker turn-dest
‘Walls are painted in black in order to make the room darker.’
b. Las paredes están pintadas de negro para

the.fem.pl wall.pl be.loc.3pl painted.fem.pl of black for
hacer la habitación más oscura.
make the room more dark
‘The walls are painted in black in order to make the room darker.’
(126) Modification with on purpose

a. Auto-a nahita konpon-du-ta dago.
car-the.abs on.purpose fix-prt-res be.loc.3sg.abs
‘The car is fixed on purpose’
126 berro
b. Horma-k nahita daude beltze-z

wall.the.abs-pl on.purpose be.loc.3pl.abs black-instr
margo-tu-ta.
paint-prt-res
‘Walls are painted in black on purpose’
c. El coche está arreglado a propósito.

the.masc car be.loc.3sg fixed.masc to purpose
‘The car is fixed on purpose.’
d. Las paredes están pintadas de negro a

the.fem.pl wall.pl be.loc.3pl painted.fem.pl of black to
propósito.
purpose
‘The walls are painted in black on purpose.’
4.2.2 Acceptance of Ergative Subjects and por-phrases

As advanced in section 2.3, in Basque adjectival participles, agent arguments
can occur explicitly marked with ergative case, the same case assigned to tran-
sitive and unergative subjects in monoclausal configurations. The ergative sub-
ject occurring with adjectival participles has several restrictions, though, that
clearly differentiate them from the ergative subjects of monoclausal configura-
tions.
(127) Gutun-a guraso-ek sina-tu-a da.

letter-the parent-erg.pl firm-prt-res be.3sg.abs
‘The letter is firmed by the parents.’
(128) Puxika-k Ane-k puz-tu-a-k dira.

balloon.the-pl Anne-erg inflate-prt-res-pl be.3pl.abs
‘The balloons are inflated by Anne.’
The acceptance of an ergative subject interpreted as the agent of the event

embedded in the participle is favored when the copula selected is izan ‘be’
(instead of egon ‘(the stage level) be’) and the resultative suffix is the adjec-
tival -a. In this case, the ergative subject does not agree with the copula and it
cannot scramble to the front of the theme of predication or to the right of the
copula (Ortiz de Urbina & Uribe-Etxebarria 1991).
(129) Ziurtagiri-a zuzendari-a-k sina-tu-a da.

certificate-the.abs director-the-erg sign-prt-res be.1sg.abs
‘The certificate is signed by the director’
(130) (*Zuzendari-a-k) ziurtagiri-a (zuzendari-a-k)

(*director-the-erg) certificate-the.abs (director-the-erg)
sina-tu-a da (*zuzendaria-a-k)
sign-prt-res be1sg.abs (*director-the-erg)
‘The certificate is signed by the director’
An agent argument can also be introduced in Spanish with a por-phrase in

participles with both a bounded and unbounded interpretation, in terms of
Bosque (2014). As can be seen, por-phrases are acceptable in both attributive
and predicative participles.
With bounded interpretation, we find the following examples:
(131) Ladrones atrapados por la policía

thieves caught.masc.pl by the police
(132) Aviones derrivados por el enemigo

aircrafts shot.down.masc.pl by the enemy
(133) Un niño castigado por su maestro

a.masc child punished.masc by his teacher
(134) Este niño está castigado por su maestro

this.masc child be.loc.3sg punished.masc by his teacher
(135) Este edificio parece diseñado cuidadosamente por un

this.masc building looks designed.masc carefully by an
architecto surrealista en una noche de insomnia
architect surrealist in a night of insomnia
‘This building seems designed by a surrealist architect in a sleepless
night.’ (Bosque 2014: 54, ex. 22)
And these with unbounded interpretation:
(136) Un criminal perseguido por la policía

a.masc criminal chased.masc by the police
128 berro
(137) Un apartamento habitado por reciéncasados

an.masc apartment inhabited.masc by newlyweds
(138) El tejado está protegido por una gruesa capa

the.masc roof be.loc.3sg protected.masc by a thick layer
de paja
of straw
‘The roof is protected by a thick layer of straw’ (Bosque 2014: 65, ex.
50a)
The data provided clearly indicates that Voice is projected in both Basque and
Spanish. Nevertheless, the acceptability of por-phrases in Spanish is appar-
ently more restricted than the ergative subjects in Basque, as noted by Gehrke
& Marco (2014). According to them, por-phrases are grammatical when they
modify the state denoted by the participle, rather than the event (i.e. when the
participle has unbounded interpretation) or, alternatively, when they modify
an event-kind, rather than an event token (see section 5). These restrictions do
not seem to hold in the Spanish variety spoken in the Basque Country, though.
5 Event-Related Modification: Restrictions
In section 4, I have shown that Basque and Spanish adjectival (and adposi-
tional) participles are compatible with event-modifiers of different sorts. They
accept manner modifiers that describe the event embedded under the par-
ticiple (such as kontu handiz ‘carefully’ in Basque and cuidadosamente ‘care-
fully’ in Spanish), and allow temporal and spatial adverbs that locate the event
at a different time and place from those of the state denoted by the partici-
ple. Additionally, both Basque and Spanish accept agent-oriented modifiers
and are compatible with controlled purpose clauses, facts that point out that
an implicit external argument is syntactically active in Basque and Spanish
adjectival (and adpositional) participles. Finally, explicit ergative arguments
that are interpreted as the agent of the event embedded in the participle are
also allowed in Basque, although these arguments are much better accepted
when the copula is izan ‘be’ and the participial ending is the adjectival -a
(and not -ta or -rik). In Spanish, too, por-phrases are accepted in both attribu-
tive and predicative adjectival participles, a fact which indicates that Voice
is projected within the participle. Nevertheless, as we are going to see in the
next subsections, the occurrence of por-phrases is not apparently totally unre-
stricted.
Gehrke (2011 and subsequent work) observes that by-phrases in German are
possible as long as they belong to the consequent state, rather than to the
underlying event, or if they have an impact on the underlying event that is
still visible in the consequent state. She refers to the latter event as an event
kind. In her analysis, adjectival participles denote consequent states of event
kinds. Event kinds cannot be modified by temporal or spatial adverbs because
event kinds lack spatiotemporal manifestation. Thus, the only modifiers that
are accepted are those which describe the consequent state, or otherwise, are
interpreted as kind-related; modifiers that create a new subkind of the event
by narrowing the event kind, and thus, having an impact on the consequent
state.
The type of manner adverbs allowed in Spanish adjectival participles seem
to point in this direction (section 4.1). Manner adverbs like cuidadosamente
‘carefully’, descuidadamente ‘sloppily’ are adverbs that, although they describe
the event, shape also the form of the consequent state. In the case of rápida-
mente ‘quickly’ in (112), the adverb is better accepted with the introduction of se
nota que ‘it is noted that’ at the beginning of the example, another piece of evi-
dence which shows that the relevance of the event-manner in the consequent
state is crucial for the acceptance of this adverb.
Nevertheless, we have seen that Spanish, like Basque, allows spatiotemporal
modification of the event, something that is unexpected if they allowed only
event-kind modifiers.
5.1 Restricted por-phrase in Spanish: Gehrke & Marco (2014)

Agentive por-phrases in Spanish seem to be better when they belong to the
consequent state denoted by the participle, rather than the event, or when
they modify an event kind. Gehrke & Marco (2014) present a corpus study on
Spanish that precisely supports this claim. They show that there is a clear con-
trast between the type of por-phrases found in adjectival participles and the
por-phrases of verbal participles. In adjectival participles, por-phrases involve
weak or non-referential nominals such as indefinites, bare nominals, weak and
generic nominals, whereas, in verbal participles, the nominals embedded in
por-phrases are more strongly referential, e.g. proper names, pronouns, demon-
stratives, regular definites. Additionally, they argue that the complements of
por-phrases in Spanish cannot have wide scope, cannot introduce discourse
referents, cannot be modified by common intersective modification and do not
involve strong determiners (Gehrke & Marco 2014: 13).
130 berro
(139) a. Todos los cuadros estaban

all.masc.pl the.masc.pl paintings be.loc.3pl.imperf.pst
pintados por un niño.
painted by a child
‘All the paintings were painted by a child.’ (>1 child possible)
NOT: ‘There was a particular child that painted all the paintings’
b. El cuadro estaba pintado por [un

the.masc painting be.loc.3sg.impr.pst painted.masc by a
niño]1.*pro1 era pelirrojo.
child was red-haired
Intended: ‘The painting was painted by a child. He had red hair’.
c. *El cuadro estaba pintado por

the.masc painting be.loc.3sg.imperf.pst painted.masc by
un niñ o pelirrojo
a child red-haired
Intended: ‘The painting was painted by a red-haired child’.
d. *El cuadro estaba pintado por

the.masc painting be.loc.3sg.imperf.pst painted.masc by
este niño
this child
Intended: ‘The painting was painted by this child’
Finally, the study also shows that the contrasts found between por-phrases that
belong to the event in verbal and adjectival participles do not find a correlate
in the por-phrases that pertain to the consequent state. Below, an example that
shows that a por-phrase modifying a state can have a strong determiner.
(140) Este edificio estaba vigilado por este

this.masc building be.loc.3sg.imperf.pst guarded.masc by this
hombre de chaqueta negra.
man of jacket black
‘This building is guarded by this man with black jacket.’
5.2 Structural Motivation for the Event Kind vs. Event Token Distinction
Gehrke (2011) and following works (Gehrke & Marco 2014; Gehrke 2015; Alexi-
adou et al. 2014; Alexiadou et al. 2015) claim that adjectival participles in lan-
guages like Spanish and German are interpreted as consequent states of event
kinds because in these languages, the verb does not get further embedded into
verbal projections, but is directly adjectivized. They generalize to verbs the
analysis of nominals (Carlson 1977; Zamparelli 1995; Chierchia 1998; Dayal 2004
among others), where it is considered that nouns start out like predicates of
kinds and get instantiated when they are embedded under Number. In the ver-
bal domain, Gehrke (2015) proposes that events also enter the derivation as
event kinds and become instantiated when they are embedded under further
verbal projections, like Aspect or T. At this level, they obtain the interpretation
of event tokens, that is, events that take place in the actual world. Since in lan-
guages like German and Spanish, participles are not further embedded under
Aspect or T, the interpretation of the event remains in the kind domain, and is
not interpreted as an instantiated event. According to the authors, the restric-
tions on event-modification, for example in languages like Spanish, follow from
the fact that the event remains as an event kind. State-modifiers of a state token
are generally accepted, because the adjectivized participle is further embedded
under T, and thus, the consequent state gets instantiated.
5.3 Event Tokens in Basque Non-verbal Participles

In Basque, ergative subjects introduced in adjectival participles are not re-
stricted, a fact that suggests that the type of event projected in Basque adjecti-
val participles is not an event kind but an event token, i.e. it is actually instanti-
ated. Additionally, we have shown that Basque adjectival and adpositional par-
ticiples allow event-related temporal and spatial modification. Within Gehrke’s
(2015) analysis, this would mean that, in Basque, participles are embedded
under further verbal projections before being categorized as adjectives or adpo-
sitions.
For Greek adjectival participles, which show a similar distribution of that
of Basque, Alexiadou et al. (2014) have proposed that vP/VoiceP is selected by
an aspectual operator (ASP PERF) and then by an adjectivizing PASS(ive) head.6
A similar analysis can be proposed for Basque (see Berro 2018), with the dif-
ference that in Basque the adjectivizing head does not absorb the external
argument, given that it can be explicitly introduced in the clause. Still, another
analysis is needed for the cases in which the adjectival or adpositional partici-
ple is interpreted as experiential.
6 In Alexiadou et al. (2015), the PASS head is substituted by an a(dj) head that has the catego-
rizing function.
132 berro
5.4 Dialectal Variation in Spanish

In relation to the facts in Basque, my perception looking at the Spanish exam-
ples in (139)—and confirmed by other speakers—is that the Spanish spoken in
the Basque Country is more flexible in allowing a wider range of por-phrases.
For instance, similar examples to those in (139bcd) but in the present tense (e.g.
(141)) are grammatical for me and other consulted speakers (Spanish/Basque
bilinguals living in the Basque Country). In the same way, in example (141a)
both interpretations are obtained. Additionally, the fact that Spanish adjectival
participles are fine with temporal and spatial modification also seems to sug-
gest that events in adjectival participles in Spanish pattern more with Basque
and Greek, rather than with German. This behavior of Spanish, that may be
restricted to the Spanish spoken in Basque Country, may be affected by the
contact between Basque and Spanish. In any case, the restrictions on man-
ner modifiers still hold, a fact that would remain unexplained if we considered
that, in the Spanish variety spoken in the Basque Country, events are instanti-
ated in adjectival participles. As observed in section 4.1, manner modifiers in
Spanish adjectival participles are better if they describe an event kind, in other
words, if they are relevant for the consequent state. In principle, if events were
instantiated in Spanish adjectival participles, as suggested in in this section,
that restrictions would be unexpected. Further investigation is needed on this
issue.
Spanish spoken in the Basque Country:
(141) a. Todos los cuadros están pintados por un

all.masc.pl the.masc.pl paintings be.loc.3pl painted by a
niño.
child
‘All the paintings are painted by a child.’ (>1 child possible)
‘There was a particular child that painted all the paintings’
b. El cuadro está pintado por [un

the.masc painting be.loc.3sg painted.masc by a
niño]1.pro1 era pelirrojo.
child was red-haired
‘The painting was painted by a child. He had red hair’.
c. El cuadro está pintado por un niño

the.masc painting be.loc.3sg painted.masc by a child
pelirrojo.
red-haired
‘The painting was painted by a red-haired child’.
d. El cuadro está pintado por este niño.

the.masc painting be.loc.3sg painted.masc by this child
‘The painting was painted by this child’
6 Conclusions
In this chapter, non-verbal participles in Basque and Spanish have been com-
pared and contrasted, in light of the literature on adjectival participles. I have
shown that non-verbal participles differ in these languages in their morpholog-
ical make-up: while Spanish adjectival participles are formally similar to verbal
passive participles, in Basque, an additional participial suffix (-a, -ta or -rik)
is added. -A participles have been considered adjectival, following the Basque
linguistic tradition, and -ta and -rik ones have been regarded as adpositional.
Non-verbal participles in both languages also differ in that Basque non-verbal
participles can give rise to experiential interpretations, which is why Krajew-
ska (2012, 2013ab) claims that Basque -a/-ta/-rik “resultative” participles are in
the middle of a grammaticalization path towards a new perfect. Additionally,
Basque adjectival participles also contrast with Spanish ones in that Basque
participles allow ergative subjects to be present, which denote the agent of the
event embedded under the participle.
Despite these differences, this chapter has made visible that Basque and
Spanish non-verbal participles are similar in a number of other aspects. For
instance, the same interaction between the lexical aspect of the event underly-
ing the participle and the aspectual interpretation of the participle is observed
in both languages, with non-dynamic events giving rise to a non-resultative
interpretation of the participle. On the other hand, even though Spanish adjec-
tival participles have been argued to admit only certain type or por-phrases, and
to involve, consequently, an event kind (and not an event token), this chapter
has shown that, at least in the Spanish spoken in the Basque Country, adjecti-
val participles allow temporal and spatial modification of the event underlying
the participle and por-phrases with wide scope, strong determiners and dis-
course referents. These data indicate that this language variety may be similar
to Basque in having non-verbal participles with an event token interpretation.
134 berro
Acknowledgements
I want to thank Ane Odria for sharing with me her linguistic intuitions and
Beatriz Fernández for her comments on the manuscript. The research lead-
ing to these results has received funding from the European Union’s Seventh
Framework Programme for research, technological development and demon-
stration under grant agreement no. 613465. Moreover, this study has been
developed thanks to several projects funded by the Basque Government (the
post-doctoral grant (POS_2015_1_0086 and POS_2016_2_0023), and the project
IT665–13) and by the Spanish Ministry of Economy and Competitiveness
(FFI2014–51878-P).
References
Alexiadou, Artemis. 2015. “On the irrealis effect on auxiliary selection”. In Rolf Kailuweit
& Malte Rosemeyer (eds.), Auxiliary Selection Revisited: Gradience and Gradualness.
Berlin: de Gruyter. 123–144.
Alexiadou, Artemis & Elena Anagnostopoulou. 2008. “Structuring participles”. In
Charles Chang & Hannah Haynie. Somerville (eds.), Proceedings of the WCCFL 26.
MA: Cascadilla Press. 33–41.
Alexiadou, Artemis, Elena Anagnostopoulou & Florian Schäfer. 2015. External Argu-
ments in Transitivity Altenations. A Layering Approach. Oxford: Oxford University
Press.
Anagnostopoulou, Elena. 2003. “Participles and Voice”. In Artemis Alexiadou, Monika
Rathert & Arnim von Stechow (eds.), Perfect explorations. Berlin: Mouton de Gruyter.
1–36.
Alexiadou, Artemis, Berit Gehrke & Florian Schäfer. 2014. “The argument structure of
adjectival participles revisited”. Lingua 149: 118–138.
Anagnostopoulou, Elena & Yota Samioti. 2014. “Domains within words and their mean-
ings: A case study”. In Artemis Alexiadou, Hagit Borer & Florian Schäfer (eds.), The
Syntax of Roots and the Roots of Syntax. Oxford: Oxford University Press.
Artiagoitia, Xabier. 2004. “The case of an enlightening, provoking and admirable
Basque derivational suffix with implications for the theory of argument structure”.
Supplements of ASJU XLVI: 147–183.
Baker, Mark C. 2003. Lexical Categories. Verbs, Nouns and Adjectives. Cambridge: Cam-
bridge University Press.
Berro, Ane. 2015. Breaking verbs. From event structure to syntactic categories in Basque.
Ph.D. diss., UPV/EHU and Université Bordeaux Montaigne.
Berro, Ane. 2017. “Resultative and non-resultative participles in Basque”. Ms, University
of the Basque Country & SFL (UMR 7023, CNRS/Paris 8).
Bosque, Ignacio. 1999. “El sistema adjectival. Complementos y modificadores del adje-
tivo. Adjetivo y participio”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática
descriptiva de la lengua española. vol. 1. Madrid: Espasa Calpe. 217–310.
Bosque, Ignacio. 2014. “On resultative past participles in Spanish”. Catalan Journal of
Linuistics 13: 41–77.
Bruening, Benjamin. 2014. “Word formation is syntactic: adjectival passives in English”.
Natural Language and Linguistic Theory 32: 363–422.
Carlson, Greg. 1977. Reference to kinds in English. Ph.D. diss., University of Massachu-
setts at Amherst, published by Garland, New York.
Chierchia, Gennaro. 1998. “Generics and concepts”. Natural Language Semantics 6: 339–
405.
Comrie, Bernard. 1976. Aspect. An Introduction to the Study of Verbal Aspect and Related
Problems. Cambridge: Cambridge University Press.
Dayal, Veneeta. 2004. “Number marking and (in)definites in kind terms”. Linguistic and
Philosophy 27: 393–450.
Demonte, Violeta. 1983. “Pasivas léxicas y pasivas sintácticas en español”. Serta Philo-
logica F. Lázaro Carreter. Madrid: Cátedra. 141–157.
Detges, Ulrich. 2000. “Time and truth: the grammaticalization of resultatives and per-
fects within a theory of subjectification”. Studies in Language 24: 345–377.
Doron, Edit. 2014. “The interaction of adjectival passive and voice”. In Artemis Alex-
iadou, Hagit Borer & Florian Schäfer (eds.), The Syntax of Roots and the Roots of
Syntax. Oxford: Oxford University Press.
Eguren, Luis. 2012. “Predication Markers in Basque”. In Urtzi Etxeberria, Ricardo Etxe-
pare & Myriam Uribe-Etxebarria (eds.), Noun Phrases and Nominalization in Basque:
Syntax and Semantics. John Benjamins Publishing Company. 243–266.
Elordieta, Arantzazu. 2001. Verb Movement and Constituent Permutation in Basque.
Utrecht: LOT.
Embick, David. 2004. “On the Structure of Resultative Participles in English”. Linguistic
Inquiry 35-3: 355–392.
Fábregas, Antonio & Rafael Marín. 2015. “How insertion in non-terminal nodes explains
two participle puzzles”. Paper presented in the 25th Coloquium on Generative Gram-
mar. Bayonne, May 21–23, 2015.
Fábregas, Antonio & Rafael Marín. 2017. “On non-dynamic eventive verbs in Spanish”.
Linguistics 55(3).
García Fernández, Luis, Ángeles Carrasco Gutiérrez, Bruno Camus Bergareche, María
Martínez-Atienza & María Ángeles García-Serrano. 2006. Diccionario de perífrasis
verbales. Madrid: Gredos.
Gehrke, Berit. 2011. “Stative passives and event kinds”. In Ingo Reich, Eva Horch & Den-
nis Pauly (eds.), Proceedings of Sinn und Bedeutung 15. Saarbrücken: Universaar—
Saarland University Press. 241–257.
136 berro
Gehrke, Berit. 2015. “Adjectival participles, event kind modification and pseudo-in-
corporation”. Natural Language and Linguistic Theory 33: 897–938.
Gehrke, Berit & Cristina Marco. 2014. The role of by-phrases in adjectival passives. Lin-
gua 149: 188–214.
Goenaga, Patxi. 1978. Gramatika bideetan. Donostia: Erein.
Hualde, José Ignacio. 1991. Basque Phonology. London and New York: Routledge.
Hualde, José Ignacio, Arantzazu Elordieta & Gorka Elordieta. 1994. The Basque dialect
of Lekeitio. Supplements of ASJU XXXIV. Bilbao/Donostia: UPV/EHU, Gipuzkoako Foru
Aldundia.
Iatridou, Sabine. 1996. “To have and to have not: on the deconstruction approach”. In
José Camacho, Lina Choueiri & Maki Watanabe (eds.), Proceedings of the 14th West
Coast Conference on Formal Linguistics (WCCFL) 14. Stanford: CSLI, 185–201.
Iatridou, Sabine, Anagnostopoulou, Elena & Roumyana Izvorski. 2001. “Observations
about the form and meaning of the perfect”. In Michael. J. Kenstowicz (ed.), Ken
Hale: a life in language. Cambridge, MA: MIT. 189–238.
Iatridou, Sabine, Anagnostopoulou, Elena & Roumyana Izvorski. 2003. “Observations
about the form and meaning of the perfect”. In Artemis Alexiadou, Monika Rathert
& Arnim von Stechow (eds.), Perfect explorations. Berlin/New York: Mouton de
Gruyter. 153–204.
Kennedy, Christopher & Louise McNally. 1999. “From Event Structure to Scale Structure:
Degree Modification in Deverbal Adjectives”. In Tanya Mathews & Devon Strolovitch
(eds.), The Proceedings of SALT IX, 1999. Ithaka, NY: CLC Publications. 163–180.
Koontz-Garboden, Andrew. 2010. “The lexical semantics of derived statives”. Linguistic
and Philosophy 33: 285–324.
Krajewska, Dorota. 2012. “Resultatives in Basque: a diachronic study”. Lingua Pos-
naniensis LIV(2): 56–67.
Krajewska, Dorota. 2013a. The diachrony of resultative constructions in Basque. Master
Thesis, UPV/EHU.
Krajewska 2013b. “Euskarazko egitura erresultatiboaren diakronia”. In Ricardo Gómez
et al. (eds.), 3rd Conference of the Luis Michelena Chair. Vitoria-Gasteiz: UPV/EHU Edi-
torial Services. 264–274.
Kratzer, Angelika. 1994. “The Event Argument and the Semantics of Voice”. Ms., Univer-
Kratzer, Angelika. 1996. “Severing the external argument from its verb”. In Johan
Roorych & Laurie Zaring (eds.), Phrase Structure and the Lexicon. Studies in Natural
Language and Linguistic Theory, 33. Dordrecht: Kluwer Academic Publishers. 109–
137.
Kratzer, Angelika. 2000. “Building Statives”. Berkeley Linguistic Society 26. 385–399.
Levin, Beth & Malka Rappaport Hovav. 1986. “The formation of adjectival passives”. Lin-
guistic Inquiry 17(1): 623–661.
Luján, Marta. 1981. “The Spanish copulas as aspectual indicators”. Lingua 54: 165–209.
Maienborn, Claudia. 2005. “On the Limits of the Davidsonian Approach: The Case of
Copula Sentences”. Theoretical Linguistics 31 (3): 275–316.
Maienborn, Claudia. 2007. “On Davidsonian and Kimian states”. In Ileana Comorovski
& Klaus von Heusinger (eds.), Existence: semantics and syntax. Springer. 107–130.
Marín, Rafael. 1997. “Participios con aspecto de adjetivos: entre la diacronía y la mor-
fología”. Moenia 3: 365–376.
Marín, Rafael. 2000. El componente aspectual de la predicación. Doctoral dissertation.
Universitat Autònoma de Barcelona.
Marín, Rafael. 2004a. Entre ser y estar. Madrid: Arco/Libros.
Marín, Rafael. 2004b. “Sobre pasivas adjetivales”. Verba 31: 455–471.
Marín, Rafael. 2009. “Del adjetivo al participio”. In Elena de Miguel et al. (eds.), Fron-
teras de un diccionario: las palabras en movimiento. San Millán de la Cogolla: Cilen-
gua. 327–348.
McFadden, Thomas & Artemis Alexiadou. 2011. “Pieces of the be perfect in German
and Older English”. In Donald Baumer, David Montero & Michael Scanlon (eds.),
Proceedings of the 25th West Coast Conference on Formal Linguistics. Somerville, MA:
Cascadilla Proceedings Project. 270–278.
McIntyre, Andrew. 2013. “Adjectival passives and adjectival participles in English”. In
Artemis Alexiadou & Florian Schäfer (eds.), Non-canonical Passives, Amsterdam:
John Benjamins. 21–42.
Mounole, Céline. 2011. Le verbe basque ancien: étude philologique et diachronique. Ph.D.
diss., University of the Basque Country (UPV/EHU) and Université Michel de Mon-
taigne-Bordeaux III.
Nedjalkov, Vladimir P. and Sergej Je. Jaxontov. 1988. “The Typology of Resultative Con-
structions”. In Vladimir Nadjalkov (ed.), Typology of resultative constructions. Ams-
terdam: John Benjamins Publishing Company. 3–62.
Ortiz de Urbina, Jon & Myriam Uribe-Etxebarria. 1991. “Participial predication in
Basque”. Supplements of ASJU 14-2: 993–1012.
Oyharçabal, Bernard. 1992. “Structural and Inherent Case-Marking; Ergaccusativity in
Basque”. In Joseba A. Lakarra & Jon Ortiz de Urbina (eds.), Syntactic Theory and
Basque Syntax. Supplements of ASJU XXVII. Donostia: UPV/EHU and Gipuzkoako
Foru Aldundia. 309–342.
Oyharçabal, Bernard. 2010. “Basque ditransitives”. In Maia Duguine, Susana Huido-
bro & Nerea Madariaga (eds.), Argument structure and syntactic relations. Amster-
dam/Philadelphia: John Benjamins. 261–282.
Pancheva, Roumyana. 2003. “The Aspectual Makeup of Perfect Participles and the
Interpretation of the Perfect”. In Artemis Alexiadou, Monika Rathert & Arnim von
Stechow (eds.), Perfect explorations. Berlin/New York: Mouton de Gruyter. 277–306.
Pinkster, Harm. 1987. “The strategy and chronology of the development of future and
138 berro
the perfect tense auxiliaries in Latin”. In Martin Harris, & Paolo Ramat (eds.), His-
torical development of auxiliaries. Berlin/New York: Mouton de Gruyter. 193–223.
Parsons, Terence. 1990. Events in the Semantics of English. Cambridge, MA: the MIT Press.
Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First Phase Syntax. Cam-
bridge: Cambridge University Press.
Rapp, Irene. 1996. “Zustand? Passiv? Überlegungen zum sogenannten “Zustandspas-
siv””. Zeitschrift für Sprachwissenschaft 15.2: 231–265.
Rapp, Irene. 1997. “Partizipien und semantische Struktur: Zu passivischen Konstruktio-
nen mit dem 3”. Status. Tübingen: Stauffenburg.
Rebuschi, George. 1984. Structure de l’énoncé en basque. Paris: SELAF.
de Rijk, Rudolf P.G. 1993. “Basque Hospitality and the suffix -ko”. In José Ignacio Hualde
& Jon Ortiz de Urbina (eds.), Generative Studies in Basque Linguistics. Amsterdam:
John Benjamins. 145–162. [reprinted in Rijk, Rudolf P.G. de. 1998. De Lingua Vas-
conum: Selected Writtings. Supplements of ASJU XLIII. UPV/EHU. 377–390]
de Rijk, Rudolf P.G. 2008. Standard Basque. A Progressive Grammar. Cambridge/Lon-
don: The MIT Press.
Rothmayr, Antonia. 2009. The Structure of Stative Verbs. Amsterdam/Philadelphia: John
Benjamins.
Sadler, Louisa & Douglas J. Arnold. 1994. “Prenominal adjectives and the phrasal/lexical
distinction”. Journal of Linguistics 30(1): 187–226.
Sleeman, Petra. 2011. “Verbal and adjectival participles: Position and internal structure”.
Lingua 121: 1569–1587.
Sleeman, Petra. 2014. “From participle to adjective in Germanic and Romance”. In Petra
Sleeman, Freek Van de Velde & Harry Perridon (eds.), Adjectives in Germanic and
Romance. John Benjamins. 171–198.
von Stechow, Arnim. 1998. “German participles II in Distributed Morphology”. Ms., Uni-
versity of Tübingen.
Wasow, Thomas. 1977. “Transformations and the lexicon”. In Peter W. Culicover, Thomas
Wasow & Adrian Akmajian (eds.), Formal Syntax. 327–360. New York: Academic
Press.
Gawron, Jean Marc. 2009. The lexical semantics of extent verbs. Ms., San Diego State
University.
Zabala, Igone. 1993. Predikazioaren teoriak gramatika sortzailean (euskararen kasua).
Ph.D. thesis, UPV/EHU.
Zamparelli, Roberto. 1995. Layers in the determiner phrase. Ph.D. thesis, University of
Rochester.
chapter 5
(In)transitive Verbs: Unergatives and Unaccusatives

Ane Berro
1 Introduction
The fact that intransitive verbs do not conform a homogeneous class is an

accepted and well-studied matter in the linguistic literature (Perlmutter 1978,
1989; Burzio 1981, 1986; Belleti 1988, 1999; Sorace 1993; Levin & Rappaport Hovav
1995 and subsequent work). According to the Unaccusative Hypothesis (Perl-
mutter 1978, 1989), intransitive verbs can be classified in two groups, depending
on the nature of the only argument of the verb. In some intransitive verbs, the
only argument corresponds to the initial direct object, whereas in others, it cor-
responds to the initial subject. The former are known as unaccusative verbs,
and the latter are called unergative.
In Basque, unaccusative and unergative verbs are clearly differentiated mor-
phologically (Levin 1983): unaccusatives occur with absolutive subjects and
izan (be) auxiliary, whereas unergatives generally appear with ergative sub-
jects and edun (have) auxiliary. Case systems that show this differentiation
among intransitive verbs have been variously called: Split S (Dixon 1979, 1994),
Split Intransitive (Merlan 1985), Extended Ergative (Dixon 1979; Ortiz de Urbina
1989), Active (Bittner and Hale 1996; see also Sapir 1917), Agentive (Mithun 1991)
or Semantically Aligned (Donohue & Wichmann 2008).
In Romance languages, be auxiliary selection has also been proposed as a
diagnostic of unaccusativity, particularly for Italian (Burzio 1981, 1986; Perl-
mutter 1989; Rosen 1981). However, not all Romance languages have auxil-
iary alternation, and those displaying it do not show the same distribution
across verb classes (e.g. Sorace 1993, 2000, 2004; Bentley & Eythórsson 2003;
MacKenzie 2006; Kayne 2008). In this chapter, I am going to analyze the unac-
cusative/unergative distinction in Basque and some Romance languages, pay-
ing special attention to the verb classes that are aligned in the unaccusative
or the unergative way in both Basque and Romance, and considering also
border-line cases such as Romance loan verbs in Basque and some new stative
unergative verbs. I will propose that intransitive stative verbs are the most vari-
able group of verbs, both cross-linguistically (in terms of auxiliary alternation,
Sorace 2000, 2004; Keller & Sorace 2003) and in Basque (regarding auxiliary
and subject case). The reason for this variability comes from the fact that the

140 berro
subject in stative verbs—a holder argument—can be introduced either as an

external subject or an internal subject, without creating a significant semantic
contrast in its interpretation and the interpretation of the event configuration.
The structure of the chapter is the following. In section 2, I will present the
basic event decomposition that I assume, which shares many aspects with the
system developed in Mateu (2002 et seq.) and Ramchand (2008 et seq.). Section
3 and 4 will describe and analyze unaccusative and unergative verbs in Basque,
attending to factors such as morphological derivation, the simplex and com-
plex contrast in unergatives, and dialectal variation. In section 5, I will focus
on the distribution of auxiliary selection across verbs classes in Romance lan-
guages like Italian and French (mentioning also Old Catalan and Old Spanish),
and I compare it to the distribution in Basque. In section 6, two aspects related
to the contact between Basque and Romance will be considered, namely, the
use of Romance loan verbs in Basque, and the emergence of new stative uses
of certain verbs. In section 7, I will provide a syntactic explanation for the great
variability exhibited by stative verbs, and section 8 will present the conclusions.
2 Unergatives and Unaccusatives in Basque
In this chapter, I am assuming a constructionist model of the event configura-

tion, where the meaning of verbs is derived from the syntactic structure (Hale
& Keyser 1993, 2002, 2005; Harley 1995, 2005; Mateu 2002; Cuervo 2003; Folli
& Harley 2005; Ramchand 2008 among others). This way, I propose that unac-
cusative and unergative verbs have different syntactic configurations, and that
this contrast is responsible for their different meaning and morphological real-
ization. In Basque unaccusative verbs, the subject has absolutive (zero) case
and the auxiliary selected is izan (be).
(1) Jon jausi da.1

John.abs fall be.3sg.abs
‘John fell’
(2) Mahai-a apurtu da.

table-det.abs break be.3sg.abs
‘The table broke’
1 Abbreviations: abs = absolutive case, cl = se clitic, det = determiner, erg = ergative case,
gen = genitive, imprf = imperfective, ine = inessive, nom = nominative case, part = parti-
tive, pl = plural, prog = progressive, prt = participle, pst = past.
(in)transitive verbs: unergatives and unaccusatives 141
Unergative verbs, in contrast, have an ergative subject and edun (have) aux-
iliary.
(3) Jon-ek ondo dantza-tzen du.

John-erg well dance-imprf have.3sg.erg
‘John dances well’
I propose that, in unergative verbs, the subject is introduced in the specifier

of Voice, a functional head that is projected outside the verbal phrase (Kratzer
1994, 1996).
(4)
The subject of unaccusatives, on the other hand, is introduced within the ver-
bal phrase.
2.1 Unaccusative Verbs

In unaccusative verbs, the sole argument of the verb is introduced internally to
the verbal phrase. Semantically, it has been associated with the role of under-
goer of a directed change or of holder of a state whose existence is asserted or
denied (Levin & Rappaport Hovav 1995).
Unlike in the majority of unergative verbs, in Basque, unaccusative verbs
consist of a single phonological word (leaving the auxiliary verb aside). The
verbs grouped within unaccusatives in Basque are the following: verbs of
change of state or location—telic and atelic ones—, verbs of appearance or
occurring, verbs of existence and (some) aspectual verbs.
(5) Verbs of change of state

apurtu ‘break’, gorritu ‘redden’, handitu ‘get bigger’, txikitu ‘get smaller’,
zabaldu ‘open’, itxi ‘close’, zahartu ‘age’, gaztetu ‘rejuvenate’, amatu ‘be-
come a mother’, mutildu ‘become a lad’ etc.
142 berro
(6) Verbs of change of location

igo ‘go up’, jaitsi ‘go down’, etorri ‘come’, joan ‘go, leave’, heldu ‘arrive’, aile-
gatu ‘arrive’, hurbildu ‘approach, come closer’, urrundu ‘move away’, etxer-
atu ‘go home’, lehorreratu ‘land, go ashore’ etc.
(7) Verbs of appearance or occurring

agertu ‘appear’, desagertu ‘disappear’, jaio ‘be born’, bukatu ‘run out’, ger-
tatu ‘happen’
(8) Verbs of existence

existitu ‘exist’, egon ‘be, stay’, izan ‘be’
(9) Aspectual verbs

hasi ‘begin’, ari ‘be engaged in’, jarraiki ‘continue’
The majority of the verbs of change of state and change of location are derived,
in the sense that they are built on elements that exist outside of this ver-
bal complex, like nouns (e.g. ama ‘mother’ and mutil ‘boy’), adjectives (gorri
‘red’, handi ‘big’, txiki ‘small’ etc.) in (5), adverbs (hurbil ‘near’, urrun ‘far’)
and adpositional phrases (e.g. etxe-ra ‘to house’ and lehorr-era ‘to land’) in
(6).
Most of them, and particularly derived ones, can have a causative variant, in
which case the causer argument is marked ergative and the auxiliary switches
to have. The ability to enter the causative alternation is considered a diagnos-
tic for unaccusativity in Levin & Rapport Hovav (1995). Another test that has
been suggested to differentiate unaccusatives from unergatives verbs in Basque
involves the use of the partitive case. Only the subjects of unaccusative verbs
(as well as the objects of transitive verbs) can occur bearing partitive marking
(10) (11), contrasting, this way, with the subjects of unergative verbs (and also
the subjects of transitive verbs) (12) (13) (Levin 1983; Ortiz de Urbina 1989; Sal-
aburu 1992).
(10) a. Ez da haurr-a etorri. ☑ Unaccusative

no be.3sg.abs child-det.abs come
‘The child has not come’
b. Ez da haurr-ik etorri.
no be.3sg.abs child-part come
‘No child has come’
(11) a. Ez dut haurr-a ikusi. ☑ Transitive object

no have.3sg.abs.1sg.erg child-det.abs see
‘I have not seen the child’
b. Ez dut haurr-ik ikusi.

no have.3sg.abs.1sg.erg child-part see
‘I have not seen any child’
(12) a. Ez du haurr-a-k deitu. ☒ Unergative

no have.3sg.erg child-det-erg call
‘The child has not telephoned’
b. *Ez du haurr-ik deitu

no have.3sg.erg child-part call
‘No child has telephoned’
(13) a. Haurr-a-k ez du ogi-a jan. ☒ Transitive

child-det-erg no have.3sg.erg bread-det.abs eat subject
‘The child has not eaten (the) bread’
b. *Haurr-ik ez du ogi-a jan

child-part no have.3sg.erg bread-det.abs eat
Unaccusative subjects and transitive objects pattern alike: they can be marked
with the partitive. In contrast, unergative subjects and transitive subjects also
behave in a similar fashion: they cannot bear partitive marking.
2.2 Unergative Verbs

The sole argument of unergative verbs is considered the initial (Perlmutter
1978) or deep subject (Burzio 1981, 1986) as well as the external argument in
the sense of Kratzer (1994, 1996). As indicated in section 2, I assume that it is
introduced in the specifier of Voice (Kratzer 1994, 1996). Semantically, it is usu-
ally related to internal causation, as in agentive verbs or in non-agentive verbs
of emission or bodily processes (Levin & Rappaport Hovav 1995). Unergativity
has also been related with lack of telicity (Tenny 1987). However, with respect
to Basque, as shown in section 2.1, telicity cannot be considered a discriminat-
ing factor, since many unaccusative verbs are actually atelic (see the list in (5)
and (6)).
Unergative verbs in Basque—which have ergative subjects and select for
edun (have) auxiliary—are mainly verbs expressing volitional acts, verbs of
144 berro
emission, involuntary bodily processes, some verbs of existence of a state and

some aspectual verbs:
(14) Volitional acts

a. Manner of motion
dantzatu ‘dance’, saltatu ‘jump’, paseatu ‘stroll’, biratu ‘turn’, nabigatu
‘navigate’, eskiatu ‘ski’, eskalatu ‘climb’ (and many other Romance
loans).
b. Animate activities
jolastu ‘play’, jokatu ‘play’, borrokatu ‘fight’, entrenatu ‘train’, bidaiatu
‘travel’ etc.
c. Dining
bazkaldu ‘have lunch’, afaldu ‘have dinner’, gosaldu ‘have breakfast’,
askaldu ‘have an afternoon snack’.
(15) Non-agentive manner of motion

irakin ‘boil’
(16) Non-animate activities

Funtzionatu ‘work properly (e.g. a machine)’
(17) Emission
Distiratu ‘shine’, dirdiratu ‘shine, sparkle’, argitu ‘light’, usaindu ‘smell’.
(18) Involuntary bodily processes

Zurrungatu ‘snore’, izerditu ‘sweat’.
(19) Existence of a state

Iraun ‘last’, biziraun ‘survive’, balio izan ‘cost, value’, pisatu ‘weigh’.
(20) Aspectual verbs

Bukatu ‘finish’, jarraitu ‘continue’
The verbs shown above are all simplex, consisting of a single phonological
word. See, for example, the sentence in (21). Note, however, that many unerga-
tive verbs in Basque are morphologically complex, consisting of the light verb
egin ‘do’ and a non-verbal element (22).
(21) Jon-ek ondo dantza-tzen du.

John-erg well dance-imprf have.3sg.erg
(22) Jon-ek ondo dantza egi-ten du.

John-erg well dance do-imprf have.3sg.erg
In the following section, I am going to deal briefly with complex unergative

verbs, and in section 4 with simplex ones. Regarding simplex unergatives, I will
show that there is dialectal variation in the subject marking and auxiliary selec-
tion of some verbs, particularly, in verbs expressing volitional acts like those in
(14) (Oyharçabal 1992; Fernández 1997; Etxepare 2003; Aldai 2006, 2009; Berro
2010, 2012).
2.3 Complex Unergative Verbs

As observed by Sarasola (1977), Levin (1983), Uribe-Etxebarria (1989), Laka
(1993a), Hale & Keyser (1993), Levin & Rappaport Hovav (1995), Etxepare
(2003), Oyharçabal (2006) and Aldai (2009), most unergative predicates in
Basque are complex, built on the basis of the light verb egin ‘do’ and a non
verbal element, usually a bare noun. These unergative verbs belong to many
different semantic classes (Etxepare 2003; Zabala 2004). Note that some verbs
also have a simplex counterpart, either as an unergative verb or as a transitive
one.
(23) Emission
a. Sound emission
deiadar egin ‘scream’, oihu egin ‘scream’, barre egin ‘laugh’, negar egin
‘cry’, zaunka egin ‘bark’ etc.
b. Light emission
Dir-dir egin ‘shine’, diz-diz egin ‘glow, sparkle’, nir-nir egin ‘twinkle,
flicker’.
c. Verbal emission
hitz egin ‘talk’, solas egin ‘talk, chat’, marmar egin ‘grunt’, dei egin ‘call’,
otoitz egin ‘pray’, errieta egin ‘reprimand, scold’, burla egin ‘mock, make
fun of’ etc.
(24) Mental activities

gogoeta egin ‘meditate’, duda egin ‘doubt’, amets egin ‘dream’ etc.
(25) Behavioral activities

planto egin ‘stop by refusing to follow a game’, paso egin ‘be uninterested’,
uko egin ‘refuse, reject’.
146 berro
(26) Internal body motion

dar-dar egin ‘to tremble’, bor-bor egin ‘to boil’
(27) Physical activities

a. Actions against an object or an individual
zizt egin ‘puncture’, putz egin ‘blow’, bultza egin ‘push’, laztan egin
‘caress’, tiro egin ‘shoot’, min egin ‘hurt’ etc.
b. Motion verbs
alde egin ‘leave’, ihes egin ‘flee’, salto egin ‘jump’, dantza egin ‘dance’,
laprast egin ‘slip’, ospa egin ‘leave’.
c. Bodily functions
eztul egin ‘cough’, aharrausi egin ‘yawn’, kaka egin ‘shit’, botaka egin
‘vomit’, izerdi egin ‘sweat’.
(28) Other types of activities

lan egin ‘work’, huts egin ‘fail’.
(29) Weather verbs

euria egin ‘rain’, eguzkia egin ‘be sunny’, elurra egin ‘snow’, hotz egin ‘be
cold’, beroa egin ‘be warm’.2
As can be seen, complex unergative verbs can be either volitional or non-

volitional, but they can all conform to the property of internal causation, as
formulated in Levin & Rappaport Hovav (1995).
Regarding their complex morphological realization, Hale & Keyser (1993)
argued, on the basis of languages like Basque, that unergative verbs are uni-
versally transitive in nature. The morphological contrast between simplex and
complex unergatives like in dantzatu ‘dance’ and dantza egin [do dance] ‘dance’
would reflect the absence vs. presence of incorporation of the nominal object
onto the verb. For instance, some analyses (Uribe-Etxebarria 1989; Bobaljik
1993; Laka 1993a and Fernández 1997) have proposed that the complement
of the light verb egin represents actually a non-incorporated object, which
may incorporate after spell out (Fernández 1997). As Uribe-Etxebarria (1989)
pointed out (see also Oyharçabal 2006), some complex unergatives allow hav-
ing egin and its complement separated in some contexts (in interrogative sen-
2 Weather verbs are different from other complex unergatives in a number of aspects. For
instance, in most of them the non-verbal (nominal) element is headed by the determiner
-a and they can be modified by adjectives like handi ‘big/a lot’ as in bero handia egin [heat big
do] ‘be very warm’. I want to thank Jon Ortiz de Urbina for bringing these facts to my attention.
tences and in focalized structures) and the complement may be marked parti-
tive in negative sentences, as common direct objects.
(30) a. Nor-k egin behar du lan?

who-erg do need have.3sg.erg work
‘Who has to work?’
b. Nor-k egin du lan?

who-erg do have.3sg.erg work
‘Who has worked?’
(31) Oso ondo egin duzu lan.

very well do have.2sg.erg work
‘Very well you have worked’
(32) Ez dut lan-ik egin.

no have.1sg.erg work-part do
‘I haven’t worked’
Furthermore, the bare nominal may also be quantified, like common direct
objects, and egin can be silenced in sentences with negative contrastive focus
(Oyharçabal 2006: 792–793):
(33) a. Lo gutxi egin dut.

sleep little do have.1sg.erg
‘I have slept a little’
b. Lan gehiegi egin dut.

work too.much do have.1sg.erg
‘I have worked too much’
(34) a. Lan egin dut, ez lo.

work do have.1sg.erg no sleep
‘I have worked, not slept’
b. Irri egin dut, ez oihu.

laugh do have.1sg.erg no shout
‘I have laughed, not shouted’
148 berro
Note, however, that the bare noun of complex unergative predicates does not
behave as a common direct object in other aspects: e.g. in its lack of determiner
and its position with respect to manner adverbials. See Oyharçabal (2006) for a
discussion on the (non)-incorporation of the noun, and the variation in dialects
and predicates.
Another aspect in favor of the transitivity of complex unergative verbs is
that, unlike simplex verbs, they are incompatible with a direct object (other
than the non-verbal element selected by egin ‘do’).
(35) a. Mikel-ek (tango-a) dantzatu du.

Michael-erg (tango-det.abs) dance have.3sg.abs.3sg.erg
‘Michael has danced (a tango)’
b. Mikel-ek (* tango-a) dantza egin du.

Michael-erg ( tango-det.abs) dance do have.3sg.abs.3sg.erg
Intended: ‘Michael has danced (a tango)’
Many works have accounted for this contrast claiming that in complex unerga-
tives, the complement of egin ‘do’ is assigned absolutive case, so that another
direct object cannot be case-licensed (e.g. Laka 1993; Fernández 1997). In Berro
(2015a), I additionally claim that the non-verbal element in complex verbs
(dantza in (35b)) occupies exactly the same position of the direct object in
simplex verbs (tangoa in (35a)), accounting in this way for the measuring prop-
erties of the direct objects in simplex verbs.3
3 The Building Blocks of the Event Configuration
The system of event configuration that I am assuming in this paper shares many
similarities with that of Mateu (2002 et seq.) and Ramchand (2008 et seq.). In
my system, the verbal phrase consists of subevents—Events and States—and
their non-subeventive complements, namely Roots. Event provides eventivity
to the predicate, introducing an event argument (Davidson 1967). State is a cen-
tral coincidence relation, which relates a Figure (its specifier) with a Ground
3 Simplex verbs can be bounded if the direct object denotes a bounded scale (e.g. 35a). In a
similar way, the event denoted in complex verbs is also measured by the scalar (aspectual)
properties of its non-verbal complement (the Root). See Harley (2005) and Berro (2015a) for
accounts about the measuring properties of Roots (Berro 2015a about Basque in particular).
(its complement) centrally. The subject of unaccusative verbs is introduced in

the specifier of a State subevent. This is the configuration of an eventive tran-
sitive verb, with both Voice and State projected:
(36) Eventive transitive verb
The different flavors of the subevents, and the theta roles of the subjects are
derived from the whole event configuration (Wood 2012). Event can have a
cause flavor, like in (36), if it is selected by Voice and if it selects for State. It
can also have a do flavor (37), if it is selected by Voice, and if it selects straight-
away for an Event naming Root (see Harley 2005; Berro 2015a). This is the case
of unergative eventive verbs.
(37) Eventive unergative verb
Event can also have a transitional (or go, become) flavor, in the case that
it selects for a State (like in (36)), but Voice is not projected above (38). In this
case, also like in (36), the Root selected by State is a State naming Root. This is
the configuration of unaccusative verbs.4
4 Following Mateu (2002 et seq.), I consider that both telic and atelic unaccusatives have the
configuration depicted in (3).
150 berro
(38) Eventive unaccusative verb
The theta-role of the subjects also varies depending on the configuration. When
Voice selects for Event, the subject introduced in its specifier has an originator
theta role, which can be further specified as causer or agent. In a configura-
tion like (36), the subject has a causer theta-role, and in (37), it may have an
agent role, but only if the entity occupying this position is animate and if the
encyclopedic content related to the Root allows for that. On the other hand,
the subject in the specifier of State has a patient role (undergoer and resultee),
because the State is selected by Event. In the case it is not (as in the stative
verbs that I will present in section 6), it has a holder role. As will be shown, the
subject in Voice can also be a holder, if Voice selects for a State instead of for an
Event.
4 Variation in Simplex Unergative Verbs
A subset of simplex unergative verbs shows dialectal variation. In eastern

dialects, certain simplex verbs occur with absolutive subjects and izan (be) aux-
iliary, just like unaccusative verbs.
(39) a. Antton bazkaldu da.

Antton.abs have.lunch be.3sg.abs
‘Antton has had lunch’
b. Maialen borrokatu da.

Maialen.abs fight be.3sg.abs
‘Maialen has fought’
The verbs that are used in the unaccusative fashion in eastern dialects comprise
volitional verbs expressing speech (mintzatu ‘speak’, solastatu ‘chat’), man-
ner of motion (promenatu ‘stroll’, jauzi ‘jump’), dining (bazkaldu ‘have lunch’,
afaldu ‘have dinner’), and animate activities (borrokatu ‘fight’, jokatu ‘play, act’,
jolastu ‘play’). As I have mentioned above, from this group, verbs of manner of
motion, dining and animate activities occur with an ergative subject and edun
‘have’ auxiliary in western and central varieties.
(40) a. Antton bazkaldu da. Eastern dialects

Antton.abs have.lunch be.3sg.abs
b. Antton-ek bazkaldu du. Western & central dialects

Antton-erg have.lunch have.3sg.erg
Aldai (2008 2009) suggests that the dialectal divide within Basque corresponds
to the divide between two case-marking systems: one based on the ergative
alignment of case, represented by eastern dialects, and a semantically aligned
one, represented by western and central dialects. However, Aldai’s characteri-
zation of the eastern case-alignment is questioned by the behavior some verbs,
such as inanimate manner of motion verbs (41ab), verbs directly expressing
duration (41c) and emission verbs (41d), which take an ergative subject and
have auxiliary in eastern dialects (see Berro & Etxepare 2017).
(41) a. Ur-a-k kurri-tzen du. Eastern dialects

water-det-erg flow-imprf have.3sg.erg
‘The water flows/is flowing’
b. Honek ez du funtziona-tzen.
this.erg no have.3sg.erg work-imprf
‘This does not work/is not working’
c. Beran-tzen zuen.
delay-imprf have.3sg.erg.pst
‘He/she was taking long’
d. Izarr-a-k distira-tzen du.

star-det-erg shine-imprf have.3sg.erg
‘The star shines/is shining’
Therefore, not all unergative verbs are aligned in the unaccusative fashion in
eastern dialects, so that the distinction between eastern and western/central
dialects in terms of a semantically based alignment versus an ergative align-
ment is not as clear as suggested in Aldai’s proposal.
152 berro
Another line of analysis that has been put forward is that the verbs that take
absolutive subjects in eastern dialects are somehow related to an aspectual
head of telicity or boundedness (Berro 2010 2012). This head would be respon-
sible for the absolutive marking of the subject and would also make it possible
for those unergative verbs to become telic or bounded by selecting an incre-
mental theme.
A third type of explanation, developed in Pineda & Berro (2018), is that
these unergative verbs are just associated to different syntactic-eventive con-
figurations in eastern and western/central dialects. It is a well-known fact that
verbs having very close meaning can correspond to different event configu-
rations across languages. For example, the verb ‘blush’ is a change of state in
Italian (arrossire, lit. ‘become red’) and it is a verb of internal causation in Dutch
(bloezen) (McClure 1990, apud Levin & Rappaport Hovav 1995). Since both lan-
guages have auxiliary alternation, in Italian arrossire combines with essere (be)
auxiliary, whereas in Dutch, bloezen occurs with hebben (have), as expected.
On the other hand, in a closer Romance language like Spanish, the verb sonreír
‘smile’ is subject to intra-linguistic variation: although ‘smile’ is usually consid-
ered an unergative predicate (also in Basque, irribarre egin, lit. ‘do smile’), in
some Spanish varieties it occurs with se clitic in non-reciprocal contexts, like
some unaccusative verbs.5
(42) El niño se ha sonreído.

det child.nom cl have.3sg smile.prt
‘The child has smiled’
From this perspective, sonreír and sonreírse are the externalizations of two dif-
ferent eventive configurations. Sonreír is an internally caused verb, whereas
sonreírse is a change of state verb, where the subject undergoes a transition
to a state of having a smile on the face, in the same way as the subject of the
Italian verb arrossire. In Basque, too, we find similar cases, for example with
the verbs leher egin and lehertu, which both mean ‘explode’. Leher egin is an
unergative verb of internal causation, and lehertu, in its intransitive variant, is
an unaccusative verb of change of state. These two different eventive configu-
rations are morphologically externalized: the former having an ergative subject
and have auxiliary, and the latter, an absolutive subject and be auxiliary.
5 This dialectal alternation is also attested in other verbs, like in engordar/engordarse ‘fatten’
and adelgazar/adelgazarse ‘lose weigh’.
(43) a. Auto-a-k leher egin du.

car-det-erg explode do have.3sg.erg
‘The car blew up’
b. Gurpil-a lehertu da.

wheel-det explode be.3sg.abs
‘The wheel exploded’
I propose that in leher egin, the verbal root is merged to an Event head with an
internal cause flavor (do), whereas in lehertu, the Event head has a transitional
flavor (trans).
(44) a. leher egin
b. lehertu
The verbs that take an absolutive subject in eastern varieties could also be
involved in an alternation of this sort. Dining verbs like bazkaldu ‘have lunch’
may be associated to a transition to a state of being finished with lunch. For
instance, some speakers judge these verbs as telic in the progressive-perfect
entailment test.
(45) Bazkal-tzen ari naiz. ⇏ Bazkaldu dut.

have.lunch-imprf be.engaged be.3sg.abs have.lunch have.3sg.erg
‘I am having lunch’ ⇏ ‘I have had lunch’
154 berro
In telic verbs, the progressive does not entail the perfect, and this is actually
what happens in the verb bazkaldu. Telicity could perhaps lead to the change of
state configuration (with Event + state) of this verb in eastern dialects. On the
other hand, other verbs like mintzatu ‘speak’ and borrokatu ‘fight’ could be con-
sidered similar to some pronominal verbs in Spanish and French (expresarse
and pelearse in Spanish, and s’exprimer and se battre in French), which have
a rather reflexive or reciprocal flavor. In French, particularly, these predicates
occur with être (be) auxiliary, as French is an auxiliary alternation language.
(46) a. Je me suis mal exprimé.

I.nom cl be.1sg wrong express.prt
‘I have not expressed myself well’ lit. ‘I have expressed myself wrong’
b. Nous nous sommes battus pour cela.

we.nom cl be.1pl fight.prt for that
‘We have fought for that’
Pineda & Berro (2018) propose that several verbs which show dialectal alter-
nation in Basque and that are aligned in the unaccusative fashion in eastern
dialects are unaccusative verbs in (Middle) French and Occitan. These lan-
guages are actually the Romance languages that are in contact with north-
eastern dialects. The unaccusativity of these verbs is supported by several diag-
nostics, such as the use of the se clitic in previous stages of the language,
occurrence in a causative variant, lack of direct object and available resulta-
tive participles (see Pineda & Berro 2018 for further details).
Similarly to what is found in (Middle) French and/or Occitan, the configu-
ration of these verbs in eastern dialects may be unaccusative (see section 6.1
in relation to this). The fact that this behavior is not generalized to all “unerga-
tive verbs” but that it is, actually, quite restricted would also suggest that we are
dealing with certain lexical exceptions.
5 Unergatives and Unaccusatives in Basque and Romance
In Romance languages, unergative and unaccusative verbs have been differen-

tiated on the grounds of several linguistic phenomena, such as prenominal pas-
sive participles, ne-cliticization in Italian (Bruzio 1981 1986), post-verbal subject
position (Torrego 1989, Belleti 1988 1999) and auxiliary selection. In this section,
I will focus on auxiliary selection, paying close attention to cross-linguistic vari-
ation and to the position that Basque occupies in this picture.
5.1 Unaccusativity and Auxiliary Alternation in Romance

All Romance languages have or have had at some point of their history aux-
iliary alternation with intransitive verbs (MacKenzie 2006), some intransitive
verbs taking have auxiliary and others taking be. have auxiliary has displaced
be in Spanish, Catalan, Portuguese and Galician (although it survives in some
dialects of Catalan and Aragonese). On the contrary, alternation between have
and be exists nowadays in French and it is very productive in Italian. In Euro-
pean French, there are fewer verbs that select auxiliary be in comparison to
Italian (Kayne 2008), and even fewer in Canadian French (Sankoff & Thibault
1977; apud Bentley & Eythórsson 2003).
Sorace (2000 2004) and Keller & Sorace (2003) present an Auxiliary Selection
Hierarchy where classes of predicates are distributed according to their auxil-
iary selection pattern.
(47) Auxiliary Selection Hierarchy (ASH)

Change of location Selects be (least variation)
Change of state
Continuation of a pre-existing state
Existence of state
Uncontrolled process
Controlled process (motional)
Controlled process (non-motional) Selects have (least variation)
With this hierarchy, Sorace (2000 2004) accounts for the delimited and sys-
tematic variation found in intransitive predicates with respect to the auxiliary
selected. According to Sorace, some predicates are consistent in showing aux-
iliary alternation across languages like Italian, French, German or Dutch, and
also in language varieties, while others consistently select for be or have, both
synchronically and diachronically. The verbs at the top and at the bottom of
the hierarchy are the ones having least variation and most consistent behav-
ior (either selecting be or have), whereas the ones in the middle are the most
variable.
Regarding the distribution of auxiliaries in unaccusative verbs in French and
Italian, three groups of verbs can be distinguished (Bentley & Eythórsson 2003:
450–451).
Firstly, change of location verbs select be in both Italian and European French
(some of these verbs take have in Canadian French).
156 berro
(48) a. Maria è venuta alla festa. Italian

Mary.nom be.3sg come.prt to.det party
‘Mary came to the party’
b. Marie est venue en retard. Eu. French

Mary.nom be.3sg come.prt in late
‘Mary arrived late’ (Sorace 2000: 863)
Sorace (2000) includes, in this group, verbs that are inherently telic. Neverthe-
less, in certain atelic change of location verbs like ‘descend’, French selects for
avoir (have) auxiliary, whereas Italian selects essere (be).
(49) l’ actif net de la Société a baissé sous le

det assets net of det society have.3sg descend.prt under det
minimum.
minimum
‘The net assets of the society descended to the minimum’
On the other hand, change of state verbs show internal variation, as they can
be classified in two groups: those selecting be in both Italian and European
French, and those selecting be in Italian and have in European French. The
verbs of the former group, that is to say, the ones that occur with be also in Euro-
pean French, are telic (50a) (51a). This means that, in French, atelic change of
state verbs (e.g. grandir ‘increase, become bigger’, viellir ‘become older’) occur
with avoir (have) auxiliary (51b).
(50) a. La temperature è salita improvisamente. Italian

det temperature.nom be.3sg rise.prt suddenly
‘The temperature suddenly rose’ (Sorace 2000: 865)
b. Gianni è invecchiato.
John.nom be.3sg become.older.prt
‘John has become older’ (Kayne 2008: 1)
(51) a. L’enfant est devenu triste. Eu. French

det-child.nom be.3sg become.prt sad
‘The child became sad’
b. Cette inquietude a grandi en raison de la

this worry.nom have.3sg increase.prt in reason of det
circulation incontrôlée.
circulation uncontrolled
‘This worry has increased due to the uncontrolled circulation’
c. Jean a vielli.
John.nom have.3sg become.older.prt
‘John has become older’ (Kayne 2008: 1)
Finally, verbs of continuation of a pre-existing state (e.g. ‘last’, ‘survive’ etc.) and
existence of a state (e.g. ‘be’ ‘exist’, ‘belong’, ‘seem’ etc.) select generally be in Ital-
ian (although, according to Sorace 2000, judgments are weaker and have is not
uncommon), and consistently have in European French.
(52) a. La guerra è / ?ha durato a lungo. Italian

det war.nom be.3sg / have.3sg last.prt to long
‘The war lasted for long’ (Sorace 2000: 867)
b. I dinosaur sono esistiti / ??hanno esistito 65

det dinosaurs.nom be.3pl exist.prt have.3pl exist.prt 65
milioni di anni fa.
million P years ago
‘The dinosaurs existed 65 million years ago’ (Sorace 2000: 869)
(53) a. Mes parents ont survécu au

my parents.nom have.3pl survive.prt to.det
tremblement de terre. Eu. French
earthquake
‘My parents survived the earthquake’
b. Les dinosaures ont existé il y a 65 millions

det dinosaurs.nom have.3pl exist.prt there is 65 million
d’ans.
of years
‘The dinosaurs existed 65 million years ago’ (Sorace 2000: 869)
As can be seen, the replacement of be auxiliary by auxiliary have takes place

gradually, more prominently in the groups of verbs that are in the middle of
Sorace’s (2000 2004) hierarchy. Mateu (2009) shows that this tendency is also
158 berro
attested in the diachronic evolution of have replacement in Old Catalan and

in Old Spanish. On the basis of the diachronic studies of Old Catalan done by
Batlle (2002) and of Old Spanish done by Aranovich (2003) and Castillo (2003),
Mateu points out that the verbs that are more variable in Italian are the ones
that lost be auxiliary earlier in both Old Catalan and Old Spanish. According to
Batlle (2002), in Old Catalan, the most innovative verbs—replacing ésser (be)
with haver (have)—were verbs of appearance and existence (‘happen’, ‘occur’,
‘become’), and copulative verbs (‘stand’, ‘rest’, ‘remain’). Telic verbs of change
of state or location were the most reluctant to appear with haver (have), and
consistently chose ésser (be). As Mateu (2000) shows, the same gradual pro-
cess took place in Spanish (Aranovich 2003; Castillo 2002). The process of
replacement of ser by haber started affecting the same classes of verbs: verbs of
appearance and existence, which are actually the verb classes in the middle of
the hierarchy. Therefore, the hierarchy put forward by Sorace (2000 2004) also
holds for diachronic variation, and for the gradual spread of have auxiliary
over be across verb classes in languages that do not have auxiliary alternation
these days.
5.2 Comparison with Basque

Coming back to Basque, this language behaves more like Italian than like
French in the use of the auxiliary in intransitive verbs (see section 3). In Basque
verbs expressing change of location and change of state occur with izan (be)
auxiliary, regardless of their inherent telicity.
(54) a. Prezio-a-k nabarmen jaitsi dira azken

price-det-pl.abs notoriously descend be.3pl.abs last
egun-etan.
day-ine.pl
‘Prices have notoriously descended in the last days’
b. Jon zahartu zen.

John.abs become.older be.3sg.abs.pst
‘John became older’
Regarding the third group of unaccusative verbs outlined before—verbs

expressing continuation of a pre-existing state and existence of a state—there
is more variation in auxiliary selection and subject marking. Certain verbs of
this group occur with edun have auxiliary and an ergative subject: iraun ‘last’,
biziraun ‘survive’, irudi ‘seem, look like’, eman ‘seem’.
(55) a. Guda-k bost urte iraun zuen.

war.det-erg five year last have.3sg.erg.pst
‘The war lasted five years’
b. Oso gaztea ema-ten duzu.

very yound seem-imprf have.2sg.erg
‘You seem very young’
Other verbs of this group, in contrast, select izan (be) auxiliary, like the verb
gustatu ‘like’ (55) and faltatu ‘lack’, and others such as existitu ‘to exist’ can be
found either with izan (be) (57a) or with edun (have) (57b) (selecting edun
(have) particularly in non-standard uses).
(56) Jon-i makarroi-a-k gusta-tzen zaizkio.

John-dat macaroni-det-pl like-imprf be.3pl.abs.3sg.dat
‘John likes macaroni’
(57) a. Munstro-a-k ez dira existi-tzen.

monster-det-pl no be.3pl.abs exist-imprf
b. Munstro-ek ez dute existi-tzen.

monster-erg.pl no have.3pl.erg exist-imprf
‘Monsters do not exist’
As can be seen, Basque intransitive verbs also conform to the hierarchy pro-
posed in Sorace (2000, 2004), as stative verbs are the ones showing more vari-
ation in the selection of have and be auxiliary. In the next section, I will con-
tinue analyzing the variation existing in stative verbs, particularly focusing on
the role of Romance se clitic and its influence on loan verbs in Basque.
6 Romance Loan Verbs and Contact Driven Uses
Basque is in permanent contact with Spanish and French, as all Basque speak-
ers are these days Spanish-Basque or French-Basque bilingual. In this section,
two consequences of this contact will be analyzed, namely, the use of Romance
loan verbs, particularly intransitive verbs, and the emergence of new uses
of certain stative predicates. In both cases, we will see that the presence or
absence of the se clitic in the original Romance verbs is closely related to the
subject case marking and auxiliary selection of the verbs in Basque.
160 berro
6.1 Loan Intransitive Verbs

Alberdi (2003) analyzes Romance loan verbs in Basque and concludes that
there are two factors influencing the marking of the subject and auxiliary selec-
tion of these verbs in Basque: (i) the semantic notion of internal causation
(Levin & Rappaport Hovav 1995), and (ii) the presence of the se clitic in the
original Romance verb. Verbs like bidaiatu ‘travel’ and abortatu ‘abort’ occur
with an ergative subject and with have auxiliary, because both verbs are inter-
nally caused. On the other hand, deskuidatu ‘get confused, lose the head’ and
mutinatu ‘mutiny’ are used with an absolutive subject and be auxiliary, since
the original verbs descuidarse and amotinarse take the se clitic in Spanish.6
Certain predicates are used differently in Spanish and French, and also show
dialectal variation in Basque. For instance, the verb divorce is used with the se
clitic in Spanish (i.e. divorciarse), and without se in French (i.e. divorcer). This
contrast is not surprising, given that ‘divorce’ can be configured as a verb of
change of state—become divorced—or as an internally caused verb—conduct
a divorce.
(58) a. Maria y Juan se han divorciado. Spanish

Mary.nom and John.nom cl have.3pl divorce.prt
‘Mary and John have divorced.’
b. Marie et Jean ont divorcé. French

Mary.nom and John.nom have.3pl divorce.prt
In relation to this contrast, in the dialects that are in contact with Spanish,
dibortziatu ‘divorce’ is used in the “unaccusative” pattern, selecting an absolu-
tive subject and be auxiliary (59a). In contrast, in the dialects that are in contact
with French, dibortziatu takes an ergative subject and have auxiliary (59b).
(59) a. Miren eta Jon dibortziatu dira. Basque (cont. with Sp.)
Mary.abs and John.abs divorce be.3pl.abs
b. Miren-ek eta Jon-ek dibortziatu dute. Basque (cont. with Fr.)

Mary-erg and John-erg divorce have.3pl.erg
6 It must be noted that se influences the alignment only in the case of intransitive loan verbs.
Uses of se in transitive clauses (comerse la manzana [eat-se the apple]) do not seem to alter
the alignment in Basque.
A reversed, but similar, pattern is attested with the verb ‘stroll’. In French,
when it is intransitive, it is a pronominal verb (i.e. se promener), whereas in
Spanish, it is non-pronominal (i.e. pasear).7 As expected, in the Basque dialects
that are in contact with French, promenatu ‘stroll’ combines with an absolutive
subject and be auxiliary, whereas in the dialects that are in contact with Span-
ish, paseatu occurs with an ergative subject and have auxiliary.
(60) a. Xabier promenatu da. Basque (cont. with Fr.)

Xavier.abs stroll be.3sg.abs
b. Xabierr-ek paseatu du. Basque (cont. with Sp.)

Xavier-erg stroll have.3sg.erg
‘Xavier has strolled.’
Another verb that shows (dialectal) variation is entrenatu ‘train’. The verb ‘train’
can be used in the pronominal and non-pronominal form in both Spanish (i.e.
entrenar and entrenarse) and French (s’entraîner). In the dialects that are in
contact with French, speakers prefer to use the intransitive entrenatu in the
unaccusative pattern, namely, with an absolutive subject and be auxiliary. In
contrast, in other dialects (central and western ones), the intransitive entrenatu
is used either in the unergative fashion—with an ergative subject and have
auxiliary—or in the unaccusative one. As in the previous cases, I suggest that
these differences in the subject marking and auxiliary selection are the exter-
nalizations of different event configurations.
(61) a. entrenatu [do training]
7 There is also a variant pasearse in Spanish, with the se clitic, but this one has a slightly differ-
ent meaning, namely, ‘stroll without any clear intention’.
162 berro
b. entrenatu [become trained]
In the unergative pattern, entrenatu would correspond to an internally

caused verb (do training) (61a), whereas in the unergative one, entrenatu would
be a change of state (become trained) (61b).
6.2 New Stative Verbs

In Basque, there is a number of stative verbs that are not accepted or recom-
mended in the standard language (Berro 2015b, in press): pisatu ‘weigh’, neurtu
‘be long’, erre ‘burn, be very hot’, labaindu/irristatu/lerratu ‘be slippery’, usaindu
‘smell’, pikatu ‘be spicy’ and kubritu ‘be deep’. Acceptance varies depending on
the speakers, but young Basque-Spanish bilinguals or bilinguals with strong
Spanish influence accept and use these verbs naturally.
(62) Asko pisa-tzen du [ez dakit eraman ahal izango duzun].

a.lot weigh-imprf have.3sg.erg
‘It is very heavy, I do not know if you are going to be able to carry it.’
(A. Cano, Belarraren Ahoa: 63)
(63) 27 kilometro neur-tzen ditu alde

27 kilometer measure-imprf have.3sg.erg.3pl.abs side
zabal-ean […]
wide-ine
‘It is 27 kilometers long in its wide side.’ (Berria, 2004-12-16)
(64) Plater-a-k erre-tzen du.

dish-det-erg burn-imprf have.3sgerg
‘The dish burns.’
(65) Lurr-a-k laban-tzen du.

floor-det-erg slip-imprf have.3sg.erg
‘The floor is slippery.’
(66) Piper hauek asko pika-tzen dute.

pepper these.erg a.lot burn-imprf have.3pl.erg
‘These peppers are very spicy.’
(67) Lantegi horre-tako ke-a-k oso gaizki usain-tzen

factory that-gen smoke-det-erg very wrong smell-imprf
du.
have.3sg.erg
‘The smoke from that factory smells very bad.’
As can be seen in the examples, these verbs are used in the unergative pat-
tern (with ergative subject and have auxiliary) and they are morphologi-
cally simplex (they do not consist of the light verb egin ‘do’). Remarkably,
these verbs are stative: they denote a property of their sole argument—like
its weigh, length, temperature etc.—and they behave like stative verbs in a
number of linguistic contexts, such as in their incompatibility with process
adverbs and the degree interpretation of the modifier apur bat ‘a little’ (see
Berro 2015b, 2016). The ergative subject of these verbs is not an agent and
causation is not involved, but the verbs are still aligned in the unergative pat-
tern.
Most of these verbs are used standardly in the language, but with eventive
meanings. In order to express these stative meanings, predicative constructions
are used instead:
(68) Hogeita bost bat milimetro luze ziren.

twenty five about millimeter long be.3pl.abs.pst
‘They were about twenty five millimeters long.’ (J. Gartzia, Sistema Peri-
odikoa. Primo Levi: 190)
(69) a. Piper hauek oso pikante-a-k dira.

pepper these.abs very spicy-det-pl be.3pl.abs
b. Piper hauek min-a-k dira.

pepper these.abs spicy-det-pl be.3pl.abs
‘These peppers are very spicy.’
(70) Lurr-a laban-a dago.

floor-det.abs slippery-det be.3sg.abs
164 berro
(71) Jertse-a-k usain txarr-a dauka.

jumper-det-erg smell bad-det.abs have.3sg.erg.3sg.abs
‘The jumper smells bad’; lit.: ‘The jumper has bad smell’.
The emergence of simplex unergative verbs in order to express these stative

meanings seems to be related to the contact with Romance languages. For
instance, the verbs quemar ‘burn’ and resbalar ‘slip’ are also polysemic between
an eventive and a stative interpretation (cf. (72a) and (73a) with (72b) and
(73b)) in Spanish.
(72) a. Juan se resbala en el suelo.

John.nom cl slip.3sg in det floor
‘John slips on the floor.’
b. El suelo resbala.
det floor.nom slip.3sg
(73) a. La casa se quema.

det house.nom cl burn.3sg
‘The house burns.’
b. La sopa quema.
det soup.nom burn.3sg
‘The soup burns’ or ‘The soup is burning hot.’
As it can be observed, in the eventive intransitive use of the verbs, the se clitic is
present. In Basque, in the eventive intransitive use, the verb gets realized in the
unaccusative pattern. Compare the eventive unaccusative verbs in (74a) and
(75a) with the stative unergative uses in (74b) and (75b).
(74) a. Jon lurr-ean irrista-tzen da.

John.abs floor-ine slip-imprf be.3sg.abs
‘John slips on the floor.’
b. Lurr-a-k irrista-tzen du.

floor-det-erg slip-imprf have.3sg.erg
(75) a. Etxe-a erre-tzen da.

house-det.abs burn-imprf be.3sg.abs
‘The house burns.’
b. Zopa-k erre-tzen du.

soup.det-erg burn-imprf have.3sg.erg
‘The soup burns’ or ‘the soup is burning hot.’
Once again, the meaning that has the se clitic in Romance is related to the unac-
cusative pattern in Basque, and similarly, the se-less meaning in Romance is
related to the unergative pattern. Actually, se is ungrammatical in the stative
uses of the verbs in Spanish.8
(76) a. *El suelo se resbala

det floor.nom cl slip.3sg
Intended: ‘The floor is slippery’
b. #La sopa se quema.

det soup.nom cl burn.3sg
Intended: ‘The soup burns’ or ‘The soup is burning hot.’
I propose that, in these cases too, the difference in the subject marking and aux-
iliary selection is derived from different eventive configurations in the syntax.
The unergative stative verbs have the structure in (75a), whereas the eventive
unaccusative verb has the configuration in (75b).
8 In French, there is a dynamic and intransitive use of the verb gliser ‘slide’ which does not take
the se clitic and which requires have auxiliary.
(i) Pierre a glissé sur la glace (et est tombé)
Pierre has slid on the ice (and is fallen)
Pierre slid on the ice (and fell)
Labelle & Doron (2010) claim that, in this case, the root has merged with a dynamic verbal
head (v), instead of with V. As a consequence, the construction focuses on the process, instead
of on the result.
166 berro
(77) a. irristatu [stative]
b. irristatu [eventive]
In the stative variant, Event is not projected, and the subject is introduced in the
specifier of Voice. There, it is interpreted as a holder argument. Recall that the
subject of Voice is only interpreted as an originator when Voice selects for
Event. On the other hand, in the eventive variant, the subject is introduced in
the specifier of State, and after Event is projected, it is interpreted as a patient
(undergoer/resultee).
7 Stative Verbs as Border-Line Cases
In this section, I will propose that stative verbs are the most variable predi-
cates both in Basque and cross-linguistically because their subject—a holder
argument—can be introduced either in the specifier of Voice or in the speci-
fier of State, without that creating an important semantic contrast. This hap-
pens precisely because, in states, the Event head is not projected. Additionally,
I will argue that, in Romance loan verbs, the presence of the se clitic in the
original Romance language forces an unaccusative configuration of the verb
in Basque, because se-variants necessarily project Event + State (Cuervo 2003,
2014).
In section 4.1 and 5.3, we saw that stative verbs are the most variable group
of verbs in terms of auxiliary alternation (and subject case marking) both in
Basque and cross-linguistically. As explained in section 4.1, verbs of existence

of state are in the middle of Sorace’s (2000, 2004) hierarchy, since these are
the verbs showing more cross-linguistic differences and less consistent judg-
ments. On the other hand, we have seen that, in Basque, there are unaccusative
and unergative verbs within the group of stative verbs, and additionally, I have
shown that there is a group of new stative verbs that are aligned in the unerga-
tive pattern, presumably, created under the influence of the Basque-Romance
contact.
In this section, I will offer a syntactic explanation for the great variability
exhibited by stative verbs, based on the event configuration assumed in this
chapter. As explained in section 2, subjects of intransitive verbs can be intro-
duced either in the specifier of Voice or in the specifier of State.
(78) a. [VoiceP Subj Voice [EventP Event [Root]]] Subj = originator, agent
b. [EventP Event [StateP Subj State [Root]]] Subj = patient
It can be observed in (78) that the subject introduced by Voice is interpreted

as an originator or an agent (the agent role being subsumed under the origi-
nator role), and that the subject introduced by State is interpreted as a patient.
However, as explained in section 2, these interpretations are derived from the
whole event configuration. The subject in (78a) is interpreted as an originator
only if Voice selects for Event, and the subject in (78b) is a patient only if State
is selected by Event. If these conditions are not met, the subjects of both Voice
and State are interpreted as holders.
(79) a. [VoiceP Sub Voice [StateP State [Root]]] Subj = holder

b. [StateP Subj State [Root]] Subj = holder
In stative verbs, Event is not projected. Thus, the subject of a stative verb can
be introduced either in the specifier of Voice (79a) or in the specifier of State
(79b), without that creating a significant semantic difference. Note that in even-
tive predicates, having an external or internal subject really makes a difference:
if it is external, it is interpreted as the entity originating the event, whereas if it
is internal, it gets the role of undergoing the event and ending in a final state.
Since there is no Event in state verbs, stative verbs can have external or internal
subjects without much semantic contrast.9
9 There is a contrast that, in Berro (2015a), I claim to be one between containing vs. being
included, or have vs. be In unergative stative verbs, the subject contains or possesses the
168 berro
Regarding the role of the se clitic in Romance and its relation with the
Basque unaccusative pattern, I propose that se unaccusatives in Romance
involve the projection of both Event and State, in the spirit of Cuervo (2003
2014) (see also Cuervo 2003 2014 for Spanish, Folli & Harley 2005 for Italian,
and Labelle 1992, Labelle & Doron 2010 for French).
In the literature, se has been related to telicity in different ways: it has been
argued that se is involved in a configuration having or emphasizing a final state
(Labelle 1992, Cuervo 2003 2014, Folli 2002, Folli & Harley 2005, Legendre &
Smolensky 2009, Labelle & Doron 2010) or a bounded path (Basilico 2010).
Cuervo (2003 2014) analyzes unaccusative verbs that have a se and a se-less vari-
ant (e.g. caer(se) ‘fall’, salir(se) ‘come out/off’, morir(se) ‘die’, ir(se) ‘go’ etc.). On
the basis of several systematic interpretive and formal contrasts between the se
and the se-less variant of these verbs (79), she proposes that these two variants
have different eventive structures, i.e. a bi-eventive vs. a mono-eventive config-
uration. Depending on the structure, the Root is interpreted as Manner or as
Result (Mateu & Acedo-Matellán 2012, see also Labelle & Doron 2010). The se-
less variants are verbs of change or motion (telic or atelic), whereas se variants
are verbs of change of state (necessarily telic). This difference is observed in the
availability of durative and frame modifiers.
(80) a. El avión cayó durante tres minutos / * en tres

det plane.nom fall.3sg.pst for three minutes / * in three
minutos antes de estrellarse.
minutes before crashing
‘The plane fell for three minutes /*in three minutes before crashing.’
b. El avión se cayó * durante tres minutos / en

det plane.nom cl fall.3sg.pst * for three minutes / in
tres minutos (# antes de estrellarse).
three minutes (# before crashing)
‘The place fell down *for three minutes/ in three minutes (before crash-
ing).’
According to Cuervo (2003, 2014), the se-less variant of these verbs is mono-
eventive, consisting of only vGO, whereas the se variant is bi-eventive, having
both vGO and vBE.
property denoted by the state, whereas in unaccusative states, the subject is included within
the set of things having the property denoted by the state.
(81) a. se-less variant

Salieron muchos yuyos
‘There appeared many weeds’
b. se variant
Se salieron tres clavos
‘Three nails came off’ (Cuervo (2014: 51–52))
The light verbs vGO and vBE can be paralleled to the Event and State heads
assumed in this chapter.10 The basic idea is that se-less variants are verbs of
change without entailing a final state, where the subject undergoes the pro-
cess ‘naturally’ (as in the case of snow, rain or leaves in the verb caer ‘fall’). In
contrast, se variants are composed of two subevents, a dynamic subevent and a
result. The subject is the holder the final result. According to Cuervo (2014), the
se clitic is the realization of the unvalued phi-features of vGO, which are valued
with the subject seated in the specifier of vBE.
Assuming Cuervo’s analysis that se variants are bi-eventive, having an Event
and a State subevent, we can understand why originally Romance se-verbs
are analyzed in Basque as unaccusatives (selecting be auxiliary and having an
absolutive subject). In Basque, they are structured in a configuration where the
subject is introduced in the specifier of State, like unaccusatives.
10 Go or become are derived flavors of Event in the assumed event configuration. For
instance, Event is interpreted as trans(ition) if it selects for State and if it is not selected
by Voice.
170 berro
(82) Romance loan verbs with the se clitic
It must be noted, though, that the relation between the unaccusative pattern in
Basque and the presence of the se clitic in Romance is not bi-directional, partic-
ularly in the domain of non-loan verbs. Many verbs that show an unaccusative
alignment in Basque do not have se clitic in their Romance counterparts, for
example, heldu ‘arrive’, etorri ‘come’, handitu ‘increase’, txikitu ‘decrease’, hazi
‘grow’, jaio ‘be born’, gertatu ‘happen’, faltatu ‘lack’ etc. Actually, the contrast
between mono-eventive and bi-eventive unaccusatives proposed in Cuervo
(2003, 2014) does not have a phonological reflex in Basque, since both show the
same alignment pattern (be auxiliary and absolutive subject). The conclusion
that can be drawn is that, for loan verbs, having a se clitic forces an unaccusative
alignment of predicates in Basque, but that, those verbs not having a se clitic
may be subject to other discriminating factors, such as transition, eventivity,
internal causation etc.
8 Conclusion
In this chapter, I have described and analyzed unaccusative and unergative

verbs in Basque, comparing the distribution of these verbs in several Romance
languages, such as French, Italian and Spanish. I have provided a syntactic
explanation for verbs that show unaccusative/unergative alternation (e.g. di-
bortziatu ‘divorce’, entrenatu ‘train’), basically proposing that they are the real-
izations of different syntactic event-configurations. Regarding the great vari-
ability exhibited by stative verbs, both cross-linguistically and in Basque, I have
argued that intransitive stative verbs are most variable because holder subjects
can be introduced either in the specifier of Voice or in the specifier of State.
Since Event is not projected in stative verbs, the specifier of both Voice and
State gets interpreted as a holder, and not as an originator or a patient. Thus,
stative verbs are more variable basically because two syntactic event configu-
rations are compatible with a similar semantic interpretation.
Acknowledgements
This work has been supported by the Basque Government (the research group
IT665–13) and the Spanish Ministry of Economy and Competitiveness
(FFI2014–51878-P). The research leading to these results has also received
funding from the European Union’s Seventh Framework Program for research,
technological development and demonstration under grant agreement no.
613465.
References
Alberdi, Xabier. 2003. “Euskal aditz mailegatuen erregimena: hurbilketa”. In Jesus Mari
Makazaga & Bernard Oyharçabal (eds.), P. Lafitteren sortzearen mendemugako bil-
tzarra. Gramatika gaiak. Iker 14 (I). Bilbao: Euskaltzaindia. 37–60.
Aldai, Gontzal. 2006. “Is Basque morphologically ergative? The semantic split-intransi-
tive case marking system of Western Basque”. In Beatriz Fernandez & Itziar Laka
(eds.), Andolin Gogoan. Essays in honour of Professor Eguzkitza. Leioa: UPV/EHU.
117–138.
Aldai, Gontzal. 2008. “From ergative case marking to semantic case marking: The case
of historical Basque”. In Mark Donohue & Søren Wichmann (eds.), The Typology of
semantic alignment. Oxford: OUP. 197–218.
Aldai, Gontzal. 2009. “Euskararen kasu markaketaren aldakortasun dialektala”. In Beat-
riz Fernández, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: aldako-
rtasun sintaktikoa aztergai [Supplements of ASJU LII]. Donostia: UPV/EHU and
Gipuzkoako Foru Aldundia. 5–19.
Aranovich, Raúl. 2003. “The semantics of auxiliary selection in Old Spanish”. Studies in
Language 27-1: 1–37.
Basilico, David. 2010. The se Clitic and its Relationship to Paths. Probus 22: 271–302.
Batlle, M. 2002. L’expressió dels temps compostos en la veu mitjana i la passive pronomi-
nal. El process de subsitutció de l’auxiliar ‘ésser’ per ‘haver’. Barcelona: Institut d’Estu-
dis Catalans / Publicacions de l’Abadia de Montserrat.
Belleti, Adriana. 1988. “The case of unaccusatives”, Linguistic Inquiry 19: 1–34.
Belleti, Adriana. 1999. ““Inversion” as focalization and related questions”, Catalan Work-
ing Papers in Linguistics 7: 9–45.
Bentley, Delia & Thórhallur Eythórsson. 2003. “Auxiliary selection and the semantics of
unaccusativity”. Lingua 114: 447–471.
Berro, Ane. 2010. “Unergative Predicates in Basque Varieties: Consequences for the
Ergative Case Assignment”. DEA research work. University of the Basque Country
(UPV/EHU).
172 berro
Berro, Ane. 2012. “Three Levels of Root Insertion in Basque Intransitive Verbs”. In
Ernestina Carrilho & Beatriz Fernández (eds.), Journal of Portuguese Linguistics.
Special Issue, 11-1. Ediçoes Colibrí-Universidade de Lisboa. 7–22.
Berro, Ane. 2015a. Breaking verbs. From event structure to syntactic categories in Basque.
Ph.D. diss., University of the Basque Country (UPV/EHU) and Université Bordeaux
Montaigne (UBM).
Berro, Ane. 2015b. “Egoera predikatu berriak euskaraz”. In Irantzu Epelde (ed.), Euskal
hizkera eta dialekotoak gaur. Special Issue of Lapurdum 3: 11–25.
Berro, Ane & Ricardo Etxepare. 2017. “Ergativity in Basque”. In Jessic Coon, Lisa M.
Travis & Diane Massam (eds.), The Oxford Handbook of Ergativity. Oxford University
Press.
Bittner, Maria & Ken Hale. 1996. “The structural determination of case and agreement”.
Linguistic Inquiry 27: 1–68.
Bobaljik, Jonathan D. 1993. “On Ergativity and Ergative Unergatives”. In Collin Phillips
(ed.), MIT Working Papers in Linguistics 19: Papers on Case and Agreement II. Cam-
bridge: MIT. 45–88.
Burzio, Luigi. 1981. Intransitive verbs and Italian auxiliaries. Ph.D. diss., MIT. Cambridge,
MA.
Burzio, Luigi. 1986. Italian Syntax. Dordrecht: Reidel.
Castillo, María Elena. 2002. Inacusatividad y aspecto léxico en los verbos de movimiento.
Estudio diacrónico, Series Documenta Universitaria. Scripta I, Girona: Edicions a Peti-
ció SL:
Cuervo, María Cristina. 2003. Datives at Large. Ph.D. diss., MIT.
Cuervo, María Cristina. 2014. “Alternating unaccusatives and the distribution of roots”,
Lingua 141: 48–70.
Davidson, Donald. 1967. “The Logical Form of Action Sentences”. In Nicholas Rescher
(ed.), The Logic of Decision and Action. Pittsburgh: University of Pittsburgh Press.
[Reprinted in Davidson 2006.]
Davidson, Donald. 2006. The Essential Davidson. Oxford: Oxford University Press.
Dixon, Robert M.W. 1979. “Ergativity”, Language 55: 59–138.
Dixon, Robert M.W. 1994. Ergativity. Cambridge: Cambridge University Press.
Donohue, Mark & Søren Wichmann. 2008. The typology of semantic alignment. Oxford:
OUP.
Etxepare, Ricardo. 2003. “Valency and Argument Structure in the Verb”. In José Igna-
cio Hualde & Jon Ortiz de Urbina (eds.), A grammar of Basque. Berlin: Mouton de
Gruyter. 369–465.
Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. Ph.D. thesis, Univer-
sity of the Basque Country (UPV/EHU).
Folli, Rafaella. 2002. Constructing telicity in English and Italian. Oxford: Oxford Univer-
sity Press.
Folli, Rafaella & Heidi Harley. 2005. “Flavors of v”. In Paula Kempchinsky & Roumaya
Slabakova (eds.), Aspectual Inquiries. Dordrecht: Springer. 95–120.
Hale, Ken & S. Jay Keyser. 1993. “On Argument Structure and the Lexical Expression of
Syntactic Relations”. In Ken Hale and S. Jay Keyser (eds.), The View from Building 20.
Cambridge, Massachusetts: The MIT press. 53–109.
Hale, Ken & S. Jay Keyser. 2002. Prolegomenon to a Theory of Argument Structure. Cam-
bridge, Massachusetts: The MIT Press.
Hale, Ken & S. Jay Keyser. 2005. “Aspect and the Syntax of Argument Structure”. In Nomi
Erteschik-Shir & Tova Rapoport (eds.), The Syntax of Aspect: Deriving Thematic and
Aspectual Interpretation. New York: Oxford University Press. 11–42.
Harley, Heidi. 1995. Subjects, Events and Licensing. Ph.D. diss., MIT.
Harley, Heidi. 2002. “Possession and the Double Object Construction”. In Linguistic Vari-
ation Yearbook 2: 29–68. Amsterdam: John Benjamins.
Harley, Heidi. 2005. “How Do Verbs Get Their Names? Denominal Verbs, Manner
Incorporation and the Ontology of Roots in English”. In Nomi Erteschik-Shir &
Tova Rapoport (eds.), The Syntax of Aspect. New York: Oxford University Press. 42–
64.
Kayne, Richard S. 2008. “A note on auxiliary alternations and silent causation”. Ms., New
York University, New York.
Keller, Frank & Antonella Sorace. 2003. “Gradient auxiliary selection and impersonal
passiviation in German: an experimental investigation”. Journal of Linguistics 39:
57–108.
Kratzer, Angelika 1994. “The Event Argument and the Semantics of Voice”. Ms., Univer-
Kratzer, Angelika. 1996. “Severing the External Argument from its Verb”. In Johan Roo-
rych & Laurie Zaring (eds.), Phrase Structure and the Lexicon. Studies in natural
language and linguistic theory, vol. 33. Dordrecht: Kluwer Academic Publishers. 109–
137.
Labelle, Marie. 1992. “Change of state and valency”. Journal of Linguistics 28: 375–414.
Labelle Marie & Edit Doron. 2010. Anticausative derivations (and other valency alter-
nations) in French. Probus 22–2: 303–316.
Laka, Itziar. 1993. “Unergatives that Assign Ergative, Unaccusatives that Assign Accu-
sative”. In Jonathan D. Bobaljik & Collin Phillips (eds.), Papers on Case and Agree-
ment I, MITWPL 18. Cambridge, MA: MITWPL. 149–172.
Legendre, Géraldine & Paul Smolensky. 2009. “French inchoatives and the Unaccusativ-
ity Hypothesis”. In Donna B. Gerdts, John Moore & Maria Polinsky (eds.), Hypothesis
A/Hypothesis B: Linguistic Explorations in Honor of David M. Perlmutter. MIT Press.
Levin, Beth. 1983. On the Nature of Ergativity. Ph.D. diss., MIT.
Levin, Beth & Malka Rappaport Hovav. 1995. Unaccusativity. At the Syntax-Lexical Se-
mantic Interface. Cambridge, MA: the MIT Press.
174 berro
Mackenzie, Ian. 2006. Unaccusative verbs in Romance Languages. Hampshire/New

York: Palgrave.
Mateu, Jaume. 2002. Argument structure: relational construal at the syntax-semantic
interface. Ph.D. diss., Universitat Autònoma de Barcelona.
Mateu, Jaume. 2009. “Gradience and auxiliary selection in Old Catalan and Old Span-
ish”. In Paola Crisma & Giuseppe Longobardi (eds.), Historical syntax and linguistic
theory. Oxford: Oxford University Press. 176–193.
Mateu, Jaume & Víctor Acedo-Matellán. 2012. “The manner/result complementarity
revisited: a syntactic approach”. In María Cristina Cuervo & Yves Roberge (eds.),
The end of argument structure? Vol. 38. Syntax & Semantics. Bingley: Emerald. 209–
228.
McClure, William. 1990. A lexical semantic explanation for unaccusative mismatches.
In Katarzyna Dziwirek, Patrick Farrell & Errapel Mejias-Bikandi (eds.), Grammatical
Relations. A cross-theoretical perspective. Stanford: CSLI Publications.
Merlan, Francesca. 1985. “Split intransitivity: functional oppositions in intransitive
inflection”. In Johanna Nichols & Anthony Woodbury (eds.), Grammar inside and
outside the clause: some approaches to theory from the field. Cambridge: Cambridge
Ortiz de Urbina, Jon. 1989. Parameters in the Grammar of Basque. Studies in Generative
Grammar 33. Dordrecht: Foris Publications.
Oyharçabal, Bernard. 1992. Structural and Inherent Case-Marking; Ergaccusativity in
Basque. In Joseba A. Lakarra & Jon Ortiz de Urbina (eds), Syntactic Theory and
Basque Syntax. [Supplements of ASJU XXVII]. Donostia: UPV/EHU and Gipuzkoako
Foru Aldundia. 309–342.
Oyharçabal. Bernard. 2006. “Basque Light Verb Constructions”. ASJU XL: 787–806.
Perlmutter, David M. 1978. Impersonal Passives and the Unaccusative Hypothesis. Pro-
ceedings of the 4th Annual Meeting of the Berkeley Linguistics Society. 157–190.
Perlmutter, David M. 1989. “Multiattachment and the unaccusative hypothesis: the Per-
fect Auxiliary in Italian”. Probus 1–1: 63–119.
Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First Phase Syntax. Cam-
Rosen, Carol. 1981. The relational structure of reflexive clauses: evidence from Italian.
Ph.D. diss., Harvard University, Cambridge, MA.
Salaburu, Pello. 1992. “Euskara, hizkuntza ergatiboa ote da?”. In J.H. Duhalde (ed.), Luis
Villasanteri omenaldia. Bilbao: Euskaltzaindia. 417–433.
Sankoff, Gillian & Pierrette Thibault. 1977. “L’alternance entre les auxiliaries avoir et
être en français parlé à Montréal”. Langue Français 34: 81–108.
Sapir, Edward. 1917. Review of C.C. Uhlenbeck “Het passieve karakter van het verbum
transitivum of van het verbum actionis in talen van Noord-Amerika”, International
Journal of American Linguistics 1: 82–86.
Sarasola, Ibon 1977. “Sobre la bipartición inicial en el análisis en constituyentes”, ASJU

XI: 51–90.
Sorace, Antonella. 1993. “Unaccusativity and auxiliary choice in non-native grammars
of Italian and French: asymmetries and predictable indeterminacy”. Journal of
French Language Studies 3: 71–93.
Sorace, Antonella. 2000. “Gradients in auxiliary selection with intransitive verbs”. Lan-
guage 76: 859–890.
Sorace, Antonella. 2004. “Gradience at the Lexicon-Syntax Interface: Evidence from
Auxiliary Selection”. In Artemis Alexiadou, Elena Anagnostopoulou & Martin Ever-
aert (eds.), The Unaccusativity Puzzle. Oxford: Oxford University Press. 243–268.
Tenny, Carol Lee. 1987. Grammaticalizing aspect and affectedness. Ph.D. diss., MIT. Cam-
bridge, MA.
Torrego, Esther. 1989. “Unergative-unaccusative alternations in Spanish”. In Itziar Laka
& Anoop Mahajan (eds.), Functional heads and clause structure: MIT working papers
in linguistics, vol. 10. Cambridge, MA. 253–272.
Uribe-Etxebarria, Myriam. 1989. “Some Notes on the Structure of IP in Basque”. Ms.,
UConn, Storrs.
Zabala, Igone. 2004. “Los predicados complejos en vasco”. In Igone Zabala, Elixabete
Pérez Gaztelu & Lluïsa Gràcia Sole (eds.), Las fronteras de la composición en lenguas
románicas y en vasco. Bilbao: Deustuko Unibertsitatea, Argitalpen Zerbitzua. 445–
534.
chapter 6
Light Verb Constructions in Basque and Romance

Víctor Acedo-Matellán and Anna Pineda
1 Introduction
1.1 What is a Light Verb Construction?

When Virginia Woolf writes in her diary I took a walk by myself, instead of I
walked by myself, she is choosing to linguistically construe a particular con-
ceptual scene by means of a L(ight) V(erb) C(onstruction). In these construc-
tions the verb—took—bears little semantic weight, this being encoded in a
N(on-)V(erbal) E(lement)—a walk. LVC s, such as take a walk, have a bath, or
give a cuddle, are certainly not a quirk of English. In fact, as research over the
last 25 years has shown, they are abundant crosslinguistically and in some lan-
guages like Basque they are actually the unmarked way to encode certain types
of events.1
While the format of LVC s varies both intra- and cross-linguistically, it dis-
plays certain constant features. As their name suggests, LVC s are headed by a
verb that, at least within the construction, has a very abstract semantic content.
In fact, the semantic contribution of the L(ight) V(erb) seems to be limited to
functional notions related to argument and event structure. On the other hand,
the NVE, whose lexical category may change across LVC s and also depending on
the language, largely encapsulates the conceptual content of the construction,
that is, the particular event evoked.2 To show how this is actually the division
of labour between the LV and the NVE, it is useful to compare, on the one hand,
two constructions sharing the same LV but differing in their NVE and, on the
1 See Butt (2010) for an overview, and for useful terminological precisions, notably the dis-
tinction between LVC and complex predicate. See also Bowern (2008). On the other hand, as
pointed out by Lin (2001:22), the use of light verb as related to LVC s is not exactly equivalent
to what is known as Chomskyan light verb (Chomsky 1995), that is, a functional verb that takes
the VP as complement and that is responsible for the projection of the external argument and
for the assignment of case to the object (see Larson 1988 and Hale & Keyser 1993 for seminal
proposals).
2 Although we use NVE throughout this work, since it constitutes a good label considering the
kind of LVC s that we encounter in Basque and Romance, it should be noted that in other lan-
guages like Urdu the element containing the conceptual portion of the LVC can be of verbal
category (see Butt 1995, 2010).

light verb constructions in basque and romance 177
other, two constructions sharing the same NVE but being different with respect
to their LV. We can illustrate the former case with the minimal pair involving
the Spanish LVC s dar miedo “give fear” ‘frighten’ and dar vergüenza “give embar-
rassment” ‘embarrass’. These LVC s share the same LV, dar ‘give’, and the abstract
semantics of transferal inherent to this verb is preserved in these constructions.
Thus, both LVC s are construed with a dative understood as the experiencer of
the emotion:3
(1) a. A Marisa le dan miedo las películas de terror.

dat Marisa dat.3sg give.3pl fear the.pl films.pl of horror
‘Horror films frighten Marisa.’
b. A Marisa le da vergüenza hablar en público.

dat Marisa dat.3sg give.3sg embarrassment talk.inf in public
‘Marisa feels embarrassed to talk in public.’
The difference between the two LVC s is exclusively of conceptual nature, and
corresponds to the difference between the emotions of fear and embarrass-
ment, as encoded in their NVE s. On the other hand, to exemplify the case of
two LVC s differing in their LV but not in their NVE we may keep dar miedo and
compare it with pasar miedo “pass fear” ‘be afraid’:4
(2) a. A Marisa le dan miedo las películas de terror.

dat Marisa dat.3sg give.3pl fear the.pl films.pl of horror
‘Horror films frighten Marisa.’
b. Marisa pasa miedo con las películas de terror.

Marisa pass.3pl fear with the.pl films.pl of horror
‘Marisa feels fear with horror films.’
In this case the same emotion (fear) is involved, but the argument-structure
properties are different. Thus, pasar miedo requires a subject experiencer and
3 The abbreviations used in glosses are the following ones: abs ‘absolutive’, adv ‘adverbial suf-
fix’, aux ‘auxiliary’, compl ‘complementizer’, dat ‘dative’, det ‘determiner’, erg ‘ergative’,
F ‘feminine’, ines ‘inessive’, inf ‘infinitive’, instr ‘instrumental’, loc ‘locative’, M ‘masculine’,
pl ‘plural’, prs ‘present’, pst ‘past’, ptve ‘partitive’, sg ‘singular’. Also, note that although we
generally provide a full-fledged gloss for Basque auxiliaries, in some instances we will just
gloss aux for reasons of space.
4 See, however, section 2.1, where we provide an account of several pairs of synonymous LVC s
sharing a NVE and differing in their LV s.
178 acedo-matellán and pineda
the stimulus can be expressed as a PP headed by con ‘with’, while dar miedo
appears with a subject stimulus and a dative experiencer. This difference, as
is evident, can only depend on the choice of LV. Interestingly, however, this
choice is not completely free: research has shown that the semantic properties
of the NVE may determine, at least to some extent, which LV it combines with
(Alba-Salas 2002: 51, and references cited therein). At the extreme of seman-
tic cohesion between the LV and the NVE lie, of course, the many LVC s with
idiomatic meaning, that is, a meaning that cannot be computed straightfor-
wardly from the sum of the meanings of the LV and the NVE, as that shown
by the Basque LVC s iskin egin “corner do” ‘elude’ or hanka egin “leg do” ‘escape
rapidly’.
A prominent fact about LVC s is that they tend to correlate with a verb that
expresses more or less the same meaning, as was shown above with the pair
take a walk and walk, and can be further illustrated with the pair dar vergüenza
“give embarrassment” and avergonzar ‘embarrass’ in Spanish. It is important to
emphasize that the semantic equivalence between these pairs is not complete,
as seminally pointed out by Wierzbicka (1982). In any case, LVC s do not always
count with synthetic counterparts. In fact, in languages like Basque most of
them do not (Martinez 2015: 30, Aldai 2009: 823). The generality of the ana-
lytic, LVC-based expression of a large set of events in Basque is, as we will see,
one of the major differences when we compare this language with modern
Romance.
1.2 A Brief State of the Art

LVC s have come to occupy an important place in the agenda of linguistic
descriptions and syntactic theory ever since Jespersen (1954) coined the term
light verb.5 This label was taken quite literally in the first papers dedicated
specifically to LVC s. Indeed, with some exceptions such as Wierzbicka’s (1982),
who studies the semantic contribution of have in constructions such as have
a bath, a leading idea in the 70’s and 80’s was that the LV is completely devoid
5 In this short introduction we cannot do justice to the vast literature on LVC s, which encom-
passes a considerable number of languages—Asian ones featuring very prominently. Works
providing general remarks on LVC s, or comparing different languages include Butt & Geuder
(2001), Butt (2003, 2010), Bowern (2008), Butt & Lahiri (2013) and Ramchand (2014). For par-
ticular languages, see, among others, Jespersen (1954), Jackendoff (1974), Wierzbicka (1982),
Cattell (1984) and Kearns (1988), on English; Grimshaw & Mester (1988), on Japanese; Ahn
(1991), on Korean; Mohanan (1994), on Hindi; Butt (1995), on Urdu; and Lin (2001), on Chi-
nese. For references on LVC s in Basque and Romance (particularly on French and Spanish)
see the following sections.
of predicational force and only contributes verbhood, and that the argument-
structure properties of the LVC are inherited from the NVE through some mech-
anism. For instance, in an influential paper on Japanese constructions based on
the LV suru ‘do’, Grimshaw & Mester (1988) propose that the LV does not have an
argument structure of its own, and that it’s the nominal NVE that transfers some
or all of its arguments to the LVC. In a similar vein, Cattell (1984) argues that
English LV s are predicationally vacuous. In turn, the lightness of LV s have made
some authors propose that they are the historic result of a process of semantic
bleaching of fully predicational verbs, and consequently, that they can be fur-
ther grammaticalized into auxiliary verbs and, finally, into inflectional markers
(Hook 1991, Hopper and Traugott 1993, 2003).
A different analysis is provided by Kearns (1988) within a lexicalist frame-
work distinguishing between Lexical Conceptual Structure and Syntactic Argu-
ment Structure (Hale & Keyser 1986, Rappaport & Levin 1988, among others).
An LV like give provides a Syntactic Argument Structure, with a particular num-
ber of argument slots in a particular configuration, and also an event argument,
but since the verb is delinked from its Lexical Conceptual Structure, no theta
roles are available. Complementarily, the NVE is associated with a Lexical Con-
ceptual Structure and a thematic array, but it lacks a full Syntactic Argument
Structure. In the LVC these two defective components converge to contribute
their lexical properties, yielding a full predicate.6 In any case, Kearns’s (1988)
work is best known for its contribution to differentiating between T(rue) LV s
like those in Take a walk, Give a groan or Have a bite, and what she calls Vague
Action Verbs, like those heading constructions such as Make an inspection, Give
a demonstration or Do the ironing. The basic difference is that True Light Verbs
combine with NVE s that are not argumental, since they do not pattern with
arguments: they cannot become passive subjects, they cannot be extracted,
they cannot be pronominalized, and they cannot be definite. By contrast, Vague
Action Verbs combine with full-fledged argumental DP s. However, as we will
see below, the range of variation within Basque LVC s (see section 2.2.) casts
some doubts on Kearns’ distinction between the two types of verbs found in
LVC s.
More recently, some of the abovementioned ideas have been to a large extent
disputed. First, the predicational lightness or emptiness of the LV has been
qualified, and standard analyses take the LV to be a predicate and to contribute
to a joint predication with the NVE through some mechanism. For example, in
Butt’s (1995) Lexical-Functional-Grammar approach, an operation called Argu-
6 See Jackendoff (1974) for a pioneering analysis along similar lines.

ment Fusion combines the LV predicate (for instance, of “give” or “let” seman-
tics) and the NVE predicate to yield a single predicate. The new predicational
unit hosts, on the one hand, argument slots directly contributed by either of
the two basic predicates (e.g., an Agent from the LV, and a Theme from the NVE)
and, on the other, argument slots that are the result of the fusion of two argu-
ments slots of the basic predicates (for instance, a Causee that subsumes a Goal
of the LV and an Agent of the NVE). Even more recently, Ramchand (2014) has
taken a step further in this direction. In her analysis, LV s possess argument and
event structures, which are, moreover, identical to those of their non-LV coun-
terparts. Thus the LV give of LVC s like John gave Mary a cuddle is an item that
lexicalizes an event structure containing the three subeventive heads proposed
in Ramchand (2008)—Init(iation), Proc(ess) and Res(ult)—plus a Poss(essive)
head merged under Res and projecting a phrase whose specifier is the recipient
(Mary) and whose complement is the NVE (a cuddle). This structure provides
the skeletal meaning of caused possession and is actually the very same one
proposed for the non-light use of give (see Ramchand 2008: 103 for details).
The difference between light and non-light give boils down to the nature of
the complement: if the complement involves a bare event noun like cuddle,
the interpretation of the construction is that of an LVC. We witness a type of
approach, therefore, in which the emphasis is put on the similarities, rather
than the differences, between the light and the non-light versions of the same
verb.7
Ramchand’s (2014) work on LV s in English, Bengali, and Persian actually
aims at further exploring and validating an empirical claim made by Butt
(2003) and Butt & Lahiri (2013) on the basis of diachronic evidence from Indo-
Aryan languages. Ramchand (2014: 217) dubs this claim Butt’s Generalization:
“Unlike auxiliaries, which may become grammaticalized over time to have a
purely functional use, light verbs always have a diachronically stable corre-
sponding full or “heavy” version in all the languages in which they are found.”
From the strictly diachronic point of view, Butt & Lahiri (2013:7) state that
they “depart from the received view that the existence of a light verb is due
7 Cf. Ramchand’s (2014:218) Semantics of Structure Conjecture on the Limits of Lightness: “The
meaning of a light verb and its corresponding heavy alternant are in a subset-superset rela-
tion in their conceptual semantics, the light version being a proper subset of the heavy. Only
non-syntactic or conceptual information is systematically negotiable within the “same” lexi-
cal item. Anything that is present in the heavy version but not in the light must therefore be
a species of Type B meaning. At its most pared down, a light verb can only be as light as the
structural semantics corresponding to the type A meaning of the pair.”
to a historical process of semantic bleaching” (see also Bowern 2008). Their

results—and those of Butt & Geuder’s (2001) and Ramchand’s (2014)—thus
cast doubt on the position, represented by Hook (1991) and Hopper & Trau-
gott’s (1993, 2003), that LV s may be an intermediary link in a grammatical-
ization cline, sitting between full-fledged verbs and auxiliaries. As we will see
below, synchronic data from Basque and Romance supports Butt’s Generaliza-
tion.
1.3 About This Chapter

In this chapter we will describe LVC s in Basque and compare them to Romance
LVC s, focusing on those Romance languages that are in direct contact with
Basque: French and Spanish.
The structure of the chapter is as follows. In section 2, we provide a descrip-
tive comparison of LVC s in Basque and Romance: we first describe the range
of verbs that can function as LV s in these languages, and, taking Butt’s Gen-
eralization and Ramchand’s (2014) remarks as background, we explore their
argument- and event-structure properties. We first focus on the LV (section 2.1).
In section 2.2, we turn to the NVE, paying attention to its categorial proper-
ties (lexical and phrase-structural category) and its relational properties (case-
related properties and the like). In section 3 we present a discussion of the
(different degrees of) syntactic and semantic cohesion between the LV and
the NVE, with a special focus on idiomaticity. Section 4 concludes and points
out issues that deserve attention in future research. As mentioned above, the
discussion is based on Basque, and Romance, mainly French and Spanish, is
brought in for comparison.
2 A Descriptive Comparison of Light Verb Constructions in Basque

and Romance
While Basque and Romance LVC s share a number of properties that are also
common to LVC s crosslinguistically, they also show some differences, partly
due to independent syntactic properties of the two linguistic systems, like case.
In this section, we offer a description of the morphosyntactic and semantic
properties of LVC s in Basque and compare them to Romance LVC s. We focus
first on the LV and then on the NVE and its cohesion with the LV.
2.1 The Light Verb

In Basque, most LVC s are formed with the LV egin ‘do’:8
(3) agur egin “do goodbye” ‘greet / say goodbye’, alde egin “do side / zone”
‘leave’, aharrausi egin “do yawn” ‘yawn’, aipu egin “do mention” ‘mention’,
aitor egin “do confession” ‘confess’, amets egin “do dream” ‘dream’, argi egin
“do light” ‘light up / enlighten’, arnas(a) egin “do breath” ‘breathe’, barre
egin “do laugh” ‘laugh’, behaztopa egin “do impediment” ‘bump into / com-
plicate’, bekatu egin “do sin” ‘sin’, berba egin “do word” ‘talk’, bultza egin “do
push” ‘push’, dantza egin “do dance” ‘dance’, dei egin “do call” ‘call’, domin-
istiku egin “do sneeze” ‘sneeze’, duda egin “do doubt” ‘doubt’, elurra egin
“do snow” ‘snow’, erregu egin “do plea / prayer” ‘beg / pray’, euria egin “do
rain” ‘rain’, eztul egin “do cough” ‘cough’, gogoeta egin “do reflection” ‘think,
reflect’, hitz egin “do word” ‘talk’, huts egin “do mistake / shortage” ‘fail /
miss’, ihes egin “do escape” ‘run away’, irain egin “do insult” ‘insult’, irrib-
arre egin “do smile” ‘smile’, iruzur egin “do trick” ‘mislead, deceive’, jauzi
egin “do jump” ‘jump’, kasu egin “do attention” ‘pay attention’, keinu egin
“do gesture / sign” ‘gesture, sign’, kontu egin “do care” ‘look after, take care’,
korrika egin “do running (adv.)” ‘run’, lan egin “do work” ‘work’, laztan egin
“do cuddle / caress / kiss” ‘caress, kiss’, lo egin “do sleep” ‘sleep’, min egin “do
pain” ‘hurt’, negar egin “do weep” ‘cry’, oihu egin “do shout, yell” ‘shout, yell’,
ospa egin “do “get out!”” ‘leave’, salto egin “do jump” ‘jump’, tiro egin “do
gunshot” ‘shoot’, topo egin “do stumble” ‘run into’, txalo egin “do applause”
‘applaud’, zin egin “do oath” ‘swear under oath’, among many others.9
However, Basque has also some LVC s with hartu ‘take’ (4), eman ‘give’ (5) and
eduki ‘have’ (Martinez 2015:113–114, Zabala 2004:486–488):10
8 In the Basque tradition, LV s are referred to as egin inergatiboak ‘do unergatives’ or

I[zen]+egin inergatiboak ‘do+N[oun] unergatives’. In the French linguistic tradition the
term verbe support ‘support verb’ is the one most widely used, while in the Spanish tradi-
tion we find verbo soporte ‘support verb’, verbo de apoyo ‘verb of support’, and verbo liviano
‘light verb’.
9 As seen, Basque LVC s describe daily, general activities (viewed as recognizable, unitary
concepts), a well-described tendency among this kind of constructions cross-linguistically
(Mithun 1984: 850, Martinez 2015:42, fn. 43).
10 Rodríguez & García Murga (2003) add to this sample LVC s with jo ‘hit, touch, give with
no delicacy’, such as adarra jo “give the horn” ‘tease’. Zabala (2004:461, 465, 488–493) pro-
vides an even wider range of “complex predicates” including verbs such as bete ‘fill’ as
in hitza bete ‘keep one’s word’, hautsi ‘break’ as in hitza hautsi ‘not keep one’s word’, jan
‘eat’ as in hitza jan ‘fail to keep one’s word’ or burua jan ‘wash someone’s brain’, and so
on. These expressions, combining a fully contentful verb and an absolutive case-marked
(4) arnasa hartu “take breath” ‘breathe’, atseden hartu “take repose, rest” ‘rest’,
atsegin hartu “take pleasure” ‘feel happy, enjoy’, damu hartu “take pity,
regret” ‘feel sad, regret’, esku hartu “take hand” ‘participate, take part’, gogo
hartu “take mind, will” ‘decide’, hitz hartu “take word” ‘compromise’, indar
hartu “take strenght” ‘strengthen’, kontu hartu “take care” ‘take care of,
keep in mind’, lur hartu “take land” ‘land’, min hartu “take pain” ‘get hurt’,
parte hartu “take part”, pena hartu “take pity” ‘feel sad’, plazer hartu “take
pleasure” ‘enjoy’,11 poz hartu “take happiness” ‘feel happy’, su hartu “take
fire” ‘burn’, and few others.
(5) amore eman “give love” ‘yield, give in, collapse’, argi eman “give light” ‘light
up / enlighten’, arrazoi eman “give reason” ‘agree with’, aurpegi eman “give
face” ‘face’, begi eman “give eye” ‘look at’, bihotz eman “give heart” ‘encour-
age’, bizkar eman “give one’s back” ‘turn one’s back on’, buru eman “give
head” ‘finish, accomplish / face’, damu eman “give pity” ‘sadden’, errieta
eman “give scolding” ‘tell off’, esker eman “give thanks” ‘thank’, fede eman
“give faith” ‘trust, vouch for’, gogo eman “give mind, will” ‘look, pay atten-
tion / think, consider’, hats eman “give breath” ‘blow’, ikara/beldur eman
“give fear” ‘frighten’, kontseilu eman “give advise” ‘advise’, kontu eman “give
account” ‘give an explanation, give account’, laztan eman “give cuddle,
hig” ‘hug, caress’, leku eman “give place” ‘allow’, lur eman “give land” ‘bury’,
min eman “give pain” ‘hurt’, musu eman “give kiss” ‘kiss’, parte eman “give
report” ‘report, inform’, pena eman “give pity” ‘make someone feel sad’, poz
eman “give happiness” ‘make happy’, su eman “give fire” ‘set fire to’, and few
others.
(6) arrazoi eduki “have reason” ‘be right’, esku eduki “have hand” ‘help, sup-
port’, kontu eduki “have care” ‘be careful’, min eduki “have pain” ‘feel pain’,
on eduki “have good (adjective)” ‘maintain, claim’, pena eduki “have
pity” ‘feel sad’, and few others.
DP, clearly fall beyond the scope of this chapter, which is mainly devoted to LVC s formed
by a LV and a bare noun (hitz egin ‘talk’)—or by a full verb and a bare noun (hitz eman
‘promise’). Moreover, since such cases feature absolutive-marked DP s (adarra ‘(the) horn’,
hitza ‘(the) word’ …), it seems more appropriate to view them as the result of a metaphor-
ical (idiomatic) interpretation of expressions which are not bona fide LVC s.
11 This example and the two preceding ones correspond to psych predicates, which fre-
quently showed a similar pattern in Old Romance languages, as in Old Catalan prendre
delit, Old French prendre plaisir or Old Spanish tomar plazer “take pleasure”, ‘enjoy’. See
Acedo-Matellán & Mateu (to appear) for a detailed description of analytic psych predi-
cates in Old Catalan.
There are also some expressions involving eman ‘give’ and hartu ‘take’ that
show an extreme degree of coalescence between the LV and the NVE, since they
are actually not separable. They are arguably not LVC s anymore, but plain, syn-
thetic verbs:
(7) onartu “take good” ‘accept’, atzeman “give hoof” ‘catch, understand’, haute-
man “give choice” ‘realise, notice’ / choose, hitzeman “give word” ‘promise’
For French and Spanish, Alba-Salas (2002: 636–637) provides the following
respective lists, in addition to faire (French) and hacer (Spanish) ‘do’:
(8) asséner ‘deal’ (asséner un coup ‘hit’), avoir ‘have’ (avoir une maladie ‘be
sick’), coller ‘give’ (coller une claque ‘smack’), commetre ‘commit’ (commet-
tre une aggression ‘commit an assault’), donner ‘give’ (donner un conseil
‘give some advice’), dire ‘say’ (dire des compliments ‘pay some compli-
ments’), diriger ‘direct’ (diriger une offensive ‘attack’), effectuer ‘carry out’
(effectuer un paiement ‘make a payment’), entamer ‘initiate’ (entamer une
discussion ‘initiate a discussion’), ficher ‘give’ ( ficher une giffle ‘slap’), flan-
quer ‘give’ ( flanquer un coup de pied ‘kick’), foutre ‘give’ ( foutre une claque
‘smack’), jeter ‘throw’ ( jeter un regard ‘cast a glance’), lancer ‘launch’
(lancer une attaque ‘launch an attack’), livrer ‘carry out’ (livrer une bataille
‘engage a battle’), passer ‘make’ (passer un coup de téléphone ‘make a
phone call’), porter ‘carry’ (porter un coup ‘strike’), poser ‘put’ (poser une
question ‘ask a question’), prendre ‘take’ (prendre une décision ‘make a
decision’), prêter ‘lend’ (prêter attention ‘pay attention’), procéder (à) ‘pro-
ceed (to)’ (procéder a une lecture ‘proceed to a reading’), subir ‘undergo’
(subir une métamorphose ‘undergo a metamorphosis’)
(9) asestar ‘give’ (asestar un golpe ‘hit’), coger ‘catch’ (coger envidia ‘get envi-
ous’), cometer ‘commit’ (cometer un asesinato ‘commit murder’), contraer
‘contract’ (contraer matrimonio ‘marry’), correr ‘run’ (correr peligro ‘be
in danger’), dar ‘give’ (dar un paseo ‘take a walk’), decir ‘say’ (decir un
cumplido ‘pay a compliment’), dirigir ‘direct’ (dirigir una crítica ‘criticize’),
echar ‘throw’ (echar una mirada ‘give a look’), efectuar ‘carry out’ (efectuar
un pago ‘make a payment’), ejercer ‘make’ (ejercer presión ‘put pressure’),
experimentar ‘experience’ (experimentar una mejora ‘improve’), llevar a
cabo ‘carry out’ (llevar a cabo una privatización ‘privatize’), meter ‘put’
(meter una bofetada ‘slap’), pegar ‘give’ (pegar una patada ‘kick’), perder
‘lose’ (perderle el respeto (a alguien) ‘lose respect (for someone)’), pillar
‘catch’ (pillar celos ‘get jealous’), presentar ‘present’ (presentar la dimi-
sión ‘resign’), prestar ‘lend’ (prestar atención ‘pay attention’), proceder

(a) ‘proceed (to)’ (proceder a una votación ‘proceed to a vote’), propinar
‘give’ (propinar una paliza ‘give a beating’), realizar ‘carry out’ (realizar
una transacción ‘carry out a transaction’), soltar ‘let out’ (soltar un grito
‘shout’), sufrir ‘undergo’ (sufrir un accidente ‘have an accident’), tener
‘have’ (tener hambre ‘be hungry’), tomar ‘take’ (tomar represalias ‘take
reprisals’)
The above lists, composed in alphabetical order, mask crucial asymmetries

among the verbs: notably, they put basic LV s like French donner ‘give’ or Span-
ish echar ‘throw’ at the same level as other “heavier” verbs like French diriger
‘direct’ or Spanish ejercer ‘exert’. Thus, for the more restricted LV s, there is fre-
quently a true lexical, i.e., non-semantic relation between the LV and the NVE
(see also Alonso-Ramos 1998). This is clearly the case with LV s like Spanish cor-
rer ‘run’, selected by the NVE peligro ‘danger’ in correr peligro ‘be in danger’.
Other verbs can arguably be rendered as register-marked alternates of more
basic LV s. For instance, Spanish ejercer presión ‘exert pressure’ constitutes a
formal rendition of hacer presión ‘make pressure’. French and Spanish, with
Basque, converge, therefore, with the rest of languages in what constitutes the
basic repertoire of LV s: DO, GIVE, HAVE, TAKE, etc. Nonetheless, as we will see
below, there is a difference between French and Spanish as regards the choice
of their most basic activity verb, which is faire ‘do’ for the former and dar ‘give’
for the latter. Another important difference between, this time, French on the
one hand and Basque and Spanish on the other, is that DO may head stative,
attributive predicates in the former, but not in the latter: Ce monsieur fait vieux
(Giry-Schneider 1986: 50) “This gentleman does old” ‘This gentleman has an
old appearance’. This brings French faire closer to Catalan fer than to Spanish
hacer or Basque egin.
In section 1.1, we mentioned that the choice of a given LV is not completely free,
but that the semantic properties of the NVE may determine the LV chosen. This
can be seen when comparing, for example, English and Basque LVC s (see (10)):
as Martinez (2015: 78) points out, most English LVC are also found in Basque.
However, this is only a tendency, and English and Basque semantically equiva-
lent LVC s may also bear different LV s—cf. English have and give vs Basque egin
‘do’ in (11)—without any relevant difference regarding event- and argument-
structure properties:
(10) English Basque
take a break atseden hartu

take part parte hartu
give a choice aukera eman
give a kiss musu eman
give a sight begi eman
give (an) account kontu eman
have an intention intentzio eduki
do a trick tranpa egin
(11) English Basque
have a cry negar egin “do weep”

give a shout garrasi egin “do shout”
have a thought hausnar egin “do meditation”
have a dream amets egin “do dream”
give a call dei egin “do call”
This random comparison between English and Basque, and particularly the
mismatches in (11), actually lead us to discuss another important issue that
was introduced in section 1.1: the division of syntactic and semantic labour in
LVC s. Thus, whereas the semantic contribution of the LV is limited to functional
notions related to argument and event structure, the NVE encapsulates the con-
ceptual content of the LVC (the particular event evoked). Interestingly, Basque
offers some examples where the same NVE can combine with different LV s, the
resulting LVC s showing, however, no difference in meaning, at least at the con-
ceptual level:
(12) a. argi egin/eman “do/give light” ‘light up, enlighten’

b. arnasa egin/hartu “do/take breath” ‘breathe’
c. atsegin egin/eman “do/give pleasure” ‘please’
d. dei egin/eman “do/give call” ‘call’
e. esker egin/eman “do/give thanks” ‘thank’
f. huts egin/eman “do/give failure” ‘fail’
g. min egin/eman “do/give pain” ‘hurt’
Moreover, in the Romance area, similar examples are found when compar-
ing Spanish and French:
(13) Spanish French Meaning
dar miedo “give fear” faire peur “do fear” ‘frighten’

dar vergüenza “give shame” faire honte “do shame” ‘shame’
dar pena “give sorrow” faire de la peine “do of ‘provoke sorrow’
the sorrow”
However, these facts do not necessarily constitute counterexamples to the for-

mer claim regarding the division of labour between the LV and the NVE. Note
that in all these pairs from Basque and Romance the event and argument struc-
ture remains the same: for example, both dar miedo “give fear” and faire peur
“do fear” require a source of the emotion (the stimulus) as well as a dative-
marked argument designating the experiencer of the emotion. Thus, it seems
that these exceptional cases can be explained via the semantic closeness of DO
and GIVE, attested in many languages, both verbs being usually interchangeable
in the context of LVC s.12 The particular case of (12)b, where the alternation is
between egin ‘do’ and hartu ‘take’, can easily be accounted for if we consider
that arnasa, when combined with the former, means ‘breathing’, and when
combined with latter, means ‘exhaled air’.
In other cases, though, egin ‘do’ and eman ‘give’, when combined with a given
NVE, yield LVC s with very different meanings:
(14) a. damu egin ‘hurt’ ~ damu eman ‘cause regret’

b. errieta egin ‘argue / tell off’ ~ errieta eman ‘tell off’
c. hitz egin ‘talk’ ~ hitz eman ‘promise’
d. hots egin ‘call / sound’ ~ hots eman ‘guide / stimulate’
e. ikara egin ‘tremble’ ~ ikara eman ‘scare’
f. kontu egin ‘look after, take care / suppose’ ~ kontu eman ‘give an expla-
nation, give account’
12 See, for example, the almost systematic equivalence between Spanish dar ‘give’ and Cata-
lan fer ‘do’: dar un beso ~ fer un petó ‘kiss’, dar un abrazo ~ fer una abraçada ‘hug’, dar miedo
~ fer por ‘frighten’, dar pena ~ fer pena ‘provoke sorrow’, dar rabia ~ fer ràbia ‘anger’, dar
lástima ~ fer llástima ‘provoke pity’, dar envidia ~ fer enveja ‘provoke jealousy/envy’, and so
on. Moreover, both verbs turn out to be even more similar, as far as argument structure is
Actually, obtaining two LVC s with different meanings and different event-
and argument-structure properties is what normally occurs in Basque when
different LV s are combined with one same NVE (see also Martinez 2015: 114;
Zabala 2004: 457–458, 483). And this is what one expects given the above-
mentioned division of labour between the two components of the LVC. Thus, to
the examples in (14) for the pair egin ‘do’/eman ‘give’, we can add the following
ones with other LV s, including hartu ‘take’ and eduki ‘have’:
(15) a. arrazoi ‘reason’ eman ‘agree with’ ~ arrazoi eduki ‘be right’
b. atsegin eman ‘please’ ~ atsegin hartu ‘enjoy, take pleasure in’
c. damu hartu ‘regret’ ~ damu eman ‘cause regret’
d. esku ‘hand’ hartu ‘participate, take part’ ~ esku eman ‘help’
e. hats hartu ‘breathe’ ~ hats eman ‘blow’
f. kontu hartu ‘take care of, keep in mind’ ~ kontu eduki ‘be careful’
g. min hartu ‘get hurt’ ~ min eman ‘cause pain’
h. pena hartu ‘get sad’ ~ pena eman ‘sadden’
The examples in (15), where verbs other than egin ‘do’ are at play, clearly sup-
port Ramchand’s (2014) approach to LVC s, in the sense that the argument and
event structure of a given LV is identical to those of their non-LV counterpart.
For example, min hartu ‘get hurt’ behaves like liburua hartu ‘get, receive the
book’; and min eman ‘give, cause pain’ behaves like liburua eman ‘give the book’.
In other words, light eman ‘give’ and hartu ‘take’ display a meaning very close
to that of their non-light counterparts, and preserve the number of arguments
and their categorial requirements, generally selecting for NP s.13 By contrast, as
will be shown in section 2.2, egin ‘do’ allows a wide range of NVE s, including
instrumental and locative case-marked DP s, adverbials, and even words not
used outside the particular LVC (Zabala 2004: 481–482). Thus, it seems clear
that there is a difference between instances with egin ‘do’, clearly a light verb,
and instances with eman ‘give’ or hartu ‘take’, which, even when appearing in
this type of expressions, have been considered verbs with full semantic content
(Zabala 2014: 458, 461). Actually, Zabala’s difference between DO-verbs (light)
and GIVE/TAKE-verbs (fully contentful) could account for the many intra- and
concerned, once we adopt the view that the dative, commonly appearing with GIVE, is an
“added argument” in the sense of Pylkkänen (2002, 2008)—see McFadden (2006) for rel-
evant argumentation in the context of Germanic. Thus, under the applicative hypothesis,
it is not the case that GIVE has a slot for the dative object which DO lacks, but instead both
verbs can combine with a dative-introducing head, the applicative.
13 And occasionally DP s: bide(a) eman “give (the) way”, hats(a) eman “give (the) breath”,
gogo(a) hartu “take (the) mind”.
cross-linguistic equivalences found, as the ones in (12)–(13). An alternative way

to deal with this difference is in terms of the greater or lesser amount of struc-
ture lexicalized by the verb (Ramchand 2008): egin ‘do’ is more abstract and
thus probably lexicalizes only the eventive head v. This is in tune with the fact
that egin ‘do’ instantiates the least marked option to form LVC s, given its wide
range of meanings and its flexibility in taking different types of complements;
on the contrary, the more marked options are those that resort to the verbs
eman ‘give’ or hartu ‘take’, which also form LVC s but are indistinguishable from
their “heavy” counterparts, along the lines of Ramchand (2014).
Particular LV s thus impose their argument- and event-structure properties
to the resulting LVC. The predicative nature of the LV is also reflected in the fact
that it combines with a particular semantic type of complement: one denoting
an event. According to Rodríguez & García Murga (2003), when the NVE com-
bines with egin ‘do’, there is a semantic process a la Jackendoff (1990) that turns
a noun referring to a “thing”, e.g. ‘smile’ or ‘word’, into a noun referring to an
event: ‘smiling, talking’.14 Interestingly, as will be shown in section 2.2, LV s can
combine with different categories, including DP s, instrumental case-marked
DP s, inessive case-marked DP s or adverbials (16), to which, according to these
authors, the very same semantic process applies yielding the final meaning of
the LVC. In other words, the NVE is always interpreted as denoting an event, no
matter what syntactic category it bears.15 The fact that the very same event can
be expressed by different categories (17) seems to provide evidence for this:
(16) a. eztula egin “do the cough” ‘cough’

b. hega-z egin “do flight-instr” ‘fly’
c. dantza-n egin “do dance-ines” ‘dance’
d. lasterka egin “do running (adv)” ‘run’
(17) a. eztul / eztula egin “do cough/the cough” ‘cough’

b. hega-z / hega-n egin “do flight-instr/flight-ines” ‘fly’
c. dantza / dantza-n egin “do dance.abs/dance-ines” ‘dance’
d. laster / lasterka egin “do run/running (adv)” ‘run’
14 We use the term thing following Rodríguez & García Murga’s (2003: 433) terminology:
“Baina barre gauza bati egiten dio erreferentzia, ez egoera bati.” ‘But barre refers to a thing,
not to a situation.’ It seems to us that this concept must be understood in a broad sense,
beyond material objects.
15 The “semantic coercion” of the NVE as denoting an event is reminiscent of Marantz’s
(2005) account of creation predicates like Bake a cake, in which the DP a cake is also taken
to denote an event. An updated view of this type of syntactically driven semantic coercion
is that expounded by Wood & Marantz (2017) in terms of contextual allosemy.
Other data seem to contradict the above-mentioned division of labour be-

tween the LV (argument- and event-structure properties) and the NVE (concep-
tual content). Zabala (2004: 478–479), for instance, argues that the argument
structure of the LVC largely depends on the particular NVE that the LV com-
bines with. Thus, in line with Grimshaw & Mester’s (1988) claims for Japanese,
she assumes that the LV is thematically incomplete and that the NVE thus trans-
fers to it some or all its arguments, yielding the resulting LVC. An argument put
forward in favour of this “fusion” of the argument structure of the NVE and the
LV is that the NVE imposes restrictions on the type of external argument of the
LVC (Zabala 2004: 479–480). In particular, with meteorological LVC s the sub-
ject is a quasi-argument that cannot be phonologically realised, and that is only
reflected in the ergative agreement in the auxiliary:
(18) a. Euri-a egin du.

rain-det.abs do aux.prs.erg3sg.abs3sg
‘It has rained.’
b. *Horr-ek / *Zeru-a-k euri-a egin

this-erg / sky-det-erg rain-det.abs do
du.
aux.prs.erg3sg.abs3sg
‘This/The sky has rained.’
Still, it is worth noting that examples such as the next one, while certainly
marked, are possible:
(19) “Nor-k egi-ten du euri? […]” egi-ten

who-erg do-inf aux.prs.erg3sg.abs3sg rain do-inf
dio galde Strepsiades zaharr-a-k
aux.prs.erg3sg.abs3sg.dat3sg ask Strepsiades old-det-erg
Sokrates-i, eta filosofo-a-k honela erantzu-ten: “Ez
Socrates-dat and philosopher-det-erg thus answer-inf not
behintzat Zeus-ek, hodei-ek baizik”
at_least Zeus-erg cloud-det.erg.pl but
“Who rains? […]”, asked old Strepsiades for Socrates, and the philosopher
answered in this way: “Certainly not Zeus, but the clouds do.” (Dakiguna
ikasten, Patziku Perurena, Alberdania, 2007, p. 132)
What is worth noticing about examples like (19) is that, while the NVE cer-
tainly shapes the conceptual scene evoked (i.e., the appearance of rain is usually
not conceived as a caused event), it cannot grammatically preclude an option

licensed by the LV (i.e., egin allows the projection of an external argument).
Moreover, the other contrasts (20)–(22) upon which Zabala bases her defence
of such a “fusion” actually do not differ much from what is found in any transi-
tive construction, where the verb “expresses a range of predicates depending on
the choice of direct object” (Marantz 1984: 25). Thus, it is expected that the LV
egin ‘do’ requires different arguments depending on the NVE that it combines
with: in (20) egin combined with hitz ‘word’ admits a dative argument, in (21)
egin combined with lo ‘sleep’ does not admit a dative argument, and finally in
(22) egin combined with laztan ‘hug, kiss’ generally requires a dative argument.
(20) a. Aduna-k hitz egin du.

Aduna-erg word do aux.prs.erg3sg.abs3sg
‘Aduna has talked.’
b. Aduna-k gizon-a-ri hitz egin

Aduna-erg man-det-dat word do
dio.
aux.prs.erg3sg.abs3sg.dat3sg
‘Aduna has talked to the man.’
(21) a. Aduna-k lo egin du.

Aduna-erg sleep do aux.prs.erg3sg.abs3sg
‘Aduna has slept.’
b. *Aduna-k gizon-a-ri lo egin

Aduna-erg man-det-dat sleep do
dio.
aux.prs.erg3sg.abs3sg.dat3sg
‘Aduna has slept to the man.’
(22) a. *Aduna-k laztan egin du.

Aduna-erg hug do aux.prs.erg3sg.abs3sg
‘Aduna has hugged.’
b. Aduna-k laztan egin dio

Aduna-erg hug do aux.prs.erg3sg.abs3sg.dat3sg
gizon-a-ri.
man-det-dat
‘Aduna has given the man a hug.’
Zabala’s (2004) position for Basque is similar to that defended by other

authors dealing with LVC s in French and Spanish. For Spanish, Alonso-Ramos
(1998) claims that the diathesis of LV s is special, since their syntactic valency
changes according to the nominal NVE with which they combine—much as
has been shown for Basque in the above examples. She illustrates with the con-
trast between tener miedo ‘have fear’ ‘be afraid’ vs tener esperanza ‘have hope’;
while the former takes a dative encoding the stimulus, the latter can only take
a genitive PP:
(23) a. Rosa le tiene miedo a la oscuridad.

Rosa dat has fear at the darkness
‘Rosa is afraid of darkness.’
b. Rosa tiene esperanza de un cambio.

Rosa has hope of a change
‘Rosa hopes for a change.’
In a smilar vein, Rafel (2004: 404) also discusses how the valency of tener ‘have’
changes according to its use in a LVC, as has been shown above, or in a run-of-
the-mill transitive construction:
(24) *Juan le tiene un coche (a su madre).

Juan dat has a car at his mother
Rafel points out that the licensing of the dative depends on the use of the NVE
miedo ‘fear’, and that the dative is not licensed in a normal transitive predi-
cate such as the one in (24). An extreme case in the position that the argument
structure of the LVC basically depends on that of the NVE is that represented
by Gross (1981) or Danlos (1992: 2), the latter considering that the function of
the LV (“the support verb”) is “mainly to bear the tense and aspect values of the
sentence”.
There are several observations to be made in relation to the above facts,
which we hope to show to conform with, rather than jeopardize, Ramchand’s
(2014) conjecture on the division of labour between the LV (= argument and
event-structure) and the NVE (= conceptual content), as adhered to in this
paper. The first thing to notice is that, for instance, tener miedo may also surface
with a genitive PP encoding the stimulus, undermining the more deterministic
stance taken by Alonso-Ramos (1998):
(25) Rosa tiene miedo de la oscuridad.

What the alternation between this example and (23)a shows, we submit, is
that, while the use of the NVE miedo ‘fear’, as opposed to esperanza ‘hope’, does
indeed determine the presence of a stimulus in the conceptual scene evoked,
that “argument” may find different ways of realization in the actual sentence,
either as a genitive or as a dative. Perhaps more importantly, it can be dropped
altogether, the grammaticality of the sentence remaining unaltered:16
(26) Rosa tiene miedo.
Quite crucially, even in (26) the presence of the stimulus does not fail to be
inferred: when we are afraid we are certainly always afraid of something. Gram-
matically, however, there is no need for the stimulus to be expressed in the LVC,
and this, we note, must have to do exclusively with the syntactic properties of
the construction, as headed by the LV.17 It is interesting, in this respect, to con-
trast the LVC at hand with the corresponding synthetic verb temer ‘fear’:
(27) Rosa teme *(la oscuridad).
16 In some cases the complement of the NVE does seem obligatory:

(i) Pepa hizo alusión *(a los problemas financieros).
Pepa did hint at the problems financial
‘Pepa hinted at the financial problems.’
In these cases, however, we think that the NVE functions a complex event nominal in the
sense of Grimshaw (1990), and its complement (a los problemas financieros ‘at the finan-
cial problems’) is a true structural complement that cannot, thefore, be dropped.
17 This discussion is related to observations made by Gross (1976) and Giry-Schneider (1978)
for French and taken up, among others, by Alonso-Ramos (1998) and Rafel (2004) for
Spanish, namely, that the argument introduced by the NVE—cf. la oscuridad ‘darkness’
in (23)—can easily be reanalysed as a constituent of its own (Danlos 1992:10):
(i) a. Jean a commis [une aggression contre Marie].
Jean has committed an aggression against Marie
b. Jean a commis [une aggression] [contre Marie].
Constituency tests like pronominalization, cleft formation, or relativization indeed show
that, e.g., contra Marie ‘against Marie’ above may be analysed as an independent con-
stituent (see the mentioned works for details). However, this does not make contra Marie
‘against Marie’ a true argument, at the same level as une aggression ‘an aggression’, as pro-
posed by Alba-Salas (2002: 13) and Rafel (2004: 424) for Spanish, since while the former
can be dropped, the latter cannot. Moreover, “le faire”, i.e., do-so substitution, can strand
the PP contre Marie, further revealing its adjunct status:
(ii) Jean a commis une aggression et il l’ a fait contre Marie.
Jean has committed an aggression and he it=has done against Marie
All in all, while “secondary arguments” like contra Marie ‘against Marie’ are certainly par-
ticipants of the conceptual scene, they are however not licensed as true structural argu-
ments of the construction.
While tener miedo “have fear” ‘be afraid’ and temer ‘fear’ mean more or less
the same, only the former can drop the constituent encoding the stimulus,
suggesting that, indeed, argument structure qua argument realization is a pre-
rogative of the head of the construction, i.e., the verb, either heavy or light,
rather than of the NVE.
With respect to (22), we observe that the dative is not completely out in pred-
icates headed by non-light tener ‘have’:
(28) La madrina les tiene regalos a sus ahijados.

the godmother dat.pl has presents at her godchildren
‘The godmother has presents in store for her godchildren.’
This shows, again, that the structural licensing of the dative does not really
depend on the complement of tener ‘have’, but on some independent configu-
rational condition.18
It should have become clear, we hope, that we are making a crucial distinc-
tion between the arguments licensed structurally, by the configuration headed
by the LV, and the participants of the event, encoded as a conceptual property
of the NVE, and which are not obligatorily realized.19
2.2 The Non-verbal Element

Basque LVC s are usually formed by the combination of a LV and a bare nom-
inal element, as shown in all the examples above. Occasionally, the NVE can
also be a postpositional phrase (29) or an adverbial expressing an activity (30)
(Etxepare 2003: 402–403; Zabala 2004: 466–467; Martinez 2015: 114–126):
(29) a. Hega-n/Hega-z egin

wing-ines/wing-instr do
‘Fly’
b. Gain-ez egin
top-instr do
‘Overfly’
18 See Pineda (2016) for an implementation of the theory of applicatives (Pylkkänen 2002,
2008; Cuervo 2003) to the analysis of datives in Romance and Basque.
19 We follow in this respect the different between structure and content arguments in Grim-
shaw’s (2005 [1993]) terms and structure and content participants in Rappaport Hovav
& Levin’s (1998) terms. And see also, within Construction Grammar, Goldberg (1995).
Notwithstanding our criticism of Alonso-Ramos’s (1998) take on tener miedo/esperanza
“have fear/hope”, we acknowledge that the difference that we are pointing out roughly
corresponds to her distinction between actant syntaxique and actant sémantique.
c. Atze-ra egin
back-adl do
‘Go back’
d. Borroka-n egin
fight-ines do
‘Fight’
e. Aitzurr-ean egin
hoe-det-ines do
‘Dig’
f. Ardatz-ean egin
spindle-ines do
‘Spin’
g. Kart-etan egin
cards-ines do
‘Play cards’
h. Futbol-ean egin
football-ines do
‘Play football’
(30) a. Hagin-ka egin

tooth-adv do
‘Bite’
b. Adar-ka egin
horn-adv do
‘Gore’
c. Aitzur-ka egin
hoe-adv do
‘Dig’
Some LVC s alternate between a version with a bare nominal NVE and a ver-
sion with a locative or adverbial NVE, as the ones shown in (29) and (30)
respectively—note that -ka is an adverbializing suffix. Unlike the versions
involving a bare nominal NVE, those involving a non-nominal NVE show an iter-
ative meaning (Etxepare 2003: 402–403) or a nuance of a frequently repeated

activity (Martinez 2015: 120–126). In addition, the versions with bare nominal
NVE s are generally more frequent, except in the case of korrika egin “running
(adv.) do”, according to Martinez’s corpus study:
(31) a. Arrantza(-n) egin

fishing(-ines) do
‘Fish’
b. Arraun(-ean) egin
oar/rowing(-ines) do
‘Row’
c. Borroka(-n) egin
fight(-ines) do
‘Fight’
d. Dantza(-n) egin
dance(-ines) do
‘Dance’
e. Ehiza(-n) egin
hunting(-ines) do
‘Hunt’
f. Errieta(-n) egin
dispute(-ines) do
‘Dispute’
(32) a. Atzapar(-ka) egin

claw(-adv) do
‘Scratch’
b. Borbor(-ka) egin
Bubbling(-adv) do
‘Bubble’
c. Eztul(-ka) egin
cough(-adv) do
‘Cough’
d. Jauzi(ka)/Salto(ka) egin
jump.adv do
‘Jump’
e. Korri(-ka) egin
running(-adv) do
‘Run’
f. Laster(-ka) egin
run(-adv) do
‘Run’
g. Musu(-ka) egin
kiss(-adv) do
‘Kiss’
h. Oihu(-ka) egin
scream-adv do
‘Scream’
i. Zotin(-ka) egin
hiccup(-adv) do
‘Hiccup’
Additionally, the NVE can also be an adjective, as in the LVC on hartu ‘accept’
(lit. “good take”) and a particle, as in goiti egin ‘vomit’ (lit. “do up”), although
these two options are “relatively rare and present a high degree of lexicali-
sation” (Oyharçabal 2006: 787).20 It is also worth mentioning that the NVE of
meteorological LVC s bero egin “hot do” ‘be hot’ and hotz egin “cold do” ‘be cold’
can function as a noun or as an adjective (Martinez 2015: 54–56).
Focusing now on nominal NVE s, they are inanimate nouns, and they are
non-referential (Martinez 2015: 44; Zabala 2004: 470).21 Other than that, even
though the range of nominals that can appear in Basque LVC is a closed set,
there is no semantic property that can apply to all of them: nominals can be
abstract (gogoeta egin “do meditation” ‘meditate’) or concrete (buru egin “do
20 Note that on hartu, actually written onartu ‘accept, admit’, is fully lexicalized and behaves
as a transitive verb: legea onartu ‘accept the law’.
21 A few NVE s, such as turrut (egin) ‘mock / fail’ or laprast (egin) ‘slip’, are not found in syn-
tactic environments other than LVC s (Zabala 2004:449).
head” ‘face’), countable (hitz egin “do word” ‘talk’) or uncountable (min egin
‘do harm’ ‘hurt’), and eventive (arrantza egin ‘do fishing’) or resultative (laprast
egin “do slip” ‘slip’) or with no argument or event structure at all (elurra egin
“do snow” ‘snow’) (Rodríguez & García Murga 2003: 420). Thus Basque con-
trasts with Japanese, where, according to Grimshaw & Mester (1988), NVE s are
always event nouns that transfer their argument structure to the LVC, as men-
tioned in section 1.2 (see Fernández 1997 for a comparison between Japanese
and Basque LVC s). However, Grimshaw & Mester’s proposal can be maintained
for Basque under a slightly different view: as argued above, regardless of the
type of noun and the readings it may have in other syntactic environments, the
NVE, in the context of the LVC, always expresses an event. Recall that consid-
ering all Basque NVE s as denoting events goes in line with Rodríguez & García
Murga’s (2003: 432–434) semantic analysis, according to which the NVE may
refer to a thing, but the final interpretation is achieved at the conceptual level
by means of a semantic process (Jackendoff 1990) ensuring that the NVE no
longer refers to a thing, but to an event. In addition, given that events inherently
lack referenciality, the above-mentioned non-referenciality of Basque NVE s is
accounted for (Zabala 2004: 470).
The lack of referentiality is also a property of a small group of Basque NVE s
that are not bare nominals but bear a determiner (Martinez 2015: 33–34, 184–
185). These LVC s mostly refer to meteorological phenomena and their NVE
bears the definite article -a (the only determiner possible in these structures):22
22 Thus, other determiners such as bat ‘a, one’, the demonstratives hau/hori ‘this, that’ (which
induce a necessarily specific interpretation) or the plural definite article -ak are not
allowed in NVE s. Examples can be found where instead of -a the quantifier asko ‘very
much, a lot’ is used, as in (i), but according to Martinez (2015:58, fn. 64) asko refers in this
case to the whole meteorological situation, rather than quantifying the noun euri:
(i) Euri asko egin du.
rain a_lot do aux.prs.erg3sg.abs3sg
‘It has rained a lot.’
This contrasts with its use with common direct objects, such as in (ii), where asko gen-
erally triggers the plural absolutive agreement -it- in the verb (actually, there is dialectal
variation in this point, with some dialects accepting (ii) with the singular auxiliary du):
(ii) Liburu asko irakurri ditu.
book a_lot read aux.prs.erg3sg.abs3pl
‘(S)he has read a lot of books.’
However, note that agreement does not constitute a reliable test here: even if asko quanti-
fied euri, we would not see the morphological effect on the verb, since euri is uncountable
and thus never triggers plural agreement on the verb. Actually, in our view, the null hypoth-
esis involves assuming that asko is quantifying directly over euri in Euri asko egin du. This
scenario is reminiscent of what Bosque & Masullo (1998) argue is involved in Spanish syn-
thetic predicates like sangrar mucho ‘to bleed a lot’. These authors propose that the degree
eguzkia egin “do sun”, elurra egin “do snow”, euria egin “do rain”, haizea egin
“do wind”, izotza egin “do ice”, trumoia egin “do thunder”, and so on (Martinez
2015: 56–60).23 In a few other LVC s, again mostly meteorological, the deter-
miner is optional: hotz(a) egin “do cold” ‘be cold’ and bero(a) egin “do hot” ‘be
hot’, eztul(a) egin “do cough” ‘cough’ and arnas(a) egin “do breath” ‘breath’. In
any case, regardless of whether the NVE is determined or not, it is not refer-
ential, but has a generic interpretation (Zabala 2004: 470; Martinez 2015: 56–
60).
Since the NVE in meteorological LVC s is usually determined but still non-
referential, non-referentiality cannot be directly related to the absence of the
determiner. Another reason that prevents us from establishing this connec-
tion is that in common transitive structures a-marked DP s can be found that
are not referential. If we compare the LVC lan egin “do work” and the com-
mon transitive structure lana egin (lana bearing a determiner -a), the latter
has to be interpreted with a definite reference (“the job, that job”), but if we
look at other common transitive structures such as etxea egin ‘build a/the
house’ or ogia jan ‘eat (the) bread’, whose direct objects cannot appear in an
LVC, then the determined direct objects can be either definite (‘the house,
that house’, ‘the bread, that bread’) or indefinite (‘a house’, ‘bread’) (Martinez
2015: 63).24 In short, Basque arguments always need to be determined, even if
they are generic. Thus, as Himmelmann (2001: 832) points out: “Definiteness,
though undeniably of central importance to the grammar and typology of the
articles, is only one of a number of meanings that can be conveyed by arti-
cles”.
Under a wider conception of LVC s, one can find in Basque some construc-
tions where the LV egin combines with a determined DP, such as bizarra egin
“do the beard” ‘shave’, harrikoa egin “do the washing up”, erroak egin “do the
roots” ‘settle down’, bakeak egin “do the peaces” ‘make peace’. We think that
these few cases are the ones that can be considered constructions involving
Vague Action Verbs in Basque, where the NVE seems to pattern with full-fledged
argumental DP s—see Kearns’s (1988) distinction in section 1.2. In turn, these
modifier mucho ‘a lot’ quantifies over a nominal predicate SANGRE ‘blood’, embedded in
the lexico-syntactic representation of the verbal phrase, involving the abstract LVC-like
structure DO BLOOD (Hale & Keyser 1993, 2002). The difference between Spanish and
Basque would lie in the synthetic vs analytic character of the predicate. On the quantifi-
cation of the NVE in Basque LVC s, see examples (58) below and Oyharçabal (2006: 792).
23 It is not surprising that meteorological LVC s are considered a particular class of Basque
LVC s by several authors (Zabala 2004; De Rijk 2008; Martinez 2015).
24 An anonymous reviewer points out that the contrast may need to be nuanced, as some-
thing like lan-a egin “work-a do” can be non-referential too in central dialects of Basque.
Vague Action Verbs constructions should be differentiated from others featur-

ing egin as a non-light verb: ogia egin ‘make the bread’, lana egin ‘do the work’,
ohea egin ‘make the bed’ (Rodríguez & García Murga 2003: 419, 424; Zabala
2004: 476, Martinez 2015: 115–126).
These particular instances aside, and as mentioned in section 1.2, Basque
LVC s are generally considered constructions headed by True Light Verbs. Thus,
even if, as will be shown in section 3, in some LVC s the NVE may pattern
with full-fledged argumental DP s in some respects (word order, partitive case,
causative formation), NVE s and argumental DP s are fundamentally different.
Most importantly, the former are overwhelmingly bare nominals, whereas the
latter need to be determined to be licensed.
Turning to Romance languages, we observe that some LVC s have a synthetic
correspondent, whereas others do not. The lists below also show that, unlike
Spanish (33), in French LVC s (34) the NVE is usually introduced by a determiner
or a partitive (Martinez 2015: 77–78):
(33) a. hacer fiesta “do holiday” ~ festejar ‘celebrate’, hacer mención “do men-
tion” ~ mencionar ‘mention’, hacer juramento “do oath” ~ jurar ‘swear’,
hacer elogio “do praise” ~ elogiar ‘praise’, hacer chantaje “do backmail” ~
chantajear ‘blackmail’, hacer daño “do damage” ‘hurt’ ~ dañar ‘damage’,
hacer ayuno “do fast” ~ ayunar ‘fast’
b. hacer deporte ‘do sport’, hacer mérito “do merit” ‘mention’, hacer juerga
“do fun” ‘have fun’, hacer luto “do mourning” ‘mourn’, hacer turismo ‘do
tourism’, hacer dinero ‘make money’
(34) a. faire mention “do mention” ~ mentionner ‘mention’, faire (de) la cuisine
“do (of) the cooking” ~ cuisiner ‘cook’, faire la recolte “do the harvest”
~ recolter ‘harvest’, faire du progrès “do of the progress” ~ progrésser
‘progress, advance’, faire le nettoyage “do the cleaning” ~ nettoyer ‘clean’,
faire des études “do of the studies” ~ étudier ‘study’, faire de l’ ombre “do
of the shadow” ~ ombrager ‘cast a shadow on’
b. faire du sport “do of the sport” ‘do sport’, faire du tourisme “do of the
tourism” ‘do tourism’, faire la ronde “do the round” ‘go the rounds’, faire
la revue “do the review” ‘review’, faire l’amour “do the love” ‘make love’,
faire la guerre “do the war” ‘wage war’, faire fortune “do the fortune”
‘make one’s fortune’.
The correspondence between LVC s and synthetic predicates in these languages

is far from straightforward. Following Piera & Varela’s (1999) discussion on
Spanish, we can distinguish the next four possible situations:
1) Morphological and semantic affinity: hacer transbordo “do transfer” ~

transbordar ‘transfer’; French: faire mention “do mention” ~ mentionner
‘mention’; some doublets exist whose members differ in register value, as
in French faire la guèrre “do the war” ‘wage war’ (less formal) ~ guerroyer
‘wage war’ (more formal)
2) Morphological affinity without (complete) semantic equivalence: hacer
fiesta “do holiday” ~ festejar ‘celebrate’, hacer daño “do damage” ‘hurt’
~ dañar ‘damage’; French: faire le récit “do the account” ‘narrate, tell’ ~
réciter ‘recite’ (Giry-Schneider 1986)
3) Semantic equivalence without morphological affinity: hacer punto “do
stitch” ‘knit’ ~ tricotar ‘knit’, dar asco “give disgust” ‘disgust’ ~ repugnar
‘disgust’; in some cases there is a semantic equivalence involving how-
ever a difference in register, as in French faire l’ amour ‘make love’ ~ baiser,
niquer ‘fuck’.
4) LVC s that do not correspond to any synthetic verb, cf. (33)b and (34)b)
Similarly to what has been pointed out for Basque above, we observe that many
LVC s in French present an NVE that is truly bare, and may even not be a nominal
at all:
(35) avoir faim/ froid/envie/besoin “have hunger/fancy/need” ‘be hungry/cold,

fancy, need’
(36) faire peur/honte/envie “do fear/shame/envy” ‘provoke fear/shame/pro-

voke envy’
(37) prendre plaisir “take pleasure” ‘enjoy’
(38) prêter attention “lend attention” ‘pay attention’
As pointed out by Acedo-Matellán (2014) for the avoir cases, these NVE s are
truly bare, since they are caseless (cf. failure of en-substitution) and cannot sus-
tain nominal quantification (cf. availability of très ‘very’, but not of beaucoup de
‘a lot of’):
(39) —Tu as faim / froid? —* Oui, j’en ai.

—you has hunger / cold? — yes, I=ptve have.1sg
(40) J’ai {très / *beaucoup de} faim / froid.

I=have very / a lot of hunger / cold
(41) —Les serpents font peur à Magali? —* Non, ils ne lui en

—the snakes do fear at Magali? — no they neg dat ptve
font pas.
do.3pl neg
(42) Les serpents font {très / *beaucoup de} peur à Magali.

the snakes do very / a lot of fear at Magali
(43) —Thomas prend plaisir à marcher? —* Oui, il en y prend.

—Thomas takes pleasure at walk.inf — yes he ptve loc takes
In Spanish the corresponding LVC s do admit nominal quantification, as in

Tengo mucha hambre “I have much.f hunger” ‘I am very hungry’. Even more
conspicuously, and in spite of its proximity to French, the Catalan counterparts
behave differently, in allowing both partitive case (as shown by ne-cliticization)
and nominal quantification (cf. the use of molta ‘much.f’):
(44) —Que tens gana / fred? —Sí, en tinc.

—compl have.2sg hunger / cold? —yes, ptve have.1sg
(45) Les serps li’n fan molta, de por, a l’Antoni.

the snaked dat=ptve do.pl much.f.sg of fear at the=Antoni
‘Antoni is very much afraid of snakes.’
As also observed by Acedo-Matellán (2014), these bare nouns must occupy an

argumental (object) position in the structure, a position that can, therefore, not
be occupied by any other DP:
(46) Thomas avait envie *(d’) une glace.

Thomas had fancy of an ice-cream
‘Thomas fancied an ice-cream.’
(47) Anne-Catherine a faim *(de) fromage.

Anne-Catherine has hunger of cheese
‘Anne-Catherine craves some cheese.’
From Acedo-Matellán’s (2014) perspective, these facts would be a problem for

an analysis of such predicates in French as involving incorporation, i.e., fusion
of the NVE and the LV at some level of representation (but see also section
3). We turn to incorporation analyses and other approaches to the relation
between the LV and the NVE in the next section.
3 The Syntactic Relationship between the LV and the NVE
In the generative tradition, Basque LVC s (48) have often been identified with
syntactically unergative verbs, which show a synthetic correlate in other lan-
guages,25 such as Romance (49)–(50):
(48) hitz egin ‘talk’, barre egin ‘laugh’, negar egin ‘cry’, lo egin ‘sleep’, amets egin
‘dream’, salto egin ‘jump’, dantza egin ‘dance’, lan egin ‘work’
(49) Sp. hablar ‘talk’, reír ‘laugh’, llorar ‘cry’, dormir ‘sleep’, soñar ‘dream’, saltar
‘jump’, bailar ‘dance’, trabajar ‘work’
(50) Fr. parler ‘talk’, rire ‘laugh’, pleurer ‘cry’, dormir ‘sleep’, rêver ‘dream’, sauter
‘jump’, danser ‘dance’, travailler ‘work’
Basing on facts like the analytic realization of unergatives in Basque (and other
languages), Hale & Keyser (1993) postulated a universal transitive nature for
unergative verbs: that is, unergative (i.e. activity) predicates take an internal
argument that is not the subject. Thus, the only difference between the Basque
analytic patterns in (48) and the Romance synthetic patterns in (49)–(50) is
whether overt incorporation of the NVE into the LV takes place or not. Actu-
ally, even in Basque many complex unergatives have a simplex counterpart, as
shown by dantza egin “do dance” and dantzatu, both meaning ‘dance’, and here
too the difference between the two options has to do with whether nominal
incorporation has taken place or not. If it does not take place (dantza egin ‘do
dance’, ‘dance’), the derivation will parallel that of a transitive clause (zerbait
egin ‘do something’).
However, there is another dimension where syntactic incorporation becomes
a controversial issue in the theoretical treatment of Basque and, to some extent,
25 See Martinez (2015: 21–32) for a very detailed semantic classification of Basque LVC s
that takes into account several previous semantic classifications, especially De Rijk’s
(2008), but also Rodríguez & García Murga’s (2003: 423), Etxepare’s (2003: 400–402) and
Zabala’s (2004: 472–476) among others. As a matter of fact, Etxepare (2003) provides a
rich sample of LVC s, distinguishing verbs of emission (sound emission oihu egin ‘yell’,
light emission dir-dir egin ‘shine’, verbal emission errieta egin ‘argue / tell off’), inter-
nal body motion (dar-dar egin ‘tremble’), physical activities (actions against an object
or an individual tiro egin ‘shoot’, motion verbs igeri egin ‘swim’, bodily functions izerdi
egin ‘sweat’), mental activities (duda egin ‘doubt’) and behavioral verbs (axut egin ‘chal-
lenge’).
Romance LVC s.26 Indeed, beyond the obvious analytic/synthetic morphologi-

cal dichotomy pointed out above, there is a debate as to whether the NVE in
Basque LVC s (dantza egin, hitz egin) patterns with DO s in common transitive
structures or it undergoes syntactic incorporation. The debate on this issue has
been intense and has led to a variety of analyses, some emphasizing the par-
allelism of LVC s with regular transitive structures, and others focusing on the
particularities of the NVE in LVC s and arguing instead that Basque LVC s are a
subgroup of unergative verbs with noun incorporation (see Lafitte 1944; Sara-
sola 1977; Levin 1983; Ortiz de Urbina 1986; Uribe Etxebarria 1989; Hale & Keyser
1993, 1997; Bobaljik 1993; Laka 1993; Austin & López 1995; Fernández 1997; Etxe-
pare 2003; Zabala 2004; Oyharçabal 2006; Martinez 2015, a.o.). The lack of a uni-
fied view with respect to this issue surely has to do with the non-homogenous
behaviour of NVE s in Basque LVC s. Indeed, they pattern with absolutive objects
in some cases, but differ from them in others. This is thus directly related to
their case properties: some tests indicate that NVE s are case-marked (with
absolutive case), but other tests point in the opposite direction, suggesting that
the noun holds a connection with the LV that goes beyond that of a run-of-the-
mill object, along the lines of noun incorporation. In what follows, some pieces
of apparently contradictory evidence as regards this issue are presented.27
26 The term incorporation has been used to refer to certain phenomena found in a variety
of languages in which verbs overtly host material arguably belonging to their comple-
ments, like prepositions (as in Chichewa, Baker 1988) or nouns (as in Siberian Koryak,
Mithun 1984). Within generative approaches, and since the extremely influential work of
Baker (1988), incorporation also refers to the syntactic mechanism proposed to account
for these phenomena, to wit, the movement of the head of the verb complement to the
immediately c-commanding verbal head (see Haugen 2015 for a recent overview and a
new proposal on the nature of incorporation). Since incorporation, in the standard gen-
erative approaches, involves head movement, it is worth pointing out that this operation
has come under considerable scrutiny at least since Chomsky (1995), who first pointed
out the theoretical inconsistencies associated to this operation and proposed to relegate
it to the purely morphophonological component of grammar. The debate is far from being
settled, with proposals that actual movement (Matushansky 2006) or lack thereof (Brody
2000; Bye & Svenonius 2012) is involved in the phenomena traditionally dealt with under
this rubric. For a recent overview, see Barrie & Mathieu (2014), and for interesting theo-
retical remarks, Haugen (2015).
27 Due to space restrictions, we must refrain from dealing with some other (dis-)similari-
ties between LVC s and common transitive structures, such as the position of NVE s with
respect to manner adverbials such as ondo, ongi ‘well’, or gaizki ‘badly’ (Oyharçabal 2006:
794; Martinez 2015: 51, fn. 53), the (non-)acceptability of adjectives and other type of mod-
ifiers (García Murga 2003: 421; Zabala 2004: 452, Oyharçabal 2006: 796; Martinez 2015:
47–48) or the (im)possibility of ellipsis in different contexts (Oyharçabal 2006: 793; Mar-
tinez 2015: 48–49), among others.
One first argument for the transitive analysis is the incompatibility of LVC s
with a DO: if the NVE (dantza) bears absolutive case, a “second” direct object
(tango) is expected not to be case-licensed (51)a, in contrast with the synthetic
version of the predicate (51)b:28,29
(51) a. *Aduna-k tango-a dantza egin du.

Aduna-erg tango-det.abs dance do aux.prs.erg3sg.abs3sg
b. Adunak tango-a dantzatu du.

Aduna-erg tango-det.abs dance aux.prs.erg3sg.abs3sg
‘Aduna has danced the tango.’
This is also the argument that Acedo-Matellán (2014) wields to defend that the
case-less NVE in the French LVC s dealt with at the end of the previous section
(e.g., avoir faim “have hunger” ‘be hungry’) occupies the object position, thus
constituing an “unincorporated root”.
That said, the incorporation analysis, whereby the NVE is caseless and under-
goes incorporation, can also account for the contrast in (51). Indeed, invoking
Chomsky’s (1981) Empty Category Principle, a “second DO” would fail to be
licensed, since the trace left behind by the incorporated DO cannot be occu-
pied by another argument. Following Martinez (2015: 177), the unfeasibility of
licensing “second DO s” actually shows that, even for those cases in which there
is actually phonological evidence of incorporation (such as hitz egin “do word”
28 Berro (2015) argues that the NVE occupies the same position as DO s in simplex verbs,
basing on the aspectual properties of the predicate. Thus, just like simplex verbs can be
aspectually bounded if the DO denotes a bounded scale, the event denoted in complex
verbs is also measured by the scalar (aspectual) properties of its NVE (i.e., the root), thus
indicating that they occupy the same position—see Harley (2005) for a seminal work on
the parallelism between internal argument DP s and roots as regards their contribution to
Aktionsart.
29 A “second DO” is possible with LVC s expressing mental or communicative activities, when
it is a subordinate clause (Martinez 2015: 74, 376–377; Zabala 2004:480):
(i) Gau-ez, [bere armairuan borrokan zihardutela] amets egi-ten
dream do-inf
zuen.
aux.pst.erg3sg.abs3sg.
‘At night, (s)he used to dream [that they were fighting inside his/her closed].’
However, these apparent double object constructions (which are also found with ditran-
sitive verbs, such as abisatu ‘tell [do-abs me] [do subordinate clause]’) can easily be
accounted for under the assumption that completive subordinate clauses don’t need case,
or are self-sufficient as far as case is concerned (Pesetsky & Torrego 2001).
‘talk’, see below), it is a syntactic process, not a lexical one, unlike that argued
to be involved in English laugh, formed in the lexical syntax out of do+laugh
(Hale & Keyser 1993, 2002; Haugen 2015). In this latter kind of incorporation no
trace is left behind allowing for a DO (a cognate object) to be projected—cf.,
e.g., laugh a good laugh.30
Both the proponents of the transitive analysis and those of the incorporation
analysis have put forward many other arguments, based on a variety of syntac-
tic tests (see below). However, since not all Basque LVC s show a homogenous
behavior, most syntactic tests have actually been used to justify both views. In
particular, the similarities and contrasts between NVE s and regular DO s arise
when the focus is put on the degree of cohesion between the LV and the NVE.
As we shall see, the apparently contradictory evidence in fact points out that
while some Basque LVC s may be the result of an incorporation process, others
cannot be. To makes matters worse, the non-homogenous syntactic behaviour
of Basque LVC s does not only depend on the particular LVC, since differences
across dialects and registers are also at play (see Oyharçabal 2006), thus mak-
30 Two LVC s in Basque are exceptional in that they allow a DO, namely huts egin “do fail-
ure/empty” ‘miss’ and hots egin “do noise/rumor” ‘call, propagate’ (Oyharçabal 2006: 791,
fn. 6; Martinez 2015: 172, 377–378):
(i) Ume-a-k eskol-ak huts egin ditu.
kid-det.abs-erg class-abs.pl failure/empty do aux.prs.erg3sg.abs3pl
‘The kid missed the classes’
(ii) Gizon-a-k berrion-a hots egin
man-det.abs-erg good_news-det.abs noise/rumor do
du
aux.prs.erg3sg.abs3sg
‘The man spread the good news’
In these cases a lexicalization process has taken place, so that the argument structure of
the LVC includes a slot for a DO with the Theme-role. That a reanalysis has taken place
in these cases is also evinced by phonology, huts egin and hots egin being generally pro-
nounced as single words with a single accent: hutségin, hotségin. In other words, there is no
longer any trace preventing the merger of a DO. Rather, there has been a reanalysis yielding
a transitive verb able to take a full-fledged DP as complement that can check the structural
absolutive case (Martinez 2015: 172, 377–378). The cases in (i) and (ii) could thus be seen
as instances of Mithun’s (1984:857) case-manipulating type of incorporation, whereby the
case left by the incorporated argument is available for another argument (until then an
oblique), which is thus promoted.
We think that huts egin—but not hots egin—is amenable to a different analysis,
namely, that huts, which can also be an adjective meaning ‘empty’, is a secondary pred-
icate rather than an argument. Specifically, huts would be a predicate of the DO, which
would then be its external argument: trena huts egin “make the train empty, missed”, i.e.,
‘miss the train’. If this analysis is on the right track, huts egin would pattern with the above-
mentioned onartu, whose NVE is also an adjective, on ‘good’.
ing it difficult to find systematic patterns that allow to tip the scales towards an
analysis or the other. It seems therefore that the most adequate way to proceed
is by assuming the existence of syntactically different types of LVC s in Basque.
One first type of Basque LVC involves constructions that behave as com-
mon transitive structures, formed by a transitive verb and its DO. Evidence of
their run-of-the-mill transitive behaviour includes the separability of the LV
and the NVE in interrogative sentences and focus structures (see (53)–(56)),
and availability of partitive case marking on the NVE in negative sentences (57)
(Etxepare 2003: 403):
(52) Aduna-k dantza egin du.

Aduna-erg dance do aux.prs.erg3sg.abs3sg
‘Aduna has danced.’
(53) Nor-k egin du dantza?

who-erg do aux.prs.erg3sg.abs3sg dance
‘Who has danced?’
(54) ADUNA-K egin du dantza.

Aduna-erg do aux.prs.erg3sg.abs3sg dance
‘ADUNA has danced.’
(55) DANTZA egin du Aduna-k.

Dance do aux.prs.erg3sg.abs3sg Aduna-erg
‘Aduna DANCED.’
(56) Dantza, Adunak EGITEN du.

‘As for dancing, Aduna does dance’
(57) Aduna-k ez du dantz-ik egin.

Aduna-erg not aux.prs.erg3sg.abs3sg dance-ptve do
‘Aduna has not danced.’
These LVC s pattern with regular transitive constructions also in that their
NVE, qua incremental theme, can be quantified (58)a just like any incremen-
tal theme of a non-light verb (58)b (Oyharçabal 2006: 792):
(58) a. Aduna-k lo gutxi egin du.

Aduna-erg sleep little do aux.prs.erg3sg.abs3sg
‘Aduna has slept a little.’
b. Aduna-k ogi gutxi jan du.

Aduna-erg bread little eat aux.prs.erg3sg.abs3sg
‘Aduna has eaten some bread.’
An additional piece of evidence for the transitive analysis comes from causative
contexts (Ortiz de Urbina 1989: 47). The NVE behaves as a canonical argument
in the formation of causatives, since the subject of the LVC (59) takes the dative
case in the causative (60), just like the causee of a causativized transitive pred-
icate (61):
(59) Aduna-k dantza egin du.

Aduna-erg dance do aux.prs.erg3sg.abs3sg
‘Aduna has danced’
(60) Aduna-ri dantza egin-arazi diote.

Aduna-dat dance do-cause aux.prs.erg3pl.abs3sg.dat3sg
‘They have made Aduna dance’
(61) Aduna-ri afari-a antol-arazi

Aduna-dat dinner-det.abs prepare-cause
diote.
aux.prs.erg3pl.abs3sg.dat3sg
‘They have made Aduna prepare the dinner’
However, not all Basque LVC s display the same behaviour in the battery of syn-
tactic contexts just presented. For example, in contrast with (57), not all NVE s
can be partitive-marked, as pointed out by Etxepare (2003: 404) and Zabala
(2004: 451–452) and thoroughly assessed by Martinez (2015: 247–277):
(62) a. *Aduna-k ez du laprast-ik egin

Aduna-erg not aux.prs.erg3sg.abs3sg slip-ptve do
b. *Aduna-k ez du alde-rik egin

Aduna-erg not aux.prs.erg3sg.abs3sg side-ptve do
(63) a. Aduna-k ez du laprast egin.

Aduna-erg not aux.prs.erg3sg.abs3sg slip do
‘Aduna has not slipped.’
b. Aduna-k ez du alde egin.

Aduna-erg not aux.prs.erg3sg.abs3sg side do
‘Aduna has not left.’
Moreover, even those LVC s showing partitive case-marking do not exactly repli-
cate the behaviour of a common transitive structure; indeed, for LVC s such case
marking is not compulsory (64), whereas for regular transitive structures it is
(65):
(64) a. Ez dut lo egin.

not aux.prs.erg1sg.abs3sg sleep do
‘I haven’t sleep.’
b. Ez dut lo-rik egin.

not aux.prs.erg1sg.abs3sg sleep-ptve do
‘I haven’t sleep.’
(65) a. *Ez dut liburu irakurri.

not aux.prs.erg1sg.abs3sg book read
‘I haven’t read a/any book.’
b. Ez dut liburu-rik irakurri.

not aux.prs.erg1sg.abs3sg book-ptve read
‘I haven’t read a/any book.’
Thus, the existence of the case-less alternates above, whether general (63)
or circumscribed to a particular dialect or register (Oyharçabal 2006) (64)a,
favours an incorporation analysis.31
Such an incorporation analysis would also allow us to account for the dif-
ferences regarding the adjacency between the NVE and the LV. As a matter of
fact, in contrast to what was shown in (53)–(56) for dantza egin, some NVE s
cannot be separated from the LV at all, unlike common DO s (Zabala 2004: 452–
453). Thus, the following examples, involving interrogative sentences (66) and
focus structures (67)–(69), should be contrasted with (54)–(56). Also, another
31 There is no exhaustive dialectal survey on this topic. Oyharçabal (2006) observes some
tendencies within dialects (and registers) but, in spite of some remarks on how some other
dialects behave, he mainly focuses on Basque eastern varieties, i.e., those spoken in the
French Basque Country (Iparralde).
sign of incorporation in the case of hitz egin (actually often written hitzegin)
comes from its phonological pattern, bearing one single accent: hitzégin, to be
contrasted, for instance, with that of negar egin, negár egín.
(66) *Nor-k egin du hitz?

who-erg do aux.prs.erg3sg.abs3sg word
Intend. ‘Who has talked?’
(67) *ADUNA-K egin du hitz.

Aduna-erg do aux.prs.erg3sg.abs3sg word
Intend. ‘ADUNA has talked.’
(68) *HITZ Aduna-k egin du.

word Aduna-erg do aux.prs.erg3sg.abs3sg
Intend. ‘Talk, has Aduna done.’
(69) *Hitz, Aduna-k EGI-TEN du.

word, Aduna-erg do-prs aux.prs.erg3sg.abs3sg
Intend. ‘As for talking, Aduna does talk.’
Likewise, as regards quantification, LVC s and transitive structures do not neces-

sarily behave alike, in spite of what we have shown in (58). Indeed, in LVC s the
quantifier may occupy a different position, in such a way that the noun+quan-
tifier sequence we saw in (58)a, repeated here as (70)a, alternates with (70)b,
where the quantifier gutxi targets the whole VP instead of just the NVE, the lat-
ter option arguably signalling incorporation. Here too variation is dialectal or
idiolectal, and it even may depend on the particular LVC. For example, with
hitz egin (71) the quantifier+LVC sequence seems to be preferred all over the
linguistic domain (Oyharçabal 2006: 802):
(70) a. Aduna-k lo gutxi egin du.

Aduna-erg sleep little do aux.prs.erg3sg.abs3sg
b. Aduna-k gutxi lo egin du.

Aduna-erg little sleep do aux.prs.erg3sg.abs3sg
‘Aduna has not slept much.’
(71) a. ??Aduna-k hitz gutxi egin du.

Aduna-erg word little do aux.prs.erg3sg.abs3sg
b. Aduna-k gutxi hitz egin du.

Aduna-erg little word do aux.prs.erg3sg.abs3sg
‘Aduna has not talked much.’
It is interesting at this point to go back to the French LVC s involving a case-less

NVE, like avoir faim ‘be hungry’ or faire peur ‘scare’, since they behave differ-
ently from Basque hitz egin ‘talk’ and pattern with lo egin in allowing for degree
quantifiers—albeit non-nominal!—to intervene between the LV and the NVE:
(72) Marjorie a très faim.

Marjorie has very hunger
‘Marjorie is very hungry.’
(73) Les serpents font très peur à Magali.

the snakes do very fear at Magali
‘Magali is very afraid of snakes.’
The position of manner adverbials such as ondo, ongi ‘well’ or gaizki ‘badly’ also
reveal differences between LVC s and ordinary transitive structures. Whereas
with the latter the unmarked order is S O Adv V (74)a, with the former two
options are possible: in (75)a, available at least in Western dialects, the LVC is
treated like a transitive structure, whereas in the more general (75)b (cf. Oyhar-
çabal 2006: 794), the LVC is treated like an intransitive structure—thus, by
hypothesis, signalling incorporation. Here again, we find variation depending
on the particular LVC. For example, hitz egin is not grammatical if treated as a
transitive structure (76)a, only the option involving incorporation is good (76)b:
(74) a. Aduna-k liburu-a ongi irakurri du.

Aduna-erg book-det.abs well read aux.prs.erg3sg.abs3sg
‘Aduna read the book well.’
b. *Aduna-k ongi liburu-a irakurri du.

Aduna-erg well book-det.abs read aux.prs.erg3sg.abs3sg
(75) a. Aduna-k lan ongi egin du.

Aduna-erg work well do aux.prs.erg3sg.abs3sg
b. Aduna-k ongi lan egin du.

Aduna-erg well work do aux.prs.erg3sg.abs3sg
‘Aduna has worked well’
(76) a. *Aduna-k hitz ongi egin du.

Aduna-erg word well do aux.prs.erg3sg.abs3sg
b. Aduna-k ongi hitz egin du.

Aduna-erg well word do aux.prs.erg3sg.abs3sg
‘Aduna has talked well.’
So far we have seen that variation depends on the particular LVC (with cases
such as hitz egin favoring the incorporation analysis in general) but also on
geographic variation. In particular, Oyharçabal (2006) finds quite a general ten-
dency in Eastern dialects to avoid the treatment of the NVE as an object, thus
preferring incorporation—which otherwise is also possible in other dialects or
even idiolects. This observation nicely fits with Berro’s (2010:4) claim that “com-
plex unergatives are much more common in southern varieties [that is, those
in contact with Spanish: our clarification, VA&AP], whereas simplex are pre-
ferred in northern ones [that is, those in contact with French: our clarification,
VA&AP].”
The tendency shown by Eastern dialects can also be witnessed in the treat-
ment of LVC s in causative contexts: in contrast to what we saw in (60), Eastern
dialects do not always treat the causee of a LVC as the causee of a transitive
verb (with dative) (77), since it may also appear with absolutive case (78), as
the causee of an intransitive verb (thus indicating that incorporation has taken
place) (Epelde & Oyharçabal 2012):32
32 Epelde & Oyharçabal (2012) propose a contact-based explanation: in French, with which
Eastern dialects are in contact, the causee of unergative predicates is accusative (i); in
turn, Western dialects are in contact with Spanish, and in particular with the Spanish spo-
ken in the Basque Country, which is leísta and thus marks the causee of unergatives with
dative case (Tu le haces reír, Yo le hago bailar). Actually, with simplex unergatives (ii) the
dialectal divide is even more marked: whereas in Western dialects dative is more gener-
ally used (iii), in Eastern dialects absolutive is normally found (iv) (Ricardo Etxepare, p.c.;
Oyharçabal & Epelde 2012).
(i) Tu le fais rire, Je le fais danser
you cl.3sg.m.acc make.prs.2sg laugh, I cl.3sg.m.acc make.prs.1sg dance
‘You make him laugh, I make him dance’
(ii) Aduna-k dantzatu du.
Aduna-erg dance aux.prs.erg3pl.abs3sg
‘Aduna has danced.’
(iii) Aduna-ri dantz-arazi diote.
Aduna-dat dance-cause aux.prs.erg3pl.abs3sg.dat3sg
‘They have made Aduna dance.’
(iv) Aduna dantz-arazi dute.
Aduna dance-cause aux.prs.erg3pl.abs3sg
(77) Aduna-ri dantza egin-arazi diote.

Aduna-dat dance do-cause aux.prs.erg3pl.abs3sg.dat3sg
(78) Aduna dantza egin-arazi dute.

Aduna dance do-cause aux.prs.erg3pl.abs3sg
Last, but not least, NVE s present a strong dissimilarity with DO s: whereas com-
mon DO s must bear a determiner (79) NVE s (generally) lack it (80) (Oyharçabal
2006: 793–794).
(79) *Liburu irakurri dut.

book read aux.prs.erg1sg.abs3sg
‘I have read book.’
(80) Barre / Lo / Dantza33 egin dut.

laugh / sleep / dance do aux.prs.erg1sg.abs3sg
‘I have laughed, slept, danced.’
Thus, the lack of a determiner is argued to signal incorporation (Martinez 2015).

Nonetheless, this difference can be accounted also in the transitive analysis, for
instance by postulating that NVE s and plain DO s receive case differently (cf.
Laka 1993).
To summarize, NVE s in Basque LVC s pattern with absolutive objects in some
cases, but differ from them in others. What one can conclude from all the pre-
vious tests and studies on Basque LVC s is that there is no single analysis for
all of them; rather, whereas some LVC s pattern with transitive structures, some
others show what can be taken as signals of syntactic incorporation. As Oyhar-
çabal (2006) suggests, Basque LVC s constitute a non-homogeneous set and the
degree of cohesion between the LVC and the NVE, and thus whether there is
incorporation or not, is subject to dialectal and idiolectal variation, with some
particular LVC s such as hitz egin “do word” ‘talk’ being a generalized case of
incorporation across varieties. Basque LVC s can be thus seen as forming a con-
tinuum, in Martinez’s (2015: 350) terms, comprising the three groups of LVC s
that we next describe.
33 The -a ending of dantza does not correspond to the determiner, but is part of the nomi-
nal root, making ‘dance’ and ‘the dance’ formally equal. Compare: hitz ‘word’ ~ hitz-a ‘the
word’ ~ hitz bat ‘one word’ and dantza ‘dance’ ~ dantza ‘the dance’ ~ dantza bat ‘one dance’.
At the beginning of the continuum, we find transitive-like structures, such

as dantza egin “do dance” ‘dance’, negar egin “do weep” ‘weep’, irribarre egin “do
smile” ‘smile’ and so on. As in common transitive structures, in Basque LVC s
the auxiliary is *edun, the subject bears ergative case, and the NVE functions as
a DO, which can bear partitive case and can be separated from the verb. How-
ever, although this type of LVC s has a transitive syntactic structure, they show
morphological, semantic and syntactic properties that distinguish them from
common transitive structures (Martinez 2015: 337, 373).
At the end of the continuum, we find the fossilized LVC s, such as alde egin
(aldegin), hanka egin, ospa egin, hitz egin (hitzegin), hots egin and a few others.
These can be syntactically accounted for as instances of incorporation and, as a
result, they do not admit partitive case, they show adjacency to the verb, etc.—
actually, these LVC s are generally phonologically treated as one single word, e.g.
aldégin, hitzégin and so on. As expected, many LVC s with a very tight relation
between their components have idiomatic readings, as clearly exemplified by
alde egin “do side/zone” ‘leave’, hanka egin “do leg” ‘leave’ (Martinez 2015: 337,
347–348), and they contrast with LVC s whose word order can be altered and
have a compositional reading (e.g. dantza egin “do dance” ‘dance’).
Finally, there is a group of LVC s presenting a behaviour ambiguous between
these two extremes (e.g. amets egin ‘dream’, barre egin ‘laugh’) (Martinez 2015:
337).
4 Conclusions
We have presented a description of LVC s in Basque, and we have compared

them with LVC s in Romance, mainly French and Spanish, keeping in mind
Ramchand’s (2014) position that there is no fundamental difference between
LVC s and the constructions headed by the same verbs in their non-light use
(cf., e.g., Basque lo egin “do sleep” ‘sleep’ vs ogia egin ‘make bread’). Rather, it is
the nature of the NVE, as opposed to other DO s, that makes LVC s special. This
hypothesis is paired with the assumption that in LVC s the LV, as head of the
construction, is responsible for its event and argument-structure properties,
while the NVE encodes the conceptual scene evoked.
We have dealt with prominent discussions on the syntax and semantics of
LVC s in Basque, and also French and Spanish. With respect to the nature of the
LV, for instance, we have remarked that the choice of LV is generally meaning-
ful, and that it is only DO and GIVE that seem to give rise to synonymous LVC s, at
least at the conceptual level, both within Basque and when we compare French
(favouring DO) and Spanish (favouring GIVE).
We have also paid attention to the NVE and its categorial properties. Basque,
unlike French and Spanish, often allows a type of LVC in which the nominal NVE
appears in the inessive case, or endowed with an adverbial suffix, the result-
ing LVC displaying peculiar aspectual properties. As regards unmarked nominal
NVE s, we have shown that they can be bare, i.e., determinerless and caseless,
in Basque and French—unlike Spanish and Catalan. For Romance, we have
also explored the different kinds of relations—semantic and morphological—
between LVC s and their synthetic counterparts.
Finally, a discussion has been dedicated to the degree of cohesion between
the LV and the NVE. It has been shown that there is empirical evidence pointing
in diverse directions as regards this issue: the NVE is in some cases quite syn-
tactically independent of the LV, in other cases it seems to be somehow fused
with it (i.e., incorporated into it), and, finally, there are cases that lie in between
these two extremes. Dialectal and lexical factors have been pointed out as pre-
dictors of this variation in Basque.
It should be clear from these conclusions that, much in the spirit of Butt’s
(2010) title, Basque (and Romance) LVC s constitute a jungle, and we are still
hacking away.
Acknowledgements
This work has been supported by the postdoctoral research fellowship Beatriu
de Pinós 2014 BP_A 00165 (Secretaria d’Universitats i Recerca del Departament
d’Economia i Coneixement de la Generalitat de Catalunya) (second author),
the postdoctoral research fellowship Juan de la Cierva-incorporación (IJCI-
2016–30474, Ministerio de Economía y Competitividad, Spanish Government)
(second author) and the research project FFI2014–56968-C4–1-P (Ministerio de
Economía y Competitividad, Spanish Government) (first and second author).
The research leading to this work began at Queens’ College, University of Cam-
bridge, in the case of the first author, and at CNRS-IKER, in the case of the
second author. We are grateful to Jaume Mateu and an anonymous reviewer
for very valuable comments. All errors remain our own.
References
Acedo-Matellán, Víctor. 2014. “Can we dance without doing a dance? Two opposite
views on the integration of roots in the syntactic structure of the vP”. In Asaf
Bachrach, Isabelle Roy & Linnaea Stockall (eds.), Structuring the argument. Mul-
tidisciplinary research on verb argument structure. Amsterdam/Philadelphia: John

Benjamins. 23–43.
Acedo-Matellán, Víctor, & Jaume Mateu. To appear. “El verb i el sintagma verbal [‘The
verb and the verb phrase’]”. In Josep Martines & Manuel Pérez-Saldanya (eds.), Gra-
màtica del català antic.
Ahn, Hee Don. 1991. Light Verbs, VP-Movement, Negation and Clausal Architecture in
Korean and English. PhD thesis. University of Wisconsin–Madison.
Alba-Salas, Josep. 2002. Light verb constructions in Romance: A syntactic analysis. PhD
thesis. Ithaka (NY): Cornell University.
Aldai, Gontzal. 2009. “Is Basque morphologically ergative?: Western Basque vs. Eastern
Basque”. Studies in Language 33(4): 783–831.
Aldai, Gontzal. 2010. “Euskararen kasu-markaketaren aldakortasun dialektala”. In Beat-
riz Fernández, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: Aldako-
rtasun sintaktikoa aztergai, [Supplements of ASJU 52]. Bilbao: Universidad del País
Vasco. 5–19.
Alonso-Ramos, Margarita. 1998. Étude sémantico-syntaxique des constructions à verbe
support. PhD thesis. Université de Montreál.
Austin, Jennifer & Luis López. 1995. Nominative, absolutive, and dative languages. In Jill
N. Beckman (ed.), Proceedings of the Twenty-fifth North Eastern Linguistic Society.
Amherst, MA: GLSA. 1–16.
Baker, Mark. 1988. Incorporation: a theory of grammatical function changing. Chicago:
University Press of Chicago.
Barrie, Michael & Éric Mathieu. 2014. “Head movement”. In Andrew Carnie, Yosuke Sato
& Daniel Siddiqi (eds.), The Routledge Handbook of Syntax. Milton Park, Abingdon,
Oxon; New York: Routledge. 133–149.
Berro, Ane (this volume). (In)transitivity. In Ane Berro, Beatriz Fernández & Jon Ortiz
de Urbina (eds.), Aligning grammars: Basque and Romance.
Berro, Ane. 2015. Breaking verbs. From event structure to syntactic categories in Basque.
Ph.D. thesis, University of the Basque Country (UPV/EHU) and Université Bordeaux
Montaigne (UBM).
Bobaljik, Jonathan David. 1993. “On ergativity and ergative unergatives”. In Collin Phil-
ips (ed.), MIT Working Papers in Linguistics 19: Papers on Case and Agreement II.
Cambridge: MIT, 45–88.
Bosque, Ignacio & Pascual Masullo. 1998. “On verbal quantification in Spanish”. In
Francesc Roca and Olga Fullana (eds.), Studies on the syntax of central romance
languages. Proceedings of the III Symposium on the Syntax of Central Romance Lan-
guages [held in Girona on the 15th and 16th of November 1996], 9–63. Girona: Uni-
versitat de Girona.
Bowern, Claire. 2008. “The diachrony of complex predicates”. Diachronica 25(2), 161–
185.
Brody, Michael 2000. “Mirror theory: Syntactic representation in perfect syntax”. Lin-
guistic Inquiry 31: 29–56.
Butt, Miriam. 1995. The structure of complex predicates in Urdu. Standord, California:
CSLI Publications.
Butt, Miriam. 2003. “The light verb jungle”. In Gulsat Aygen, Claire Bowern, Conor
Quinn (eds.), Harvard Working Papers in Linguistics, vol. 9, 1–49. Papers from the
GSAS/Dudley House Workshop on Light Verbs.
Butt, Miriam. 2010. “The light verb jungle: Still hacking away”. In Mengistu Amberber,
Brett Baker, Mark Harvey (eds.), Complex Predicates: Cross-linguistic Perspectives on
Event Structure, 48–78. Cambridge: Cambridge University Press.
Butt, Miriam, & Wilhelm Geuder. 2001. “On the (semi)lexical status of light verbs”. In
Norbert Corver & Henk van Riemsdijk, H. (eds.), Semi-lexical Categories: On the Con-
tent of Function Words and the Function of Content Words. Berlin: Mouton de Gruyter.
323–370.
Butt, Miriam & Aditi Lahiri. 2013. “Diachronic pertinacity of light verbs”. Lingua 135:
7–29.
Bye, Patrick, & Peter Svenonius. 2012. “Non-concatenative morphology as epiphenom-
enon”. In Jochen Trommer (ed.), The Morphology and Phonology of Exponence. Ox-
ford: Oxford University Press. 427–495.
Cattell, Ray. 1994. Composite Predicates in English. Sydney: Academic Press Australia.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Cuervo, Mª Cristina. 2003. Datives at large. PhD thesis, MIT
Danlos, Laurence. 1992. “Support verb constructions: linguistic properties, representa-
tion, translation”. French Language Studies 2: 1–32.
De Rijk, Rudolf. 2008. Standard Basque: A Progressive Grammar. Cambridge, MA: MIT
Press.
Epelde, Irantzu & Beñat Oyharçabal. 2012. Resultados lingüísticos del contacto francés-
euskera: el cambio dativo > absolutivo en algunas construcciones causativas. Sociedad
Española de Lingüística. Universitat de València.
Etxepare, Ricardo. 2003. “Valency and Argument Structure in the Verb”. In Jose Igna-
Gruyter. 369–465.
Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. PhD thesis, Euskal
Herriko Unibertsitatea / Universidad del País Vasco.
Giry-Schneider, Jacqueline. 1978. “Les Nominalisations En Français: L’opérateur ‘faire’
Dans Le Lexique”. Genève: Droz.
Giry-Schneider, Jacqueline. 1986. “Les noms construits avec “faire”: compléments ou
prédicats?”. Langue française 69, 49–63.
Goldberg, Adele. 1995. Constructions: A Construction Grammar Approach to Argument
Structure. Chicago; London: University of Chicago.
Grimshaw, Jane. 1990. Argument Structure. Cambridge, Mass: MIT.

Grimshaw, Jane. 2005 [1993]. “Semantic Structure and Semantic Content in Lexical
Representation”. In Jane Grimshaw (ed.), Words and Structure. Stanford, CA: CSLI
Publications. 75–89.
Grimshaw, Jane & Armin Mester. 1988. “Light Verbs and θ-Marking”. Linguistic Inquiry
19(1): 205–232.
Gross, Maurice. 1976. “Sur quelques groupes nominaux complexes”. In Chevalier, Jean-
Claude and Maurice Gross (eds.). Méthodes en grammaire française. Paris: Klinck-
sieck, 97–119.
Gross, Maurice. 1981. “Les bases empiriques de la notion de prédicat sémantique”. Lan-
gages 63: 7–52.
Hale, Kenneth and Samuel Jay Keyser. 1986. “Some transitivity alternations in English”,
Lexicon Project Working Papers #7. Cambridge, MA: Center for Cognitive Science,
Massachusetts Institute of Technology.
Hale, Kenneth & Samuel Jay Keyser. 1997. “On The Complex Nature of Simple Predica-
tors”. In Alex Alsina Joan Bresnan & Peter Sells (eds.), Complex Predicates. Stanford,
Calif.: CSLI Publications. 29–65.
Hale, Kenneth & Samuel Jay Keyser. 1993. “On argument structure and the lexical
expression of syntactic relations”. In Kenneth Hale and Samuel Jay Keyser (eds.),
The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger. Cam-
bridge, MA: MIT Press. 53–109.
Haugen, Jason. 2015. “Incorporation”. In Peter Müller, Ingeborg Ohnheiser, Susan Olsen
& Franz Rainer (eds.), Word Formation: An International Handbook of the Languages
of Europe. Berlin: Mouton de Gruyter, 413–434.
Himmelman, Nikolaus. 2001. “Articles”. In Martin Haspelmath et al. (eds.), Language
Typology and Language Universals. Handbücher ders Sprach und Kommunikation-
swissenschaft, vol. 20, 1. Berlin, New York: Walter de Gruyter. 831–841.
Hook, Peter E. 1991. “The Emergence of Perfective Aspect in Indo-Aryan Languages”.
In Elizabeth Traugott and Bernd Heine (eds.), Approaches to Grammaticalization,
vol. II: Types of grammatical marker. Amsterdam: Benjamins. 59–89.
Hopper, Paul J. & Elizabeth C. Traugott. 1993. Grammaticalization. Cambridge: Cam-
bridge University Press.
Jackendoff, Ray. 1974. “A Deep Structure Projection Rule”. Linguistic Inquiry 5(4), 481–
505
Jespersen, Otto. 1954. A Modern English Grammar, vol. 4. New York: Barnes and Noble.
Kearns, Kate. 1988/2002. Light Verbs in English. Ms., MIT.
Lafitte, Pierre 1944. Grammaire Basque. Baiona: Elkar.
Laka, Itziar. 1993. “Unergatives that Assign Ergative, Unaccusatives that Assign Accusa-
tive”. In Jonathan D. Bobaljik & Collin Phillips (eds.), Papers on Case and Agreement I,
MITWPL 18. Cambridge, MA: MITWPL, 149–172.
Larson, Richard K. 1988. “On the double object construction”. Linguistic Inquiry 19(3),
335–391.
Levin, Beth & Malka Rappaport. Hovav. 1995. Unaccusativity: At the Syntax-lexical Se-
mantics Interface. Cambridge, Mass.; London: MIT, 1995.
Lin, Tzong-Hong. 2001. Light Verb Syntax and the Theory of Phrase Structure. PhD thesis,
University of California, Irvine.
Marantz, Alec. 2005. “Objects Out of the Lexicon: Objects as Events”. Ms., MIT.
Martinez, Arantzazu. 2015. “[Izen + egin]a aditz-lokuzioak: inkorporazio mailak”. PhD
thesis, Deustuko Uniberstitatea.
Matushansky, Ora. 2006. “Head movement in linguistic theory”. Linguistic Inquiry 37,
69–109.
McFadden, Thomas. 2006. “German inherent datives and argument structure”. In
Daniel P. Hole, Abraham Werner & André Meinunger (eds.), Datives and Other Cases:
Between Argument Structure and Event Structure. Amsterdam; Philadelphia: John
Benjamins. 49–77.
Mithun, Marianne. 1984. “The evolution of Noun Incorporation”. Language 60, 847–893.
Mohanan, Tara. 1994. Argument Structure in Hindi. Stanford, California: CSLI Publica-
tions.
Ortiz de Urbina, Jon. 1989. Parameters in the Grammar of Basque. Dordrecht: Foris.
Oyharçabal. Beñat. 2006. “Basque Light Verb Constructions”. ASJU XL: 787–806.
Pesetsky, David & Esther Torrego. 2001. “T-to-C Movement: Causes and Consequences”.
In Michael Kenstowicz (ed.), Ken Hale: A Life on Language. Cambridge, MA: MIT
Press. 355–426.
Piera, Carlos & Soledad Varela. 1999. “Relaciones entre morfología y sintaxis”. In Vio-
leta Demonte & Ignacio Bosque (ed.), Gramática descriptiva de la lengua española,
vol. 3: Entre la oración y el discurso. Morfología. Madrid: Real Academia Española /
Espasa-Calpe. 4367–4422.
Pineda, Anna. 2016. Les fronteres de la (in)transitivitat. Estudi dels aplicatius en llengües
romàniques i basc [published and revised version of the PhD dissertation]. Barcelo-
na: Institut d’Estudis Món Juïc. Col·lecció Cum Laude, 6.
Pylkkänen, Liina. 2002. “Introducing Arguments”. PhD thesis, MIT.
Pylkkänen, Liina. 2008. “Introducing Arguments”. Cambridge, Mass.; London: MIT.
Rafel, Joan. 2004. “Los predicados complejos en espanyol”. In Elixabete Pérez Gaztelu,
Igone Zabala & Lluïsa Gràcia (eds.), Las fronteras de la composición en lenguas
románicas y en vasco. Donostia: Deustuko Unibertsitatea. 393–443.
Ramchand, Gillian C. 2008. Verb meaning and the lexicon: A first phase syntax. Cam-
Ramchand, Gillian C. 2014. “On structural meaning vs conceptual meaning in verb
semantics”. Linguistic Analysis 39(1–2), 207–243.
Rappaport, M. & Levin, B. 1988. “What to do with theta-roles”. In Wendy Wilkins (ed.)
Syntax and Semantics 21: Thematic relations. New York: Academic Press, 7–36.
Rappaport Hovav, Malka & Beth Levin. 1998. “Building verb meanings”. In Miriam Butt
& Wilhelm Geuder (ed), The Projection of Arguments: Lexical and Compositional Fac-
tors. Stanford, Calif.: CSLI Publications, 97–134.
Rodríguez, Sonia & Fernando García Murga. 2003. “IZEN + EGIN predikatuak euskaraz”.
In Jesús Makatzaga & Beñat Oyharçabal (eds.), Euskal gramatikari eta literaturari
buruzko ikerketak XXI. mendearen atarian, Gramatika gaiak, Iker-14. Bilbo: Euskal-
tzaindia. 417–436.
Sarasola, Ibon. 1977. “Sobre la bipartición inicial en el análisis en constituyentes”,
ASJU XI: 51–90.
Uribe-Etxebarria, Myriam. 1989. “Some Notes on the Structure of IP in Basque”. Ms.,
UConn, Storrs.
Wierzbicka, Anna. 1982. “Why Can You Have a Drink When You Can’t *Have an Eat?”
Language, Vol. 58, No. 4, 753–799.
Wood, Jim & Alec Marantz. 2017. “The interpretation of external arguments”. In Roberta
D’Alessandro, Irene Franco & Ángel Gallego (ed.), The verbal domain. Oxford: Oxford
Zabala, Igone. 2004. “Los predicados complejos en vasco”. In Elixabete Pérez Gaztelu,
Igone Zabala & Lluïsa Gràcia (eds.), Las fronteras de la composición en lenguas
románicas y en vasco. Donostia: Deustuko Unibertsitatea. 445–534.
chapter 7
On Non-selected Datives: Ethical Datives in Basque

and Spanish
Beatriz Fernández
1 Introduction
There is a long tradition of research dealing with datives in general and non-
selected datives in particular. These non-selected datives include external pos-
sessors, benefactives and ethical datives, among others (see for example, Bosse,
Bruening & Yamada 2012). In this chapter, I will analyze some of these datives,
particularly, ethical datives. In doing so, one of the major difficulties to over-
come is the extreme variation of designations and the wide range of properties
(not necessarily coherent with each other) attributed to them. For instance,
ethical datives have been considered to refer to the participants of the speech
act and mostly to the speaker (sometimes also to the hearer). If this were the
case, then ethical datives would be restricted to 1st (or 2nd) person and couldn’t
be doubled by a DP, among other properties (see Jaeggli 1986 for Spanish). Nev-
ertheless, not all the so-called ethical datives seem to be restricted to clitics or
1st and 2nd person. Hence, a more-fine grained typology seems to be necessary
in order to clarify this puzzling scenario.
In this chapter, I will follow Franco & Huidobro (2008) in distinguishing
three types of ethical datives in Spanish. First of all, Class I, reflexive non-
argumental clitics as in me comí la manzana ‘I ate the apple’, alternatively ana-
lyzed as aspectual by Miguel & Fernández Lagunilla 2000, and reminiscent of
personal datives (Horn 2008, 2013). Second, Class II of the ‘non volitional se +
dative clitic type’, as in se me caló el coche ‘the car stalled on me’. Ethical datives
from Class II correspond to affected experiencers in Bosse, Bruening & Yamada’s
(2012) typology or affected datives (dativos de afectación) in Maldonado (1994)
for Spanish and in Fernández (2010) for Basque. Third, Class III, as in me le
pegaron una paliza a mi primo ‘they beat my cousin up’ with a dative clitic clus-
ter including the ethical dative me and the goal dative le. This Class III shows
the most restricted version of ethical datives, as they seem to be necessarily 1st
and 2nd person (me le/te le/*se le) whereas ethical datives from Class I and II do
not. Besides, ethical datives from either Class III or Class I cannot be doubled
(me/*a mí) whereas those from Class II can.

222 fernández
On the other hand, unlike in Hispanic linguistics, ethical datives have been
only briefly mentioned but not analyzed in depth in Basque. The sole exis-
tence of ethical datives in Basque is not uncontroversial, as they are neither
restricted to 1st and 2nd person nor reduced to clitics (Fernández 2010). As a
consequence, some of the properties attributed to ethical datives in general
and Spanish ethical datives in particular are not met in Basque.
In this chapter, I will review classical syntactic tests (as in Borer & Grodzin-
sky 1986 or Jouitteau & Rezac 2007) and discuss more novel criteria proposed
by Bosse, Bruening & Yamada (2012) and Horn (2008, 2013). I will claim that
the only ethical datives attested in Basque are those related to Spanish ethical
datives from the Class II as neither Class I nor Class III is attested in Basque.
I will also claim that only Spanish (and not Basque) exhibits ethical datives in
their most restricted definition (Class III), i.e. 1st and 2nd person clitics referred
to the participants of the speech act with no DP-doubling among other prop-
erties. As we will see, ethical datives of the Class II contribute truth conditional
meaning—or at issue meaning in Potts’ (2005) designation—but contrary to
what is argued by these authors, they do not contribute non-truth conditional
meaning—or not-at-issue meaning—in either Basque or Spanish.
Although no particular analysis will be provided here, I will claim that ethi-
cal datives of Class II are merged in the Specifier of an Applicative head above
v, as argued in Odria (2017) for experiencers of psychological verbs and exter-
nal possessors. Thus, the similarity in the behavior of ethical datives of Class II,
external possessors and also benefactives follows. Besides, ethical datives from
Class I and III need further explanation that go beyond the limits of this
paper.
This chapter will be structured as follows. First of all, I will briefly present
non-selected datives in general and ethical datives in particular (section 2). Sec-
ondly, I will review and discuss some classical and novel tests on Basque and
Spanish ethical datives of Class II (section 3). Section 4 will present the main
conclusions and a brief discussion.
2 Ethical Datives
2.1 Some Clarifications before Starting

Although the theoretical analysis of datives has been exponentially increased
especially since the nineties and the first decade of this century, there are still
some controversial issues in the literature of datives, particularly, the distinc-
tion between selected and non-selected ones. This distinction goes back to
some classical papers such as Borer & Grodzinsky (1986) on Hebrew. In this
on non-selected datives: ethical datives in basque and spanish 223
paper, some of the most salient properties of non-selected datives are pre-
sented and discussed in depth, as far as external possessors and ethical datives
are concerned. The basis of the typology of these non-selected datives and the
properties attributed to each type are still studied as a classical benchmark test.
Other studies such as Jouitteau & Rezac (2007) have explored the literature on
ethical datives in great detail in order to characterize this particular type of
non-selected datives in French.
Paradoxically, one of most influential recent works on datives, i.e. Pylkkä-
nen 2008 [2002] and her well-known high v low applicative distinction does
not solve the issue, as non-selected datives such as ethical datives and bene-
factives can be analyzed as high applicatives, blurring the distinction between
both of them. Besides, although external possessors are treated as low applica-
tives in her proposal, they behave similarly to ethical datives and benefactives
(high applicatives) in many respects.
An attempt to typify non-selected datives properly is due to Bosse, Bruen-
ing & Yamada (2012), an analysis that combines both syntactic and semantic
properties of non-selected datives. In particular, these authors analyze affected
experiencers which seem to correspond to a particular type of ethical datives
in Spanish, i.e. Class II in Franco & Huidobro’s (2008) typology. Unfortunately,
some of the properties attributed to affected experiencers, in particular their
contribution to non-truth conditional meaning and sentientness, are not met
in Spanish and Basque ethical datives of this class.
In this chapter, I will concentrate on ethical datives and I will leave aside
external possessors and benefactives, although I will refer to these non-selected
datives in order to compare the properties of one another. Both external posses-
sors and benefactives are also attested in the languages under discussion—see
for instance Arregi (2003), Fernández (2010), Fernández & Ortiz de Urbina
(2010) and Odria (2017) for Basque external possessors, and Maldonado (1994)
for a discussion on the distinction between benefactives and ethical datives
(named as datives of interest by the author) in Spanish.1 As we will see, the
1 Correspondence between designations attested in some of the papers mentioned in this

chapter: ethical datives Class I (Franco & Huidobro 2008) = personal datives (Horn 2008,
2013) = subject co-referent datives (Bosse, Bruening & Yamada 2012); ethical datives Class II
= affected datives (Maldonado 1994, Fernández 2010) = affected experiencers (Bosse, Bruen-
ing & Yamada 2012) = experiencer datives (Bosse & Bruening 2010); ethical datives Class III
(Franco & Huidobro 2008) = datives of interest Maldonado (1994) = attitude holders and also
affected experiencers (Bosse, Bruening & Yamada 2012). In referring to ethical datives, Strozer
(1978) uses the term datives of interest. Although no particular designations are provided in
her paper, the three dative types distinguished by this author can accommodate to Franco &
Huidobro’s (2008) three-way typology assumed here.
224 fernández
label ethical datives involves quite different types of datives. Hence, a more fine-
grained typology of ethical datives is necessary in order to characterize them
properly.
2.2 Ethical Datives in Basque and Spanish: Some Data and a Three-Way
Typology
Let us present some basic data on Basque and Spanish ethical datives. In prin-
ciple, an ethical dative is a non-selected one, i.e. a dative not included in the
argument structure of the verb. This ethical dative experiences the positive
or negative effect of an event, thus, it seems to be affected by the event. For
instance, Basque examples in (1) show a dative of this sort:
(1) a. (Niri) haurra lokartu zait.

I.dat kid (abs) fall asleep aux + 1sg.dat2
‘The kid has fallen asleep.’
b. Haurrak (niri) liburua apurtu dit.

kid.erg I.dat book (abs) break aux + 3sg.dat
‘The kid has broken a book.’
The sentences in (1) include a 1st person singular dative clitic -t in both the
bivalent unaccusative form (zait) and in the ditransitive one (dit).3 The datives
in these sentences are non-selected arguments, as their counterparts with no
dative are equally available and grammatical in Basque:
(2) a. Haurra lokartu da.

kid (abs) fall asleep aux
2 Abbreviations: abs absolutive, acc accusative, adl adlative, allo allocutive, aux auxiliary,
cl clitic, dat dative, d determiner, ed ethical dative, ed-ii ethical dative of Class II, ed-iii
ethical dative of Class III, erg ergative, fam familiar, fem feminine, mas masculine, nom
nominative, pl plural, psr possessor, pst past, sg singular, 1, 2, 3 for first, second and third
person.
3 Basque is a language with auxiliary alternation: BE is selected with unaccusative predicates
and HAVE with unergative and (di)transitive predicates. Besides, Basque verbal forms include
clitics/agreement markers for absolutive, ergative and dative arguments. For ease of exposi-
tion, I will gloss the verbal form as aux (auxiliary) + cl.dat (dative clitic) when it includes a
dative clitic, and as aux, when it does not. For ease of understanding, further details will be
omitted.
b. Haurrak liburua apurtu du.

kid.erg book (abs) break aux
The datives in (1) experience the effect of the kid falling asleep (1a) and the kid
breaking the book (1b). The examples in (1) are equivalent to Spanish examples
in (3).
(3) a. Se me ha dormido el niño.

cl cl.1sg.dat aux fall asleep the kid
(adapted from Maldonado 1994)
b. El niño me ha roto el libro.

the kid cl.1sg.dat aux break the book
Ethical datives of the ‘(non volitional) se + dative clitic type’ type as the one in
example (3a) are classified as ethical datives within Class II in Franco & Huido-
bro’s (2008) three-way typology. I will name them as Ethical Datives-Class II
(ED-II). Other designations have been used for the very same datives, such as
affected datives (Spanish dativos de afectación) in Maldonado (1994) for Spanish
and Fernández (2010) for Basque. These ED-II seem to correspond to affected
experiencers in Bosse, Bruening & Yamada (2012).
The interpretation of the ethical dative in (1a) and (3a) can be both negative
and positive, whereas it seems to be negative in (1b) and (3b). Thus, the desig-
nation affected experiencer is used instead of malefactive or adversative in some
works, as explicitly said in Bosse & Bruening (2010).
Nevertheless, the malefactive interpretation of datives involved in construc-
tions similar to (1b) and (3b) is not obligatory. For instance, if instead of ‘break
the book’, we say ‘pass the exam’, the interpretation of the dative will be, in
principle, positive. Therefore, I will assume that both benefactive and malefac-
tive interpretations are available for examples such as (1) and (3), an assump-
tion also made by Maldonado (1994) and Franco & Huidobro (2008) for Span-
ish.
It is worth pointing out that that the (di)transitive structure in (3b) is not
included in Franco & Huidobro’s Class II, as it does not involve the clitic se.
Leaving aside this clitic se, the main difference between both constructions
resides in the fact that the former example (3a) is intransitive whereas the
latter is transitive. As far as Basque is concerned, I see no reason to exclude
transitive structures in this ED-Class II. Actually, Maldonado (1994) includes
both intransitive and transitive structures when dealing with ED-Class II. More-
226 fernández
over, the second type of datives distinguished by Strozer (1978), as in me ha

comido la manzana (a mí) ‘he/she has eaten the apple’ fits well within this ED-
Class II.
The other two types of ethical datives distinguished by Franco & Huidobro
(2008) (and also in Strozer 1978) are on the one hand, those including an aspec-
tual or reflexive (non-argumental) clitic (Class I), as in example (4a), and on the
other hand, dative clitics such as the first singular clitic in (4b) (Class III).
(4) a. Yo me he comido toda la sopa para cenar.

cl cl.1sg.dat aux eat.pfv all the soup to dine
‘I have eaten all the soup for dinner.’ (Franco & Huidobro 2008)
b. El niño me le ha roto el libro.

the kid cl.1sg.dat cl.3sg.dat aux break the book
The dative in (4a) is reminiscent of the so-called personal datives attested in

Southern and Appalachian U.S. English as in I love me some him (Horn 2008: 176,
2010)—see also Christian (1991) and Webelhuth & Dannenberg (2006) among
others. I will not include the reflexive non-argumental dative in this chapter, as
it has no dative counterpart in Basque. Besides, the clitic cluster me-le in exam-
ple (4b) and any other are not attested in Basque—see section (3.3). Thus, if
a construction with a clitic cluster is needed in order to typify ethical datives
in Class III, then Basque shows no evidence of these datives. Nevertheless,
Franco & Huidobro (2008) mentions another structure with a sole clitic in this
very same Class III, exemplified by (5a). Its Basque counterpart is provided in
(5b).
(5) a. A mi primo, me lo llevaron preso.

to my cousin cl.1sg.dat cl.3sg.acc take prisoner
‘My cousin was taken prisoner.’ (Franco & Huidobro 2008)
b. Nire lehengusua, preso eraman zidaten.

my cousin prisoner take aux + 1sg.dat
The Spanish example (5a) shows some restrictions that lead Franco & Huido-
bro (2008) to typify it as ED-Class III. However, its Basque counterpart shows
exactly the same behavior of ED-Class II. In this point Spanish and Basque split.
I will come back to this discussion in section 3.2.
2.2 Typifying Ethical Datives of Class II

Now let us concentrate on (1) and (3). As I have said before (section 2.1), the
datives in these examples correspond to ED-Class II in Franco & Huidobro
(2008) along with some (di)transitive structures including an ethical dative.
These ethical datives show the following properties: 1) it can be 1st, 2nd and
3rd person; 2) it can be doubled by a DP; 3) it cannot be combined with any
dative other than ethical dative of Class I and III in Spanish or allocutive cli-
tics in Basque; 4) it contributes truth conditional meaning, as it can be ques-
tioned and bind a variable in the truth-conditional tier; and finally, 5) it does
not contribute non-truth conditional meaning, as the experience itself can be
negated and the meaning of experiencers does not survive beyond yes/no ques-
tions.
This ED-Class II contrasts with the ED-Class III attested as a first person sin-
gular clitic me in example (4b) repeated here as (6):
(6) El niño me le ha roto el libro.

‘The kid has broken the book.’
Compared to ED-Class II, the ED-Class III exemplified in (6) shows the oppo-
site properties: 1) it can only be 1st or 2nd person; 2) it cannot be doubled by a
DP (it is a weak pronoun); 3) it can be freely combined with any other dative
argument; 4) it does not contribute truth conditional meaning, as it cannot be
questioned or bind a variable in the truth-conditional tier; and 5) it only con-
tributes non-truth conditional meaning (an implicature).
Although Basque allocutives (Oyharçabal 1993) fall outside the scope of this
chapter, they share all the properties attributed here to by ED-Class III, with
the particularity that allocutives can only be 2nd person, as in (7). This exam-
ple includes an allocutive marker: a 2nd person singular morpheme that can
be either masculine -k or feminine -n and corresponds to a familiar addressee.
(7) Haurrak liburua irakurri ziok/n.

I.erg book (abs) read aux + 3sg.dat + 2sg.allo.mas/fem
‘The kid has read a book (it affects him/her + I tell you)’
In what follows, I will analyze the properties attributed here to ED-Class II in

comparison to other non-selected datives in general and ED-Class III in partic-
ular.
228 fernández
3 Ethical Datives of the Class II: Basque and Spanish in Comparison
In section 2, I have presented Franco & Huidobro’s (2008) three-way typology

of Spanish ED. This typology is a good start to review some of the properties
attributed to ED in general and ED-Class II in particular. In this section, I aim
at revisiting some of these properties in order to characterize properly Spanish
and Basque ED-Class II in comparison to each other.
3.1 Ethical Datives-Class II and Person

One of the properties repeatedly mentioned in the literature on ethical datives
has to do with person: ethical datives can only be 1st and 2nd person (see Jaeg-
gli 1982; Cuervo 2003: 195; Jouitteau & Rezac 2007: 2.1 among many others). In
a more refined typology of ethical datives, it is worth analyzing which partic-
ular class shows this restriction. As far as ED-Class II is concerned, there is no
person restriction at all.
(8) a. Se me/te/le ha dormido el niño.

cl cl.1/2/3sg.dat aux fall asleep the kid
b. Haurra lokartu zait/zu/o.

kid fall asleep aux + 1/2/3sg.dat
As discussed in Fernández (2010), Basque shows that either external posses-

sors or ED-Class II can be 1st, 2nd or 3rd person. The same behavior has been
observed by Maldonado (1994: 259) for Spanish ED-Class II.4 This pattern is also
attested in transitive structures:
(9) a. El niño me/te/le ha comido la manzana.

kid.erg cl.1/2/3sg.dat aux eat the apple
‘The kid has eaten the apple.’
4 The datives involved in Maldonado’s (1994) examples can be argued to be external possessors
(and not ED-Class II), as there is a relationship of inalienable possession between the dative
(the possessor) and a PP including a body part (the possessum), as in (i).
(i) Se me/te/le murió en los brazos.
cl cl.1/2/3SG.dat die in the arms
‘He/she died on me/you/him-her / He/she died in my/your/his-her arms.’
Nevertheless, similar examples including a possessor inside the PP would be also grammati-
cal: en los brazos de Juan ‘in John’s arms’. Hence, the datives involved in (i) can be ED and not
necessarily external possessors.
b. Haurrak (niri) sagarra jan dit/zu/o.

kid.erg I.dat apple eat aux + 1/2/3sg.dat
This property is also shared by other non-selected datives, such as benefactives

and ED-Class I (the class involving the reflexive non-argumental dative clitic,
as in Spanish Yo me como la manzana or Southern and Appalachian English
instances such as I love me some him). Thus, only one class remains: ED-Class III.
By definition, if ethical datives correspond to the participants of the utter-
ance, then the restriction to be 1st and 2nd person follows. Nevertheless, leaving
a priori definitions aside, the impossibility for ED-Class III to be 3rd person
seems to be real. Take for instance, one of the examples provided for ED-
Class III, repeated here as (10a). As we will see, the combination of a hypothet-
ical 3rd person ED-Class III and a 3rd person argumental clitic is out, whereas
1st and 2nd person are fine (10a,b):
(10) a. El niño me le / te le / *se le ha roto el libro.

kid.erg ed-iii 1/2/3 + cl.3sg.dat aux break the book
b. El niño me le / te le / *se le ha comido la manzana.

kid.erg ed-iii 1/2/3 + cl.3sg.dat aux eat the apple
‘The kid has eaten the apple.’
Note that the clitic cluster se le would be grammatical in a context where the
clitic se were co-referential with the subject, that is, ED-Class I (reflexive non-
argumental dative clitic) combined with a ED-Class II. To the contrary, a se clitic
not co-referential with the subject would be out.
Other ED-Class III are not so easily explained, but seem to exhibit the same
constriction:
(11) A mi primo, me/te/*se lo llevaron preso.

to my cousin cl.1/2/3sg.dat cl.3sg.acc take prisoner
‘My cousin was taken prisoner.’
The clitic se cannot be interpreted as an ED, although the same cluster se lo

would be grammatical, assuming (11) to be an aspectual se construction (de
Miguel 2000). Nevertheless, the ungrammaticality of *se lo in a construction
other than aspectual seems to be a side-effect of a Me-lui Constraint (see Odria
2017 and references therein), as also pointed out by Franco & Huidobro (2008:
footnote 4). Interestingly enough, Basque shows no person constraint in similar
structures:
230 fernández
(12) Nire lehengusua, preso eraman zidaten/zizuten/zioten.

my cousin prisoner take aux + 1/2/3sg.dat
In Basque, nothing prevents the ED from being 3rd person. Hence, we can spec-
ulate with the idea that the inability for the ED in (11) to be 3rd person cannot be
only due to its nature but to the morphological and syntactic constraints that
govern particular clitic combinations. Be that as it may, Basque and Spanish
seem to separate in this respect.
Summing up, nothing prevents ED-Class II from being 3rd person. With
regard to this, ED-Class II behaves in the same way external possessors, bene-
factives and any other argumental datives such as experiencers do. Being non-
argumental, ED-Class I also aligns with ED-Class II in this respect. It remains an
open question whether ED-Class III is restricted or not to 1st (and 2nd person),
as frequently discussed in the literature. The examples provided in (10) point
in that direction.
3.2 Ethical Datives-Class II and Clitic Doubling

Another property mentioned in the literature in order to distinguish ethical
datives from other datives (selected or not) has to do with their impossibility
to be doubled—see Jaeggli (1986: 22–23), Borer & Grodzinsky (1986: 180), and
Jouitteau & Rezac (2007: section 2.4).
(13) a. Juan me le arruinó la vida a esa chica (* a

John ed-iii cl.1/2/3sg.dat ruin the life to that girl ( on
mí).
me)
‘John ruined her life on me.’ (Perlmutter 1971)
b. Te me ensuciaste el pantalón (* a mí).

cl.2sg.dat ed-iii get dirty the trousers ( to me)
‘You got your trousers dirty (on me).’ (Jaeggli 1986: 22)
As can be seen in Perlmutter and Jaeggli’s classical examples the 1st person
singular dative clitic cannot be doubled (me … *a mí)—note that the 3rd per-
son singular dative clitic le in Perlmutter’s example (11a) can be doubled, as
the presence of the DP (a esa chica) reveals. This clitic corresponds to the ED-
Class III of the typology assumed in this chapter. To the contrary, ED-Class II
does allow doubling dative clitics in Spanish:
(14) a. Se le quedó dormido a su madre.

cl cl.3sg.dat fall asleep to his/her mother
‘He/she felt asleep (and it affected his/her mother).’
b. Se le escapó al policía.
cl cl.3sg.dat get away to the police
‘He/She got away from the police.’ (Maldonado 1994: 260)
The same behavior can be observed in their Basque counterparts: doubling

does not entail ungrammaticality.
(15) a. Bere amari lokartu zitzaion.

his/her mother.dat fall asleep aux + 3sg.dat
b. Poliziari ihes egin zion.

Police.dat get away aux + 3sg.dat
This property is closely related to the dative being or not an argument. Thus,
for instance, Franco & Huidobro (2010: 218, section 3) assume that the possibil-
ity of having a DP in a case position doubled by a clitic reflects the argumental
status of that clitic. Being this the case, ED-Class II are argumental, as shown
in (12), whereas ED-Class I and ED-Class III are not—(4a) and (4b) respectively
repeated here as (14a) and (14b):
(16) a. Yo me he comido toda la sopa para cenar (*a

cl cl.1sg.dat aux eat.pfv all the soup to dine to
mí mismo).
myself
‘I have eaten all the soup for dinner.’
b. El niño me le ha roto el libro (*a mí).

The proper approach to the argumental vs non argumental nature of ED is open

to discussion and depends on the particular theory on non-selected datives. For
instance, Franco & Huidobro (2008) assume that ED-Class II are arguments
whereas ED-Class I and III are adjuncts. Other novel approaches such as Bosse
(2015) take the (German) ED to be an applied argument with only not-at issue
meaning (or non-truth conditional meaning). Actually, in the particular per-
232 fernández
spective taken by this author in her (2015) work and in Bosse, Bruening and
Yamada (2012) among others, ED are non-selected arguments of particular
heads.
In this chapter, I will assume ED-Class II to be arguments in their more gen-
eralized meaning. In particular, I will propose that these ED-Class II behave
exactly as benefactives, experiencers and external possessors. I will follow
Odria (2017) in assuming the very same merging position for experiencers and
external possessors, i.e. the Specifier of an Applicative (Appl) head above v.
Moreover, I will claim that these ED-Class II merge in this position. Hence,
the properties attributed here to ED-Class II follow directly. In particular, as
observed by Franco & Huidobro (2008), it is worth noting that the presence of
the dative clitic doubling the DP is not optional but obligatory in Spanish inver-
sion predicate structures, a particular instance of ED-Class II. The designation
refers to structures with a subject like dative and a nominative theme with an
agreeing verbal form, as in examples (17). The absence of the dative clitic entails
ungrammaticality:
(17) a. A Juan *(le) falta dinero.

to Juan cl.3sg.dat lack money
‘Juan is missing some money.’
b. A Juan *(le) gustan los libros.

to Juan cl.3sg.dat like the books
‘Juan likes the books.’ (Franco & Huidobro 2008: 219)
With regards to these forms, Spanish and Basque look alike.
(18) a. Joni dirua falta zaio / *da.

to Juan lack money lack aux + 3sg.dat / aux
‘Jon is missing some money.’
b. Joni liburuak gustatzen zaizkio / *dira.

to Juan the books like aux + 3sg.dat / aux
‘Jon likes the books.’
Note that the Basque examples include an absolutive marked argument instead
of the corresponding nominative theme in Spanish counterparts. In order to
be grammatical, the presence of the dative clitic is obligatory, as pointed out by
Fernández (2010) Fernández & Ortiz de Urbina (2010) and recently Odria (2017)
among others. This is not surprising, as leaving aside north-eastern varieties
of Basque, datives clitics are obligatory elsewhere. However, in north-eastern

varieties of Basque where dative clitic dropping is attested, dative clitics are
obligatory with experiencers of psychological verbs of the gustatu ‘like’-type
as in (18b), and ED-Class II in particular (18a)—external possessors and bene-
factives behave alike.
Nevertheless, the designation inversion predicate structure does not seem
accurate in Basque. The word inversion implies that the word order in Spanish
examples (17a,b) is an inverted and thus, derived dative-nominative word order,
but their Basque dative-absolutive word order counterpart seems to be the
canonical and underived order. If ED-Class II are introduced by an Appl head
above v, as proposed here, then the dative-absolutive order does not seem to
be consequence of inversion. Under the same hypothesis, even Spanish exam-
ples (17a,b) show the order expected, if an Appl head above v is assumed to be
responsible for introducing the ED-Class II, as I claim here.
3.3 Ethical Datives-Class II and Clitic Clusters

In order to distinguish ethical datives from other datives, their ability to be
combined with any other dative clitic seems to be crucial (see Jouitteau &
Rezac 2007 and references therein). In particular, ED-Class III can co-occur
with selected dative goals (19a) and non-selected datives such as benefactives
(19b), ED-Class I (19c), ED Class II (19d) and external possessors (19e)—see also
Maldonado (1994: 254–255). Example (19a) is provided by Franco & Huidobro
(2008: 218) and examples (19b–d) by Maldonado (1994):
(19) a. A mi primo me le pegaron una paliza en una fiesta.

to my cousin ed cl.3sg.dat hit a beating at the party
Intended: ‘They beat my cousin up at the party (on me).’ (ED-Class III +
Goal)
b. Nacho me le hizo la tarea a Valeria.

Nacho ed cl.3sg.dat do the homework to Valeria
‘Nacho did the homework for Valeria.’ (ED-Class III + Benefactive)
c. Juan se me compró una moto.

Juan cl ed buy a motorbike
‘Juan bought a motorbike.’ (ED-Class I + ED-Class III)
d. Me les rompieron el libro a los niños.

ed cl.3pl.dat break a book to the kids
Intended: ‘They broke the book to the kids (on me).’ (ED-Class III + ED-
Class II)
234 fernández
e. Me le rompió la nariz.
ed cl.3sg.dat read break the nose
Intended: ‘He/she broke his/her nose (on me).’ (ED-Class III + External
possessor)
This is particularly interesting as the presence of two dative clitics shows that
ED-Class III, whatever they are, need to be separated from the rest, specifi-
cally, from ED-Class II (19d). On the other hand, ED-Class II, benefactives and
external possessors cannot be combined with one another. Assuming, as I do,
that all these non-selected datives are merged in Spec, Appl, as Odria (2017)
do for Basque experiencers, external possessors and causees,5 the inability of
ED-Class II to be combined with all these non-selected datives follows.
On the other hand, Basque can never combine two dative clitics, whether
they be selected or not (Fernández 2010), as there is a sole slot for a dative clitic
in the morphological configuration. Therefore, the Basque counterpart (20a) to
Perlmutter’s example in (13a) is ungrammatical. Besides, the ungrammaticality
affects both ditransitive auxiliary forms (20a) and intransitive auxiliary forms
(20b): two dative clitics never combine with each other. In (20a) the morpheme
-t can only be interpreted as an ergative and never as a dative.
(20) a. *Juanek neska horri bizitza hondatu

Juan.erg girl that.dat life ruin
diot
aux + 3sg.dat + 1sg.dat
Intended: ‘Juan ruined that girl’s life (on me).’
b. *Neska horri bizitza hondatu zaiot

girl that.dat life ruin aux + 3sg.dat + 1sg.dat
Intended: ‘That girl’s life is ruined (on me).’
As a consequence, this test does not apply in Basque. Nevertheless, the exam-
ples in (20) indirectly show that there is nothing comparable to ED-Class III in
Basque.
5 I am deliberately ignoring other double dative constructions, such as those including dif-
ferentially marked objects and goals or any other including derived datives (mainly goals),
exhaustively analyzed by Odria (2017). For an analysis of other double dative constructions
including an experiencer and an applied dative without verb, see in Berro & Fernández (to
appear).
3.4 Ethical Datives-Class II Involving Truth-Conditional Meaning:

Ability to be Questioned and Binding
Contrary to the properties discussed so far, Bosse, Bruening & Yamada (2012)
take an approach to non-selected datives that combines both syntactic and
semantic aspects. Basing on Potts (2005) among others, they distinguish truth-
conditional meaning (at-issue meaning in Potts’ designation) and non-truth
conditional meaning (not-at-issue meaning), which are represented on sepa-
rate tiers of meaning. As they show, only material on the truth conditional tier
can be questioned or bind a variable. Thus, in their typology of non-selected
datives, external possessors and benefactives are entirely truth conditional
whereas attitude holders and subject co-referential datives are entirely non-truth
conditional. Besides, in their proposal, affected experiencers may or may not
contribute truth-conditional meaning. Hence, this is a parameterized property
with yes value for German and Albanian and no value for Hebrew, as argued by
the authors.
ED can fit in three categories in their proposal: they can be either subject
co-referential, attitude holders or affected experiencers. As far as Basque and
Spanish are concerned, ED-Class I correspond to subject co-referential datives;
ED-Class II correspond, roughly speaking, to affected experiencers and ED-
Class III to attitude holders.
Let us explore the ability of ED-Class II to be questioned in the two languages
under discussion. The property of some non-selected datives, specifically exter-
nal possessors, being able to be questioned by a Wh-question and extracted is
also mentioned and discussed by Borer & Grodzinsky (1986: 1981–1982). Thus,
the test is not a new one, and it has been also explored by Arregi (2003) and
Fernández (2010) with regard to Basque external possessors. In Bosse, Bruen-
ing & Yamada (2012), if ED are part of the truth-conditional tier, then they can
be questioned by a Wh-phrase and extracted. As far as Basque is concerned,
ED-Class II seem to contribute truth-conditional meaning, as they can be ques-
tioned by Wh-questions and extracted.
(21) a. Nori lokartu zaio haurra?

who.dat fall asleep aux + 3sg.dat kid
Lit. ‘To whom has the kid fallen asleep?’
b. Nori apurtu dio liburua?

who.dat break aux + 3sg.dat book
Lit. ‘To whom has he/she broken the book?’
Spanish ED-Class II show the very same behavior.

236 fernández
(22) a. ¿A quién se le ha dormido el niño?

to who cl cl.3sg.dat aux fall asleep the kid
b. ¿A quién le ha roto el libro?

to who cl.3sg.dat aux break the book
Notice that ED-Class III—exemplified in (4b) and repeated here as (23a)—

show the opposite behavior, as they cannot be questioned. Hence, a hypothet-
ical Wh-question on the ED-Class III would be ungrammatical (23b).
(23) a. El niño me le ha roto el libro

b. *¿A quién le ha roto el libro?

to who cl.3sg.dat aux break the book
Intended: To whom has he/she broken the book to someone else?
Notice that (23b) would be grammatical when asking about the one affected
by the book being broken, but not about the one involved in the book being
broken to him/her. It is worth pointing out that (23b) would be perfect if the
clitic le were coreferential with ‘to whom’.
Besides, ED-Class II can be a quantifier that binds a truth conditional ele-
ment as a variable. Thus, ED-Class II contribute truth conditional meaning with
regard to binding too. Basque examples show the pattern:
(24) a. Ama guztiei beren haurrak lokartu zaizkie.

mother every.dat their kids fall asleep aux + 3pl.dat
‘Kids have fallen asleep to every mother.’
b. Haur guztiei apurtu diete beren liburua.

child every.dat break pfv aux + 3pl.dat their book
‘They have broken the book to every children.’
Spanish shows no difference regarding binding.
(25) a. A todas las madres se les han dormido sus niños.

to every mother cl cl.3pl.dat aux fall asleep their kids
b. A todos los niños les han roto su cuento.

to every kid cl.3pl.dat aux break their book
If this is on the right track, then the value for both Basque and Spanish is yes,
ED-Class II contribute truth conditional meaning.
3.5 Ethical Datives-Class II and Non-truth Conditional Meaning:

Negation
Till now, I have shown, following Bosse, Bruening & Yamada (2012), that Basque
and Spanish ED-Class II contribute truth conditional meaning. This is not sur-
prising, as these datives are arguments (within Inflection Phrase) and behave
as such. However, affected experiencers, equivalent to our ED-Class II, are
also claimed to contribute non-truth conditional meaning, as their meaning
projects beyond negation and yes-no questions and they cannot be directly
negated. In this regard, Spanish and Basque show an unexpected pattern in
these authors’ proposal, as ED-Class II do not contribute non-truth conditional
meaning in either Spanish or Basque.
Let us start with Basque and Spanish negative examples, similar to those in
Bosse & Bruening (2010).
(26) a. Joni ez zitzaion haurra lokartu (Aneri bai).

Jon.dat not aux + 1sg.dat kid fell asleep Ane.dat yes
‘The kid didn’t fall asleep.’
b. A Jon no se le ha dormido el niño (A Ane sí).

to Jon not cl cl.3sg.dat aux fell asleep the kid to Ane yes
As argued by Bosse & Bruening (2010) and Bosse, Bruening & Yamada (2012),
the experience itself cannot be negated: it is the whole verbal event, but not
the experience, that is negated. However, at least in Basque and Spanish, ED-
Class II can be negated. Hence, the meaning of the experience does not go
beyond negation (against what has been observed by these authors for lan-
guages such as Albanian and German among other languages). If this is so,
ED-Class II do not contribute non-truth conditional meaning. Hence, it does
not confirm the yes value attributed to affected experiencers by the authors
with regard to this property. This unexpected value is also attested in ditransi-
tive structures such as (27a) and (27b), negative counterparts of examples (1b)
and (3b). Once again, it seems that not only the verbal event but also the expe-
rience itself can be negated:
238 fernández
(27) a. Haurrak ez dit liburua apurtu.

kid.erg not aux + 1sg.dat book break
‘The kid hasn’t broken a book (but if s/he had, it would have affected
me).’
b. El niño no me ha roto el libro.

the kid not cl.1sg.dat aux break the book
The reason for that might be that, contrary to Maldonado (1994) and this chap-
ter, the datives in (1b) and (3b) and their corresponding negation in (26a) and
(26b) do not show an affected experiencer but a benefactive—assuming that
there is a distinction between affected experiencers and benefactives, some-
thing not so obvious attending to our results (see Table 7.1 below). If they
were benefactives and not affected experiencers, then, as in other benefactives
analyzed by Bosse & Bruening (2010), the benefactive relation itself could be
negated, as benefactives are entirely truth conditional.
(28) a. Haurrak liburua apurtu du, baina ez dit niri

kid.erg book break aux but not aux + 1sg.dat me.dat
apurtu.
break
‘The kid has broken a book, but not for me.’
b. El niño ha roto el libro, pero no me lo

the kid aux break the book but not cl.1sg.dat cl.3sg.acc
ha roto a mí.
aux break to me
Nevertheless, I see no clear reason to treat these examples as not showing an

ED. Although much research is needed in principle, the contribution to non-
truth conditional meaning does not seem to apply in Basque and Spanish as
far as ED-Class II are concerned.
4 Discussion and Conclusions
So far, I have presented and analyzed some significant properties of non-

selected datives in general and ED in particular. These properties and their
yes/no values are represented in Table 7.1. This table has been inspired by those
in Bosse & Bruening (2010) and Bosse, Bruening & Yamada (2012), but has been
table 7.1 Typology of non-selected datives inspired by Bosse & Bruening (2010), extended
and modified
ED-Class I ED-Class III ED-Class II Benefactives External

(allocutives) possessors
Person restriction no yes no no no

Doubling no no yes yes yes
Clitic Cluster yes yes no no no
Truth conditional no no yes yes yes
meaning
Non-truth condi- yes yes no no no
tional meaning
Languages Spanish Spanish, Basque, Spanish Basque, Basque,
(Basque) Spanish Spanish
Other designations Subject co-referential, Attitude Affected experiencers,
Personal datives holders Affected datives
extended and modified mainly, but not only, on the basis on Franco & Huido-
bro (2008: 223). Besides, I have maintained the general designation along with
its particular classes, ED-Class I, III and II, instead of subject co-referential, atti-
tude holders and affected experiencers respectively.
The table also includes allocutives, which share the values of the properties
attributed to ED-Class II, although this only covers a part of the intricate nature
of allocutives not necessarily involving dative-like clitics.
One more property could be added to the table, that is, subject co-reference
attested in ED-Class I with a yes value, and with a no value in the remaining
cases. Nevertheless, person restriction distinguishes both ED-Class I and III.
Thus, it has not been included in the table.
Besides, there is a striking (and undesirable) scenario with regard to ED-
Class II, benefactives and external possessors: they show exactly the same
properties and cannot be distinguished from one another. The property ‘NP
must be sentient’ analyzed by Bosse, Bruening & Yamada (2012) could in prin-
ciple distinguish ED-Class II and benefactives, being ED-Class II sentient and
benefactives not necessarily. Unfortunately, ED-Class II do not seem to be nec-
essarily sentient in either Basque or Spanish, as shown in the following exam-
ples:
(29) a. Aitita zenari hilarria apurtu diote.

grandfather late.dat gravestone break aux + 1sg.dat
‘They have broken the gravestone to the late grandfather.’
240 fernández
b. Le han roto la lápida al difunto abuelo.

cl.3sg.dat aux break gravestone to the late grandfather
As a consequence, ED-Class II and benefactives share exactly the same proper-

ties in Basque and Spanish, and cannot be distinguished from each other, since
both choose the value no for the property NP must be sentient and the same
no value for the property Non-truth conditional meaning (vs Bosse, Bruening &
Yamada 2012).
On the other hand, sharing the same properties discussed here, the claim
that ED-Class II merge in [Spec Appl], as proposed in Odria (2017) for Basque
experiencers and external possessors, seems to be on the right track. More
research is needed in order to distinguish syntactically ED-Class II, benefactives
and external possessors from one another. Their merging position can explain
their uniform behavior with regard to the properties mentioned in this chap-
ter, but still more syntactic tests are needed in order to explain their divergent
nature. Bosse, Bruening & Yamada (2012) is a good attempt to do so, although
unfortunately it does not solve some of the main issues, at least with regards to
Basque and Spain.
Acknowledgements
The author would like to thank Jon Ortiz de Urbina for his valuable comments
on a previous version of this paper and to Solveig Bosse for some feedback.
All errors remain my own. The author also acknowledges the research funding
received from the Basque Government (IT665–13) and the Spanish Ministry
of Economy and Competitiveness (FFI2014–51878-P). The research leading to
these results has also received funding from the European Union’s Seventh
Framework Program for research, technological development and demonstra-
tion under grant agreement no. 613465.
References
Arregi, Eurídice. 2003. “On possessor raising effects in Basque”. Ms., UPV/EHU.
Authier, Marc & Lisa Reed. 1992. “On the syntactic status of French affected datives”.
The Linguistic Review 9: 295–311.
Berro, Ane & Beatriz Fernández. To appear. “Applicatives without verbs”. Natural Lan-
guage and Linguistic Theory (NLLT).
Bosse, Solveig & Benjamin Bruening. 2010. “Benefactive versus experiencer datives”. In
Mary B. Washburn, Katherine McKinney-Bock, Erika Varis, Ann Sawyer & Barbara
Tomaszewicz. (eds.), Proceedings of 28th WCCFL. 69–77.
Bosse, Solveig, Benjamin Bruening & Masahiro Yamada. 2012. “Affected experiencers”.
Natural Language and Linguistic Theory (NLLT), 30(4): 1185–1230.
Bosse, Solvig. 2015. Applicative arguments. Peter Lang.
Christian, Donna. 1991. “The personal dative in Appalachian English”. In Peter Trudgill
& J.K. Chambers (eds.), Dialects of English. London: Longman. 11–19.
Borer, Hagit & Yosef Grodzinsky. 1986. “Syntactic cliticization and lexical cliticization:
The Case of Hebrew Dative Clitics”. In Hagit Borer (ed.), Syntax and Semantics 19.
New York: Academic Press. 175–215.
Fernández, Beatriz. 2010. “Goi datiboak eta are goragokoak euskaraz: jabe datiboak
eta datibo hunkituak vs. datibo etikoak”. Gogoa 10, 1–2. San Sebastián/Donostia:
UPV/EHU. 1–20.
Fernández, Beatriz & Jon Ortiz de Urbina. 2010. Datiboa hiztegian. Bilbo: UPV/EHU.
Franco, Jon & Huidobro, Susana. 2008. “Ethical Datives, Clitic doubling and the Theory
of pro” Joyce Bruhn de Garavito & Elena Valenzuela (eds.), Selected Proceedings of
the 10th Hispanic linguistics Symposium. Somerville: Cascadilla Proceedings Project.
215–224.
Gutierrez Ordoñez, Salvador. 1999. “Los dativos”. In Ignacio Bosque & Violeta Demonte
(eds.), Gramática descriptiva de la lengua española II. Madrid: Real Academia Espa-
ñola / Espasa Calpe. 1855–1930.
Horn, Laurence R. 2008. “I love me some him: The landscape of non-argument datives”.
In Olivier Bonami & Patricia Cabredo Hofherr (eds.), Empirical Issues in Syntax
and Semantics 7, CSSP, Paris, 169–192. [On-line publication, http://www.cssp.cnrs.fr/
eiss7].
Horn, Laurence R. 2013. “I love me some datives: Expresive meaning, free datives, and
F-implicature”. In Daniel Gutzmann & Hans M. Gärtner (eds.), Beyond Expressives:
Explorations in Use-Conditional Meaning. Leiden/Boston: Brill. 153–201.
Jaeggli, Osvaldo A. 1982. Topics in Romance Syntax. Dordrecht: Foris.
Jaeggli, Osvaldo A. 1986. “Three issues in the theory of clitics: case, doubled NP s and
extraction”. In Hagit Borer (ed.), Syntax and Semantics 19. New York: Academic Press.
15–42.
Jouitteau, Mélanie & Milan Rezac. 2007. “The French ethical dative, 13 syntactic tests”.
Bucharest Working Papers in Linguistics, IX (1): 97–108.
Maldonado, Ricardo. 1994. “Dativos de interés sin intereses”. Verbo y estructuras frási-
cas [Actas do IV Coloquio Internacional de Lingüística Hispánica]. Porto, Facultad
de Letras. 241–264.
Miguel, Elena de & Marina Fernández Lagunilla. 2000. “El operador aspectual se”.
Revista Española de Lingüística 30 (1). 13–43.
Odria, Ane. 2017. The syntax of Differential Object Marking and datives in Basque. Ph.D.
thesis, UPV/EHU.
242 fernández
Oyharçabal, Beñat. 1993. “Verb agreement with non-arguments: On allocutive agree-

ment”. In José Ignacio Hualde & Jon Ortiz de Urbina (eds.), Generative studies in
Basque linguistics. Amsterdam: John Benjamins. 89–114.
Perlmutter, David. 1971. Deep and surface constraints in syntax. New York: Holt, Reinhart
and Winston.
Potts, Chris. 2005. “The logic of conventional implicatures”. Oxford: Oxford University
Press.
Pylkkänen, Liina. 2008 [2002]. Introducing arguments. Cambridge/London: the MIT
Press.
Strozer, Judith. 1978. “On the so-called ‘dative of interest’ ”. Hispania 61: 117–123.
Webelhuth, Gert & Clare Dannenberg. 2006. “Southern American personal datives: The
theoretical significance of syntactic variation”. American Speech 81: 31–55.
chapter 8
Differential Object Marking in Basque and Spanish

Dialects
Ane Odria
1 Introduction
Many southwestern varieties of Basque display Differential Object Marking

(DOM) (Bossong 1985, 1991; Aissen 2003). In these varieties, human and (gener-
ally) definite objects—especially first and second person pronouns—tend to
bear dative marking, instead of the canonical absolutive expected in an erga-
tive language like Basque (Fernández & Rezac 2010, 2016; Mounole 2012; Odria
2012, 2014, 2017, to appear; Rodríguez-Ordóñez 2013, 2016).1 Southwestern vari-
eties of Basque are those which are in contact with Spanish, and thus, also with
Spanish DOM—i.e., a-marking (Pensado 1995, Torrego 1998, Leonetti 2004, a.o.).
The Basque varieties situated in the French speaking area are referred as north-
eastern varieties and, as happens in Standard Basque, they show no DOM, since
only the canonical absolutive marking is available for them.
This paper compares the syntactic behavior of Basque and Spanish DOM,
paying special attention to the Spanish variety spoken in the Basque speak-
ing area—i.e., Basque Spanish (Landa 1995). Additionally, the paper distin-
guishes the syntax of DOM from dative objects in bivalent unergative predi-
cates. It argues that the distinction between these kinds of objects is not only
configurational—as has been argued in previous work—, but also categorical.
Basque and Spanish DOM share significant commonalities. The differential
marking is determined by animacy and specificity in both of them, and it is
morphologically identical to the dative marking of indirect objects. Besides,
Basque Spanish is particular in showing DOM both in the nominal—i.e., a-
marking—and in the clitic system of the verbal complex—i.e., leísmo (Franco
1993, Landa 1995, Fernández-Ordóñez 1994, 1999). This makes Basque Spanish
even closer to Basque DOM, as in Basque the differential marking is attested not
only in the nominal, but also in the finite verbal form.
1 DOM is optional for many southwestern Basque speakers and its actual distribution is subject
to dialectal as well as idiolectal variation. Besides, its appearance is highly reduced with third
person objects.

244 odria
In both Basque and Spanish, DOM objects show the same morphology as
the dative objects of bivalent unergative predicates of the lagundu (Basque) /
ayudar (Spanish) ‘accompany, help’ type: (i) nominal as well as verbal dative
marking in Basque, (ii) a-marking in Spanish, and (iii) cliticization with le(s) in
(Basque) Spanish. However, in the case of bivalent unergatives, the marking of
the object happens to be independent from factors like animacy and specificity
and, semantically speaking, the object patterns more akin to the goal dative of
ditransitive predicates. As a consequence, it has been argued that, contrary to
the direct object configuration of DOM objects, the dative objects in bivalent
unergatives have an indirect object configuration—see Fernández & Ortiz de
Urbina (2010, 2012) and Ortiz de Urbina & Fernández (2016) for Basque, Tor-
rego (2010) and Fábregas (2013) for Spanish, and Pineda (2016) for Romance
languages in general.
This paper makes a further step in the syntactic distinction between Basque
and Spanish DOM on the one hand, and dative objects in bivalent unerga-
tives on the other. It claims that apart from their syntactic configuration, these
objects are also distinguished by their categorical status. While DOM objects
pattern with causee, experiencer and possessor datives in exhibiting a DP
syntactic category, the datives in bivalent unergatives behave more akin to
goal datives and thus show a PP-like behavior. This contrast is evidenced by
the licensing of depictive secondary predication, which—along with the rest
of DP s—is allowed with DOM datives, but—as happens with PP s and goal
datives—is generally rejected with datives in bivalent unergatives (Odria 2015,
2017).
The paper is structured as follows. In section 2, I lay out the main properties
of the DOM found in Basque (section 2.1), in Spanish (section 2.2) and in Basque
Spanish (section 2.3). Section 3 compares DOM objects with dative objects in
bivalent unergatives in both Basque (section 3.3) and Spanish (section 3.4). Sec-
tion 4 deepens on the syntactic distinction exhibited by these types of objects.
Based on the licensing of depictive secondary predication, it argues that, while
DOM objects are DP s categorically, the datives in bivalent unergatives show a
PP-like—i.e., a dual DP/PP—character. This is again proved with both Basque
(section 4.1) and Spanish (section 4.2) data. Finally, chapter 5 closes the paper
by summing up the main conclusions.
differential object marking in basque and spanish dialects 245
2 DOM in Basque, Spanish and Basque Spanish
2.1 Basque DOM

2.1.1 Canonical vs. Non-canonical Object Marking
In Standard as well as in other varieties of Basque, the subject of a transitive
predicate is marked with ergative case (-k) and the object with absolutive case
(-ø). Likewise, as illustrated in (1), the finite verbal form agrees with the two
arguments by means of their respective markers: -zu for the second person
ergative and n- for the first person singular absolutive.
(1) Zu-k ni ikusi nauzu.

you-e I.a see aux[1sg.a-2sg.e]2
‘You have seen me.’
In contrast to the canonical configuration, southwestern Basque varieties dis-

playing DOM tend to mark human and (generally) definite objects dative, as in
(2).
(2) Zu-k ni-ri ikusi didazu.

you-e I-d see aux[1sg.d-2sg.e]
In (2), the subject is marked with ergative case (-k), but the object appears with
dative case (-(r)i) rather than with the absolutive. Besides, the two arguments
are coded by their respective markers in the finite verbal form: the second per-
son ergative by -zu and the first person singular dative by -da-.
Contrary to what happens in the DOM of widely studied accusative lan-
guages, in Basque the differential marking co-appears with the ergative mark-
ing of the transitive subject, leading thereby to a configuration with two argu-
ments bearing a morphologically marked case: the ergative and the dative.3
This is an unexpected pattern from a typological point of view, because the
canonical configuration in transitive clauses is supposed to bear a single
2 The following abbreviations will be used in the text: e (ergative), d (dative), a (absolutive),
ine (inesive), acc (accusative), all (allative), dest (destinative), caus (causative), fut
(future), prog (progressive), nom (nominalizer), comp (complementizer), m (masculine),
f (feminine), aux (auxiliary verb), sg (singular), pl (plural), 1/2/3 (first, second and third per-
son).
3 Basque is not an exception in this regard. Other languages like Hindi also display DOM
together with an ergative subject (Mahajan 1990, Bhatt & Anagnostopoulou 1996).
246 odria
marked case, either in the subject—as happens in ergative languages—or in

the object—as in accusative languages.
In addition, as occurs in many DOM languages, in Basque the differen-
tial marking happens to be morphologically to the dative marking in indirect
objects. This is a common pattern cross-linguistically (Bossong 1991: 154, 157–
158; Aissen 2003: 446), as other DOM languages governed by both animacy and
specificity show the same behavior as well.4 The morphological identity shared
by DOM and indirect objects can be observed by comparing the example in (2)
with that in (3). In (3) we see that the indirect object of a ditransitive predicate
is marked dative in Basque.
(3) Zu-k ni-ri liburua eman didazu.

you-e I-d book.a give aux[3sg.a-1sg.d-2sg.e]
‘You have given me the book.’
In (3), the subject is marked with ergative case (-k), the direct object with the
absolutive (-ø) and the indirect object with the dative (-(r)i). Likewise, the three
arguments are cross-referenced by the finite verbal form, which shows the same
shape as that in (2): the second person ergative is coded by -zu and the first per-
son singular dative by -da-.5
Despite being morphologically identical to indirect objects, DOM objects
share the same thematic as well as argumental relationships with absolutive
direct objects. Hence, it seems reasonable to assume that DOM objects generate
in the same syntactic position as canonical absolutives—i.e., in the comple-
ment position of V. This is common ground in the literature on the topic, given
that, in spite of the differential marking, DOM objects are considered to be
direct objects configurationally—see, among others, Bárany (2018).
2.1.2 The Main Factors Conditioning Basque DOM

As happens in many DOM languages, in Basque the dative marking of the object
is favored by both animacy and specificity, as the object has to be human and
(generally) definite in order to be differentially marked (Fernández & Rezac
2010, 2016; Mounole 2012; Odria 2012, 2014, 2017; Rodríguez-Ordóñez 2013,
2016).
4 With few exceptions, Romance and Semitic languages are clear examples of this typolog-
ical tendency, and so are other modern Indo-Aryan languages like Hindi or Punjabi and
Amerindian languages like Guaraní and Aymara (Bossong 1991: 157).
5 In Basque the third person absolutive is not overtly marked in the verbal form.
On the one hand, dative marking is only accepted with human-referring

objects. This is illustrated in the examples in (4), where only the human object
Jon (4a) is able to display dative marking.6
(4) a. Jon-i ikusi diot.

Jon-d see aux[3sg.d-1sg.e]
‘I have seen Jon.’
b. *Telebista-ri ikusi diot.

television-d see aux[3sg.d-1sg.e]
‘I have seen the television.’
c. Telebista ikusi dut.

television.a see aux[3sg.a-1sg.e]
‘I have seen the television.’
Besides, even though the main cutting point between DOM and non-DOM
objects is determined by humanness, person deserves special attention too.
In fact, many DOM speakers make a further distinction within human objects
and distinguish between first and second person objects on the one hand, and
third person objects on the other (Fernández & Rezac 2010, 2016; Mounole 2012;
Odria 2012, 2014, 2017; Rodríguez-Ordóñez 2016). As a consequence, it is easier
to find dative marking with first and second person objects than with third per-
son. This distinction is realized in different manners. For some speakers, only
first and second person objects can carry dative marking with a verb like ikusi
‘see’. As shown in (5) and (6), this is the case of a speaker from Zumaia (Central
Basque).7
(5) a. Kali-an ikusiyazu.

street-ine see.aux[1sg.d-2sg.e]
‘You have seen me in the street.’
6 Some DOM speakers consider examples like (4a) quite marginal, and add that the dative mark-
ing of a third person object would be more natural with other verbs. This may be due to the
social stigmatization of the DOM phenomenon with the verb ikusi ‘see’, as it is one of the most
corrected errors at school as well as in Basque academies.
7 The dialectal data presented in this paper were collected in the fieldwork carried out by
myself in previous work—see especially Odria (2017).
248 odria
b. Kali-an ikusizut.
street-ine see.aux[2sg.d-1sg.e]
‘I have seen you in the street.’
(6) a. *Jon-i kali-an ikusiyot.

Jon-d street-ine see.aux[3sg.d-1sg.e]
‘I have seen Jon in the street.’
b. Jon kali-an ikusiet.

Jon.a street-ine see.aux[3sg.a-1sg.e]
‘I have seen Jon in the street.’
For other speakers, the distinction between first and second vs. third per-
son objects is not so rigid. A speaker from Larrabetzu (Western Basque), for
instance, finds DOM with first (7a) and second (7b) person very natural, but
prefers the absolutive for the third one—although in present tense the dative
is acceptable in this case too (8).
(7) a. Ez dostezu ikusi?

not aux[1sg.d-2sg.e] see
‘Haven’t you seen me?’
b. Ikusi dotsut kontzertu-en.

see aux[2sg.d-1sg.e] concert-ine
‘I have seen you in the concert.’
(8) a. Leire-ri ikusi dotset kotxe barri-en.

Leire-d see aux[3sg.d-1sg.e] car new-ine
‘I have seen Leire in the new car.’
b. Leire ikusi dot kotxe barri-en.

Leire.a see aux[3sg.a-1sg.e] car new-ine
‘I have seen Leire in the new car.’
The different behavior attested between first and second vs. third person
objects is also mentioned in some descriptions of dialectal varieties. Hualde,
Elordieta & Elordieta (1994: 125–127), for instance, notice that in Lekeitio
Basque (a Western variety), DOM occurs more frequently with first and sec-
ond person than with the third. Likewise, Ibarra (1995: 427) reports that in
Ultzama Basque (a Navarrese variety), first and second person objects are gen-
erally marked dative, while DOM with third person objects is just optional.
Furthermore, it is important to note that, generally speaking, the DOM exam-
ples provided by dialectal grammars involve usually objects of first and second
person.
Along with animacy—or even person—, specificity is also an important
conditioning of Basque DOM. As argued by Mounole (2012), only human
objects with a referential interpretation are able to display the differential
marking—see also Fernández & Rezac (2010, 2016) and Odria (2012, 2014, 2017).
Mounole (2012: 368–369) shows that DOM is mostly incompatible with the
indefinite determiner bat ‘a’, the indefinite quantifier asko ‘many’ or the indef-
inite personal pronoun inor ‘nobody’. This is illustrated by the examples in
(9), (10) and (11), provided by a speaker from Itsasondo (Central Basque). This
speaker admits DOM with first, second and—less commonly—third person
objects. However, in order to bear dative marking with the verb ikusi ‘see’, those
of third person have to be not only human, but also definite. As a consequence,
as happens with many other speakers, human objects containing indefinite
determiners like bat ‘a’ (9), asko ‘many’ (10) and inor ‘nobody’ (11) tend to bear
absolutive marking.
(9) a. Neska bat ikusi det.

girl one.a see aux[3sg.a-1sg.e]
‘I have seen a girl.’
b. *Neska bat-i ikusi diot.

girl one-d see aux[3sg.d-1sg.e]
‘I have seen a girl.’
(10) a. Neska asko ikusittut.

girl many.a see.aux[3pl.a-1sg.e]
‘I have seen many girls.’
b. *Neska asko-i ikusi diet.

girl many-d see aux[3pl.d-1sg.e]
‘I have seen many girls.’
(11) a. Ez det inor ikusi.

not aux[3sg.a-1sg.e] anyone.a see
‘I haven’t seen anyone.’
250 odria
b. *Ez diot inor-ri ikusi.

not aux[3sg.d-1sg.e] anyone-d see
‘I haven’t seen anyone.’
In her description of DOM in Tolosa Basque (a Central variety), Mounole (2012;

369) observes that DOM can additionally distinguish between specific and non-
specific interpretations, arguing that indefinite nominals bearing dative mark-
ing entail a specific interpretation when combined with the verb bilatu ‘look
for’.8 However, I have found no speaker admitting the dative marking with an
indefinite nominal in that context; an example like (12a) is considered ungram-
matical by all my consultants. The indefinite object can only carry absolutive
marking, as in (12b)—the same results are obtained by Arraztio (2010) too in
Araitz-Betelu Basque (a Central variety).9
(12) a. *Idazkari bat-i bila-tzen diot.

secretary one-d look for-prog aux[3sg.d-1sg.e]
‘I look for a secretary.’
b. Idazkari bat bila-tzen dut.

secretary one.a look for-prog aux[3sg.a-1sg.e]
‘I look for a secretary.’
Contrary to Mounole (2012), I thus conclude that instead of specificity per se,
the main factor conditioning the marking of the object along with animacy is in
fact definiteness. Basque DOM is then more restrictive than Romance languages
like Spanish, where specificity rather than definiteness seems to be relevant.
All in all, despite its significant impact, it should be noted that—as happens
in many Romance languages (Bossong 1991: 160–161)—, in Basque definite-
8 Instead of ‘specificity’, Mounole uses the term ‘referentiality’. However, for the sake of sim-
plicity, I continue using the notion ‘specificity’, assuming that both share the same meaning.
9 At least in some southwestern varieties, the sentence in (12b) would sound more natural with
the verb behar ‘need’, which—as happens with bilatu ‘look for’—requires absolutive marking
with indefinite objects (i). In these varieties, bilatzen dut ‘look for.prog aux’ would occur in
a verbal periphrasis formed with the verb ibili ‘be’ (ii):
(i) Idazkari bat/*-i behar dut / *diot.
secretary one.a/bat-d need aux[3sg.a-1sg.e] / aux[3sg.d-1sg.e]
‘I need a secretary.’
(ii) Idazkari bat-en bila nabil.
secretary one-ine look for be.aux[1sg.a]
‘I’m looking for a secretary.’
ness does not condition the marking of the object as rigidly as animacy does.
Whereas DOM is never attested with non-human or inanimate objects, in some
cases indefinite objects can occasionally occur with the differential marking.
This groups the DOM found in Basque with Spanish DOM, where animacy is also
known to be more influential than specificity (Brugger & Brugé 1996, Leonetti
2004).
Besides, although their influence is not as generalized as that caused by the
referential properties of the object, clausal properties like tense and finiteness
affect the marking of the object as well, given that DOM is sometimes reduced
in present tense as well as non-finite contexts (Fernández & Rezac 2010, 2016;
Odria 2017). This is a particular aspect of Basque DOM, as cross-linguistically
clausal factors like tense and finiteness do rarely condition the distribution of
DOM. Be that as it may, the influence of these factors is far from being a system-
atic pattern and seems to be related to morphological facts. Likewise, it should
be noted that the actual distribution of Basque DOM is also determined by the
nature of the verb, since the differential marking is more common with some
verbs than with others (Mounole 2012, Rodriguez-Ordóñez 2016, Odria 2017).
2.2 Spanish a-marking

In Spanish, animate and specific objects are marked by the preposition a,
whereas inanimate and non-specific ones are left unmarked—see, among
many others, Pensado (1995), Brugè & Brugger (1996), Torrego (1998), Leonetti
(2004), Rodríguez-Mondoñedo (2007), Fábregas (2013) and Ormazabal &
Romero (2013ab). This is illustrated in the examples in (13), provided by Orma-
zabal & Romero (2013a: 222). In (13a), the animate and specific object la niña
‘the child’ bears a-marking. On the contrary, although specific, the inanimate
el libro ‘the book’ is unable to be preceded by a in (13b), and the same thing
happens with the animate generic object niñas ‘children’ in (13c).
(13) a. He encontrado *(a) la niña.

have found.1sg dom the child.f
‘I have found the girl.’
b. He encontrado (*a) el libro.

have found.1sg dom the book
‘I have found the book.’
c. He encontrado (*a) niñas.

have found.1sg dom children.f
‘I have found girls.’
252 odria
As in Basque, Spanish DOM involves a marker that is morphologically identi-

cal to that found with indirect objects. This is shown in (14), where the indirect
object is preceded by the same a-marker carried by the animate and specific
direct object in (13a).
(14) Le he comprado un libro a la niña.

3sg.d have bought.1sg one book to the child.f
‘I have bought the child a book.’
Despite their similarities, Spanish behaves different to Basque in four main

aspects. First, being an accusative language, DOM leads to a typologically
expected configuration, with only one of the two arguments of a transitive
construction being overtly marked. Second, it has no person distinction when
assigning the differential marking—i.e., in Spanish a-marking is equally avail-
able with first, second and third person objects. Third, the differential marking
appears only in the noun phrase, and not in the clitic system of the verbal com-
plex. And fourth, clausal properties like tense and finiteness do not affect the
distribution of DOM, given that there is no contrast between present vs. past
tenses and inflected vs. non-inflected clauses.
2.3 Basque Spanish leísmo

In contrast to Standard as well as other varieties of Spanish, Basque Spanish
exhibits a pattern which makes its DOM closer to that found in southwestern
Basque, as the differential marking is not only attested in the nominal, but also
in the clitic system of the verbal complex.
Basque Spanish is a leísta dialect: the dative clitic le(s) is used instead of the
accusative lo(s)/la(s) when referring to an animate object.10 This is exemplified
in (15) and (16), where the inanimate object el coche ‘the car’ is coded by the
accusative clitic lo (15), whereas the dative le pronominalizes both the animate
masculine object el chico ‘the boy’ and the feminine la chica ‘the girl’ (16).11
10 Contrary to what happens in Basque Spanish, the leísmo in Standard Spanish refers to the
coding of masculine, singular (usually personal) direct objects by the dative clitic le, sub-
stituting this way the accusative lo—and only exceptionally la (Fernández-Ordóñez 1994:
7, 1999: 1319; Landa 1995: 152). Consequently, the Basque Spanish leísmo is also referred as
real leísmo (Fernández-Ordóñez 1999) or animated leísmo (Urrutia-Cárdenas 1995, 2003).
11 Urrutia-Cárdenas (2003: 292) mentions that the use of le(s) can also be extended to inan-
imate objects. Nevertheless, I agree with Landa (1995: 8) in claiming that such pattern is
hardly acceptable among Basque Spanish speakers.
(15) a. Vi el coche aparcado.

saw.1sg the car parked
‘I saw the car parked.’
b. Lo / *le vi aparcado.
3sg.acc / 3sg.d saw.1sg parked
‘I saw it parked.’
(16) a. Vi al chico / a la chica contento/-a.

saw.1sg dom the boy / dom the girl happy.m/-f
‘I saw the boy/girl happy.’
b. Le vi contento/-a.
3sg.d saw.1sg happy.m/-f
‘I saw him/her happy.’
Moreover, the leísmo in Basque Spanish is special in allowing clitic doubling

with a-marked objects in their canonical object position (Franco 1993, Landa
1995, Fernández-Ordóñez 1994, 1999). Consider now the examples in (17).12
(17) a. Le he llevado al niño a casa.

3sg.d have carried.1sg dom the child.m to house
‘I have carried the child at home.’
b. Le conozco a Jon desde pequeño.

3sg.d know.1sg dom Jon since child
‘I know Jon since he was a child.’
As noted by Landa (1995: 117), in Basque Spanish clitic doubling depends on

a-marking. That is, those objects that are doubled by the le(s) clitic must be
a-marked, which means that the objects allowing clitic doubling are also ani-
mate and specific—or more precisely, animate and presuppositional, as argued
by Franco (1993) and Landa (1995).13 Take the sentences in (18). In (18a), the
12 This is rejected in the Standard leísmo, where the same object would only be differentially
marked in the nominal by means of a-marking.
13 Landa (1995: 162) claims that, instead of specificity, in Basque Spanish clitic doubling with
le(s) is constrained by presuppositionality and gives the example in (i) to support her
hypothesis.
254 odria
object la niña ‘the child’ is a-marked, and the doubling with the dative clitic
le is acceptable. On the contrary, in (18b), the object el libro ‘the book’ is inan-
imate, and thus, non-a-marked. As a consequence, doubling it with a dative
or accusative clitic is ruled out. The same is true for the animate non-specific
object camareros ‘waiters’ in (18c).
(18) a. Le he visto a la niña.

3sg.d have seen.1sg dom the child.f
‘I have seen the child.’
b. (*Le / lo) he visto el libro.

3sg.d / 3sg.acc have seen.1sg the book
‘I have seen the book.’
c. (*Les / los) necesito camareros que sepan inglés.

3pl.d / 3pl.acc need.1sg waiters that know English
‘I need waiters that know English.’
Note that this is also the case in Basque DOM, as only those objects bearing
dative case are able to be coded by dative marking in the agreement complex
(19).
(19) Zu-k ni-ri / *ni ikusi didazu.

you-e I-d / I.a see aux[1sg.d-2sg.e]
(i) Los burócratas le llegan a cansar a uno.

the bureaucrats 3sg.d manage to tire.3pl dom one
‘Bureaucrats end up boring you.’
The object bearing clitic doubling in (i) cannot be interpreted as specific –uno is a generic
animate pronoun, and the sentence is still acceptable in Basque Spanish. Based on this
fact, Landa concludes that clitic doubling in Basque Spanish is only possible with animate
and presuppositional objects. This would also explain the ungrammaticality of examples
like (ii) and (iii), which contain non-presuppositional objects (Landa 1995: 165).
(ii) ???Les vi a bastantes hombres
3pl.d saw.1sg dom enough men
‘I saw quite a few men.’
(iii) *Les necesito a seis hombres
3pl.d need.1sg dom six men
‘I need six men.’
Summing up, in spite of their differences, we see that Basque DOM and
(Basque) Spanish DOM coincide in a great extent: (i) DOM is triggered by ani-
macy and specificity, (ii) the differential marking is morphologically identical
to the marking of the indirect object, and (iii) together with the noun phrase,
the differential marking is also realized in the finite verbal form. These shared
aspects are relevant because, as mentioned in section 1, DOM in Basque only
exists in the dialects that are in contact with Spanish.
In the literature on the topic, some authors have suggested that Basque
influences the leísmo in Basque Spanish. As Basque makes no gender dis-
tinction on direct objects, authors like Fernández-Ordóñez (1994, 1999) and
Landa (1995) have suggested that this could have reinforced the use of the
dative le(s) instead of lo(s) for masculine objects and la(s) for feminine ones.
Be that as it may, the so-called contact situation has also been analyzed in
the opposite direction. Mounole (2012), for instance, attributes the spread
of Basque DOM to the influence of Spanish, claiming that Basque DOM is
at least reinforced by the contact with Spanish. Other authors like Austin
(2006, 2015) or Rodríguez-Ordóñez (2013, 2016) go even further and claim
that Basque DOM is not only reinforced, but also induced by the contact with
Spanish—and more precisely, by the leísta Basque Spanish. This is, thus, an
open issue.
3 DOM vs. Datives in Bivalent Unergative Predicates
3.1 Basque DOM vs. Datives in Bivalent Unergatives

In addition to the transitive predicates exhibiting DOM, in Basque, dative
objects can also occur with other predicates involving an ergative-dative con-
figuration. As can be observed in (20), this is what happens with bivalent
unergative predicates like lagundu ‘accompany, help’, jarraitu ‘follow’ or begi-
ratu ‘look at’ among speakers that do not necessarily accept DOM, both in
southwestern and northeastern Basque (Etxepare 2003, Fernández & Ortiz de
Urbina 2010).14
14 Along with lagundu ‘accompany, help’, jarraitu ‘follow’ and begiratu ‘look at’ bivalent
unergatives involve other verbs like abisatu ‘notify’, barkatu ‘forgive’, bultzatu ‘push’, deitu
‘call’, entzun ‘hear, listen to’, eskertu ‘thank’, heldu ‘hold’, obeditu ‘obey’ or ukitu ‘touch’—
see Etxepare (2003) and Fernández & Ortiz de Urbina (2010) for a complete classification
of this group of predicates.
256 odria
(20) Jon-i lagundu / jarraitu / begiratu diogu.

Jon-d help / follow / look at aux[3sg.d-1pl.e]
‘We have helped/followed/looked at Jon.’
As happens with DOM objects, along with the ergative-dative configuration, in

bivalent unergatives, the dative object alternates with the absolutive, forming
thereby a canonical transitive configuration. This is illustrated in (21a) and (21b)
with the verb deitu ‘call’. In (21a), the object is marked absolutive and a canon-
ical ergative-absolutive alignment emerges. On the contrary, in (21b) the same
object receives dative marking, giving rise to an ergative-dative frame.15
(21) a. Goiz-ean deitu zaitut.

moring-ine call aux[2sg.a-1sg.e]
‘I have called you in the morning.’
b. Goiz-ean deitu dizut.

morning-ine call aux[2sg.d-1sg.e]
‘I have called you in the morning.’
In spite of their similarities, robust evidence has been provided distinguish-

ing the syntactic behavior of DOM objects on the one hand and the datives in
bivalent unergatives on the other (Fernández & Ortiz de Urbina 2010, 2012;
Fernández & Rezac 2010, 2016; Odria 2012, 2017; Ortiz de Urbina & Fernán-
dez 2016, Pineda 2016). While DOM objects are semantically similar to canon-
ical absolutives, the datives in bivalent unergatives pattern more akin to indi-
rect objects—i.e., goals of ditransitive predicates. This has lead Fernández
& Ortiz de Urbina (2012), Ortiz de Urbina & Fernández (2016) and Pineda
(2016) to propose that the ergative-dative frame with bivalent unergative pred-
icates contains a silent direct object and that the sole dative argument is in
fact an indirect object introduced by a Low Applicative head à la Pylkkänen
(2008). In fact, as reported by these authors, many of these verbs are verbs of
communication—i.e., abisatu ‘notify’ or deitu ‘call’—or obey verbs—i.e., obed-
itu ‘obey’. According to them, they involve a transfer of message or order, which
is equivalent to a silent direct object.16
15 Generally speaking, in bivalent unergatives dative marking is preferred in southwestern

dialects, while the absolutive is more frequent in northeastern ones (Fernández & Ortiz
de Urbina 2010).
16 Nevertheless, note that other bivalent unergatives correspond to verbs of relative mo-
In addition to their syntactic configuration, DOM and the datives in bivalent

unergatives also differ with regards to the factors lying behind the dative mark-
ing. Contrary to what happens in DOM, animacy and specificity are irrelevant
when marking the object dative in bivalent unergatives (Fernández & Ortiz
de Urbina 2012, Ortiz de Urbina & Fernández 2016). The examples in (22), for
instance, show that the object of verbs like jarraitu ‘follow’ (22a), itxaron/itx-
oin ‘wait for’ (22b) and begiratu ‘look at’ (22c) is assigned dative regardless of
animacy, given that dative marking with inanimates is grammatical in all of
them.
(22) a. GPSa-ri jarraitu diot.

GPS-d follow aux[3sg.d-1sg.e]
‘I followed the GPS.’
b. Hemen itxoin-go diogu trena-ri.

here wait for-fut aux[3sg.d-1pl.e] train-d
‘We will wait here for the train.’
c. Koadroa-ri begira geratu naiz.

painting-d look at stay aux[1sg.a]
‘I have stayed looking at the painting.’
It is true that in the case of some verbs animacy has a say in the dative marking
of the object. As noted by Fernández & Ortiz de Urbina (2010: 164), this is the
case of itxaron/itxoin ‘wait for’. These authors show that, with few exceptions,
in the General Basque Dictionary Orotariko Euskal Hiztegia (OEH) (Mitxelena &
Sarasola 1989–2005), the object of itxaron/itxoin ‘wait for’ is generally marked
dative when animate (23a), and either dative (23b) or absolutive (23c) when
inanimate.17
(23) a. Beste asko-ri etxeden ez ta ni-ri etxeden.

other many-d wait for not and I-d wait
‘Not wait for many others and wait for me.’ (Cb Eg II 71)
tion—i.e., jarraitu ‘follow’ or lagundu ‘accompany, help’—and verbs of physical con-

tact—i.e., heldu ‘hold’ or ukitu ‘touch’. Therefore, in those cases it would be harder to
assume the indirect object hypothesis including a silent direct object.
17 Of course, this does not mean that absolutive marking is rejected with animate objects. As
258 odria
b. Bihar joan-go naz Bilbora ontzia-ri itxaro-tera.

tomorrow go-fut aux[1sg.a] Bilbao-all ship-d wait for-nomi
‘Tomorrow I will go to Bilbao to wait for the ship.’ (Echta Jos 929)
c. Ikastetxe-ra sartzeko ordua itxaro-ten zaude-la.

school-all to enter time wait for-prog aux[2sg.a]-comp
‘While you are waiting for the time to enter school.’ (Osk Kurl 73)
However, in spite of its influence in verbs like itxaron/itxoin ‘wait for’, a closer
look to the different types of bivalent unergatives indicates that animacy is not
a conditioning trigger for the dative marking in these verbs. Its influence is only
particular of certain verbs or even dialects/speakers, and it is not as system-
atic as in DOM. Likewise, person is neither determinant in marking the object
dative, and the same thing happens with specificity. Consider now the exam-
ples in (24).
(24) a. Lagun asko-ri deitu diet, baina ez dit

friends many-d call aux[3pl.d-1sg.e] but not aux[1sg.d-3sg.e
inor-k erantzun.
anyone-e answer
‘I have called to many friends, but nobody has answered me.’
b. Ez zion inor-i begiratu.

not aux[3sg.d-3sg.e] anyone-d look at
‘He/she didn’t look at anyone.’
c. Ikasle bat-i lagun-tzen geratu naiz.

student one-d help-prog stay aux[1sg.a]
‘I have stayed helping to a student.’
In (24), we see that the same indefinite particles that are generally excluded
with DOM are in fact acceptable with dative objects of bivalent unergatives.
3.2 Spanish a-marking vs. Datives in Bivalent Unergatives

As happens in other Romance languages (Pineda 2016), Spanish also has some
verbs whose objects bear a-marking regardless of specificity (Pensado 1995, To-
I have already pointed out, northeastern dialects show a preference to mark both animate
and inanimate objects absolutive in bivalent unergative predicates.
rrego 1998, 2010; Leonetti 2004, Fábregas 2013). Similar to the Basque bivalent
unergatives, in Spanish the animate object of certain verbs carries a-marking
even with a non-specific reading. This is exemplified by Leonetti (2004: 85) with
the verbs entrevistar ‘interview’ (25a) and admirar ‘admire’ (25b).18
(25) a. Cada estudiante entrevistará a un personaje conocido.

each student will interview.3sg to a person well-known
‘Each student will interview a well-known person.’
b. Todas las niñas admiraban a algún cantante.

every children.f admired.3pl to some singer
‘Every child admired some singer.’
Likewise, Fábregas (2013: 29) adds that a-marking can also be assigned to inan-
imate objects with other verbs. In (26a), for instance, it is possible for the
inanimate object of acosar ‘harass’ to bear a-marking, and the marking is even
obligatory for the inanimate object of ayudar ‘help’ in (26b) (Fábregas 2013:
29).
(26) a. La policía acosa (a)l narcotráfico.

the police harass.3sg to the drug-trafficking
‘The police harass drug-trafficking.’
b. La policía ayuda *(a)l narcotráfico.

the police help.3sg to the drug-trafficking
‘The police help drug-trafficking.’
Similar to what happens in Basque, verbs like acosar ‘harass’ or ayudar ‘accom-
pany, help’ show a parallelism with pure transitives involving DOM, as both are
bivalent and both contain an a-marked object. Notwithstanding, contrary to
the object of pure transitives, animacy and specificity do not necessarily hold
in these verbs. Therefore, as in Basque, not all a-marked objects that occur in
verbs entailing a bivalent argument structure involve DOM.
18 A more complete list of this group of verbs includes acusar ‘accuse’, admirar ‘admire’, afec-
tar ‘affect’, ayudar ‘accompany, help’, castigar ‘punish’, entevistar ‘interview’, golpear ‘beat’,
insultar ‘insult’, odiar ‘hate’, offender ‘offend’, saludar ‘greet’ and sobornar ‘bribe’ (Fábregas
2013: 27–28).
260 odria
3.3 Basque Spanish leísmo vs. Datives in Bivalent Unergatives

In Basque Spanish, the apparent similarity shared by DOM and a-marked
objects in verbs like ayudar ‘accompany, help’ is also visible in the clitic sys-
tem of the verbal complex, since the dative clitic le(s) is used in both of them.
Consider the examples in (27). (27a) corresponds to a DOM construction, as its
object is semantically a theme and both a-marking and the dative clitic le(s)
depend on the animacy and specificity of the object. In contrast, (27b) belongs
to the class of verbs whose (usually) goal object receives a-marking regardless
of animacy and specificity—i.e., bivalent unergatives.
(27) a. Le llevé (al niño) a casa.

3sg.d brought.1sg dom the child.m to house
‘I brought the child home.’
b. Le ayudé (al niño) a casa.

3sg.d helped.1sg to the child.m to house
‘I helped/accompanied to the child home.’
As reported by Fernández-Ordóñez (1999: 123–1339), the dative clitic in sen-

tences like (27b) should not be grouped with the leísmo attested in sentences
like (27a). This author observes that in Spanish the dative clitic can code the
object of verbs that tend to omit their direct object—i.e., tocar ‘touch’—as
well as the object of verbs whose valency has been reinterpreted, as in ayudar
‘accompany, help’, avisar ‘notify’ or obedecer ‘obey’. In order to distinguish the
two types of bivalent configurations involving le(s), Fernández-Ordóñez (1999)
refers to the pattern in (27a) as real leísmo—see footnote 10—, while that in
(27b) is considered as apparent leísmo.
Besides, Torrego (2010), Fábregas (2013) and Pineda (2016) note that, in many
cases, the configuration involving a-marking regardless of animacy and speci-
ficity corresponds to a ditransitive construction. In those cases, the verb hap-
pens to be related to a noun; such a noun would correspond to the direct object
and the a-marked object to the indirect object, as illustrated in (28a) with gol-
pear ‘beat’ and in (28b) with ayudar ‘help’ (Fábregas 2013: 29).
(28) a. Golpear a uno ~ dar un golpe a uno

beat to someone give a blow to someone
b. Ayudar a uno ~ dar ayuda a uno

help to someone give help to someone
Recall that this is also the main insight in Fernández & Ortiz de Urbina
(2012), Ortiz de Urbina & Fernández (2016) and Pineda (2016). These authors
claim that in Basque the dative object in bivalent unergatives is syntactically
equivalent to the indirect object in ditransitive constructions involving a silent
direct object. In fact, the Spanish correspondences in (28a) and (28b) are
equally obtained in Basque bivalent unergatives. Consider, for instance, the
‘bivalent unergative/ditransitive’ correspondence between deitu/dei egin ‘call’
(29a), bultzatu/bultza egin ‘push’ (29b) or lagundu/laguntza eman ‘help’ (29c).
(29) a. Norbait-i deitu ~ norbait-i dei egin

someone-d call someone-d call do
b. Norbait-i bultzatu ~ norbait-i bultza egin

someone-d push someone-d push do
c. Norbait-i lagundu ~ norbait-i laguntza eman

someone-d help someone-d help give
Hence, it seems reasonable to think that in both Basque and Spanish the single
object that appears with dative morphology is in fact an indirect object in these
cases.
Having coupled the syntactic configuration of DOM objects with canonical
absolutives and that of the datives in bivalent unergatives with indirect objects,
the following section makes a further step in the syntactic distinction attested
between these kinds of objects. It argues that the difference between DOM
and dative objects in bivalent unergatives is basically reduced to their DP vs.
PP-like—i.e., dual DP/PP—categorical status (Odria 2017). While DOM objects
exhibit a DP category, the datives in bivalent unergatives behave more akin to
the rest of PP-like goals. This is evidenced by the possibility or not to license
depictive secondary predication.
4 The Licensing of Depictive Secondary Predication
4.1 The Dual DP vs. PP-like Category of Dative Arguments

In Basque, some datives behave under certain circumstances more akin to
PP s than to DP s, as they can occur without finite verbal agreement—see spe-
cially Albizu (2001) on this point. This is what happens with goal datives in
northeastern Basque, which—contrary to experiencers and possessors—are
able to appear as non-agreeing—see Masullo (1992), Demonte (1995), Romero
262 odria
(1997) and Cuervo (2003) for a similar pattern in Spanish.19 In northeastern

Basque, causees are able to occur without dative markers too, but in this case
the possibility to appear as non-agreeing does not seem to be as extended
as with goals (Fernández, Ortiz de Urbina & Landa 2009).20 In southwest-
ern Basque, non-agreeing datives are limited to specific configurations such as
ditransitives affected by the PCC and double dative constructions involving a
causee and a goal. Generally speaking, only goals are allowed to occur as non-
agreeing in them, indicating that causees pattern more akin to experiencers
and possessors in this regard. Based on these facts, in what follows I will refer
to causee, experiencer and possessors as ‘DP datives’ and to both agreeing non-
agreeing goals—as ‘PP-like datives’ (Odria 2017, to appear).
The contrast between DP and PP-like datives has additionally been sup-
ported by the ability to occur as adnominals in headlines, which—along with
the rest of PP s—is only possible for goal datives—see de Rijk (2008: 378),
Albizu (2001: 63, 2009: 13), and specially, Fernández & Sarasola (2010) and Berro
& Fernández (2018). As demonstrated by Fernández & Sarasola (2010), posses-
sors and experiencers pattern with ergative and absolutive DP s in requiring the
genitive marker -ren in order to function as adnominals.
In this section, I claim that DP and PP-like datives can also be distinguished
by their ability to license depictive secondary predication (Odria 2015, 2017, to
appear). In Basque (Zabala 1993, 2003; Oyharçabal 2010) and Spanish
(Demonte 1987, 1988; Romero 1997; Demonte & Masullo 1999), datives and PP s
have been considered to be unable to control depictive secondary predica-
tion.21 Such a restriction has been illustrated by sentences like (30) and (31)
19 For non-agreeing datives in northeastern Basque, see Ortiz de Urbina (1995), Albizu (2001,
2009), Fernández (1997, 2010, 2014), Fernández & Landa (2009), Fernández et al. (2009),
Fernández & Ortiz de Urbina (2010), Etxepare & Oyharçabal (2009ab, 2013), Etxepare
(2014) and Ormazabal & Romero (2017).
20 Fernández, et al. (2009: 214–215) report that non-agreeing datives in causative construc-
tions should be analyzed more carefully, given that in this case the lack of dative mark-
ers could be influenced by external factors like indefiniteness, which is known to foster
dative agreement drop (Ortiz de Urbina 1995). Besides, these authors note that, in their
corpus—which is based on classical writings from the 19th and 20th centuries—some
of the non-agreeing datives in causative constructions represent the goal of the embed-
ded predicate, instead of the causee itself. Likewise, they add that the availability to have
non-agreeing datives in causatives could be affected by idiolectal variation as well, as non-
agreeing causees are found in the writings of some but not all north-eastern classical
authors that make use of agreement drop when it comes to goals. In addition, as noted
by Javier Ormazabal (p.c.), in northeastern Basque, dative agreement drop with causees
appeared chronologically later than with goals.
21 See Williams (1980) and Rothstein (1983) for the same pattern in English, McFadden
(2004) for German and Koizumi (1994) for Japanese.
adapted from Zabala (1993: 255, 258)—, where neither the dative (30) nor the
PP (31) is able to control the depictive secondary predicate.
(30) a. *Miren-ii liburua urdurii oparitu diot.

Miren-d book.a nervous gift aux[3sg.a-3sg.d-1sg.e]
‘I have gifted the book to Miren nervous.’
b. *A Maríai lei he regalado el libro nerviosai .

to María 3sg.d have gifted.1sg the book nervous.m/-f
‘I have gifted the book to María nervous.’
(31) a. *Jostailu bat erosi dugu haurra-rentzati poziki .

toy one.a buy aux[3sg.a-1pl.e] child-dest happy
‘We have bought a toy for the child happy.’
b. *Hemos comprado un juguete para el niñoi contentoi .

have bought.1sg a toy for the child happy
‘We have bought a toy for the child happy.’
So far, the restriction in question has been mainly tested with goal datives.
Notwithstanding, a closer look at the different nature of dative arguments
reveals that not all kind of datives are equally reluctant when it comes to
licensing depictive secondary predication. In fact, apart from the well-known
exception of the causee, possessor and experiencer datives are also able to
license this kind of predication. This indicates that, instead of the syntactic
configuration—i.e., subject / direct object vs. indirect object—or case mark-
ing—i.e., ergative / absolutive vs. dative—, the restriction on secondary predi-
cation depends basically on the syntactic category of the controller, as DP—i.e.,
causee, experiencer and possessor—but not PP-like—i.e., goal—datives are
able to do so.
Let us first focus on the behavior of the causee. As illustrated in (32) (Zab.ala
1993: 269) and (33) (Demonte 1987: 154), causees are able to control depictive
secondary predication.22
22 Comparing to intransitive and transitive causatives, secondary predication with ditransi-

tive causatives seems to be harder to process both in Basque and Spanish, as apart from
the subject—i.e., causer—and the direct object, they involve two dative arguments: the
causee and the indirect object. For this reason, contrary to what I did in Odria (2014), I
give no example of ditransitive causatives involving secondary predication.
264 odria
(32) a. Ama-k haurra-rii gaixoriki joan-arazi zion

mother-e child-d ill go-caus aux[3sg.d-3sg.e]
eskola-ra.
school-all
‘The mother made the child go to school ill.’
b. Irakasle-ak Mikel-ii guzti-en aurrean bakarriki dantza-arazi

teacher-e Mikel-d in front of everyone alone dance-caus
dio.
aux[3sg.d-3sg.e]
‘The teacher has made Mikel dance alone in front of everyone.’
c. Ama-k haurra-rii indabak gogorik gabei jan-arazi

mother-e child-d beans.a without wanting eat-caus
zizkion.
aux[3pl.a-3sg.d-3sg.e]
‘The mother made the child eat the beans without wanting.’
(33) Juan (la)i hizo bailar a Maríai desnudai.

Juan 3sg.acc made.3sg dance to María naked
‘Juan made María dance naked.’
Possessors show a similar behavior, as in spite of their dative marking, depic-

tives are allowed to be hosted by them. Consider now the Basque examples in
(34) and the Spanish examples in (35)—the examples in (35a) and (35b) are
provided by Demonte & Masullo (1999: 2466) and that in (35c) by Hernanz
(1988: 12).23
(34) a. (Ni-ri)i udaleku-etan ilea lokartutai moztu

I-d summer camp-ine hair.a sleeping cut
23 Demonte & Masullo (1999: 2466–2467) explain that the licensing of depictive secondary
predication by possessor datives is quite restricted in Spanish. On the one hand, they say
that the depictives allowed in these contexts tend to be replaceable with gerundive verbal
periphrases like estando dormida ‘being slept’ for dormida ‘slept’, or estando anestesiada
‘being anesthetized’ for anestesiada ‘anesthetized’. On the other hand, they note that in
these constructions the possessor tends to be placed in a preverbal position and the depic-
tive in a post-verbal one. Taking these and other facts into account, these authors conclude
that in this case the secondary predicate appears in a position distinct to the predicates
controlled by subjects and direct objects.
zidaten.
aux[3sg.a-1sg.d-3pl.e]
‘In the summer camp, they cut me my hair sleeping.’
b. Jon-ii kistea anestesiatutai kendu zioten.

Jon-d cyst.a anesthetized remove aux[3sg.a-3sg.d-3pl.e]
‘They removed the cyst to Jon anesthetized.’
(35) a. A Maríai, lei operaron el quiste dormidai.

to María 3sg.d operated.3pl the cyst sleeping
‘María was operated the cyst sleeping.’
b. Lei extirparon el lunar a Consueloi anestesiadai.

3sg.d removed.3pl the mole to Consuelo anesthetized
‘They removed the mole to Consuelo anesthetized.’
c. A la enfermai, los cirujanos lei extirparon el quiste

to the patient the surgeons 3sg.d removed.3pl the cyst
anestesiadai.
anesthetized
‘The surgeons removed the cyst to the patient anesthetized.’
Besides, the sentences in (36) illustrate that in Basque experiencers can license
depictives too, a pattern that is again replicated in Spanish subject-like datives
(37) (Fernández-Soriano 1999: 124).24
(36) a. (Ni-ri)i mozkortutai mundu guztia gustatzen zait,

I-d drunk world whole.a like-prog aux[3sg.a-1sg.d]
baina mozkortuta ez nago-en-ean ez
but drunk not be.aux.[1sg.a]-comp-ine not
dut inor aguanta-tzen.
aux[3sg.a-1sg.e] no one stand-prog
‘When I’m drunk I like everyone, but when I’m not drunk I can’t stand
anyone.’
24 The example in (37a) was gathered by Fernández-Soriano (1999: 124) in La Dorotea by Lope
de Vega.
266 odria
b. (Ni-ri)i mozkortutai mundu guztia jaus-ten zait

I-d drunk world whole.a get on-prog aux[3sg.a-1sg.d]
ondo.
well
‘When I’m drunk I get on well with everyone.’
c. (Ni-ri)i lotsatutai gaizki atera zitzaidan aurkezpena, beraz,

I-d shy bad go aux[3sg.a-1sg.d] presentation hence
hurrengoan lotsarik gabe joa-ten saiatu-ko naiz.
next time with no shame go-comp try-fut aux[1sg.a]
‘The presentation went (to me) wrong shy, so next time I’ll try to do it
with no shame.’
(37) a. Mal mei fue ausentei pero peor presentei.

bad 1sg.d went.1sg absent but worse present
‘Things were bad for me while absent but worse when present.’
b. Nosi dieron las dos borrachosi.

1pl.d gave.3pl two o’clock drunk
‘It got as late as 2 o’clock on us and we were drunk.’
c. Lei ocurrió un accidente borrachai.

3sg.d happened.3sg an accident drunk
‘An accident happened to her while being drunk.’
Overall, contrary to goals and PP s, causee, experiencer and possessors seem

to be compatible with depictive secondary predication, a pattern that is again
attested not only in Basque, but also in Spanish.
Taking this into account, the following sections examine the behavior of
depictive secondary predication with the datives in DOM and in bivalent
unergatives. Focusing on both Basque and Spanish, I show that only DOM
objects allow the licensing of depictive secondary predication. Hence, given
that depictives are only rejected by PP s and PP-like datives, this implies that,
while datives in bivalent unergatives show a PP-like behavior, DOM objects
behave categorically as the rest of canonical absolutives, that is, as DP s (Odria
2015, 2017, to appear).
4.2 DOM and Depictive Secondary Predication

As has been demonstrated by Fernández & Rezac (2010, 2016) and Odria (2012,
2014, 2017, to appear), in Basque, DOM objects are able to control depictive sec-
ondary predication. This is illustrated in (39) with examples from Larrabetzu

Basque (a Western variety). The examples in (38) belong to a speaker with
no DOM (in present tense) and those in (39) to a speaker with DOM (in both
present and past tenses). What is important for the purpose of this section is
that the dative objects in (39) can control depictive secondary predication in
the same way as the absolutive objects in (38).
(38) a. (Zu-k) (ni)i ostondutei harrape nozu.

you-e I.a hidden catch aux[1sg.a-2sg.e]
‘You have caught me hidden.’
b. (Ni-k) (zu)i tontotutei ikusi zaitut.

I-e you.a silly see aux[2sg.a-1sg.e]
‘I have seen you silly.’
c. Lagun-ek Mireni mozkortutei grabe deure.

friends-e Miren.a drunk record aux[3sg.a-3pl.e]
‘Her friends have recorded Miren drunk.’
(39) a. (Zu-k) (ni-ri)i ostondutei harrape dostezu.

you-e I-d hidden catch aux[1sg.d-2sg.e]
‘You have caught me hidden.’
b. (Ni-k) (zu-ri)i tontotutei ikusi dotsut.

I-e you-d silly see aux[2sg.d-1sg.e]
‘I have seen you silly.’
c. Lagun-ek Miren-erii mozkortutei grabe dotsie.

friends-e Miren-d drunk record aux[3sg.d-3pl.e]
‘Her friends have recorded Miren drunk.’
Even though in Basque DOM exhibits a great deal of dialectal as well as idi-
olectal variation, the availability to control depictive secondary predication is
a persistent pattern, given that the results obtained in Larrabetzu are general-
ized among the rest of the consulted speakers from other Basque varieties, such
as Araitz-Betelu (Central Basque) (Arraztio 2010), Lekeitio and Dima (West-
ern Basque) (Fernández & Rezac 2010, 2016), Elgoibar (Transitional Central-
Western Basque), Zumaia (Central Basque), Hondarribia (Transitional Central-
Navarrese Basque), Oñati (Western Basque), Tolosa (Central Basque), Errente-
ria (Central Basque) and Itsasondo (Central Basque) (Odria 2014, 2017).
268 odria
As happens in Basque, in Spanish a-marked objects are equally able to

license depictive secondary predication. As shown by Demonte (1987: 148),
in (40a) the secondary predicate borracha ‘drunk’ cannot be controlled by
the goal indirect object—i.e., Maria. Contrarily, the same author says that in
(40b) the secondary predicate turns out to be controlled by the DOM object a
María.
(40) a. *Juan lei habló a Maríai borrachai

Juan 3sg.d talked.3sg to María drunk
‘Juan talked to María drunk.’
b. Juan lai encontró a Maríai borrachai.

Juan 3sg.acc found.3sg dom María drunk
‘Juan found María drunk.’
In order to justify the possibility for DOM objects to license depictive secondary
predication, Demonte (1987) claims that, as in causee datives, a is not a true
preposition in DOM objects. Instead, this author takes a in examples like (40b)
to be a dummy preposition, a preposition that does not project a PP maxi-
mal projection and does not therefore impede c-command relations. This way,
Demonte states that the DOM object a María in (40b) can control the depic-
tive borracha ‘drunk’ because there is no P head that blocks the structural c-
command relation between the argument and the predicate.
Interestingly, the possibility for DOM objects to license depictive secondary
predication persists even when the object is coded—or clitic doubled—by the
dative clitic le(s) in the leísta Basque Spanish. Consider, for instance, the exam-
ples in (41), which are the correspondent Spanish examples of those in (39).25
(41) a. Juan lei pilló (al niño)i escondidoi.

Juan 3sg.d caught.3sg dom child.m hidden.m
‘Juan caught the child hidden.’
b. Juan lei vio (a Pedro)i atontadoi.

Juan 3sg.d saw.3sg dom Pedro silly.m
‘Juan saw Pedro silly.’
25 In order to avoid interferences, in the Spanish examples in (41) the object bears the same
gender as well as number specification as the subject. Likewise, so as to test the behavior
of the dative clitic le(s), I have changed the objects in (41a) and (41b) to third person.
c. Su amiga lei grabó (a María)i borrachai.

her friend 3sg.d recorded.3sg dom María drunk.f
‘Her friend recorded María drunk.’
Overall, assuming that depictives are only incompatible with PP s and PP-like
datives, I conclude that both Basque and (Basque) Spanish DOM objects are
DP s categorically.
4.3 Datives in Bivalent Unergatives and Depictive Secondary Predication

Contrary to DOM objects, the datives in bivalent unergatives are generally
unable to control depictive secondary predication (Fernández & Rezac 2010,
2016; Odria 2012, 2017; Ortiz de Urbina & Fernández 2016). As illustrated in (42),
in this case, the only possible interpretation is that the secondary predicate ton-
totuta ‘silly’ (42a) and mozkortute ‘drunk’ (42b) modifies the subject Jon and not
the dative object Mikel.26
26 Even being a bivalent unergative predicate, entzun ‘hear, listen to’ is exceptional allowing
secondary predication (Fernández & Rezac 2010: 134, Odria 2017). This is illustrated in (i):
(i) Ni-ki Mikel-ij mozkortutai/j entzun nion.
I-e Mikel-d drunk hear/listen to aux[3sg.d-1sg.e]
‘I heard to Mikel drunk.’
I believe that the distinction between the two different meanings of entzun ‘hear, listen
to’ might have an influence in this fact. Actually, when entzun means ‘listen to’, dative
marking is generally attested in southwestern dialects, while the absolutive is preferred
in northeastern ones (Fernández & Ortiz de Urbina 2010: 129, 134). On the contrary, when
entzun means ‘hear’, the object can also be absolutive in southwestern varieties, behaving
more akin to a canonical direct object. In fact, Ortiz de Urbina & Fernández (2016: 79–
80) mention that in a sentence like I heard the bossi angryi, the object is usually marked
absolutive, and that dative marking makes the sentence slightly worse. Hence, if entzun
in (i) means ‘hear’ instead of ‘listen to’, its dative object could be closer to a DOM object,
and in that case, it would pattern similar to other perception verbs like ikusi ‘see’, which
would explain its apparently exceptional behavior in (i).
In addition, it is important to note that the dative object of lagundu ‘help, accompany’
can also control secondary predication when pragmatics forces to do so. This is the case
in (ii), where along with the ergative subject, many speakers allow the depictive to modify
the dative object.
(ii)(Haiek)i (ni-ri)j etxe-ra mozkortutai/j lagundu zidaten
they.e I-d house-all drunk help/accompany aux[1sg.d-3plE]
‘They helped/accompanied me home drunk.’
Crucially, some of the consultants add that in (ii) the possibility for the dative to control
secondary predication is logically or pragmatically conditioned. Hence, I believe that the
example in (ii) does not necessarily bring into question the fact that (under normal cir-
cumstances) datives in bivalent unergatives are unable to control secondary predication.
270 odria
(42) a. Jon-eki Mikel-ij tontotutai/*j begiratu zion.

Jon-e Mikel-d silly look at aux[3sg.d-3sg.e]
‘Jon looked at Mikel silly.’
b. Jon-eki Mikel-ij mozkortutai/*j deitu zion.

Jon-e Mikel-d drunk call aux[3sg.d-3sg.e]
‘Jon called Miren drunk.’
Spanish behaves once again like Basque, as it does not allow depictives to depict
of the dative object in verbs like mirar ‘look at’ (43a) or llamar ‘call’ (43b).27
(43) a. Juani lej miró a Pedroj atontadoi/*j.

Juan 3sg.d looked at.3sg to Pedro silly
‘Juan looked at Pedro silly.’
b. Juani lej llamó a Pedroj borrachoi/*j.

Juan 3sg.d called.3sg to Pedro drunk
‘Juan called drunk to Pedro.’
All in all, the impossibility to license depictive secondary predication links the
dative object of bivalent unergatives with PP-like goals of ditransitive construc-
tions, and thus implies that these objects have a dual DP/PP categorical status.
5 Conclusions
In this paper I have compared the syntactic behavior of Basque and (Basque)
Spanish DOM. I have highlighted the main similarities and differences attested
between the two of them and I have distinguished them from the dative object
in bivalent unergative predicates. Contrary to what happens in DOM, in biva-
lent unergatives the dative marking is independent from factors like animacy
and specificity and the object patterns more akin to indirect objects of ditransi-
tive predicates. Moreover, based on the licensing of depictive secondary pred-
ication, I have argued that these datives are additionally distinguished by their
syntactic category. Contrary to DOM objects—which exhibit a DP categorical
status—, datives in bivalent unergatives are generally incompatible with depic-
tives, a pattern that groups them with the rest of PP-like goals.
27 In (43b), borracho ‘drunk’ can also be interpreted as a primary predicate referred to the
dative Pedro. However, when behaving as a secondary predicate, borracho ‘drunk’ can only
be controlled by the subject Juan.
On a broader perspective on dative arguments, the categorical distinction

attested between DP and PP-like datives suggests that a different categorical
origin should be assigned for the different kinds of datives. That is, a DP cat-
egory for causee, experiencer, possessor and DOM datives, and a PP one for
goal datives—this implies that the so-called restriction is related to the P head
and holds at the level of argument structure. However, even being distinct in
the first merge position, both DP and PP-like datives should share the same
Agree/Case position, as both occur with dative markers in the finite verbal form
and both pattern to a certain extent in the same way in different kinds of agree-
ment restrictions (Odria 2017, to appear).
Acknowledgements
This work has been supported by the Basque Government (the research group
IT665–13) and the Spanish Ministry of Economy and Competitiveness
(FFI2014–51878-P). The research leading to these results has also received fund-
ing from the European Union’s Seventh Framework Program for research, tech-
nological development and demonstration under grant agreement no. 613465.
I am grateful to Ane Berro and Javier Ormazabal for their comments and sug-
gestions. Needless to say, all errors are mine.
References
Aissen, Judith. 2003. “Differential object marking: iconicity vs. economy”. Natural Lan-
guage and Linguistic Theory 21: 435–483.
Albizu, Pablo. 2001. “Datibo sintagmaren izaera sintaktikoaren inguruan: eztabaida-
rako oinarrizko zenbait datu”. In Beatriz Fernández & Pablo Albizu (eds.), Kasu eta
komunztaduraren gainean. On Case and Agreement. Bilbo: UPV/EHU Editorial Ser-
vices. 49–69.
Albizu, Pablo. 2009. “Construcciones inacusativas con dativos posesivos y dativos de
interés en vasco: un análisis derivacionalista”. Seminario de Lingüística Teórica, CSIC,
Madrid, May 11.
Arraztio, Kontxi. 2010. “Datibo osagarri bitxiak Araitz-Beteluko hizkeran”. In Beatriz
Fernández, Pablo Albizu & Ricardo Etxepare (eds.), Euskara eta euskarak: aldako-
rtasun sintaktikoa aztergai. [Supplements of ASJU LII]. Bilbo: UPV/EHU. 113–149.
Austin, Jennifer. 2006. “Dative overmarking in Basque”. Euskalingua 9: 136–145.
Austin, Jennifer. 2015. “Transfer and contact-induced variation in child Basque”. Fron-
tiers in Psychology 5: 1–11.
272 odria
Bárany, András. 2018. DOM and dative case. Glossa: a journal of General Linguistics 3(1):
97, 1–40.
Berro, Ane & Beatriz Fernández. 2018. “Applicatives without verbs”. Natural Language
and Linguistic Theory (NLLT).
Bhatt, Rajesh & Elena Anagnostopoulou. 1996. “Object Shift and Specificity: Evidence
from ko-phrases in Hindi”. In L. Dobrin et al. (eds.), Proceedings of CLS 32. Chicago:
Chicago Linguistics Society.
Bossong, Georg. 1985. Empirische Universalienforschung. Differentielle Objekt Markie-
rung in der neuiranischen Sphrachen. Tübingen: Narr.
Bossong, Georg. 1991. “Differential Object Marking in Romance and beyond”. In Dieter
Wanner & Douglas A. Kibee (eds.), New Analyses in Romance Linguistics. Amster-
dam/Philadelphia: John Benjamins. 143–170.
Brugè, Laura & Gerhard Brugger. 1996. “On the Accusative a in Spanish”. Probus 8(1).
1–51.
Cuervo, M. Cristina. 2003. Datives at Large. Ph.D. thesis, MIT.
Demonte, Violeta. 1987. “C-command, Preposition and Predication”. Linguistic Inquiry
18: 147–157.
Demonte, Violeta. 1988. “Remarks on secondary predicates. C-command, extraction
and reanalysis”. The Linguistic Review 6–1: 1–39.
Demonte, Violeta. 1995. “Dative Alternation in Spanish”. Probus 7. 5–30.
Demonte, Violeta & Pascual J. Masullo. 1999. “La predicación. Los complementos pred-
icativos”. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva de la
lengua española. Madrid: Espasa Calpe, RAE. 2461–2524.
Etxepare, Ricardo. 2003. “Valency and Argument Structure in the Verb”. In José Igna-
Gruyter. 369–465.
Etxepare, Ricardo. 2014. “Contact and Change in a restrictive theory of parameters”.
In Carme Picallo (ed.), Linguistic Variation in the Minimalist Framework. Oxford:
Oxford University Press. 108–139.
Etxepare, Ricardo & Bernard Oyharçabal. 2009a. “Bi datibo egitura ipar-ekialdeko zen-
bait hizkeratan”. Lapurdum XIII: 145–158.
Etxepare, Ricardo & Bernard Oyharçabal. 2009b. “Datiboa ipar-ekialdeko zenbait hizk-
eratan”. Ms., IKER, April 25.
Etxepare, Ricardo & Bernard Oyharçabal. 2013. “Datives and adpositions in northeast-
ern Basque”. In Beatriz Fernández & Ricardo Etxepare (eds.), Variation in datives: a
micro-comparative perspective. Oxford: Oxford University Press. 50–95.
Fábregas, Antonio. 2013. “Differential Object Marking in Spanish”. Borealis. An Interna-
tional Journal of Spanish Linguistics 2–2: 1–80.
Fernández, Beatriz. 1997. Egiturazko kasuaren erkaketa euskaraz. Bilbo: UPV/EHU Edi-
torial Services.
Fernández, Beatriz. 2010. “Goi datiboak eta are goragokoak euskaraz: jabe datiboak eta
datibo hunkituak vs. datibo etikoak”. Gogoa, 10: 1–2: 1–20.
Fernández, Beatriz. 2014. “Goi-eta behe-datiboak eta Datibo-komunztaduraren murriz-
tapena: ipar-ekialdeko hizkerak eta euskara estandarra”. Irantzu Epelde (ed.), Euskal
dialektologia: lehena eta oraina [Supplements of ASJU LXIX]. UPV/EHU. 37–63.
Fernández, Beatriz & Josu Landa. 2009. “Datibo komunztadura beti zaindu, inoiz
zaindu ez eta batzuetan baino zaintzen ez denean. Hiru ahoko aldagaia, datu iturri
bi, eta erreminta bat: Corsintax”. Lapurdum 13: 159–181.
Fernández, Beatriz & Jon Ortiz de Urbina. 2010. Datiboa hiztegian. Bilbo: UPV/EHU Edi-
torial Services.
Fernández, B. & J. Ortiz de Urbina, J. 2012. “Dative (First) Complements in Basque”. In
Ernestina Carrilho & Beatriz Fernández (eds.), Syntactic microvariation in Westmost
European Languages. Journal of Portuguese Linguistics (Special Issue), 11-1. Edições
Colibrí—Unviersidade de Lisboa. 83–98.
Fernández, Beatriz, Jon Ortiz de Urbina & Josu Landa. 2009. “Komunztadurarik gabeko
datiboen gakoez”. In Ricardo Etxepare, Ricardo Gómez & Joseba A. Lakarra (eds.),
Beñat Oihartzabali gorazarre. ASJU XVIII: 352–380.
Fernández, Beatriz & Milan Rezac. 2010. “Datibo osagarri bitxiak eta Datiboaren Lek-
ualdatzea: ari nai diyot eta kanta egin nazu bidegurutzean”. In Beatriz Fernández,
Pablo Albizu, & Ricardo Etxepare (eds.), Euskara eta euskarak: aldakortasun sintak-
tikoa aztergai. [Supplements of ASJU LII]. Bilbo: UPV/EHU. 113–149.
Fernández, Beatriz & Milan Rezac. 2016. “Differential Object Marking in Basque vari-
eties”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
grammar of Basque. Amsterdam/Philadelphia: John Benjamins. 93–139.
Fernández, Beatriz & Ibon Sarasola. 2010. “Marinelei abisua: izen ondoko datibo sintag-
mak izenburuen sintaxian”. Lapurdum XIV: 55–75.
Fernández-Ordóñez, Inés. 1994. “Isoglosas internas del castellano. El sistema referen-
cial del pronombre átono de tercera persona”. Revista de Filología Española 74-1/2:
71–125.
Fernández-Ordoñez, Inés. 1999. “Leísmo, laísmo y loísmo”. In Ignacio Bosque & Vio-
leta Demonte (eds.), Gramática descriptiva de la lengua española I. Madrid: Espasa
Calpe. 1317–1394.
Fernández-Soriano, Olga. 1999. “Two types of impersonal sentences in Spanish: locative
and dative subjects”. Syntax 2–2: 101–140.
Franco, Jon A. 1993. “Conditions on clitic doubling: the agreement hypothesis”. ASJU 27:
285–298.
Hernanz, Mª Lluisa. 1988. “En torno a la sintaxis y la semántica de los complementos
predicativos en español”. Estudis de Sintaxis, Estudi General, 8 (Colegi universitari
deGirona) 7–29.
Hualde, José Ignacio, Gorka Elordieta & Arantzazu Elordieta. 1994. The Basque dialect of
274 odria
Lequeitio [Supplements of ASJU XXXIV]. Bilbo & Donostia: UPV/EHU & Gipuzkoako
Foru Aldundia.
Ibarra, Orreaga. 1995. Ultzamako hizkera. Inguruko euskalkiekiko harremanak. Iruñea:
Nafarroako Gobernua.
Koizumi, Masatoshi. 1994. “Secondary predicates”. Journal of East Asian Linguistics 3:
25–79.
Landa, Alazne. 1995. Conditions on Null Objects in Basque Spanish and their relation to
Leísmo and Clitic Doubling. Ph.D. thesis, University of Southern California.
Leonetti, Manuel. 2004. “Specificity and differential object marking in Spanish”. Cata-
lan Journal of Linguistics 3: 75–114.
Mahajan, Anoop Kumar. 1990. The A/A-bar distinction and movement theory. Ph.D.
thesis, MIT.
Masullo, Pascual J. 1992. Incorporation and case theory in Spanish. A crosslinguistic per-
spective. Ph.D. thesis, University of Washington.
McFadden, Thomas. 2004. The position of morphological case in the derivation: a study
on the syntax-morphology interface. Ph.D. thesis, University of Pennsylvania.
Mounole, Céline. 2012. “Evolution of the transitive verbs in Basque and apparition of
datively marked patients”. In Gilles Authier & Katharina Haude (eds.), Ergativity,
Transitivity, and Voice. Berlin/New York: Mouton de Gruyter. 355–379.
Odria, A. 2012. What lies behind differential Object Marking: a Basque dialect survey.
Master’s thesis, UPV/EHU. In Pello Salaburu, Patxi Goenaga & Ibon Sarasola (eds.),
Sareko Euskal Gramatika.
Odria, Ane. 2014. “Differential Object Marking and the nature of dative case in Basque
varieties”. Linguistic Variation, 14-2. Amsterdam/Philadelphia: John Benjamins. 289–
317.
Odria, Ane. 2015. “Euskarazko bigarren mailako predikazioaren murriztapen sintak-
tikoez”. In Beatriz Fernández & Pello Salaburu (ed.), Ibon Sarasola gorazarre. Hom-
enatge. Homenaje. Bilbo: UPV/EHU. 449–515.
Odria, Ane. 2017. Differential Object Marking and datives in Basque syntax. Ph.D. thesis,
UPV/EHU.
Odria, Ane. to appear. “DOM and datives in Basque: not as homogeneous as they might
seem”. In A. Pineda & M.A. Irimia (eds.), DOM and datives -a homogeneous class?
Linguisticae Investigationes. Amsterdam/Philadelphia: John Benjamins.
Ormazabal, Javier & Juan Romero. 2013a. “Differential Object Marking, Case and Agree-
ment” Borealis. An International Journal of Hispanic Linguistics 2–2: 221–239.
Ormazabal, Javier & Juan Romero. 2013b. “Non Accusative Objects”. Catalan Journal of
Linguistics 12. 155–173.
Ormazabal, Javier & Juan Romero. 2017. “Historical Changes in Basque Dative Alterna-
tions: evidence for a P-based (neo)derivational analysis”. Ms., UPV/EHU & University
of Extremadura.
Ortiz de Urbina, Jon 1995. “Datibo komunztaduraren gainean”. In Ricardo Gómez &
Joseba A. Lakarra (eds.), Euskal Dialektologiako Kongresua (Donostia, 1991ko Irailak
2–6) [Supplements of ASJU XXVIII]. Donostia: Gipuzkoako Foru Aldundia. 579–588.
Ortiz de Urbina, Jon & Beatriz Fernández. 2016. “Datives in Basque bivalent unerga-
tives”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
grammar of Basque. Amsterdam/Philadelphia: John Benjamins. 67–93.
Oyharçabal, Bernard. 2010. “Basque ditransitives”. In Maia Duguine, Susana Huidobro &
Nerea Madariaga (eds.), Argument structure and syntactic relations: A cross-linguistic
perspective. Amsterdam/Philadelphia: John Benjamins. 233–260.
Pensado, Carmen. 1995. “El complemento directo preposicional: estado de la cuestión
y bibliografía comentada”. In Carmen Pensado (ed.), El complemento directo preposi-
cional. Madrid: Visor. 11–59.
Pineda, Anna. 2016. Les fronteres de la (in)transitivitat. Estudi dels aplicatius en llengües
romàniques i basc. Barcelona: Institut d’Estudis Món Juïc (Col·lecció Cum Laude, 6).
Pylkkänen, Liina. 2008. Introducing arguments. Cambridge/London: MIT Press.
de Rijk, Rudolf P. 2008. Standard Basque. A Progressive Grammar. Cambridge/London:
The MIT Press.
Rodríguez-Mondoñedo, Miguel. 2007. The Syntax of Objects: Agree and Differential
Object Marking. Ph.D. thesis, University of Connecticut.
Rodríguez-Ordóñez, Itsaso. 2013. “Contact-induced phenomena in Gernika Basque: the
case of dative over-marking”. Selected Proceedings of the 15th Hispanic Linguistics
Symposium, Athens, Georgia. 236–251.
Rodríguez-Ordóñez, Itsaso. 2016. DOM in Basque: grammaticalization, attitudes and ide-
ological representations. Ph.D. thesis, University of Illinois at Urbana-Champaign.
Romero, Juan. 1997. Construcciones de doble objeto y gramática universal. Ph.D. thesis,
Universidad Autónoma de Madrid.
Rothstein, Susan D. 1983. The syntactic forms of predication. Ph.D. thesis, MIT.
Torrego, Esther. 1998. The dependencies of objects. Cambridge, Mass.: MIT Press.
Torrego, Esther. 2010. “Variability in the case pattern of causative formation in Romance
and its implications”. Linguistic Inquiry 41: 445–470.
Urrutia-Cárdenas, H. 1995. “Morphosyntactic features in the Spanish of the Basque
Country”. In Carmen Silva-Corvalán (ed.), Spanish in four continents. Washington,
D.C.: Georgetown University Press. 243–259.
Urrutia-Cárdenas, Hernán. 2003. “Los clíticos de tercera persona en el español en el
País Vasco”. Revista de Filología y su Didáctica 26: 517–538.
Williams, Edwin S. 1980. “Predication”. Linguistic Inquiry 11: 203–237.
Zabala, Igone. 1993. Predikazioaren teoriak Gramatika Sortzailean (Euskararen kasua).
Ph.D. thesis, UPV/EHU.
Zabala, Igone. 2003. Nominal Predication: copulative sentences and secondary predi-
cation. In José Ignacio Hualde & Jon Ortiz de Urbina (eds.), A Grammar of Basque.
Berlin: Mouton de Gruyter. 426–446.
chapter 9
Complex Causative Verbs and Causees in Basque

(and Romance)
Jon Ortiz de Urbina
1 Introduction
Basque productive causatives are primarily morphological: a causative mor-

pheme arazi (also erazi in some dialects, as well as eragin, erazo or arazo in
western and especially Biscayan Basque) is added to a verbal base (the verbal
root in standard Basque), forming a derived causative verb, as in (1b):1
(1) a. Semeak medizina hartu du.

son medicine take aux
‘The son took the medicine.’
b. Jonek semeari medizina harrarazi dio.

Jon.erg son.dat medicine take.cause aux
‘Jon made the son take the medicine.’
In principle, this initial description sets the Basque data apart from the
Romance ones, if causatives in Spanish and French are of the analytic type
(Comrie 1989; Dixon 2000). Although usually bound and without the ability
to occur freely, however, the causative ‘morpheme’ in Basque is fairly transpar-
ently a verb. This brings the Basque causative construction closer to that with
causative verbs like Spanish hacer or French faire ‘make’. The type of complex
verb formed in all these languages and complementation types are addressed
first in Section 2. Section 3 will be devoted to causee marking and agreement
conflicts resulting from the presence of two datives in the same structure. As
we will see, this is also an area where new patterns converging with Romance
ones seem to be emerging or developing side by side with traditional ones.
1 The following abbreviations have been used in glosses: abs ‘absolutive’, aux ‘auxiliary’, comp
‘complementizer’, dat ‘dative’, det ‘determiner’, comp ‘complementizer’, erg ‘ergative’, fem
‘feminine’, fut ‘future’, gen ‘genitive’, inst ‘instrumental’, masc ‘masculine’, nom ‘nominal-
izing affix’, pl ‘plural’, q ‘question morpheme’, subj ‘subjunctive’.

complex causative verbs and causees in basque (and romance) 277
2 Complex Verb Formation in Causatives
As indicated above, Basque causative verbs are complex words containing two
verbal forms. There are reasons to treat this combination not as a result of
morphological derivation, but as a syntactic one. Section 2.1 provides some jus-
tification for this analysis, which, in turn, allows for a closer comparison with
Romance in general and Spanish in particular. The complex verb unit in the lan-
guages under discussion is examined in section 2.2, and section 2.3 focuses on
direct and indirect causation in these structures. Impersonal causatives and/or
faire par constructions are discussed in section 2.4.
2.1 Basque Causative Morphemes as Bound Morphemes

In the standard language and in dialects where arazi/erazi ‘cause’ is used, this
is a bound morpheme and must form a unit with the verbal root expressing the
caused event, making this look like a standard case of morphological causative
formation. However, all the Basque causative ‘morphemes’ mentioned so far
are verbs, and more precisely, verbs diachronically formed by affixation of the
archaic (probably morphological) causative marker -ra- to a verbal root (see
de Rijk 2008: 375–377). Thus, although the root to which -rain arazi/erazi
itself may have been attached is no longer extant as a verbal root, eragin ‘cause’
is clearly (to the linguist, not to the speaker) the archaic causative form of
egin ‘do’. Synchronically, moreover, these causative forms exist as independent
verbs, especially but not exclusively eragin and erazo. The General Basque
Dictionary (OEH) gives independent entries for all of them, with one mean-
ing ‘make do, cause, force’. This use is normal with eragin (2b) but definitely
marginal in the case of arazi (2a):
(2) a. Ene errenak ezion ikasinai biloaren motxtan baia

my daughter.in.law neg.aux learn.want hair.gen cut but
arazi naun.2
cause aux
‘My daughter in law did not want to learn how to cut hair, but I forced
her to.’
b. Beharrak beharra eragiten du.

need.erg need makes.do aux
‘Need makes one do what is necessary.’
2 Example from a mid 20th century letter to Basque linguist Azkue, in Roncalese, included in
the OEH under arazi.
278 ortiz de urbina
Eragin as an independent causative verb is standard in the case of causative

of light verb constructions where egin ‘do’ combines with bare nominal com-
plements in unergative-like expressions like lan egin ‘work, lit. do work’, negar
egin ‘cry, lit. do cry’, lo egin ‘sleep, lit. do sleep’, etc. (see Levin 1983; Laka 1993 or
Berro 2015):
(3) a. Umeak negar egin du.

child.erg cry do aux
‘The child cries. (lit. does crying)’
b. Umeari negar eragin diote.

child.dat cry do.cause aux
‘They have made the child cry.’
Outside causative unergatives, though, eragin ‘cause’ as an independent verb,

although clearly semantically linked to causation, has lost a transparent con-
nection with egin ‘do’; as a result, a productive causative where arazi is attached
to eragin is attested in the General Basque Dictionary, with the meaning of
‘cause to do, cause to move’ (4a), and in turn this has even spread to the light
verb in substandard unergative causatives (4b):
(4) a. egin deitzüdan gaizkiez oroz eta besteri eragin erazi

do aux.that evil.acts all.for and other.to do.cause cause
dütüdanez
aux
‘for all the evil acts I performed and for those that I made others per-
form’ (UskLiB 101)
b. ene tristeziak ihez eragiñ arazitzen baitarot gizonen

my sadness.erg escape do.cause cause since.aux men
konpaiñietarik
company.from
‘since my sadness makes me flee from the company of men’ (Gç 47)
The existence of causative morphemes as independent verbs, then, approxi-

mates these structures to Romance causatives more than if the causative mor-
pheme were a regular derivational morpheme like, for instance, chi in Quechua
or (t)at/(t)et in Hungarian. The order of the two verbs in Basque is, as expected
for a head last language, event-cause (which would however be the case both if
this is a morphological unit or a syntactic unit).
The latter is the approach taken already in DHM (1989), following Baker
(1988) and also in the spirit of Zubizarreta’s (1985) Complex Verb Hypothesis,
although the choice is made there for theoretical reasons. Some evidence for
this syntactic approach can be derived from the fact that VV compounding is
not found in Basque morphology outside of these cases. Moreover, there is a
correlation between the shape the event verb takes in causatives in any given
dialect and the availability of the root in the syntax of that dialect. That is, while
all dialects use the verbal root in morphological derivation, it is only found as a
free form in the syntax of eastern dialects. Subjunctive and potential forms are
then combined with the root in these dialects (har ‘take’ in 5a), but with the
participial form (which happens to be the neutral, citation form of the verb in
Basque) in western dialects (hartu in 5b):
(5) a. … hori har dezan

that take aux.comp
‘so s/he takes that’
b. … hori hartu dagian

that take aux.comp
‘so s/he takes that’
Similarly, in western dialects that use eragin as the causative verb, this is com-
bined not with the verbal root, as in the standard language, but with the per-
fective participle: we then find hartueragin rather than harrarazi ‘make take’,
or ikusieragin rather than ikusarazi ‘make see’. The use of the ‘syntactic’ verbal
base in causative formation, rather than the verbal root, reserved for morphol-
ogy in western dialects, strongly suggests causatives involve a syntactic combi-
natory process rather than a morphological one.3
An important difference between causative verbs in Basque and Romance
hacer/faire is that the former do not take tensed clausal complements, while
the latter do (6b), generally conveying a more indirect type of causation than
with tenseless ones:
3 According to the OEH, the easternmost use of eragin ‘cause’ can be found in the Labourdin
writer Urte, where, as expected, it combines with the verbal root, productive in the syntax of
that dialect. Central dialects also use this causative verb occasionally, and just like with their
main causative verb arazi, it seems to combine with either the verbal root of the participial
form depending on the syntactic productivity of the former.
280 ortiz de urbina
(6) a. Le hizo volver a Pedro.

him made return to Pedro
‘S/he made Pedro return.’
b. Hizo que Pedro volviera.

made that Pedro return.subj
‘S/he made Pedro return.’
Tensed complements with Basque causative arazi or eragin do not seem to

exist. Very occasionally, examples with egin ‘do’ can be found, as in (7):
(7) Egizu beraz guziak maita detzadan.

make then all love aux.comp
‘Make [me] love them all.’ (Brtc 207)
Similarly, in Sarasola’s (1985) dictionary, many entries for causative verbs with
erazi are provided with definitions which involve the verb egin ‘do’ taking a
subjunctive embedded clause, as ito erazi ‘cause to drown’, glossed as “make
that something drown”, higi erazi ‘cause to move’, given as “make that some-
thing move” or iraun erazi ‘cause to last’ as “make that something last”. The
glosses are as marked in Basque as their translations are in English, and look
more like a metalinguistic device than actual examples of productive lan-
guage.4
Distributionally equivalent to subjunctive tensed clauses in many respects,
tenseless nominalized complement clauses are also very occasionally found as
complements of a causative verb, instead of the general root or participial form.
As before, this possibility does not exist with arazi itself but rather with other
causative verbs, arazo ‘cause’ in (8a), eragin ‘cause’ in (8b) and eginarazi ‘cause
to do’ in (8c):
(8) a. Neuri be erazo egin daustazu emakumien aurka … esaten.

I.dat too cause do aux women.gen against saying
‘You have also made me say [speak] … against women.’ (Otx 29)
4 This use is not registered in the very comprehensive General Basque Dictionary (OEH). B. Fer-
nández (p.c.) points out, however, that it is found in the speech of young speakers of Ondarroa
Basque.
b. Irabazi nebanean poz handiagoa sorraraztea eragin

win aux.when happiness bigger appear.cause.nom cause
eban.
aux
‘When I won, this made a bigger happiness appear.’ (Euskonews Gaz-
tea, 94)
c. Eginerazi zeitan funtsen partajatzia.

do.cause aux fund.gen.pl share.nom.det
‘(S)he made me share the funds.’ (Etch 140)
Leaving these isolated cases aside, this paper will exclusively focus on the gen-
eral pattern, where the causative verb combines with a verb in its root or par-
ticipial form, as in (1b).
2.2 Causatives as Complex Verbal Forms

Already in DHM (1989) the combination of the two verbs in causative con-
structions is interpreted as a syntactic combination of two verbs, technically
expressed as the result of head-to-head movement following Baker’s (1988)
incorporation analysis. This analysis brings Basque causatives close to their
Romance counterparts with infinitival complements of causative verbs. This
initial parallelism between Basque and Romance is particularly clear with
respect to Spanish constructions with adjacent verbs like (9a), as opposed to
those like (9b) with the causee intervening between them. The first option, on
the other hand, is the only one available in other Romance languages like Ital-
ian or French (9c,d):
(9) a. Jon (le) hizo tomar la medicina a su hijo.

Jon him made take the medicine to his son
‘Jon Made his son take the medicine.’
b. Jon (le) hizo a su hijo tomar la medicina.
c. Jean a fait prendre le médicament à son enfant.

Jon has made take the medicine to his son
‘Jon Made his son take the medicine.’
d. *Jean a fait à son enfant prendre le médicament.

282 ortiz de urbina
The connection between the two verbs in Basque, however, is more com-
plete than what one finds in Romance. Thus, the type of complement taken by
the causative verb makes it impossible for the complement verb to be negated
(10b)5 or have aspectual modification or passive morphology, unlike Spanish
(10c,d,e):
(10) a. El alcohol te hace no caminar derecho.

the alcohol you makes not walk straight
‘Alcohol causes you not to walk straight.’
b. *Edariek zuzen ez ibilerazten zaitu

drinks.erg straight not walk.cause aux
‘(Intended) Drinks make you not walk straight.’
c. ??Su ceguera le hizo no haberse dado cuenta antes.

his blindness him made not have given notice earlier
‘His blindness made him not notice earlier.’
d. Esta píldora te hará estar durmiendo muchas horas seguidas.

this pill you make be sleeping many hours in.a.row
‘This pill will make you sleep for many hours in a row.’
e. ? El juez le hizo ser defendido por un abogado de oficio.6

the judge him made be defended by a lawyer of office
‘The judge made him be defended by a public defender.’
While perfective in (c) is very marginal (probably due to the temporal sequenc-
ing inherent in situations where someone is forced to do something), passive is
5 The presence of negation has the familiar effect of preventing ‘clause union’ effects, so that
clitic climbing is not possible:
(i) El alcohol te hizo entenderlo. El alcohol te hizo no entenderlo.
the alcohol you made understand.it the alcohol you made not understand it
(ii) El alcohol te lo hizo entender. *El alcohol te lo hizo no entender
the alcohol you it made understand the alcohol you it made not understand
As for Basque, negation blocks clause union in participial complements of nahi ‘want’ and
behar ‘need’: compare ikusi nahi ‘want to see’, with clause union effects and ez ikustea nahi
‘want not to see’, with the nominalized verbal form ikustea instead.
6 Zubizarreta (1985) stars causatives with embedded passives as in this example. As also appar-
ent from the discussion of matrix passives, there are important dialectal differences at work
here.
only slightly marked (e), while progressive seems perfectly acceptable (d). No
Basque counterparts can be formed for (10).
Some differences between Spanish and French causatives on one hand and
Italian on the other are brought forth in Zubizarreta (1985), who accounts
for them claiming Italian fare forms are monoclausal structures while Span-
ish and French causatives are associated “simultaneously with two structures:
one biclausal, the other monoclausal” (1985: 280). In essence, the difference
would imply a more close knit verbal union in the case of Italian as opposed
to French/Spanish. The difference can be observed in the availability of pas-
sivization of the causative verb itself. Zubizarreta’s examples below contrast
the grammatical Italian (11a) and ungrammatical Spanish (11b):
(11) a. Quei brani furono fatti leggere da Giovanni.

those passages were made read by Giovanni
‘Those passages were made to be read (by Giovanni).’
b. *La casa fue hecha construir (por Casimiro)

the house was made build by Casimiro
‘The house was made to be built (by Casimiro).’
The contrast is not found in Peninsular Spanish, where sentences like (b) are
acceptable. In the analysis presented by Zubizarreta (1985: 274), the causative
verb is a morphophonological word but acts as a bound morpheme from a
morphosyntactic perspective, which brings apparently syntactic causatives
closer to morphosyntactic ones, just as the facts in 1.1 bring the apparently
morphological causatives closer to the syntactic ones. In any event, we would
expect Basque to behave like Italian if, from a descriptive perspective, a mon-
oclausal analysis with a morpheme-like causative verb fare is assumed to
account for the Italian facts. This cannot be checked directly with passives,
very marginal in Basque, but parallel facts can be found in ‘long impersonal’
constructions, similar to Spanish ‘reflexive passives’ or English middles, where
the object of a verb becomes the subject of an intransitivized version of that
verb. As already pointed out in Aissen and Perlmutter (1986), this process can
sometimes apply across clauses, as a result of their proposed clause union or
reduction process. The examples below illustrate this monoclausal-like effect
in both Basque and Spanish using long mediopassives based on impersonal
(causee-less) causative structures, to which we will return below in section
1.4:
284 ortiz de urbina
(12) a. Liburu hauek eskola guztietan irakurrarazten dira.

book these school all.in read.cause aux
‘These books are made to be read in all schools.’
b. Se hacen pulir los diamantes en Holanda.

se make polish the diamonds in Holland
‘Diamonds are made to be polished in Holland.’
The logical objects of read (‘those books’) and of polish (‘diamonds’) are now
subjects of the causative (or of the complex verb), as shown by the fact that
they agree with the latter, even though, in principle, they are not its argu-
ments.
2.3 Direct and Indirect Causation in Tenseless Complement Causatives

As indicated, Romance allows for tensed subjunctive clauses as complements
of causative verbs, in sentences like (6b). Causatives with embedded sub-
junctive complements convey a less direct type of causation than those with
infinitival complements, in accordance with Comrie’s (1989) generalization. A
consequence, according to Bordelois (1974), Treviño (1994) or Tubino (2011),
is that structures without animate causees in the infinitival complement are
deviant. This is exemplified below with expletive subjects (13a), passive infini-
tive complements (13b) and dative experiencer constructions with psych verbs
(13c):
(13) a. *El mago hizo empezar a llover.

the magician made begin to rain
‘The magician made it begin to rain.’
b. *La policía hace (a) la ley ser obedecida.

the police makes (to) the law be obeyed
‘The police makes the law be obeyed.’
c. *Su ruido hace a las discotecas molestar a Pedro.

their noise makes to the discos bother to Pedro
‘Their noise makes discos bother Pedro.’
Pragmatic considerations seem to be involved. Thus, Tubino (2011) shows that

an animate causee in some constructions is acceptable with inanimate causers
but less so with animate ones as shown in (14):
(14) a. Su falta de cuidado / ? María le hizo ser descubierto.

his lack of care Maria him made be discovered
‘His lack of care/María made him be discovered.’
b. Tu ironía / ? Pedro le hizo enfadarse a Juan.

your irony Pedro him made get.mad to Juan
‘You irony/Pedro made Juan get mad.’
Conversely, the embedded expletive in (13a) improves in an impersonal context

like (15), which indicates that no such causer exists:
(15) No se puede hacer llover cuando se quiere.

not se can make rain when se wants
‘One cannot make it rain when he wants.’
Regardless of the actual constraints that play a role in the more or less deviant
status of some of the previous sentences, what is clear is that they disappear
when tensed complement clauses are used, so that (16), for instance, is accept-
able:
(16) El ruido hace que a Pedro le molesten las discotecas.

the noise makes that to Pedro him bother the discos
‘Their noise makes discos bother Pedro.’
The Basque causative with arazi ‘cause’ also expresses a fairly direct type of cau-
sation, less felicitous where pragmatic considerations do not make this type of
causation possible. In this, then, it clearly patterns with Romance cause+infini-
tive structures, rather than with tensed embedded patterns. This is shown in
(17) with stative complements of the causative verb, which Oyharçabal (2003)
identifies as excluded from causative structures. The stative predicate in the fol-
lowing examples is paired with clausal subjects for the causative verb, forcing
a non-agentive interpretation:
(17) a. ??Txikitan jogurta asko jateak Mikel(i) altua izanarazi

youth.in yoghourt much eat Mikel(dat) tall be.cause
du(dio).
aux
‘Eating yoghourt as a child made Mikel be tall.’
286 ortiz de urbina
b. ? Comer yogur de pequeño le ha hecho ser alto a Mikel.

eat yog. as small him has made be tall to Mikel
‘Eating yoghourt as a child made Mikel be tall.’
c. Comer yogur de pequeño ha hecho que Mikel sea alto.

that Mikel be tall
(18) a. ??/* Inguru eleanitz batean bizi izateak hizkuntza asko

environment multilingual one.in live be language many
jakinarazi dit.7
know.cause aux
‘Living in a multilingual environment has made me know many lan-
guages.’
(19) b. ? Vivir en un entorno multilingüe me ha hecho saber muchas

live in an env. multilingual me have made know many
lenguas.
languages
‘Living in a multilingual environment has made me know many lan-
guages.’
c. Vivir en un entorno multilingüe ha hecho que sepa muchas

that I.know many
lenguas.
languages
The corresponding Spanish (b) examples sound less deviant than the Basque
(a) examples, while the (c) sentences with tensed complements are grammat-
ical in Spanish. Similarly, weather expressions are formed in Basque with exis-
tential constructions which cannot be causativized; alternatives resort to egin
‘do’ for the root verb:
(20) a. Euria / hotza da.

rain / cold is
‘It is raining/cold.’
7 A verb like jakinarazi ‘cause to know’ is acceptable in the meaning ‘communicate’, shared
with Spanish, but not in the stative reading in (18).
b. Ipar haizeak euria / hotza *izanaraziko / eginaraziko du.8

north wind.erg rain / cold be.cause.fut / do.cause.fut aux
‘The north wind will make it rain/be cold.’
In fact, izan ‘be’ is particularly uncommon with a causative verb regardless of

whether it denotes states or activities, as noted in Oyharçabal (2003), although
they seem acceptable in Spanish. The online dictionary of Euskaltzaindia
(Royal Academy of the Basque Language), contains a handful of examples for
this entry, all of which except for one (given here as 23) actually correspond to
‘have’:9
(21) a. *Medizinak logurea izanarazi dit.10

medicine sleepyness be.cause aux
‘The medicine made me feel sleepy.’
b. La medicina me hizo tener sueño.

the medicine me made have sleep
‘The medicine made me feel sleepy.’
(22) a. Irakasleak ??arduratsua / arduratsuagoa izanarazi ninduen.

teacher careful / more.careful be.cause aux
‘The teacher made me be careful/more careful.’
b. *Aitak abokatua izanarazi ninduen.

father.erg lawyer be.cause aux
‘My father forced me to be a lawyer.’
8 In the case of hotz ‘cold’, egin ‘do’ itself is commonly used instead of izan ‘be’, so the
causative eginarazi is not totally unexpected. It is in the case of euria, though. Whether
because it is a recent use of egin or because it is not a light verb here, unlike in expressions
like negar egin ‘cry, lit. make cry’, the fact is that this egin is never causativized as eragin.
9 There is no tenseless form of the transitive auxiliary; the intransitive auxiliary izan ‘be’ is
used instead. For texts up to the 20th century, the OEH gives three examples, two from
Bible translations and a third one in a modern philosophical book in the context of gods
‘making be’, i.e., creating, without causee. The Ereduzko Prosa Gaur online corpus gives 13
instances from year 2000 to the present, all of them in translations or, two of them, in a
philosophical work. None in the written press.
10 At best, the sentence may be interpreted as the causative of the verb used in the substan-
dard expression logure(a) dut ‘I have sleep’, with a transitive ‘have’ probably patterned after
Spanish tener sueño ‘be sleepy (lit. ‘have sleep’)’ rather than the standard Basque intransi-
tive ‘be’ of logure izan ‘be sleepy’ (lit. ‘be sleep-wanting’).
288 ortiz de urbina
c. El profesor me hizo ser cuidadoso / más cuidadoso.

the teacher me made be careful / more careful
‘The teacher made me be careful/more careful.’
(23) haiek atzarriago eta kartsuago izanarazteko

they smarter and more.active be.cause.to
‘… so that they are smarter and more active’
The incremental scale in the comparatives in (22) and (23) has the effect of
making the Basque causatives of izan ‘be’ more acceptable.
As for ‘inversion’ predicates with experiencer datives as in (13c), they are
often states and, in any event, one need not expect the theme subject of such
predicates to act as causee, but rather the experiencer. The theme causee
(example (24a,b), in the intended interpretation) with gustar is as unaccept-
able in Basque as in Spanish, and the experiencer causee is slightly better
(24c,d). Again, the que clause is perfect (25):
(24) a. *Bere etxeko giroak musika gustarazi zion Mikeli.

his house.of atmosphere music like.cause aux Mikel.dat
‘His house atmosphere made music be likeable to Mikel.’
b. *Su ambiente familiar le hizo a la música gustarle a

his atm. familiar him made to the music like.him to
Mikel.
Mikel
‘His family atmosphere made music be likeable to Mikel.’
c. ??Bere etxeko giroak Mikeli musika gustarazi zion.

his house.of atmosphere Mikel.dat music like.cause aux
‘His house atmosphere made Mikel like music.’
d. ? Su ambiente familiar le hizo a Mikel gustarle la música.

his atm. familiar him made to Mikel like the music
‘His family atmosphere made Mikel like music.’
(25) Su ambiente familiar hizo que a Mikel le gustara la música.

that to Mikel him like the music
In spite of apparent differences depending on the degree of control of non-

agentive, cause-like subjects of the causative verb, it seems clear that Basque
arazi ‘cause’ causatives can be directly connected with Spanish and perhaps
Romance make+infinitive causatives.
2.4 Causative, Impersonal and Passive

Within Romance, a further split separates passive-like causatives like (26) from
the active-like ones in (9) above:
(26) a. Jean a fait prendre le médicament par son enfant.

Jean has made take the medicine by his son
‘Jean has made his son take the medicine.’
b. Juan hizo recoger el paquete (por uno de sus empleados).

Juan made pick the package by one of his employees
‘John had the package picked up (by one of his employees).’ (Tubino
2011:220)11
Examples like (b) are quite marked in Spanish, and are found at best in very
formal contexts.12 Notice that this so-called passive causative does not occur
with passive morphology on the embedded verb; the latter, as indicated above,
is possible with an animate causee:
(27) a. El juez me hizo ser defendido por un abogado de oficio.

the judge me made be defended by a lawyer of office
‘The judge made me be defended by a public defender.’
b. El médico le ha hecho a Pedro ser operado de urgencia.

the doctor him has made to Pedro be operated of urgency
‘The doctor made Pedro be operated on urgently.’
In contrast, the examples in (26) can have a by-phrase which does not co-occur
with passive morphology. The presence of a passive agent adjunct like par son
11 English does not admit ‘subjectless’ complements of make, so (i) is ungrammatical. This
is why a translation with have is given:
(i) *The City Hall made repair the house
Passive complements of make are of course possible, with overt passive morphology and
intervening NP; in those cases, however, make usually requires a to-infinitive in standard
dialects:
(ii) The City Hall made the house to be repaired
12 In this they resemble ‘reflexive passive’ impersonals with overt agentive phrases like Se
firmó la paz por los embajadores.
290 ortiz de urbina
enfant ‘by his son’ is only one aspect of the characterization of these struc-
tures, and in fact these structures stand out as a separate type even if there
is no overt agent expressed. The reason is that, crucially, the causee forced to
act by the causer is not expressed as a core argument of the faire+verb unit,
so these constructions are impersonal, causee-less in that sense (not necessar-
ily agentless), as discussed in Zubizarreta (1985). Thus, the following example
would also qualify as a ‘passive’ causative of this type:
(28) El ayuntamiento hizo reparar la casa.

the city.hall made repair the house
‘The City Hall had the house repaired.’
Since what is left unexpressed is the external argument (logical subject) of

the embedded verb, impersonal ( faire-par) causative constructions subsume
impersonal causatives of embedded intransitive verbs, where no by-phrase is
of course possible. Some examples are given below with both unaccusative and
unergatives verbs. Unaccusative verbs, ungrammatical in this construction in
French, do seem to be less acceptable than unergatives in Spanish, but not nec-
essarily ungrammatical:
(29) a. ? El hielo hacía resbalarse en las aceras.

the ice made slip in the sidewalk
‘The ice made [people] slip in the sidewalks.’
b. Demasiadas veces, las caravanas hacen llegar tarde al trabajo.

too.many times the traffic.jams make arrive late to work
‘Too often, traffic jams make [people] arrive late to work.’
c. ??Un contexto familiar pobre hace nacer con muchas

a context family poor make be.born with many
desventajas.
disadvantages
‘A poor family context make [one] be born with many disadvantages.’
(30) a. En ese caso, el reglamento hace jugar en el campo contrario

in this case the regulation makes play in the field opposite
(*por el equipo).
by the team
‘In this case, the regulations make [teams] play in the other team’s
field.’
b. Esta dirección hace trabajar hasta muy tarde (*por el

this management makes work until very late by the
personal).
staff
‘This management team makes [staff] work until very late.’
The constraints on unaccusative faire-par causatives in (29) seem to be a sub-

case of the more general difficulty in forming impersonal constructions for
verbs without external arguments, as discussed first by Perlmutter (1978) in
connection to Dutch impersonals.
There are no causative constructions in Basque where an agent/causee may
be recoverable as an adjunct, so constructions like (26) are not found.13 How-
ever, we can still draw a parallelism if impersonal causatives like (28) actually
correspond to the same type. Similar impersonal causatives constructions are
also available with arazi ‘cause’ in Basque. This is exemplified with embedded
transitive (31), unaccusative (32) and unergative (33) verbs:
(31) a. Zuzendariak txosten berri bat idatzarazi zuen.

director report new one write.cause aux
‘The director made [someone] write a new report.’
b. Haurrak katua hilarazi du.

child.erg cat die.cause aux
‘The child made [someone] kill the cat.’ (Oyharçabal 2003, ex. (2b))14
(32) a. ? Sarritan, pilaketek lanera berandu helarazten dute.

often jams.erg work.to late arrive.cause aux
‘Often, traffic jams make one arrive late to work.’
13 Ortiz de Urbina (2003: 604) brings forth the following example as a possible candidate for
this construction:
(i) Mutilez zain-arazten zitian bere arthalde handiak
boys.inst care-cause aux his flock large
‘He had his large flocks looked after by boys’ (Eliss. P.A.: 80)
The instrumental is used for passive agents in the dialect in question; it is possible, how-
ever, to regard this as a normal impersonal causative with an instrumental reading of
mutilez.
14 Oyharçabal’s example (b) is ambiguous between the intended impersonal interpretation
and an irrelevant interpretation where causative arazi is added to an inchoative unac-
cusative hil ‘die’. The causative version of an inchoative is possible, though less common
than the lexical causative pattern (meaning kill). See Berro (2015).
292 ortiz de urbina
b. ? Pobreziak bizitza negargarri baten ondoren hilarazten du.

poverty.erg life deplorable one.gen after die.cause aux
‘Poverty makes one die after an awful life.’
c. ??Sekta horrek eritetxeetan gabe, etxean jaioarazten du.

sect that.erg hospitals.in without home.at be.born.cause aux
‘That sect makes [babies] be born in hospitals rather than at home.’
d. Alkoholak ezbeharrez ahantzarazten du.

alcohol.erg mishaps.about forget.cause aux
‘Alcohol makes one forget about mishaps.’
(33) a. ? Sarritan, ikastetxeetan futbolean jokarazten dute.

often schools.in football.at play.cause aux
‘They often make [students] play football in schools.’
b. Armadetan norberaren sinesmen kontra ere borrokarazten

armies.in one.gen beliefs.gen against also fight.cause
dute.
aux
‘In armies they make [people] fight even against one’s own beliefs.’
c. Ordutegia hobetzeko, zazpietan afalarazi beharko lukete.

timetable improve.to seven.at dine.cause need.fut aux
‘To improve timetables, they should make people have dinner at seven.’
As the examples show, causativized impersonal unaccusative verbs are quite

marked in these constructions. They seem most easily constructed in generic
contexts, with the help of genericity activating phrases as in (32), but even so
the results are quite deviant with them if compared with unergative predicates
like those in (33).
There are differences between regular infinitival causatives and impersonal/
passive causatives (see Folli & Harley 2007 for Italian, adapted in Tubino 2011
for Spanish), related to the semantic role of the subject on one hand and pas-
sivization on the other. Beginning with thematic roles, causes (as opposed to
agents) are claimed to be ruled out as subjects of impersonal causatives, but are
grammatical with regular infinitival causatives. The distinction in the former is
shown by examples like the following (Tubino 2011: 227):
(34) Juan / * su enfado hizo castigar al niño (por el profesor).

John / * his rage made punish to.the child (by the teacher)
‘John/*his rage had the child punished (by the teacher).’
Other pragmatic factors may be involved, since causes are acceptable in other
sentences. In the following sentence the animate noun phrase could be inter-
preted as an agent or perhaps also as a cause, without apparent effect on the
grammaticality of the sentence; the inanimate subject would be a cause:
(35) El vino / los buenos amigos hace(n) olvidar las penas.

det wine / det good friends make forget det pains
‘Wine/good friends make one forget one’s pains.’
Similarly, although most of the Basque acceptable impersonal causatives in (31)

through (33) do have agent subjects, this does not seem like a necessary condi-
tion:15
(36) Berak / bere presentziak pena guztiak ahantzarazten ditu.

(S)he his / her presence pain all forget.cause aux
‘He/his presence makes one forget all pains.’
This sentence seems acceptable regardless of the cause or agent interpretation

of the subject. This area, therefore, requires further research, at least in the
case of Basque.16 The second difference mentioned in Folli & Harley (2007)
relates to the possibility of passivization in impersonal/ fair-par causatives.
We have already seen examples where the embedded infinitival complement
is passivized (see example (27)). However, the differences emerge when pas-
sivization of the causative verb itself is considered. This is only possible in
impersonal causatives (a), but not in ‘personal’ ones (b):
(37) a. Il pacchetto fu fatto arrivare (da Gianni).

the package was made arrive by Gianni
‘The package was made to arrive (by Gianni).’
15 The unaccusative verbs in (32) have a mixture of causes and agents; changing the former
to the latter does not have any effect on the acceptability of these sentences.
16 With respect to direct and indirect causation, Oyharçabal (2003) points out, quoting Dan-
los (2001), that “causality in the real world resembles a chain at the end of which it is always
possible to attach a further link”.
294 ortiz de urbina
b. *Maria fu fatta mandare un pacchetto (da Gianni).

Maria was made send a package (by Gianni)
‘Maria was made to send a package (by Gianni).’
Passives of causative hacer have an uncertain status in Spanish (Torrego 1998;

Treviño 1994; Tubino 2011), but, as the latter claims, there seems to be a differ-
ence so that passives in impersonal causatives (38), uncommon though they
are, seem more acceptable than those in regular infinitival complements (39):
(38) a. El palacio fue hecho construir por Carlos V.

the palace was made build by Carlos V
‘The palace was made to be built by Carlos V.’
b. El edicto fue hecho publicar en todos los diarios.

the ruling was made publish in all the newspapers
‘The ruling was made to be published in all the newspapers.’
c. El himno fue hecho cantar en todas las escuelas.

the anthem was made sing in all the schools
‘The anthem was made sing in all schools.’
(39) a. *El escuadrón fue hecho andar / correr / marchar varios

the squadron was made walk / run / march several
kilómetros.
kilometers
‘The squadron was made to walk/run/march several kilometers.’
b. Juan fue hecho venir por su hijo.

Juan was made come by his son
‘Juan was made to come by his son.’ (Trevino 1994; Tubino 2011)
c. *Juan fue hecho llegar tarde.

Juan was made arrive late
‘Juan was made to arrive late.’
d. *Los estudiantes fueron hechos cantar en todas las escuelas.

the students were made sing in all the schools
‘The students were made to sing in all schools.’
e. *Juan fue hecho reparar el coche.

Juan was made repair the car.
f. *El coche fue hecho reparar a los mecánicos.

the car was made repair to the technicians
‘The car was made to be repaired by the technicians.’
According to Tubino (2011), only internal arguments of the impersonal/ faire-

par construction can become matrix passive subjects. This excludes (39f),
where the presence of the causee indicates this is not an impersonal causative.
Since the faire-par construction is characterized as impersonal in the sense of
having no external subject argument, unaccusatives like (39b) are included in
this type and expected to be acceptable, as Tubino (2011) claims for this exam-
ple. However, other unaccusative subjects do not seem to share this behav-
ior (39c). These details aside, though, there does seem to be an asymmetry
in the behavior of the two constructions with respect to matrix passiviza-
tion.
Examples like (12) above already show that even if passives are not produc-
tive in Basque, mediopassive impersonal structures can be formed by mak-
ing the embedded object subject of the matrix verb. We give more exam-
ples in (40) and (41) below. As (41) shows, mediopassives are just as easily
acceptable when the causee is expressed as in the impersonal counterparts in
(40):
(40) a. Pena zaharrak pena berriekin ahantzarazten dira.

pain old pain new.with forget.cause aux
‘Old pains are made to be forgotten with new pains.’
b. Errefuxiatuentzako kanpamenduak hiri bazterretan eraikiarazi

refugees.for camps town margins.in build.cause
ziren.
aux
‘Refugee camps were made to be built in the town margins.’
(41) a. Kanpamenduak hiri bazterretan eraikiarazi zitzaizkien

camps town margins.in build.cause aux
langileei.
workers.dat
‘Camps were made to be built by workers.’
296 ortiz de urbina
b. Ikasleak garaiz helarazi beharko lirateke eskolara.

students on.time arrive.cause need.fut aux school.to
‘Students should be made to arrive to school on time.’
c. Intxaurrak makilez jausiarazten dira.

walnuts sticks.with fall.cause aux
‘Walnuts are made to fall with sticks.’
No contrast between the personal and impersonal causative constructions

therefore emerges in Basque mediopassive structures. The complex causative
verb behaves in them like a regular transitive verb, and the mediopassive
eliminates the causer, leaving it as an implicit argument and detransitivizes
the structure. This contrasts with Spanish, where mediopassives based on
impersonal/ faire-par seem far better than those based on objects of regular
causatives.17 Plural nominals are used throughout these examples to check eas-
ily whether the causative verb agrees with them in a (medio)passive structure;
the sentences would all be acceptable with singular agreement on hacer ‘cause’;
this is, however, a different, impersonal structure:
(42) a. Se hacían construir puentes en todas partes.

se made build bridges in all sides
‘Bridges were made to be built everywhere.’
b. Esos diamantes se hacen pulir en Holanda.

those diamonds se make polish in Holland
‘Those diamonds are made to be polished in Holland.’
(43) a. ? Se hicieron construir puentes a los obreros.

se made build bridges to the workers
‘Bridges were made to be built by workers.’
b. ??Se hacen publicar novelas románticas a los escritores.

se make publish novels romantic to the writers
‘Romantic novels are made to be published by writers.’
17 Notice that in all cases the embedded verb is in an active form, unlike in some of the
translations; the causee is given as a by phrase as a way to make the intended meaning
understood.
c. Se hacen caer las nueces con palos.

se make fall the walnuts with sticks
‘Walnuts are made to fall with sticks.’
d. Se hicieron llegar refuerzos por mar.

se made arrive reinforcements by sea
‘Reinforcements were made to arrive by sea.’
e. ??En el circo se hacen aparecen conejos en los sombreros.

in the circus se make appear rabbits in the hats
‘In the circus, they make rabbits appear inside hats.’
The last three examples contain unaccusative verbs, and although judgements
are not so clear for all examples/speakers, their inanimate causees, which
would appear as direct objects in the personal causatives, seem to be able
to become subjects of the mediopassive construction. In any event, leaving
the status of Spanish causativized unaccusatives as faire-par or impersonal
causatives aside, the contrast obtained by the presence of the causee in (a,b)
is clear and may indicate a different status/position for causees in Basque
and Spanish causatives. Summarizing, in spite of the apparent differences in
causativization strategies, if the data described here are correct, Basque and
Spanish (as well as Romance) causatives with tenseless complements share
important fine-grain features, as well as some differences. Accounting for them
goes beyond the limits of the modest descriptive goal of this paper, and the
complexity of acceptability patterns makes the enterprise more difficult. How-
ever, it seems we might be able to gain insights about both languages and about
causative structures in general from the close examination of their grammars.
In the following section we turn to an area where contact may be having a more
distinct impact on the grammar of Basque, namely causee marking, with par-
ticular emphasis on agreement conflicts created in Basque as a result of dative
doubling.
3 Causee Marking in Basque (and Spanish) Causatives
Whether formed in a strictly morphological way by adding a bound morpheme

to a verbal base, through verb/clause union processes from what at first look
like biclausal structures or by selection of VP or phrase (or even Root) com-
plements for a Cause head as in Pylkkänen (2008), complex verb formation
in causatives involves adding an extra argument to those of the event predi-
298 ortiz de urbina
cate. This can be effected in widely different ways depending on one’s theoret-
ical persuasion, but the general layout is fairly clear. As discussed in Comrie
(1989) and Dixon & Aikhenwald (2000), the causer argument of the causative
verb/morpheme/head is aligned with the subject function, and the causee can-
not be a subject unless a fully biclausal structure is formed. If this is not the
case, typologically the most widespread situation is for intransitive event sub-
jects to occur as objects of the complex causative verb and for transitive sub-
jects to be found as indirect objects. The following subsections will focus on
causee case in Basque causatives. In section 3.1 the Standard Basque pattern in
presented, focusing on causee marking and case doubling phenomena. DOM—
(Differential Object Marking) related dialectal or substandard phenomena will
be discussed in section 3.2.
3.1 Increasing Arguments by One: Causee Marking

Standard Basque presents a fairly typical situation for valency increasing via
causative formation. We examine causativization of basic intransitive and
monotransitive predicates in section 3.1.1, concentrating mostly on the appear-
ance of dative causees. Section 3.1.2 is devoted to causativization of argument
configurations which already have a dative participant, focusing on the inter-
action between dative causees and the original datives of those configurations.
3.1.1 Causativizing Intransitive and Monotransitive Verbs

Causativized verbs are always transitive, since even in the case of intransitive
verbs, the basic intransitive argument will be increased by one, the causer. The
‘object’ like argument of the unaccusative verb will no longer surface as a sub-
ject, but will remain object, so the resulting complex verb resembles a typical
monotransitive in terms of case marking and auxiliary configuration:
(44) a. Jon garaiz etorri da.

Jon.abs on.time arrive aux
‘Jon arrived on time.’
b. Edurnek Jon garaiz etorriarazi du.

Edurne.erg Jon.abs on.time arrive.cause aux
‘Edurne made Jon arrive on time.’
Unergative predicates with agentive subjects are primarily constructed with

egin ‘do’ as a light verb, taking a bare object indicating the action performed:
lo egin ‘sleep, lit. do sleep’, lan egin ‘work, lit. do work’, etc. The connection
between unergativity and Basque egin constructions was first discussed in
Levin (1983), and since then many alternative analyses have been proposed (see
Laka 1993; Etxepare 2003 or Berro 2015, among many others). Levin also noticed
that some unergative predicates also (or exclusively) exist as simplex verbs, tak-
ing then ergative subjects and transitive auxiliaries: jolastu ‘play’, saltatu ‘jump’,
distiratu ‘glitter’, iraun ‘last’, etc. There is more variation for the causative of
these verbs. De Rijk (2008: 378–380), from where (45b) is taken, states than
in northern dialects they tend to behave like unaccusatives (a), although occa-
sionally now, and more often in the past, dative causees were also found. In
the central dialects (Gipuzkoan mostly, since arazi ‘cause’ gives way to eragin,
arazo ‘cause’ further to the west), unergative causees tend to be marked dative,
behaving then in causatives just like the ergative subjects of transitive verbs
(b):
(45) a. Eguzkiak Sacre Coeur basilikaren kupula distirarazten zuen.

sun.erg basilica.of dome glitter.cause aux
‘The sun made the Sacre Coeur basilica dome glitter.’
b. Norbaitek edo zerbaitek iraunarazi dio hizkuntzari.

someone.erg or something.erg last.cause aux language.dat
‘Someone or something made the language last.’
As also pointed by de Rijk, the normative advice is to use datives with animate
causees and absolutive with inanimate ones, introducing a DOM-like distinc-
tion to which we will return in 3.2. Both unaccusative and unergative verbs
may take yet a second argument in the dative; these will be discussed in the
following section, once dative causees are introduced for regular monotransi-
tive verbs, to which we turn now.
The external argument of a transitive verb appears as a dative causee in the
causative version of that verb, as shown in (46a):
(46) a. Irakasleak ikasleei liburu hori irakurrarazi die.

teacher.erg students.dat book that read.cause aux
‘The teacher made the students read that book.’
b. Irakasleak ikaslei liburua irakurri die.

teacher.erg students.dat book.the read aux
‘The teacher read the book to the students.’
This pattern is general. The embedded verb already includes a direct object and
the logical subject of irakur- ‘read’ cannot appear as a direct object, given that
300 ortiz de urbina
direct object doubling is not found in Basque. The resulting configuration is

then identical to that of regular ditransitive predicates like eman ‘give’ or the
ditransitive use of irakurri ‘itself’ (46b).
3.1.2 Causativizing Verb Configurations Including Datives

Each of the three previous complement configurations exists in another ver-
sion including a dative. Ergative/dative configurations, absolutive/dative and,
finally, ergative/absolutive/dative ditransitive patterns will be examined in
turn.
3.1.2.1 Causativizing Basic Ergative/Dative Configurations

Beginning with ergative/dative complementation types, some unergatives
verbs take dative objects (Ortiz de Urbina & Fernández 2016). One such verb
is begiratu taking complements in the dative and, occasionally for locative
nouns, in the allative. As the following examples show, when causativized, the
causee is also marked dative. The basic pattern is exemplified in (47a), while
the causative version is found in (47b,c):
(47) a. Jonek berari begiratu zion.

Jon.erg he.dat look.at aux
‘Jon looked at him.’
b. Lurrera begirarazi zioten (berari).

floor.to look.at.cause aux he.dat
‘(They) made him look down to the floor.’
c. Atzamarrez kokotsetik heldu eta niri begirarazi

fingers.with chin.by grab and I.dat look.at.cause
nion.
1erg.3dat.pst
‘I grabbed his chin and made him look at me.’
The (b) example illustrates the causative of the basic unergative configuration
of this verb, with an allative complement; as described in the previous section,
the causee appears in the dative. When the verb takes a dative complement, a
dative doubling situation emerges, exemplified here in (c). The auxiliary gloss
has been given in more detail to point out that when two datives vie for agree-
ment, it is the causee that becomes cross-referenced; in this example dative
agreement is with the unexpressed berari ‘to him’, while the goal niri ‘to me’
does not trigger any marker on the auxiliary. If we eliminate the causee by using
an impersonal causative, the original dative is still unable to agree, so that in the
following examples the dative auxiliary is flagged as ungrammatical. If used,
the dative is interpreted as referencing an unexpressed causee:
(48) Eskola honetan irakasleari begiratu / entzunarazten dute /

school this.in teacher.dat look.at hear.cause aux /
*diote.
aux.3sg.dat
‘In this school they make [one] look at/listen to the teacher.’
(49) Hatsarre moral hauek eskubiderik gabeko jendeari lagunarazten

principle moral these right without people.dat help.cause
dute / *diote.
aux / aux
‘These moral principles help people without rights.’
The dativeless auxiliary in sentences of this type is exceptional in Standard

Basque, where datives must be cross-marked in the inflection, and these sen-
tences are slightly dubious, perhaps the result of some repair strategy. However,
the version with the dative auxiliary is clearly ungrammatical. Similar facts can
be observed in Spanish, although in that case it is the position of clitics that is
significant:
(50) En esta escuela (*les) hacen llamar(les) a los padres

in that school them.dat make look.at.them.dat to.the parents
por teléfono.
by phone
‘In this school they make (people, teachers) phone parents.’
As indicated, the clitic referring to the indirect object of llamar ‘call’ must
remain attached to the lower verb, and may not climb to the matrix causative
verb, as if ‘clause union’ had not taken place (Aissen & Perlmutter 1983).
3.1.2.2 Causativizing Absolutive/(Experiencer, Goal, Ethical) Dative

Configurations.
On top of the basic unaccusative configuration, with intransitive auxiliary and
absolutive subject, some verbs display a bivalent pattern with an added dative
participant. The dative argument may be both an optional goal with verbs
like joan ‘go’, hurbildu ‘approach’ or an obligatory experiencer argument in an
‘inversion’ predicate like gustatu ‘like’, iruditu ‘seem’ or interesatu ‘interest’. Ethi-
302 ortiz de urbina
cal/interest datives may also be found with most verbal configurations and will
be addressed briefly below.
Beginning with inversion predicates, we already mentioned that the dative
argument can only marginally be taken to be a causee, as in (24c), repeated
here as (51a); attested examples are given in (51b,c):
(51) a. ? Bere etxeko giroak Mikeli musika gustarazi zion.

his house.of atmosphere Mikel.dat music like.cause aux
‘His family background made Mikel like music.’
b. Behar da [poesia] publikoari gustarazi.

need aux poetry public.dat like.cause
‘One needs to make the public like poetry.’
c. Arren gustaraz diezadazu … kausitzen dena.

please like.cause aux find aux.comp.det
‘Please, make me like what is found …’ (Arbill 43, III 15, 22)
Dative is expected here for the causee, since there is already an absolutive
element, and this case also fits its experiencer status. In order to check for
dative doubling configurations, we would need an example where the theme is
interpreted as a causee, a situation which, as indicated in 1.3 above, is usually
unacceptable in Spanish with infinitival causatives. The type of interpretation
sought is similar to that in English (52), where the intervening NP would corre-
spond to the causee of the structures in question in Basque and Spanish:
(52) They made the teacher be liked by the students.
There is no thematic link between make and the following NP in English, while
the causee does have one with the complex verb in both Spanish and Basque;
these sentences are unacceptable:18
18 There are nonetheless occasional attested examples which may correspond to this inter-
pretation, without an auxiliary:
(i) [ikasleei] hizkuntza hurreratu behar zaie, gustarazi
students.dat language draw need aux like.cause
‘language must be brought over to the students, made be likeable’
The interpretation seems to be that something must be done to make the language more
appealing to students, rather than simply making the students like the language, and to
the extent that this is a possible (and different) interpretation, it would perhaps be an
(isolated) example of theme ‘causee’.
(53) a. ??/*Hoy en día al profesor le hacen gustarles a los

today in day to.the teacher him make like.them to the
estudiantes.
students
‘Now a days they force the teacher to be liked by students.’
b. *Gaur egun irakaslea ikasleei gustarazten diete.19

today day teacher.abs students.dat like.cause aux.3pl.dat
‘Now a days, the force the teacher to be liked by students.’
The Spanish sentence is interpretable (with some effort) thanks to the two cli-
tics and their placement. The single agreement marker in the Basque examples
makes interpretability more complicated.
The second type of ‘bivalent unaccusative’ is that where a goal dative is
added to an intransitive verb (54a).20 While the directional complement is
more commonly expressed in the allative, the dative exists as a formal register
possibility. Given ergative marking, the absolutive intransitive subject remains
absolutive, now as object of the transitive causative verb. The combination with
the ergative causer and the dative goal would produce a pattern identical to that
of ditransitive predicates (54b); however, the structure is considered deviant by
speakers:
(54) a. Jon lagunei hurbildu zaie.

Jon.abs friends.dat approach aux
‘Jon approached his friends.’
b. *Beharrak Jon lagunei hurbilarazi die.

need.erg Jon.abs friends.dat approach.cause aux.3pl.dat
‘Need made Jon approach his friends.’
c. *Mirenek sua Mikeli hurbilarazi zion.

Miren.erg fire.abs Mikel.dat approach.cause aux.3sg.dat
‘Miren brought the fire close to Mikel.’
Since the basic non-causative pattern is rather literary, judgments may be

uncertain; but agreement with the dative seems to be the problem here: not
19 Using a dativeless transitive auxiliary like dute does not improve the sentence.
20 See Rezac (2009), who follows Ormazabal & Romero’s (2007) derivational approach for
Double Object Constructions, deriving goal datives from PP s.
304 ortiz de urbina
all causees are marked dative in the standard language, but all datives seem
to be interpreted as causees, and this is not the interpretation in (54b).21 In
(54c) an inanimate causee is intended, with a goal dative; the sentence is only
interpretable as the causative of a monotransitive use of hurbildu, again with
a causee dative. An indication that unacceptability arises from the unintended
connection between dative and causee is that if dative agreement is eliminated,
the sentence is perfectly interpretable and acceptable for many speakers in the
intended sense. The connection between dative and causee is also made evi-
dent by looking at impersonal, causee-less structures. Again, their dative is not
interpreted as goal but as causee, so (55) is unacceptable in the intended inter-
pretation:
(55) Beharrak lagunei hurbilarazten *die /

need.erg friends.dat approach.cause aux.3erg.3pl.dat /
du.
aux.3erg
‘Need makes [one] approach friends.’
More specifically, as discussed above, the agreement dative corresponds to the

causee. The sentence becomes acceptable if dative agreement is eliminated
(the auxiliary du only includes information about the subject and an unspeci-
fied object). As discussed in the following section, this sentence is acceptable
in the western varieties that mark intransitive subject causees as dative. In
those varieties, this sentence is the causative of a basic monoargumental unac-
cusative, with lagunei ‘to the friends’ interpreted either as a causee coreferen-
tial with the dative in the inflection (‘make the friend approach’) or as a goal. In
this case, however, the dative in the inflection must be understood as marking
an unexpressed causee. The connection may be relaxed for some speakers, so
they find (56) acceptable:
(56) Beharrak Mikel diru eske etorrarazi zidan.

need.erg Mikel money asking approach.cause aux.1dat
‘Need made Mikel come to me asking for money.’ (Ortiz de Urbina 2003)
The scant use of the basic goal dative configuration in non-formal contexts only
adds to the complexity of factors which may help us understand speaker judge-
ment variability in these cases.
21 This applies to argumental datives. Ethical datives can be found in causative structures
without receiving a causee interpretation. We return to them below.
To finish with causativized absolutive/dative basic configurations, it may be

worthwhile considering, at least in passing, ethical datives, typically indicat-
ing external possession (alienable or inalienable) or interest relations.22 The
most widespread type of ethical dative, geographically and diachronically, is
that where the interest relation is predicated of an absolutive element, whether
the intransitive subject (57a) or transitive object (57b); there is more variation
in acceptability of a dative associated with an ergative (57c,d):23
(57) a. Umea erori zait.

child fall aux.1dat
‘The child has fallen (on me).’
b. Umea eraman didate.

child take.away aux.1dat
‘They have taken (my) child away.’
c. Umeak ondo borrokatu dit.

son.erg well fight aux.1dat
‘My son has fought well (on me).’
d. Jonek indabak gorrotatzen dizkit.

Jon.erg beans hate aux.1dat
‘(My) Jon hates beans.’ (Fernández & Ortiz de Urbina 2010: 181)
Albizu (2009) reports dialectal variation as to the possibility of associating eth-

ical datives with absolutive causees, as in (58a,b). Trying to associate ethical
datives to the absolutive object of a transitive (58c) or the causee unergative
subject of (58d) is impossible. Dative doubling results in these cases, with a
22 See Fernández (this volume) for a comparison between Spanish and Basque ethical and
other non-selected datives. See also Fernández & Ortiz de Urbina (2010: 179) for Basque
ethical and interest/possessive datives.
23 Interest datives associated with ergatives are felt, right or wrongly, to show the influ-
ence of Spanish on Basque. The association with the absolutive would be expected in a
treatment of ethical datives as low applicative heads, as in Pylkkänen (2008), where the
dative introduces a source, goal, benefactive, or, in this case, interested argument related
to the object. This approach is taken in Oyharçabal (2007) and Albizu (2009), among oth-
ers. In the case of (d), even association with the absolutive object (‘my beans’) may be
excluded in eastern dialects, as ethical datives are excluded from stative contexts (Oyhar-
çabal 2007).
306 ortiz de urbina
priority for agreement with the cause excluding the ethical marking, in ways
similar to what has been noted above for dative causees; we return to similar
cases in the following section:
(58) a. Solairuaren ezkoak umea eroriarazi dit.

floor.of wax.erg kid.abs fall.cause aux.1dat
‘The floor wax made my kid fall.’
b. Ilargi beteak semea lehenago jaioarazi digu.

moon full.erg son.abs earlier be.born.cause aux.1dat
‘The full moon has made our son to be born earlier.’
c. *Liburua Mikeli harrarazi didate.

book Mikel.dat take.cause aux.1dat
‘They have made Mikel take (my) book.’
d. *Semeari / semea ondo borrokarazi didate.

son.dat son.abs well fight.cause aux.1dat
‘They have made my son fight well.’
Although variability is rampant and the situation for each dialectal area must
be checked, it seems that we can arrive at the generalization that the exam-
ples clearly ruled out are those where ethical dative agreement on the verb is
given priority over agreement with a dative causee. This is the case in (58c)
and (58d), in dialects where the unergative subject surfaces as a dative cause.
Those examples which are acceptable for some speakers and dubious for others
(58a,b and 58d with absolutive causee) are those where no such conflict arises;
the variability in the acceptability of (58a,b) would then be unrelated to these
agreement conflict considerations.24
3.1.2.3 Causativizing Ditransitive Verbs

Let’s turn finally to causatives of ditransitive uses of verbs. Since the ditransitive
verb also contains a dative argument, and a transitive subject is causativized
as a dative causee, a dative doubling situation emerges. Such sentences are
acceptable to most speakers provided the association of the dative nominals
24 Albizu (2009) provides a technical analysis in the line of Pylkkänen (2008) and Rezac
(2009). In western dialects with DOM effects in causatives, all causees in (58) will be
marked dative, as discussed in the following section. See footnote 28 for some comments
on the dative doubling situation created by ethical datives.
with a causee and a goal is possible; if not the result is probably more unac-
ceptable because of the impossibility of disambiguation than ungrammati-
cal:
(59) Mikeli liburua emanarazi didate.

Mikel.dat book give.cause aux.1dat
‘The made me give the book to Mikel.’
Disambiguation is achieved here through agreement, where the dative marker

must correspond to the causee. Where the causee and goal have the same gram-
matical person (a), or in a tenseless context without agreement (b,c), accept-
ability decreases along with disambiguation ease:
(60) a. Lagunek Mikeli Joni liburua emanarazi zioten.

friends Mikel.dat Jon.dat book give.cause aux.3dat
‘Their friends made Mikel give a book to Jon.’
b. Niri Mikeli liburua emanaraztea ez da ideia ona.

I.dat Mikel.dat book give.cause.nom no is idea good
‘Making me give a book to Mikel is not a good idea.’
c. Mikeli Joni liburua emanaraztea ez zen ideia ona.

Mikel.dat Jon.dat book give.cause.nom no is idea good
‘Making Mikel give a book to Mikel is not a good idea.’
In a sufficiently clear context all these sentences would probably be judged

acceptable; on top of contextual information, parsing seems easier if the first
dative is interpreted as causee and the second one as a goal, especially if a
pause separates them and the second one forms an intonational unit with the
causative verb. The causee is also more salient in that most speakers prefer as a
first interpretation of an example with focalized dative like (61), one in which
the latter corresponds to the causee:
(61) Mikeli emanarazi diote liburua Joni.

Mikel.dat give.cause aux.3dat liburua Jon.dat
‘It is Mikel that they made give the book to Jon.’
Similar acceptability issues are found in the dative doubling examples (47c),
(48) and (49) examined above, where the causee dative corresponds to the log-
ical subject of an ergative verb.
308 ortiz de urbina
As discussed with regard to the (lack of) interaction between ethical datives
and causee datives in unaccusative predicates, any other dative preempts the
use of ethical datives. In a ditransitive sentence, in particular, the dative may
not be interpreted as ethical. Thus, in the basic ditransitive (62) the dative may
be interpreted as an ethical dative only if the verb eraman ‘take’ is used as a
monotransitive one:
(62) Mikeli semea eraman diote.

Mikel.dat son take aux.3dat
‘They took the son to Mikel/They took away Mikel’s son.’
If we try to add an ethical dative, corresponding to an interpretation in which

Mikel is the goal of taking and I am the interested participant or viceversa, the
result is ungrammatical, as shown in (63):
(63) *Semea Mikeli eraman didate.

son Mikel.dat take aux.1dat
‘They have taken my son to Mikel/They have taken Mikel’s son to me.’
Notice, however, that even if speakers consider the sentence ungrammatical,

they can assign it an interpretation only if the agreement corresponds to the
ethical dative. Where no auxiliary is found, as in the nominalized clause in (64),
ungrammaticality remains, as a result of the incompatibility between goal and
ethical datives.25
(64) *niri semea Mikeli eramatea …

I.dat son Mikel.dat take.nom.abs
‘taking my son to Mikel/taking Mikel’s son to me …’
We can now introduce dative causees into the picture. A sentence like (65) with
a single overt dative and a verb like eraman ‘take’, which can have both mono-
transitive and ditransitive uses, is several ways ambiguous:
(65) Mikeli semea eramanaraztea ez zen gauza ona.

Mikel.dat son take.cause.nom.abs not was thing good
‘Making Mikel take the son/Making someone take the son to Mikel/Mak-
ing someone take Mikel’s son was not a good thing.’
25 Speakers can be made to understand the meaning of this clause if the first dative corre-
sponds to the interested participant and the dative closer to the verb corresponds to the
goal.
The first and most prominent interpretation is one where the dative is inter-
preted as a causee, the agent of a monotransitive use of eraman ‘take’. A second
interpretation, more formal, corresponds to an impersonal causative of ditran-
sitive eraman ‘take’ where the dative is interpreted as a goal, i.e., making some
unspecified causee take the son to Mikel. The third interpretation is one where
the dative is interpreted as an ethical dative, and again, this is available only
if there is no causee, i.e., making some unspecified agent take Mikel’s son. If
we try to include two of these datives in the same sentence the result is quite
complex:
(66) Niri semea Mikeli eramanaraztea ez zen gauza ona.

I.dat son Mikel.dat take.cause.nom.abs not was thing good
‘Making me take the son to Mikel was not a good thing.’
Again, this can be considered grammatical, if difficult to parse in the absence

of inflection. The primary interpretation is for datives to be causee and goal,
i.e., making me take the son to Mikel. The inverse interpretation (making Mikel
take the son to me) is more difficult to obtain, as one of the interpretive strate-
gies is to associate the dative closer to the verb with an internal argument of
that verb (here goal). The causee and ethical interpretation for the datives is
definitely more remote, whether it be making me take Mikel’s son or making
Mikel take my son. Finally, interpreting the datives as ethical and goal is also
remote, and obviously only feasible if the possibility of linking a dative with
the causee is eliminated by giving the clause an impersonal causative interpre-
tation. Speakers can be made to understand the meaning, but again only if the
dative closer to the verb is interpreted as the goal (i.e. making someone take my
son to Mikel). If the subject clause in (66) is turned into a tensed one with agree-
ing auxiliary, the agreement priorities mentioned above help rule out some of
the meanings:
(67) Niri semea Mikeli eramanarazi zidaten / zioten.

I.dat son Mikel.dat take.cause aux.1dat / aux.3dat
‘They made me/Mikeli take the son.’
In both cases, the agreeing dative is interpreted as causee. The other dative may
be interpreted as goal or ethical, perhaps with different likelihood depending
on ordering alternatives which remain to be studied.
310 ortiz de urbina
3.2 DOM in Dialectal Patterns

In western varieties, and to a lesser extent in central ones, eragin ‘cause’ and,
less commonly, erazo ‘cause’ are the causative verbs, with some differences
with respect to arazi. One difference has already been mentioned in section 1,
namely, the fact that eragin combines with the participial form of the embed-
ded verb, as opposed to arazi, which combines with the verbal root. Mixed
patterns exist in central areas, combining arazi with the participle.26
A similarly distributed difference may be observed in causee marking.
According to de Rijk (2008), eragin takes dative causees regardless of transi-
tivity. This may have been the situation in earlier forms of the language, but
at present, a differential object marking (DOM) pattern seems to have spread
in the western and central varieties (roughly, those south of the France/Spain
border). DOM in causatives may be a specific case of this general DOM pattern
found in contemporary dialects south of the border (see for instance Fernández
& Rezac 2016 or Odria this volume), less conspicuous and, perhaps as a result of
that, less stigmatized. In the general pattern, similarly to Spanish, animate, spe-
cific direct objects, especially 1st and 2nd person for some speakers, are marked
differently from other objects and identically to indirect objects; in the case of
Spanish both bear the preposition a, while in Basque both are marked dative.
As indicated, while the pattern is colloquial and stigmatized in the case of reg-
ular transitive constructions, it has gained more general ground in causatives,
and can be observed both with eragin and with arazi in these dialects. We give
some examples below with both; the first two examples contain dative animate
causees, while the third one contains the absolutive inanimate one, identical
to the standard:
(68) a. Bide pendiz bat artu eragin eta goraiño joan eragin eutsan.
path steep one take cause and up go cause aux.3dat
‘[S/he] made him take a steep path and go all the way up.’ (Aug. Zubika-
rai)
26 In one of the earliest attested occurrences of eragin, from the second half of the 16th cen-
tury, this causative actually appears not only with a participial complement verb, but even
in a tensed form, suggesting it is not fused with the embedded verb, in a structure similar
to English make:
(i) Azaiteau galdu lerait / ardura usatzakeak.
nourishment lose cause often use.without.erg
‘Not using it often would make me lose this nourishment.’ (Lazarraga 1196v)
The causee is dative, again possibly as an argument of the causative verb itself.
b. Aurreko gaira itzuli eragin zion solaskideari.

before.of topic.to return cause aux.3dat speaker.dat
‘[S/he] made his conversation partner return to the previous topic.’
(J. Landa, Odolbildua)
c. Daukien bildurra desagertu eragin gura dogu.

have.which fear.abs disappear cause nahi aux
‘We want to make the fear they have disappear.’ (Larrabetzu city info)
(69) a. Eguraldi txarrak berriro portura itzularazi zion.

weather bad.erg again port.to return.cause aux.3dat
‘The bad weather made him return to port again.’ (Wikipedia, Francis
Drake)
b. Komisaldegira joanarazi zieten bi gazteri.

police.station.to go.cause aux.3pl.dat two youth.dat
‘[They] made the two young persons return to the police station.’
(Argia 3/1/2017)
c. Txetxenian homosexualak hil eta desagerrarazi

Chechnia.in gay kill and disappear.cause
dituztela.
aux.3pl.abs.that
‘… that in Chechnia gay people have been killed and made disappear.’
(Hiruka 4/7/2017)
The (a) and (b) examples above contain animate causees which would have
been marked absolutive in the standard language, as described in 3.1.2.1 above,
marked dative in published texts. The (c) examples present the types of causees
which, being inanimate or non-specific, would usually take absolutive case also
in these dialects.27 This dative interacts and overlaps with other datives in ways
which may but need not be similar to those described above for multiple dative
situations, but, like DOM patterns in general, this is an understudied area in
Basque linguistics.28
27 Fewer examples are given of the standard-like pattern precisely because it is the norma-
tive one, found in more monitored texts in the western and central area. The qualification
‘types of animate causees’ in the text is introduced because examples with inanimate
absolutive causees should belong to the same author and, preferably, text, as those with
animate causees, something which has not been attempted here.
28 For instance, double datives will also emerge with unaccusative absolutive dative config-
urations:
312 ortiz de urbina
The development of DOM patterns in dialects of Basque in contact with

Spanish may have been helped by the existence of dative causee patterns with
unaccusative subjects with causative verbs like eragin. However, this could only
account for the appearance of new dative causees, but not for the maintenance
of absolutive causees with inanimate or indefinite causees. The correspon-
dence of the DOM isogloss with the area of contact with Spanish, and the main-
tenance of the traditional pattern in the contact area with French, which does
not display DOM effects, clearly points at convergence with the former; in fact,
the pattern is stigmatized by the normative tradition and felt to be a Spanish
calque. The relation with Spanish is, however, more complex than this indi-
cates. This is so because, as Odria (this volume) points out, the Spanish variety
of the Basque Country treats animate and inanimate objects differentially not
only in the use of the preposition a, but also in the clitic system, as described
by Landa (1995) and Fernández-Ordoñez (1999), for instance. While dative case
can be associated with the prepositional marking, clitics are more connected
with inflection, so this ‘objective conjugation’ enables us to see object marking
in the verb in a way parallel to Basque agreement. In Basque Spanish, animate
direct objects, whether masculine or feminine, are associated with the clitic
le. This is etymologically the dative clitic, which in Castilian Spanish has been
extended to (mostly singular) masculine personal direct objects, and in Basque
Spanish has been further extended to feminine personal direct objects, as in
(70):
(70) a. Lei vi a Maríai / Pedroi en la calle.

him/her saw to María / Pedro in the street
‘I saw Maria/Pedro in the street.’
b. La / lo vi en la calle.
it.fem / it.masc saw in the street
‘I saw it in the street.’
(i) Joni niri itzuli eragin eutsan.

Jon.dat I.dat return cause aux.3dat
‘[S/he] made Jon return to me.’
As with doubling in bivalent unergative verbs, agreement must correspond to the causee
rather than the goal. Causees also win out over ethical datives for agreement, so that an
ethical dative would be ruled out in (ii):
(ii) Semiari gerrara joan eragin eutsan / *eustan.
son.dat war.to go cause aux.3dat / aux.1dat
‘S/he made my son to go war.’
table 9.1 Adapted from Fernández Ordóñez 1999
Masculine Feminine
+Animate -Animate +Animate -Animate
Direct object le lo le la
Indirect object le le le le
The clitic le in (a) can refer both to a feminine and masculine direct object,
while the standard feminine and masculine direct object clitics la/lo in (b) can
only refer to inanimate objects with that gender. The resulting system is illus-
trated in Table 9.1, adapted from Fernández-Ordóñez (1999); for simplicity, only
the singular forms are given.
A result of this clitic/agreement system is that animate direct objects are
marked in the same way as indirect objects,29 a distribution essentially identi-
cal to that of the preposition a with objects.30 The presence of DOM patterns
in both preposition marking and clitic form in Spanish compounds the pres-
sure of DOM marking on the Basque system, perhaps helping to account for
the causative DOM patterns discussed above.
There is a further aspect of DOM which highlights the degree of conver-
gence with Spanish. As noted earlier in this section, DOM in Basque is more
widespread and mainstream in causatives than in basic direct object mark-
ing, where it is largely stigmatized. Similarly, le as a feminine direct object
clitic, although apparently spreading, is largely restricted to colloquial speech
and lower socio-economic status (Fernández-Ordóñez 1999). However, DOM in
Spanish causatives is common even in educated speech in the Spanish of the
Basque Country and neighboring areas, so that the patterns in (71) are standard:
(71) a. Lei hice venir a Maríai / Pedroi.

him/her made come Maria / Pedro
‘I made María/Pedro come.’
29 In addition, as Odria (this volume) points out, DOM contexts are often sensitive to speci-
ficity, and both the appearance of the clitic and of the preposition a are more likely to
occur with definite, known, direct objects.
30 It is doubtful that the emergence of this clitic pattern in Basque Spanish may in turn be
due to Basque influence. Animacy does play a role in the shape of locative cases, but these
facts are quite different from DOM.
314 ortiz de urbina
b. La / lo hice caer.
it.fem / it.masc made fall
‘I made it fall.’
The feminine and masculine direct object clitic in (b) are preferably interpreted
as inanimate.
Extensive bilingualism and the growing impact of ‘new speakers’, bilingual
speakers with Spanish as the dominant language who have acquired/learned
the language in a school environment (Ortega et al. 2015) are most likely having
a deep impact on Basque. The complex interactions among the systems of dif-
ferent varieties of Basque and between Basque and the surrounding Romance
languages, as also seen in other contributions to this volume, make of this con-
tact a promising ground for linguistic research.
Acknowledgements
Research for this paper was partially funded by the Spanish MINECO project
FFI2014–51878-P. Special thanks to Ane Berro and Beatriz Fernández for their
patience and comments on this paper. Examples taken from the General
Basque Dictionary (OEH) have been left with the code given there identifying
their source.
References
Aissen, Judith & David Perlmutter. 1983 [1976]. “Clause Reduction in Spanish”. In Studies
in Relational Grammar. Chicago: Chicago University Press. 360–403.
Albizu, Pablo. 2009 “Construcciones inacusativas con dativos posesivos y dativos de
interés en vasco: un análisis derivacionalista”. Talk presented at the Seminario de
Lingüística Teórica, CSIC, Madrid, May 11, 2009.
Baker, Mark. Incorporation: A Theory of Grammatical Function Changing. Chicago: Uni-
versity of Chicago Press.
Berro, Ane. 2015. From Event Structure to Syntactic Categories in Basque. Ph.D. disserta-
tion, University of the Basque Country-University Bordeaux Montaigne.
Bordelois, Ivonne. 1974. The Grammar of Spanish Causative Complements. Ph.D. disser-
tation, M.I.T.
Comrie, Bernard. 1989. Language Universals and Linguistic Typology. Chicago: Chicago
University Press.
De Rijk, Rudolf P.G. 2008. Standard Basque. A Progressive Grammar. Cambridge, MA:
MIT Press.
Dixon, Richard M.W. 2000. “A Typology of Causatives: Form, Meaning and Syntax”. In
Richard M.W. Dixon & Alexandra Y. Aikhenwald (eds.), Changing valency. Case stud-
ies in transitivity. Cambridge: Cambridge University Press, 30–83.
Dunlos, Laurence. 2001. “Event coreference in causal discourses”. In Pierrette Bouillon
& Federica Busa (eds.), The Language of Word Meaning. Cambridge: Cambridge Uni-
versity Press.
DHM. 1989. [Deustuko Hizkuntzalaritza Mintegia: Andolin Eguzkitza, Javier Ormaza-
bal, Jon Ortiz de Urbina, Koldo Sainz & Myriam Uribeetxebarria] “Inkorporazioa
perpaus kausatiboetan” [Incorporation in causative clauses]. In Pello Salaburu (ed.),
Sintaxi Teoria eta Euskara. San Sebastián: EHU Argitarapen Zerbitzua. 87–108.
Etxepare, Ricardo. 2003. “Valency and argument structure”. In José Ignacio Hualde &
Jon Ortiz de Urbina (eds.), A Grammar of Basque. Berlin/New York: Mouton De
Gruyter. 362–426.
Fernández, Beatriz & Milan Rezac. 2016. “Differential Object Marking in Basque Vari-
eties”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
Grammar of Basque. Amsterdam: John Benjamins. 93–138.
Fernández, Beatriz. This volume. “On non-selected datives: Ethical datives in Basque
and Spanish”.
Fernández-Ordóñez, Inés. 1999. “Leísmo, laísmo y loísmo”. In Ignacio Bosque & Vio-
leta Demonte (eds.), Gramática Descriptiva de la Lengua Española. Madrid: Espasa
Calpe, 1317–1397.
Folli, Raffaella & Heidi Harley. 2007. “Causation, obligation, and argument structure:
On the nature of little v”. Linguistic Inquiry 38(2). 197–238.
Laka, Itziar. 1993. “Unergatives that Assign Ergative, Unaccusatives that Assign Accusa-
tive”. In Jonathan D. Bobaljik & Colin Phillips (eds.), Papers on Case and Agreement
I, MITWPL 18. Cambridge, MA. 149–172.
Landa, Alazne. 1995. Conditions on Null Objects in Basque Spanish and their Relation to
leismo and Clitic Doubling. Ph.D. dissertation, Univesity of Southern California.
Levin, Beth. 1983. On the Nature of Ergativity. Ph.D. dissertation, MIT
Odria, Ane. This volume. “Differential Object Marking in Basque and Spanish Dialects”.
OEH [Orotariko Euskal Hiztegia-General Basque Dictionary]. Available online at www
.euskaltzaindia.eus/oeh.
Ormazabal, Javier & Juan Romero. 2007. “The Object Agreement Constraint”. Natural
Language and Linguistic Theory 25. 315–347.
Ortega, Ane, Jacqueline Urla, Estibaliz Amorrortu, Jone Goirigolzarri & Belen Uranga.
2015. “Linguistic identity among new speakers of Basque”. International Journal of
the Sociology of Language 231. 85–105.
Ortiz de Urbina, Jon. 2003. “Causatives”. In José Ignacio Hualde & Jon Ortiz de Urbina
(eds.), A Grammar of Basque. Berlin/New York: Mouton De Gruyter. 592–607.
Ortiz de Urbina, Jon & Beatriz Fernández. 2016. “Datives in Basque Bivalent Unerga-
316 ortiz de urbina
tives”. In Beatriz Fernández & Jon Ortiz de Urbina (eds.), Microparameters in the
Grammar of Basque. Amsterdam: John Benjamins. 67–92.
Oyharçabal, Bernard. 2003. “Lexical Causatives and Causative Alternation in Basque”.
ASJU Supplement XLVI. 223–253.
Oyharçabal, Bernard. 2007. “Basque Ditransitives”. Ms, IKER-CNRS.
Perlmutter, David. 1978. “Impersonal Passives and the Unaccusative Hypothesis”. In
Fourth Annual Meeting of the Berkeley Linguistics Society, vol. 4. 157–190.
Pylkkänen, Liina. 2008. Introducing Arguments. Cambridge, MA: The MIT Press.
Rezac, Milan. 2009. “Person restrictions in Basque intransitives”. Lapurdum 2009, 13.
305–322.
Treviño, Esthela. 1994. Las causativas del español con complemento infinitivo. México: El
Colegio de México.
Tubino, Mercedes. 2011. Causatives in Minimalism. Amsterdam/Philadelphia: John Ben-
jamins.
Zubizarreta, M.L. 1985. “The relation between morphophonology and morphosyntax:
The case of Romance causatives”. Linguistic Inquiry, 16:2. 247–289.
chapter 10
Grammaticalization Processes in Causal

Subordination
José Ignacio Hualde and Manuel Pérez Saldanya
1 Introduction
Causal constructions express meanings related to the notion of “cause”, includ-

ing, not only the reason why something happens, but also motivations, expla-
nations and justifications for saying something or performing a specific speech
act. Structurally, a causal construction may contain a finite clause (e.g. John
got a fine [because he was speeding]), a nonfinite clause (e.g. John got a fine [ for
going too fast]) or a noun phrase headed by a preposition (e.g. John was pun-
ished [ for his bad behavior]). In this chapter we will concentrate only on finite
subordinate causal clauses.
A well-known tripartite classification of causal clauses is as content caus-
als (expressing the “real-world” cause of an event or state of things, e.g. The
ground is wet [because it has rained]), epistemic causals (speaker’s inference or
reason, e.g. It has rained, [because the ground is wet]) and speech-act causals
(justification for a command or another speech act expressed in the main
clause, e.g. Go home, [because it is about to rain]) (see Sweetser 1990: 76–
82).
Structurally, causals may be internal, if integrated with the predicate of the
main clause, or external, if they are sentence adjuncts, prosodically separated
from the main clause by “comma intonation” (see Santos Río 1982, Goethals
2002, Pérez Saldanya 2014: §28.4.1, among others). External causals may appear
at the left periphery (e.g. [Since it was about to rain], he decided to leave right
away), at the right periphery (e.g. He decided to leave right away, [since it was
about to rain]) or in a parenthetical position (e.g. He decided, [since it was
about to rain], to leave right away). Internal causals are always in the content
domain, whereas external causals may be in the content, epistemic or speech
act domain.
Regarding information structure, causals may be classified either as rhe-
matic (expressing new information) or as thematic or presupposed (expressing
given or assumed information). Internal causals are always rhematic, whereas
external causals may be either rhematic or thematic/presupposed.

318 hualde and pérez saldanya
table 10.1 Typology of inflected causal clauses
Properties Examples
internal causal rhematic content The ground in wet [because it has rained]
external causal Mary couldn’t see the road, [because it was so

dark]
epistemic It has rained, [because the ground is wet]
speech-act Go home, [because it is about to rain]
thematic content [Since it was about to rain], he decided to

leave right away
speech act [Since you cannot wait], do it yourself
Following the criteria just mentioned, it is possible to differentiate up to six

types of causal clauses, as shown in Table 10.1. We will make use of all these
classificatory features in our analysis and comparison of Basque and Romance
causal clauses.
2 Types of Causal Subordinators in Basque and Romance
In the Romance languages, as in English, causal clauses, like other types of sub-
ordinate clauses, may be headed by a conjunction (e.g. Sp porque ‘because’,
como ‘as’, que ‘that’, pues ‘as, so’) or by more complex expressions, which show
different degrees of grammaticalization in this function (e.g. Sp ya que ‘since’,
puesto que ‘since’, dado que ‘given that’). In Basque, on the other hand, there is
more structural heterogeneity. Finite clauses with causal value in Basque may
contain verbal suffixes, verbal prefixes and/or clause initial conjunctions, as
shown in (1) with the example ikusi dugu ‘we have seen (it)’:
(1) Causal subordination in Basque

a. ikusi dugulako ‘because we have seen it’
b. ikusi dugunez ‘as we have seen’ (-n ‘subordinator’ + -ez ‘instrumental’)
c. ikusi dugunez gero ‘since we have seen it’ (gero ‘later’)
grammaticalization processes in causal subordination 319
d. ikusi dugu eta ‘since we have seen it’ (eta ‘and’)

e. ikusi baitugu ‘since we have seen it’ (bait- ‘subordinator’)
f. zeren ikusi dugun ‘because we have seen it’ (zeren ‘of/for what’)
g. zeren (eta) ikusi baitugu ‘because we have seen it’
h. ezen ikusi dugu ‘since we have seen it’ (ezen ‘that, since’)
i. zergatik ikusi dugu(n) ‘because we have seen it’ (zergatik ‘why’)
j. nola ikusi baitugu ‘since we have seen it’ (nola ‘how’)
In spite of all the obvious differences between Basque and Romance causal
elements, there are also many similarities, especially when we consider the his-
torical origin of causal connectors. In Table 10.2, we make explicit some of these
similarities by comparing Basque causal subordinators with etymologically
equivalent expressions in the main Romance languages with which Basque
has been in contact, Spanish, Gascon and French.1 Since contact with French
is relatively recent, we indicate only the most usual forms for this language.
(Another historically important contact language was Navarro-Aragonese
Romance, which in all relevant respects coincides with Castilian Spanish). Sub-
ordinators that have become obsolete are signaled with “†”.
As shown in Table 10.2, in both Basque and Romance we find causal sub-
ordinators that have developed historically from question words, from gen-
eral complementizers (‘that’) and from time and modal expressions, as we
also find in many other language (Kortmann 1997, 1998; Pérez Saldanya 2014:
3479–3483). On the other hand, Basque shows at least one grammaticaliza-
tion phenomenon that has no parallels in Romance: the development of a
causal construction involving the coordinating conjunction eta ‘and’ as a verb
enclitic.
We also find other differences in the specific meanings and uses of different
causal structures. In modern Basque a direct answer to the question ‘why’ gen-
erally requires the construction with -(e)lako ‘because’ attached to the inflected
verb. This suffix is thus the basic equivalent of Sp porque, Fr parce que or Eng
because. However, unlike Sp porque, French parce que and Eng because, the
Basque connector is usually only employed in internal causal clauses. Other
elements such as preverbal bait- ‘since’, clause-initial zeren ‘since, because’ and
postverbal eta have a larger set of causal usages in Basque.
1 We have obtained the information about Occitan from Alibèrt (1935–1937) and Fernández
González (1985). For Gascon, we have consulted Palay (1974) and the Bearnais dictionaries in
Lexilogos (2002–2017).
table 10.2 Causal subordinators in Basque and Romance classified by etymological origin
Origin Basque Spanish Old occitan and French

contemporary
Gascon
Interrogative (zer ‘what’ >) (Lat. quia >) †ca (Lat. quāre >) †car car
zeren … bait-
zeren … -(e)n
ze (colloquial)
General bait- (Lat. quid >) que que que
complementizer ezen
‘that’+ ‘of ’ (Bq) -(e)lako porque per çò que parce que
‘for’ + ‘that’ (Rom) perquè
Locutions with nouns dela kausa/bide/medio a/por causa de que a causa que à cause que
and ‘that’ † por razón que per (a)mor/pramor
que
Time adverb -(e)nez gero (Lat.post >) pues (que) † pus (que), puix que puisque
(or other temporal ya que
items) puesto que vist que vu que
del moment que
Modal adverb nola … bait- (Lat. quomodo >) †com/coma comme
(& interrogative ‘how’) nola … -(e)n como
-(e)nez
Copulative conjunction (-la) eta n.a. n.a. n.a.
Connectors of modal origin (Bq -n-ez ‘subordinator+instrumental’, nola

‘how’> ‘as, since’; Sp como ‘how’> ‘as, since’) are essentially restricted in their
use to thematic content clauses. On the other hand, regarding causal connec-
tors related to time expressions, whereas Bq -n-ez gero seems to be used only in
thematic causal clauses, Sp pues ‘since’ (from Lat post ‘after’), ya que ‘already
that’> ‘since’, puesto que ‘given that’ are used both in thematic clauses and in
external rhematic clauses.
In the next sections we consider each of these types of causal construc-
tions in more detail. We structure the discussion around Basque, pointing out
similarities and differences with respect to Spanish and other neighboring
Romance languages.
3 Direct Answers to the Question ‘Why?’ (Internal Causal Clauses)
In this section, we consider the basic connectors used in internal causal clauses
in Basque. As mentioned, internal causal clauses are always rhematic and in
the content domain. As such, they can be used to provide a direct answer to
the question ‘why?’. With this function, modern English uses the conjunction
because, Spanish has porque (por ‘for’+ que ‘that’) and French parce que (par
‘for’+ ce ‘this’ + que ‘which’). In contemporary standard Basque, the main sub-
ordinator used in this context is the suffix -(e)lako. Morphologically, -(e)lako
can be analyzed as bimorphemic, with the relational suffix -ko attached to the
subordinating suffix -la ‘that’ (cf., e.g., ikusi dugu-la ‘that we have seen it’, ikusi
dugu-la-ko ‘because we have seen it’). Some examples illustrating the use of
-(e)lako are given in (2) (adapted from EGLU 2005: 192) and (3) (from a journal-
istic interview). We provide Spanish translations as well for comparison:2
(2) —Zergatik eman diozu dirua?

why give aux money.the
‘Why did you give him/her the money?’,
Sp ¿Por qué le has dado el dinero?
—Egunkaria erosi behar zuelako.

newspaper.the buy must aux.because
‘Because s/he had to buy the newspaper.’,
Sp Porque tenía que comprar el periódico.
(3) —Derrigorrezko galdera: Zergatik?

‘Obligatory question: Why?’
—Gauzek ez zutelako lehen bezalako naturaltasunez funtzionatzen.

‘Because things were not working with the same naturalness as before.’
(http://www.argia.eus/argia‑astekaria/2560/napoka‑iria)
Sp Pregunta obligatoria: ¿Por qué?
Porque las cosas no funcionaban con la naturalidad de (como) antes.
Unlike their Romance and English counterparts, causal clauses with -(e)lako
often appear before the main clause as in (4) (adapted from Artiagoitia 2013:
715):
2 The following abbreviations are used in the glosses: abl ‘ablative’, aux ‘auxiliary’, comp ‘com-
plementizer’, det ‘determiner’, erg ‘ergative’, fut ‘future’, gen ‘genitive’, imp ‘imperative’,
impf ‘imperfective’, inst ‘instrumental’, loc ‘locative’, neg ‘negation’, part ‘partitive’, pl ‘plu-
ral’, rel ‘relative’, subor ‘subordinator’.
(4) Beste erremediorik ez nuelako egin nuen.

other remedy.part neg I.had.because do aux
‘I did it because I did not have any other option.’
Sp Lo hice porque no tenía otro remedio.
The reason for this difference between the languages is that in Basque focal-
ized elements are generally placed in preverbal position (Artiagoitia 2003: 714,
Ortiz de Urbina 2003: 454–455).
In addition to its use as a causal connector, -(e)lako can also be used to link
an inflected clausal complement to a following head noun (a function that can
also be fulfilled by the suffix -(e)n). This is in fact the only use in which this
verbal suffix is found in early Eastern authors (again, Spanish translations are
provided for comparison):
(5) Zenbait granoz arin delako seinalea da.

some grain.inst light is.lako sign is
lit. ‘It is a sign that he is light by a few grains.’ (i.e. ‘it is a sign that he is a
little reckless’) (Axular, Gero: 270, 1643)
Sp Es una señal de que es ligero por algunos granos.
(6) erremediatuko zarelako esperantzarekin

remedy.fut aux.lako hope.with
‘With the hope that you will reform yourself’ (Axular, Gero: 145, 1643)
Sp con la esperanza de que te reformarás
The bridging context between these two rather distinct usages of -(e)lako might
be found in potentially ambiguous examples like (7), where the embedded
clause can be understood as a complement of the noun poz ‘happiness’, but
also as a causal complement of the main clause:3
(7) etorri zarelako ene poz handia

come aux.lako my happiness big
‘my great happiness that/because you have come’
In (8) we give an example (from a contemporary novel) where both uses of

-(e)lako can be observed:
3 In modern Eastern usage, there is a formal differentiation between -(e)lakotz, which is used
as causal suffix, and -(e)lako, used with complements of nouns and adjectives.
(8) Bakartia baldin banaiz, izan ere, inor ez molestatzeko da,

solitary if if.I.am be too anyone not bother.for is
ezin dudalako eraman inori traba edo kalte egiten
not.be.able aux.because bear anyone difficulty or harm do
diodalako irudipena edo kontzientzia, eta bakarrik egotea delako
aux.of suspicion or awareness and alone being is.because
hori saihesteko modu seguruena, bakarra ez bada.
that avoid.for way surest only not if.is
‘If I am solitary, it is actually not to bother anybody, because I cannot bear
the suspicion or the awareness that I am hurting or bugging anybody; and
because being alone is the surest way to avoid that, if not the only way’.
(J.L. Zabala, Ospa, 2017)
As mentioned, the Basque verbal suffix -(e)lako is transparently analyzable as

consisting of the complementizer -(e)la ‘that’ (cf. uste dut ikusi dugu-la ‘I think
that we have seen it’) and the relational suffix -ko used to link complements
to nouns (cf. mendi-ko etxea ‘the house of the mountain’, mendi-ranz-ko bidea
‘the road (of) towards the mountain’). As shown, in the Spanish translations
provided in (5)–(6) above, the preposition+complementizer sequence de que
in Spanish offers a direct parallel (in mirror image).4
It may be noted that Sp porque ‘because’ also has a somewhat similar struc-
ture, with the preposition por ‘for, by’ preceding the general complementatizer
que ‘that’. Whereas Sp porque is morphologically identical to por qué ‘why’, how-
ever, Bq -(e)lako ‘because’ and zergatik ‘why’ are obviously dissimilar. Although
Basque has a causative suffix -gatik, it is not used with inflected clauses, but
only with noun phrases and nominalizations, e.g. euri-a-gatik ‘because of the
rain’, Sp por la lluvia; Jon ikuste-a-gatik ‘because of seeing Jon’, Sp por ver a Jon;
egia esate-a-gatik ‘for telling the truth’, Sp por decir la verdad. On the other
hand, uninflected nominalized clauses with -ko (paralleling -(e)lako) have a
final rather than causal interpretation, e.g. Jon ikusteko ‘in order to see Jon’, Sp
para ver a Jon; egia esateko ‘in order to tell the truth’, Sp para decir la verdad.
Also as already mentioned, unlike Sp porque or Eng because, which have
a wider usage, the Basque suffix -(e)lako is restricted in its use essentially to
internal causal clauses, at least in some dialects (cf. EGLU 2005: 169, 179). It
is not commonly employed in epistemic causals (e.g. It has rained, because
4 Modal adverbs also take the suffix -(e)la (e.g. honela ‘thus, this way’, bestela ‘otherwise’). When
these adverbs function as a complement of a noun we find the suffix sequence -la-ko (e.g.
honelako galderak ‘this type of questions’, bestelako iritziak ‘different oppinions’, also with an
empty noun head: bestelakoa da ‘it is different’).
the ground is wet) or speech-act causals (Go now, because it is about to rain),
where other connectors are employed instead (see sections 4 and 5). Its (pos-
sible) origin in inflected complements of nouns and adjectives may account
for the restriction of -(e)lako to the internal causal context. In both cases, the
subordinate clause is an internal adjunct, and if -(e)lako causals derives from
-(e)lako noun complements, the internal character of the former may be a con-
servative property of the latter. Nevertheless, nowadays -(e)lako is sometimes
also used in external causal clauses, as in (9) and (10), both from the same
novel:
(9) Endredo bat zen enkargu hura, idazten ari nintzen nobelatik ateraraziko
ninduelako.
‘That request was a complication, because it would take me out of the
novel I was writing.’ (J.L. Zabala, Ospa, 2017)
Sp Ese encargo era un enredo, porque me sacaría de la novela que estaba
escribiendo.
(10) jemezeak deituak, GMC markakoak zirelako

‘So called “ge-em-ces”, because they were of the GMC brand’ (J.L. Zabala,
Ospa, 2017)
Sp llamados jemecés, porque eran de la marca GMC
This use of -(e)lako in external causal clauses seems to be relatively recent

(EGLU 2005: 179) and may be due to convergence with Spanish.
Internal content clauses in Basque may also be introduced by clause-initial
conjunctions such as zeren, ze, ezen and zergatik, which also have a wider
causal usage (see section 5). Although these clause-initial conjunctions are
found since the first texts, they are often avoided in more recent writing, as
they are viewed by some as calques from the Romance languages (Azkue 1923:
491).
4 External Causal Clauses
Leaving aside causal subordinators that have their origin in modal and time
conjunctions (see sections 6 and 7), we find two additional main subordinators
in external causal clauses in Basque, whether content, epistemic or speech-act:
the verbal prefix bait- (attached to the inflected verb, sometimes with elision
of t), as in (11), and enclitic -eta (usually enclitic to the inflected verb, but not
exclusively, and frequently phonologically reduced to ta), as in (12).
(11) Nik atzo hik baiño bide gehiago iragan nian inguratu bainuen
I yest. you but way more pass aux surround since.aux
mundu guztia.
world all
‘Yesterday I walked more than you did, because I went around the whole
world.’ (Axular, Gero: 21, 1642)
(12) Ez barrerik egin, egia da ta.

no laugh.part do truth is and
‘Do not laugh, because it’s true.’ (Azkue, Bein da betiko, apud EGLU 2005:
199)
As in the two examples just mentioned, both subordinators are generally used
in rhematic clauses, and are thus normally found in the right periphery of the
sentence. Nevertheless, causals with bait- may also be thematic, located at the
left periphery, as in (13).
(13) ziren zirena baitzira, zuzaz pena dizit nik

you.are.rel you.are.rel.det comp.you.are you.inst pain aux I.erg
‘Because you are who you are, I agonize over you.’ (Etxepare, Primitiae:
VIII.10, 1545)
Traditionally these two connectors had different geographical distributions,

bait- being used in eastern varieties and -eta in western and central areas. Nev-
ertheless, both are nowadays used in standard Basque.
4.1 The Subordinator bait-

The verbal prefix bait- is employed in a wide variety of subordinate clauses in
eastern dialects, including relative, completive, consecutive and causal clauses
(Oyharçabal 2003). It is in complementary distribution with the subordinating
suffixes -(e)n and -la (and those derived from them, such as -(e)nez, -(e)lako,
etc.), in the sense that it cannot co-occur with any of them. This prefix is
attested in early texts from both eastern and western regions, but it fell out
of use in western and central dialects. The main use of bait- in the modern
standard language is in causal clauses, with very much the same usages as
postverbal causal -eta. Krajewska (2016) argues that bait- was initially a relative
clause marker (which is its most common use in 16th century texts) and that
it developed as a causal subordinator in contexts where there was ambiguity
between a relative clause and a reason clause reading.
In his study on causal constructions in P. Axular’s Gero (1643), Villasante
(1982b: 364–380) provides several examples where a clause with bait- can be
interpreted as either an explicative relative clause or as an external causal

clause. In his Spanish translation, Villasante chooses a causal interpretation for
the example in (14) and a relative clause interpretation for (15), but it seems that
both interpretations would be possible in both cases. (Notice that bait- and the
verb may contract phonologically. In these examples, bait-da > baita and bait-
dira > baitira).
(14) Iakin dezagun gauza bat, sentitzekoa baita.

‘Let’s keep one thing in mind, because it is worth noticing (or: which is
worth noticing).’ (Axular, Gero: 75, 1964 [1643])
Sp Tengamos en cuenta una cosa, pues es muy digna de advertencia.
(L. Villasante’s translation)
(15) Utzten ditut plazer haren aitzingibelak, anarteraiñoko, eta ondoreko egit-
ekoak, eta atsekabeak, handiak baitira.
‘I leave aside the pros and cons of that pleasure, the preceding ones, and
those that follow, which are big (or: because they are big).’ (Axular, Gero:
251, 1643)
Sp Dejo ahora a un lado las circunstancias de dicho placer, los que-
braderos de cabeza y los disgustos que le preceden y le siguen, que son
grandes. (L. Villasante’s translation)
It is likely that the grammaticalization process that has led to bait- being used
as a causal connector is the same that is responsible for a general subordina-
tor being able to introduce external causal clauses, as is the case with Spanish
que ‘that’. For instance, in a Spanish example such as (16a), literally, ‘I am leav-
ing, that is late’ an inference of causality allows the general subordinator que
‘that’ to be interpreted as ‘because’. We may envision a similar process in the
grammaticalization of Bq bait- as a causal connector, as in (16b).
(16) a. Me voy, que es tarde.

‘I am leaving, because it is late.’
b. Banoa, berandu baita

I.leave late that.it.is
This explanation, however, does not account for why bait- has become special-
ized as a causal subordinator in modern Basque (whereas Sp que has not had
this development). It also leaves unexplained the fact that bait- can be used in
thematic causals, whereas Sp que is restricted to rhematic ones, especially in
speech-act causals (Pérez Saldanya 2014: 3510–3514).
4.2 The Subordinator -(e)ta

One of the most productive clausal constructions in the present-day language
of the western and central Basque-speaking area involves the conjunction eta
‘and’, usually cliticized to an inflected verb in sentence-final position, as in (17):
(17) Horixe esango nuke, nahikoa ezkor aritu naiz-eta.

that say.fut aux rather negative speak aux-and
‘I would say that, since I have been speaking rather negatively.’ (https://
barren.eus/albisteak/kultura/1806‑maialen‑liujanbio‑euskaratik‑eta‑
gure‑mundu‑erreferentzialetik‑egin‑behar‑dugu‑besteen‑ispiluetan‑
gure‑buruaren‑bila‑ibili‑gabe)
This causal construction, which does not have a structural parallel in Romance,
appears to have arisen from coordinated structures with some specific proper-
ties, as explained in EGLU (2005: 201) which provides the example in (18) to
illustrate the process:
(18) a. Mutil argia da eta berehala ikasiko du.

boy smart is and immediately learn.fut aux
‘He is a smart boy and will learn quickly.’
b. Berehala ikasiko du, mutil argia da eta.

‘He will learn quickly, since he is a smart boy.’
In the coordinated structure in (18a) it is possible to make the inference that

the reason for the proposition expressed in the main clause is that expressed
in the second clause. By transposition of the two clauses, in coordinated clauses
where this inference arises, we obtain (18b), with enclitization of the conjunc-
tion eta ‘and’: [S1]-eta [S2] → [S2] [S1]-eta.
The reason why this grammaticalization phenomenon has taken place in
Basque, but not in Romance (or in Germanic) may be a prosodic difference in
the parsing of conjunctions. In Basque, a postpositional language, the conjunc-
tion eta is often prosodically cliticized to the preceding word, as a postposition.
This is most evident in dialects that have certain phonological rules that apply
only within word domains. One of these rules, for instance, is the raising of
/a/ to /e/ after a high vowel, as in lagun-a → lagune ‘the friend’, which may
also affect (e)ta ‘and’ in dialects with this rule: hiru ta lau → hiru-te lau ‘three
and four’. Another such phonological rule is the palatalization of /t/ after /i/:
mendi-tik → mendittik [mendicik] ‘from the mountain’ and also, e.g. bi ta hiru
→ bi-tte hiru ‘two and three’. In fact, encliticization of eta ‘and’ to a verbal form
has also resulted in other grammaticalizations, such as resultative construc-

tions, e.g. ikusi eta ‘see and’→ ikusita (in some varieties, ikusitte) ‘having seen’
(Krajewska 2013).
As also noted in EGLU (2005: 202), we have a parallel development with the
coordinating conjunction baina ‘but’, which can also appear postposed to the
second clause, giving rise to structures with adversative meaning. The causal
construction with gero (see section 6) also has its origin in the clitization of an
adverb to the right of a verb.5
In addition, the particle eta is suffixed to a copulative verb marked with -(e)la
in causal constructions where the cause is taken to be a referential expression
(a noun phrase), e.g. katuak direla-eta ‘because of the cats’, cf. also the grammat-
icalized expression zer dela-eta ‘because of what, why?’ (lit. what that-is-and).
(EGLU 2005: 204–205).
The enclitization of conjunctions and other elements to verbal forms,
whether finite or participial, has thus been the source of a number of gram-
maticalized constructions in Basque. This path of development, of which the
causal construction with -eta is an instance, does not have any parallels in the
Romance languages.
5 Clause-Initial Causal Conjunctions
From the earliest texts in Basque we find causal clauses headed by conjunctions
such as zeren (or zerren), zeren eta and zergatik. In these subordinate clauses,
the inflected verb may take the prefix bait-, as in (19) and (20), the suffix -(e)n
or even -(e)lako. It may also be unmarked as a subordinated verb, as in (21).
(19) egiten dut ihes zeren ezagutzen baitut neure kondizinoa,

do.impf aux flee because know.impf comp.aux my condition
ez naizela iend’ arteko gai
not I.am.comp people among.of able
‘I flee because I know my condition, that I am not capable in society’ (Axu-
lar, Gero: 192, 1643)
5 The conjunction edo ‘or’ is also used as a postposition; specifically it is postposed to measure
phrases to convey the meaning ‘or so’, e.g. lauretan edo ‘at 4:00 or so’. In addition, both dis-
junctive conjuntions edo and ala are used as tags in questions, like German oder, cf. Eng or
what, Sp o qué.
(20) Zeren Jeinkoa egun oroz ongi ari baitzaigu guk ere hala
because God day all.inst well act comp.aux we.erg too thus
behar harzaz untsa orhitu.
must 3.inst well remember
‘Since God rewards as every day, we too should remember Him in grati-
tude.’ (Etxepare, Primitiae: I.17, 1545)
(21) Esarasu zurinean serren da berroago goxetaraco.

pour.imp white.abl because is warmer morning.for
‘Pour some white [wine], because it is warmer for the morning.’ (Mikoleta,
Modo breve: 208, 1653)
(Sp translation in Mikoleta’s text: Echa de lo blanco, que es mas caliente
para por la mañana.)
Zeren and the nowadays stigmatized form zergatik are polyfunctional causal
conjunctions, in the sense they may introduce internal clauses, as in (19) above,
and external clauses, either thematic, as in (20), and rhematic, as in (21). On the
other hand, zeren eta always introduces external clauses, mostly with epistemic
value, as in (22).
(22) Eta hutsuneak ere hor daude, bertzalde, zeren eta

and lacunae also there they.are other.side because and
memoriak bere zuloak eta bere haitzuloak baititu.
memory.erg its holes and its caverns comp.aux
‘And the lacunae are there too, on the other hand, because memory has
its holes and its caverns.’ (J.M. Irigoien, Lur bat haratago: 10, 2000)
As has been already mentioned, zerga(i)tik is also the interrogative word ‘why’,
so that there is an obvious parallelism with Sp por qué ‘why’ and porque
‘because’. The structure of this form is transparently zer ‘what’+ gai ‘matter’+
tik ‘through, from’. As for zer(r)en, it is the genitive form of zer ‘what’, which in
earlier times was also used as benefactive; that is, it may be literally translated
as ‘for what’. The use of zeren as ‘why’ is also found in early texts, as in (23)–
(24):
(23) Zeren, bada, erho gira gaixo bekhatariak?

why thus crazy we.are poor sinners
‘Why, then, are we poor sinners so foolish?’ (Etxepare, Primitiae: 2.60,
1545)
(24) Ene Iainkoa, Ene Iainkoa, zeren abandonatu nauk?

‘My God, my God, why hast thou forsaken me?’ (Leizarraga, Testamentu,
Matthew: 27:46, 1571)
Question words are a common source of causal conjunctions crosslinguis-

tically (Kortmann 1988: 514). In the Romance languages, in addition to the
example provided by Sp por qué ‘why’, porque ‘because’, and cognates in other
languages, French, Occitan and Catalan car ‘because’ has its origin in Latin
quāre ‘by what means?’ (Baños 2011: 207, 2014: 54). The Latin causal conjunc-
tion quia ‘why’ also derives from a question word (Hofmann & Szantyr 1965:
585, ap. Baños 2011: 206–207). The basic mechanism seems to be an equation
between question and answer particles, with a change in intonation. As argued
by Pineda & Pérez-Saldanya 2017 (based on the diachronic analysis of car in
Catalan), the evolutionary path would be as in (25):
(25) question word > conjunction in internal causal clauses > polyfunctional
causal conjunction > conjunction in external causal clauses
Within the diachronic path in (25), zeren appears to be in the third stage, as it
clearly has a polyfunctional use, as mentioned above. In this respect it is com-
parable in its range of functions to Spanish porque and to car in French, Occitan
and Old Catalan (and English because). We may distinguish a number of causal
contexts in which Bq zeren is employed.
a) First of all, it is used in internal causal clauses, with a meaning similar to that
of -(e)lako, as in (19) above, where the subordinate clause expresses the reason
for the subject fleeing, and in (26), where it expresses the reason for Peter being
grieved:6
(26) Triste zedin Pierris zeren erran baitziezon herenean, On dariztak niri?
‘Peter was grieved because he said unto him the third time, Lovest thou
me?’ (Leizarraga, Testamentu: John, 21:17, 1571)
As internal causal clauses, clauses headed by zeren can be used as a reply to

questions with zergatik or zeren, as in (27), and can also be the focus of nega-
tion, as in (28):
6 All English translations from the Bible are from the King James version, which is roughly con-
temporary with Leizarraga’s Basque translation.
(27) Zergatik ene lengoajea eztuzue aditzen? zeren ezin enzun baitirokezue
ene hitza.
‘Why do ye not understand my speech? even because ye cannot hear my
word’ (Leizarraga, Testamentu: John 8: 43, 1571)
(28) Eta haur erraiten du hunela, ez zeren dantzatzea bera hain gauza gaixtoa
den, baiña zeren dantzatik anhitz okhasino behar eztenik sortzen den.
‘And he says this this way, not because dancing is by itself such an evil
thing, but because from the dance many inappropriate occasions arise.’
(Axular, Gero: 268, 1643)
b) The conjunction zeren is also very commonly found in external causal

clauses, especially those in the content domain:
(29) eta bertan ilki zitezen, zeren ezpaitzuten lur barnerik.

‘and because they had no root, they withered away.’ (Leizarraga, Testa-
mentu: Matthew 13:6, 1571)
(30) Serren diñoe doctoreac goxeti xayguitea dala ona ossasunensaco.

Porque dicen los medicos que para la salud es bueno el leuantar de ma-
ñana.
‘Because physicians say that to get up early is good for your health.’ (Miko-
leta, Modo breve: 202, 1653)
(31) Errege Daviti eman zerautzatenean Saulen harmak, baitziren ezin hobea-
goak eta halakoak, guztiarekin ere, probetxu baiño kalte gehiago egiten
zioten, trabatzen zuten. Zeren nola anarteraiñokoan bethi arzain ibili
baitzen, ezpaitzen oraiño harmetan usatua.
‘When they gave King David Saul’s weapons, even though they were the
best, they did him more harm than good; they bothered him. Because,
since up to that moment he had always been a shepherd, he was not yet
used to weapons’ (Axular, Gero: 61, 1643)
c) It may also occur in external causals in the speech act domain, as in the
following examples. In (32) the causal clause justifies a command, in (33) it
justifies a question and in (34) it motivates an exclamation:
(32) Boz eta alegera zaitezte, zeren zuen saria handi baita zeruetan
‘Rejoice, and be exceeding glad: for great is your reward in heaven’ (Leizar-
raga, Testamentu: Matthew 5:12, 1571)
(33) zenbat gehiago eztire bihotzez estekatuak, josiak beren tokiko mintza-
iari, zeren erlisionea kasik orok haurdanik mintzai hartaz ikhasia dute,
guk emen euskaraz bezala?
‘How many more are not tied and attached in their hearts to their native
language, for almost all of them were taught religion in childhood in that
language, just like we here in Basque?’ (H.U.Gonz 193, ap. Hualde & Ortiz
de Urbina 2003:730)
(34) Ai gure urrikalkizunak, galdu gara, egin du gureak. Zeren akhabatu hur-
ran da eguna, beheititu da iguzkia, eta itzala ere luzatuago da arrat-
saldean.
‘Poor us, we are lost, our luck has finished. Because the day has almost
ended, the sun has lowered and even the shade is longer in the evening.’
d) It is also found in thematic external causal clauses, especially those in the

content domain, (35)–(36). This was also the case for Sp porque some centuries
ago, but it is less so in contemporary Spanish usage (Pérez Saldanya 2014: 3519–
3524):
(35) Gero iguzkia goratu eta, erre izan dirade, eta zeren ezpaitzuten errorik,
eiarthu izan dirade.
‘And when the sun was up, they were scorched; and because they had no
root, they withered away.’ (Leizarraga, Testamentu: Matthew: 13:6, 1571)
(36) Zeren amak diren koleratsu, halatan dira umeak ere faltatsu.
‘Because the mothers are quick to anger, the children are also defective.’
e) Finally, zeren may introduce clauses that express both cause and finality,
which is also possible in Spanish with porque:
(37) Zeren unsa orhit ziten nizaz, ama eztia, / gogo honez erranen dut zuri
Abe Maria.
‘So that you may remember me, sweet mother, with good intention I will
say Hail Mary to you’ (Etxepare, Primitiae: 93, I.445, 1545)
(38) Eçtoçu cerren berva oriec esan, Sirena

‘You have no reason to say those words, Sirena’ (Lazarraga, Eskuizkribua:
1151v, c. 1567)
The conjunction zeren is thus a polyfunctional causal element which has a

very wide range of causal usages, in some respects going beyond those of Sp
porque or Eng because.
6 Causal Subordinators of Temporal Origin
Crosslinguistically, one of the most common lexical sources of causal sub-

ordinators are time expressions referring to a point in time after the point
of reference, such as ‘after’, ‘since’, ‘already’ and even ‘when’ (in the mean-
ing of ‘at that point’, ‘and right after that’). In the Romance languages we
find several causal connectors with this origin. The most widespread ones
are those deriving from Lat post ‘after’ or its comparative variant postiu
‘more after’, which had acquired a causal usage already in Late Latin: Sp pues
(que), Port pois (que), Occ/Cat pus (que), puix (que), Fr puisque, It poiché.
Causal connectors deriving from ‘already’ are also common in the Romance
languages, but they arise later and are found in fewer languages: Sp ya que,
Port já que, Cat (and some Occ varieties) ja que, It giacché (but not in French).
Absolute participial constructions (‘given that’) can also be included in this
group: Sp puesto que, dado que, Port visto que, Cat/Occ vist que, atès que, Fr
vu que.
In Basque, the adverb gero ‘later’ (e.g. gero ikusiko dugu ‘we will see it later’)
has also developed a causal usage in postverbal position, even though, unlike
the Romance connectors just mentioned, it has kept in addition its temporal
interpretation (cf. Eng since).
With temporal value, gero is a transitive adverb and may take a complement
with the instrumental suffix -(e)z-, after the subordinator -n if the complement
is a finite clause, as in (41). In this usage it is equivalent to después de in Span-
ish. In (39) gero’s complement is a noun phrase, in (40) a participial clause,
and in (41) an inflected clause. Spanish translations are provided for compari-
son (examples from Elhuyar online dictionary):
(39) Barbaroen inbasioaz gero, ez da sekula horrelakorik

barbarian.gen.pl invasión.inst after not aux never thus.of.part
ikusi.
see
‘After the Barbarians’ invasion, nothing like that has ever been seen.’
Sp Después de la invasión de los Bárbaros, nunca se ha visto algo seme-
jante.
(40) Aita hilez gero sortu zen.

Father die.inst after be.born aux
‘S/he was born after his father had died’
Sp Nació después de morir su padre.
(41) Hanka hautsi zuenez gero, ez da eskiatzera joan.

leg break aux.inst after not aux ski.to go
‘After he broke his leg, he has not gone skiing.’
Sp Después de que se rompiera la pierna, no ha ido a esquiar.
In addition to this use as a temporal connector, gero may also have a causal or
conditional interpretation:
(42) Eta gauden erne; zer gertatuko den ez

and we.be.subor alert what happen.fut aux.subor not
dakigunez gero.
we.know.subor.inst after
‘And let’s be alert, since we do not know what will happen.’
(43) Gogor ikasiz gero, azterketa gaindituko nuke.

Hard learn.inst after exam pass.fut aux
‘If I studied hard, I would pass the test.’
The word gero itself may take the instrumental suffix (gero-z) or both the
instrumental and the ablative (gero-z-tik). We may note that a similar process of
morphological reinforcement is found in the Romance languages in its transla-
tion equivalents. Thus, in Spanish when pues (< post) acquired a causal value,
a new form después (des ‘from’+ pues) developed as a time adverb. In addi-
tion, there is a morphologically more complex nonstandard form endespués.
In Gascon as well we find similar reinforced forms: despuix (des + puix), apuix
(a +puix), d’après (de + après).
By morphological reanalysis, the sequence -(e)z gero may appear as ezkero
reflecting a pronunciation with devoicing of /g/ after a sibilant. There is also a
historical variant azkero, as in (44), where the initial vowel is from reanalysis
of the determiner sufix -a ‘the’. The reanalyzed form ezkero is alive in western
dialects, but it is no longer used in writing.
(44) a. aizkenengo konfesio azkero egin diraden pekatuak

‘the sins that have been committed after the last confession’ (Otxoa de
Arin, Doktrina, 1713)
b. konfesioaz gero > konfesio azkero

confession.det.inst after
All the forms that we have mentioned may have either temporal or non-
temporal interpretation, but there is a tendency for inflected clauses with
-(e)nez gero to be used as causal, for participial clauses with -(e)z gero to have
conditional value and for -(e)nez geroztik to be used to express time, especially
in the South (EGLU 2005: 191). To some extent, this specialization is parallel to
that between pues and después in Spanish.
The conditional and causal uses of gero are already found in the first texts
from the 15th–16th centuries. It is thus not possible to trace the grammati-
calization path of this element. Nevertheless, similar evolutions, from time
expression to causal connector, have been well-studied for other languages.
The standard view is that the causal meaning is the result of conventionaliza-
tion of an inference: post hoc, ergo propter hoc (Geis & Zwicky 1971; Levinson
1995; Traugott & Dasher 2002: 16–17, 80; Pérez Saldanya 2014: 2536–2540, among
others). By hypothesis, the causal value would arise in potentially ambigu-
ous examples like (45), where precedence in time may also be interpreted as
cause:
(45) Adanek ausi eban ezkero / Xaunaren mandamendua / Miseriazko balle

onetan / Echaku falta llantua
‘Since Adam broke the Lord’s commandment, we have not lacked tears in
this valley of sorrows.’ (Barrutia, Gabonetarako ikuskizuna, 1715)
The causal inference can only emerge in contexts like that in (45), in which the
temporal clause is external to the predicate of the main clause and appears at
the left margin of the sentence, in the thematic position (Gutiérrez Ordóñez
2000: 91; Goethals 2002: 111; Pérez Saldanya 2014: 2536–2540; Pérez Saldanya &
Hualde 2017).
The characteristics of the bridging context where the grammaticalization
process is triggered tend to persist (Heine 2002: 84). Thus, Sp pues and ya que
are only found in causal clauses external to the main predicate and this is also
true of causal clauses with gero in Basque (cf. also causal clauses with since in
English). For this reason, they cannot be used to answer the question ‘why’ in
any of these languages.
Moreover, at the initial stage in the grammaticalization process, causal con-
junctions of temporal origin are used in causal clauses providing given or pre-
supposed information, as in (46).
(46) Ene penea daucusun guero, arren, berba bat esazu

‘Since you see my pain, please, tell me a word’ (Lazarraga, Eskuizkribua:
c. 1567)
Subordinate clauses expressing given information are expected to occur pre-

posed to the main clause (which is the position where the reinterpretation from
time to cause is expected to take place), as in the example above. Nevertheless,
in Basque we find instances of postposed clauses from the first texts showing
that the grammaticalization took place some time before:
(47) Suplicaetan nachaçu deguidaçula remedioa emun, medecinea çu çarean

guero
‘I beg you that you give me the remedy, since you are the medicine’
(Lazarraga, Eskuizkribua: 1152v, c. 1567)
At an early stage, causal connectors of temporal origin tend to be used in

speech-act causals or in causals that justify some type of modality, mostly
related to volition (Santos Río 1982: 242). This is what we find in the two
examples just cited, where the subordinate clause justifies the request that the
speaker makes by means of a command, as in (46) above, or through the pred-
icate ‘to beg’, as in (47) above.
At later stages along this evolutionary path, causal clauses with conjuntions
of temporal origin may start being used in other contexts; in particular as rhe-
matic clauses in the content or even the epistemic domain (Pérez Saldanya &
Hualde 2017). This possibility is well documented in Spanish, where the con-
junction pues is nowadays only used in rhematic causal clauses (e.g. No veía
el camino, pues era de noche ‘I couldn’t see the road, since it was at nighttime’)
and ya que has added this rhematic use to its original thematic value (Pérez Sal-
danya 2014: 3551–3556). In Basque as well we may find gero in rhematic content
causal clauses in the content-domain, as in (48), and in the epistemic domain,
as in (49) (examples from the newspaper Berria have been obtained from EPG):
(48) Are handiagoa izango da gaurkoa, urteurrena mendeurren bihurtu denez

gero.
‘And today’s will be even bigger, because the anniversary has become a
centennial celebration.’ (Berria, 2004-06-16)
(49) Bere garaiko Itsasondo badoa, baina jende gehiago ere etorriko da, etxeak
eraikitzen ari direnez gero.
‘The Itsaondo of that time is going, but more people will come, because
they are building more houses.’ (Berria, 2004-10-02)
This appears to be a recent usage, perhaps due to contact with Spanish.

Nevertheless, unlike what has been observed for Spanish pues and ya que,
the extension in the use of gero in causal contexts has mostly been towards
thematic clauses in the content domain, a context where in the Romance lan-
guages we may find conjunctions of modal origin (Sp como, Fr comme):
(50) Baña aizen ezkeroz burutik maingutu artaraño, non nai dek nerekin
jokatu
‘But since you are so short of wit, you want to play with me everywhere.’
(Iturriaga, Ipuinak, 1842)
Sp ‘Pero como eres tan corto de mente, en todas partes quieres apostar
conmigo.’
(51) desengañoa arturik zegoen ezkeroz eskatu zeban mesedez laga zegiotela
edaten baso bat ur
‘As he was taken by disappointment, he asked them to please let him drink
a glass of water.’ (Iztueta, Gipuzkoako probintziaren kondaira, 1847)
Sp ‘como estaba desengañado, pidió que por favor le dejaran beber un vaso
de agua’
(52) Baina Lotek hainbeste erregutu zienez gero, berarekin joan ziren etxera.
‘But, since Lot begged them so much, they went home with him.’
(Elizen Arteko Biblia: Hasiera [Genesis] 19:3, 2008)
Sp ‘Pero como Lot les suplicó tanto, se fueron con él a casa.’
Basque causal clauses with gero have moved farther along the grammatical-
ization path than Fr puisque, which is used only in thematic clauses in the
speech-act domain or expressing modality (Ducrot 1983; Zufferey 2012), but less
than Sp pues and ya que, which have completely lost all temporal usages (Pérez
Saldanya 2014).
Two factors may have contributed to the extension of usages of gero in con-
tent thematic clauses. One of them may have been the development of adver-
bial clauses with -enez with thematic content value (e.g. ikusi dugunez ‘as we
have seen’, section 7.1), which may have created an equivalence between -enez
and -enez gero in this type of causal clause (Artiagoitia 2003: 717–719). The other
factor could be the desire to avoid the use of clause-initial nola ‘how, as’ (sec-
tion 7.2), since clause-initial subordinators have sometimes been regarded as
calques from Romance syntax.
7 Causal Constructions of Modal Origin
Modal subordinators are also a common lexical source of causal subordinators

(Kortmann 1977: 195–197). As is the case for causal subordinators of temporal
origin, the causal meaning is the result of conventionalization of an inference
that emerges in certain contexts; specifically, in contexts in which the modal
clause is external to the predicate of the main clause and is placed in the left
periphery of the sentence, in thematic position, as in, e.g., As you want it, we will
do it (Pérez Saldanya 2014: 3576–3579). In the Romance languages, the modal
conjunction deriving form the Latin adverb quomodo ‘as, like’ (> Sp/Port
como, Cat com que, Occ coma, It come, Fr comme, Rum cum) has acquired a
causal interpretation. In Basque, this grammaticalization is also the source of
the causal use of the interrogative nola and the complex verbal suffix -enez.
7.1 How > Cause

In Basque, the interrogative word nola ‘how’ can be used to introduce subor-
dinate clauses with causal interpretation (‘since, as’), in addition to other uses
of this subordinator in indirect interrogatives and comparatives. In these sen-
tences, the finite verb of the subordinate clause bears either the prefix bait- or
the suffix -en, both marking subordination. Causal constructions of modal ori-
gin tend to have different usages than those originating from time adverbials.
In particular, their main use is as thematic causals that justify the propositional
content of the main clause. This can be observed in the following 17th century
examples of nola causals, from Axular’s Gero (Spanish translations from L. Vil-
lasante’s bilingual edition are also provided for comparison):
(53) Eta nola bertze mundukoa baita geroko, erraiten dute,

and how other world.of.det comp.is after.of say aux
eztutela presenteko plazera, geroko esperantzagatik utzi
not.aux.subor present.of pleasure after.of hope.for leave
nahi
want
‘And since what pertains to the other world is for later, they say that they
do not want to leave the pleasure of the present for the hope of the future’
Sp ‘Y como lo del otro mundo es para después, dicen que no quieren dejar
el placer presente por la esperanza futura’ (L. Villasante’s translation)
(54) Zeren zu nola baitzaude deliberatua, eta bai batzutan

because you how comp.you.are decided and also sometimes
pasionatua eta itsutua ere, eztuzu den guztia ikhusten

passioned and blinded too not.aux aux.subor all see
‘Because, since you are decided, and sometimes even passionate and
blind, you do not see everything there is’ (Axular, Gero: 211, 1643)
Sp ‘Porque, como tú estás decidido y a veces hasta enardecido y cegado,
no ves toda la realidad’ (L. Villasante’s translation)
(55) Eta nola baitzihoan dudatuaz eta fidantziaren galduaz, hala

and how comp.he.went doubting and trust.gen losing thus
zihoan itsasoan barrena ere sarthuaz, estaliaz eta hondatuaz
he.went sea.loc through also entering covering and sinking
‘And, as/since he was doubting, and losing his confidence, he was thus
sinking and drowning in the sea’ (Axular, Gero: 87, 1643)
Sp ‘Y como iba dudando y perdiendo la confianza, en la misma medida
se iba metiendo dentro del mar, se iba hundiendo’ (L. Villasante’s transla-
tion)
The following 19th century example cited in Hualde & Ortiz de Urbina (2013:
834) shows the use of nola as a causal conjunction ‘since, as’ and as a compar-
ative (‘like, as’):
(56) nola sorho horiek ez baitute hainbertze lan nola

how field those not comp.aux as.much work as
laborantzek
tillage.lands.erg
‘Since those fields do not require as much work as tillage lands’ (Duvoisin,
Laborantzako liburua 1858: 79)
cf. Sp como esos campos no requieren tanto trabajo como los de labrantío
As can be seen in the Spanish translations of the examples above, the use of
nola as a subordinator is very much like in Spanish, where interrogative cómo
‘how’ has also developed an unstressed variant como with these modal and
causal values:
(57) ¿Cómo lo has hecho?

‘How did you do it?’
(58) No sé cómo lo has hecho.

‘I do not know how you did it.’ (stressed cómo, indirect interrogative)
(59) Como lo has hecho, lo tienes que pagar.

‘Since you did it, you must pay for it.’ (causal)
However, whereas Spanish unstressed como can also introduce purely modal
clauses, in Basque a different construction is used in this case, with bezala ‘like,
as’, moduan ‘in the manner’ or legez (Bizkaian, from lege ‘law’ in the instrumen-
tal case) ‘like, as’, postposed to a verb bearing the subordination marker -(e)n:
(60) a. Debes hacerlo como te digo.

‘You must do as I tell you.’ (modal)
b. Esaten dizudan bezala/moduan/legez egin behar duzu

say.impf aux.subor do need aux
‘You must do as I tell you.’ (modal)
c. *Nola esaten dizudan egin behar duzu / *Egin behar duzu nola esaten
dizudan
Nevertheless, in a relatively small western area the construction with legez ‘as,
like’ has expanded its use to cover the same space as Spanish como, including
subordinate clauses with both a modal and a causal interpretation (Zuazo 2017:
61). This seems to be a very recent development.
Notice that the use of nola as both an interrogative ‘how’ and as a causal
subordinator ‘as, since’ is parallel to that of zeren ‘why; because’. It can thus be
explained in a similar fashion, as a possible syntactic calque from Romance.7
7.2 Instrumental > Cause

Subordinate clauses in which an inflected verb bears the instrumental suffix
-(e)z (following the subordination marker -(e)n) are essentially modal, as in (61)
but they may also have a causal interpretation, as in the examples in (62)–(63)
(examples from newspapers, extracted from EPG):
(61) Jakin dugunez, beste bilera paralelo bat zegoen.

know aux.inst other meeting parallel one there.was
‘As we have found out, there was another, parallel meeting.’
7 Modal expressions may also give rise to causal conjunctions, cf. Eng how come? In Sp we find
cómo es que, cómo así and in Bq nolatan ‘why, how come’, with nola ‘how’ inflected in the
locative case.
(62) Euskal Herri osoko administraziorik ez dugunez,

Basque Country whole.of administration.part not we.have.inst
Udalbiltza ezinbesteko eragilea da.
U. necessary agent is
‘Since we do not have an administration of the whole Basque Country,
Udalbiltza (Federation of Town Governments) is a necessary agent.’
(63) Gimnasiorik ez dugunez, auzoko kiroldegira joan

gym.part not we.have.inst neighborhood.of sports.center.to go
behar izaten dugu.
must be aux
‘Since we do not have a gym, we have to go to the neighborhood’s sports
center.’
This construction is historically very recent, dating back only to the 18th cen-
tury. It appears to have resulted from the omission of gero in the older construc-
tion -enez gero (§6). The first examples of -enez without gero have modal value
and involve expressions such as dirudienez, antza denez ‘as it seems, appar-
ently’, dioenez ‘as s/he says’ and the like.
Clauses with -enez have a very similar distribution and interpretation as
Spanish subordinate modal and causal sentences headed by unstressed como
‘as, like; since’. Thus, the Spanish translation of the subordinate clause in the
three examples given in (61)–(63) would be, respectively, Como hemos sabido
‘As we have found out’, Como no tenemos una administración de todo el País
Vasco ‘Since we do not have administration of the whole Basque Country’ and
Como no tenemos gimnasio ‘Since we do not have a gym’. This thus seems to be
a case of convergence with Spanish in recent Basque usage, even if the actual
morphosyntactic mechanisms are quite different.
8 Conclusions
In this chapter, we have examined the main types of causal structures in

Basque, establishing a comparison with Romance structures and develop-
ments. Structurally, Basque has more ways to indicate subordination than
Romance, including verbal prefixes and suffixes, as well as clause-initial con-
nectors, and all these elements are used in causal constructions. Clause-initial
subordinators such as zeren ‘why; because’ and nola ‘how; as’ may have arisen
under the influence of a Romance model, but they are found from our earliest
texts.
table 10.3 Causal subordinators in Basque and Spanish
Type of causal Subordinators
internal (rhematic) (content) Bq.: -(e)lako, zeren … bait-/-(e)n

Sp.: porque
external rhematic content Bq: zeren bait-/-(e)n, bait-, -(e)ta

Sp: pues, porque, ya que, puesto que, †ca, †que
epistemic Bq: zeren bait-/-(e)n, bait-, (e)ta, zeren eta

Sp: porque, pues, †ca
speech-act Bq: bait-, eta; ze

Sp: que, porque, †ca
thematic content Bq: -nez (gero), bait-, nola … bait-/-(e)n, zeren

(or presuposed) bait-/-(e)n
Sp: como, † porque
speech-act Bq: -nez gero

Sp: ya que, † pues, puesto que, dado que
In many respects, the grammaticalization paths of causal elements are par-

allel to those found in the Romance languages, even if the specific structures
are very different in many cases, as with bait- ‘that’ and gero ‘later, since’. Other
developments in Basque, however, lack a Romance counterpart. This is the case
in the grammaticalization of enclitic eta ‘and’ as a causal connector.
From a functional point of view, in Basque, like in Spanish and French, there
is a causal subordinator that has a polyfunctional character and other causal
subordinators with more restricted usages. In Table 10.3 we include the main
causal subordinators in Basque and Spanish, classifying them according to the
main types of causal clauses in which they tend to appear. (We indicate obso-
lete Spanish conjunctions with “†”)
Basque zeren and Spanish porque are polyfunctional, since they may appear
both in internal causal clauses and in different types of external clauses. The
rest of the subordinators in Table 10.3 have a more restricted usage. Notice
that whereas Basque has a specialized subordinator in internal clauses, -(e)lako,
Spanish lacks a conjunction with these properties.
The Basque subordinators bait- and -(e)ta have a certain functional paral-
lelism with the Spanish conjunction que (and Old Spanish ca), since all of them
can appear in external rhematic causal clauses. Nevertheless, Bq bait- may also
be found in external thematic clauses.
In both Basque and Spanish (and other Romance languages) we find causal
subordinators of temporal origin: Bq gero, Sp pues (que), ya que. These subor-
dinators, in a first stage, are mostly used in thematic causals in the speech-act
domain. Later, they tend to broaden their usage to other external causal mean-
ings. This evolution has been different in Basque in Spanish. In Basque the
broadening of usage has been towards content thematic clauses. In Spanish,
instead, it has been towards rhematic clauses, in particular, those in the con-
tent domain.
Finally, in both languages we also find causal subordinators of modal origin,
employed in content thematic causal clauses: Bq -nez, nola and Sp como.
Acknowledgements
The first author acknowledges the research funding received from the Spanish
Ministry of Economy and Competitiveness (FFI2014–51878-P).
References
Alibèrt, Lois. 1935–1937. Gramatica occitana segons los parlars lengadocians. Montpel-
hièr: CEO.
Artiagoitia, Xabier. 2003. “Adjunct subordination”. In Jose Ignacio Hualde & Jon Ortiz
de Urbina (eds.). 2003. 710–762.
Azkue, Resurrección María de. 1923–1925. Morfología Vasca. Published in fascicles in
Euskera. [Repr. 1969, Bilbao: La Gran Enciclopedia Vasca.]
Baños, José Miguel. 2011. “Causal clauses”. In Philip Baldi & Pierlougi Cuzzolin (eds.),
New perspectives on the historical Latin syntax, 4: Complex sentences, grammatical-
ization, typology. Mouton de Gruyter: New York-Amsterdam. 195–234.
Baños, José Miguel. 2014. Las oraciones causales en latín: su evolución diacrónica.
Madrid: Escolar y Mayo editores.
Ducrot, Oswal. 1983. “Puisque: essai de description polyphonique”. Revue Romane 24.
166–185.
EGLU = Euskaltzaindia. 2005. Euskal gramatika lehen urratsak: Lehen urratsak-VI, (Men-
deko perpausak-2, baldintzazkoak, denborazkoak, helburuzkoak, kausazkoak, kontse-
ziozkoak eta moduzkoak). [Available online: www.euskaltzaindia.eus].
Elhuyar hiztegiak. Online: https://hiztegiak.elhuyar.eus

EPG = Ibon Sarasola, Pello Salaburu, Josu Landa & Josu Zabaleta. 2009. Ereduzko
Prosa Gaur [Model Prose Today]. Online corpus: http://www.ehu.eus/euskara‑orria/
euskara/ereduzkoa/
Fernández González, José Ramón. 1985. Gramática histórica provenzal. Oviedo: Univer-
sidad de Oviedo.
Geis, Michael & Arnold Zwicky. 1971. “On invited inferences”. Linguistic Inquiry 2: 561–
566.
Goethals, Patrick. 2002. Las conjunciones causales explicativas en castellano. Un estudio
semiótico-lingüístico. Lovaina: Peeters.
Goethals, Patrick. 2010. “A multi-layered approach to speech events. The case of Span-
ish justificational conjunctions”, Journal of Pragmatics 42. 2204–2218.
Gutiérrez Ordóñez, Salvador. 2000. “Causales”, Boletín de la Real Academia Española 80.
47–159.
Heine, Bernd. 2002. “On the role of context in grammaticalization”. In Ilse Wischer
and Gabriele Diewald (eds.), New Reflections on Grammaticalization. Amsterdam /
Philadelphia: Benjamins. 83–101.
Hopper, Paul J. 1991. “On some principles of grammaticization”. In Elizabeth Closs Trau-
gott & Bernd Heine (eds.), Approaches to Grammaticalization, vol. I. Amsterdam:
John Benjamins, 1991. 17–36.
Hualde, José Ignacio & Jon Ortiz de Urbina (eds). 2003. A Grammar of Basque, Berlin /
New York: De Gruyter Mouton.
Kortmann, Bernard. 1997. Adverbial Subordination. A Typology and History of Adverbial
Subordinators Based on European Languages. Berlin-New York: Mouton de Gruyter.
Kortmann, Bernard. 1998. “Adverbial subordinators in the languages of Europe”. In
Johan van der Auwera (ed.), Adverbial Constructions in the Languages of Europe.
Berlin and New York: Mouton de Gruyter. 457–561.
Krajewska, Dorota. 2013. The diachrony of resultative constructions in Basque. Master’s
Thesis, UPV/EHU.
Krajewska, Dorota. 2016. “Basque subordinate clauses: a diachronic study”. Presented at
49th Annual Meeting of the Societas Linguistica Europaea 31 August–3 September
2016, University of Naples Federico II, Naples.
Levinson, Stephen. 1995. “Three levels of meaning”. In Frank R. Palmer (ed.), Grammar
and meaning: Essays in honor of Sir John Lyons. Cambridge: Cambridge University
Press. 90–115.
Lexilogos. 2002–2017. Dictionnaire béarnais. [Available online: http://www.lexilogos
.com/bearnais_dictionnaire.htm]
Mounole, Céline. forthcoming. Lazarragaren eskuizkribuko gramatika, UPV/EHU.
OEH = Euskaltzaindia, Orotariko Euskal Hiztegia [Available online: http://www.
euskaltzaindia.eus/index.php?option=com_content&view=article&id=
276&Itemid=413&lang=eu]
Ortiz de Urbina, Jon 2003. “Word order”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.). 2003. 448–459.
Oyharçabal, Bernard. 2003. “Relatives”. In Jose Ignacio Hualde & Jon Ortiz de Urbina
(eds.). 2003. 762–823.
Palay, Simin. 1974. Dictionnaire du béarnais et du gascon modernes (Bassin Aquitain):
embrassant les dialectes du Béarn, de la Bigorre, du Gers, des Landes, et de la Gascogne
maritime et garonnaise. Paris: Centre National de la Recherche Scientifique.
Pérez Saldanya, Manuel. 2014. “Oraciones causales”. In Concepción Company (dir.), Sin-
taxis histórica de la lengua española. Tercera parte: Preposiciones, adverbios y conjun-
ciones. Relaciones interoracionales, vol. 3. Ciudad de México: Universidad Nacional
de México / Fondo de Cultura Económica. 3447–3610.
Pérez Saldanya, Manuel and José Ignacio Hualde. 2017. “From theme to rheme: The evo-
lution of causal conjunctions of temporal origin in Catalan”, Studies in Hispanic and
Lusophone Linguistics 10. 319–348.
Pineda, Anna & Manuel Pérez Saldanya. 2017. “Les construccions causals i finals”. In
Josep Martines & Manuel Pérez Saldanya (eds) Gramàtica del català antic. In prepa-
ration.
Santos Río, Luis. 1982. “Reflexiones sobre la expresión de la causa en español”, Studia
Philológica Salmanticensia 6. 231–277.
Sweetser, Eve. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects
of Semantic Structure. Cambridge: Cambridge University Press.
Traugott, Elisabeth C. & Richard B. Dasher. 2002. Regularity in Semantic Change. Cam-
Villasante, Luis. 1981a. “Las oraciones causales en Axular (I): La construcción de zeren
con partícula que modifica el verbo”. Fontes linguae vasconum: Studia et documenta
37. 9–18.
Villasante, Luis. 1981b. “Las oraciones causales en Axular (II): Causales subordinadas y
Causales coordinadas”. Fontes linguae vasconum: Studia et documenta 38. 9–22
Villasante, Luis. 1982a. Las oraciones causales en Axular (III): Oraciones causales fron-
terizas. Fontes linguae vasconum: Studia et documenta 39. 9–20.
Villasante, Luis. 1982b. “Las oraciones causales en Axular (IV): Oraciones causales con-
struidas a base de bait solo”. Fontes linguae vasconum: Studia et documenta 40. 359–
386.
Zuazo, Koldo. 2017. Mendebaleko euskara. Donostia: Elkar.
Zufferey, Sandrine. 2012. “‘Car, parce que, puisque’ revisited: Three empirical studies on
French causal connectives”. Journal of Pragmatics 44. 138–153.
Name Index
Abney, Stephen 21 Bosse, Solveig 4, 8, 221–223, 225, 231–232,

Acedo-Matellán, Víctor 7, 168, 183, 201–202, 235, 237–240
205 Bossong, Georg 3, 243, 246, 250
Ahn, Hee Don 178 Bowern, Claire 176, 178
Aikhenwald, Alexandra Y. 298 Brody, Michael 204
Aissen, Judith 3, 243, 246, 301 Bruening, Benjamin 4, 8, 82, 84, 124, 221–
Alba-Salas, Josep 178, 184, 193 223, 225, 232, 235, 237–240
Alberdi, Xabier 160 Brugè, Laura 251
Albizu, Pablo 261–262, 305–306 Brugger, Gerhard 251
Aldai, Gontzal 63–65, 72, 145, 151, 178 Burzio, Luigi 7, 139, 143, 154
Alexiadou, Artemis 5, 41, 82–84, 99, 108, 113, Butt, Miriam 176, 178–180, 215
119, 124–125, 130–131 Bybee, Joan 60, 67, 73
Alibèrt, Lois 319 Bye, Patrick 204
Allières, Jacques 2, 77
Alonso-Ramos, Margarita 185, 192–194 Carlson, Greg 102–103, 131
Altube, Seber 42, 47, 49–50 Castillo, María Elena 158
Anagnostopoulou, Elena 5, 7, 41, 82, 84, 94, Cattell, Ray 178–179
107, 111–113, 124 Chierchia, Genaro 131
Aranovich, Raúl 158 Chomsky, Noam 204–205
Arnold, Douglas J. 101–102 Cinque, Guglielmo 18–19, 29, 44
Arraztio, Kontxi 250, 267 Comrie, Bernard 8, 112, 114, 276, 298
Arregi, Karlos 44, 223, 235 Contreras, Heles 41
Artiagoitia, Xabier 22, 110, 321–322, 337 Cruschina, Silvio 44
Austin, Jennifer 204 Cuervo, María Cristina 140, 166, 168–170,
Azkue, Resurrección María 324 194, 228, 262
Baker, Mark 86, 204, 279, 281 Dahl, Östen 67

Baños, José Miguel 330 Danlos, Laurence 192–193, 293
Barrie, Michael 204 Dannenberg, Clare 226
Basilico, David 168 Dasher, Richard B. 335
Batlle, Mar 158 Davidson, Donald 148
Belletti, Adriana 7, 139, 154 Dayal, Veneeta 131
Bennet, Paul A. 18 de Miguel, Elena 221, 229
Bentley, Delia 139, 155 de Rijk, Rudolf P.G. 15, 17, 23, 27–31, 34, 36–
Berro, Ane 4, 5, 7, 14, 25, 83, 106, 110, 131, 145, 37, 39–40, 87, 89, 92, 110, 199, 203, 262,
148–149, 151–152, 154, 162–163, 167, 205, 277, 299, 310
212, 234, 262, 278, 291, 299 Demonte, Violeta 5, 26, 83, 261–264,
Bianchi, Valentina 44 268
Birabent, Jean-Pierre 69, 77–78 Detges, Ulrich 93
Bittner, Maria 139 DHM 279, 281
Bobaljik, Jonathan D. 146, 204 Dixon, Robert M.W. 8, 139, 276, 298
Bordelois, Ivonne 284 Doron, Edit 82, 165, 168
Borer, Hagit 4, 222, 230, 235 Dryer, Mathew 26, 29, 33
Bosque, Ignacio 5, 16, 34, 83–84, 90, Ducrot, Oswal 337
95, 105–108, 111–114, 117, 122, 127, Duguine, Maia 48
198
348 name index
EGLU 321, 323, 327–328, 335 Grodzinsky, Yosef 4, 222, 230, 235
Eguren, Luis 86 Gross, Maurice 192–193
Elordieta, Arantzazu 42–44, 46, 48, 75, 87, Gutiérrez Ordóñez, Salvador 335
248
Elordieta, Gorka 44, 46, 75, 248 Haase, Martin 66
Embick, David 7, 82–83, 87–88, 90–92, 95, Haddican, Bill 48
99, 107, 111–113 Hale, Kenneth 139–140, 145, 176, 179, 198,
Epelde, Irantzu 212 203–204, 206
Escandell Vidal, Victoria 44–46 Harley, Heidi 140, 148–149, 168, 205, 292–
Etxeberria, Urtzi 22 293
Etxepare, Ricardo 42, 46, 145, 151, 194, Harre, Catherine 73
196, 203–204, 207–208, 212, 255, 261, Haspelmath, Martin 33
299 Haugen, Jason 204, 206
Euskaltzaindia 26–27, 37, 48, 50, 287 Heine, Bernd 335
Eythórsson, Thórhallur 139, 155 Hernanz, Maria Lluïsa 264
Hidalgo, Bittor 40, 42
Fábregas, Antonio 90–91, 106, 244, 251, 259– Hofmann, Johann Baptist 330
260 Hook, Peter E. 179, 181
Fernández, Beatriz 4, 8, 14, 24–25, 82, 145– Hopper, Paul J. 179, 181
146, 148, 198, 204, 221–223, 225, 228, 232, Horn, Laurence R. 4, 8, 221–223
234–235, 243–244, 246–247, 249, 251, Hualde, José Ignacio 2–3, 6, 9, 14, 44, 46, 51,
255–257, 261–262, 266–267, 269, 280, 75, 87, 92, 110, 114, 116, 248, 332, 335–
300, 305, 310 336, 339
Fernández González, José Ramón 319 Huidobro, Susana 8, 221, 223, 225–229, 231–
Fernández Lagunilla, Marina 221 233, 239
Fernández Ordóñez, Inés 243, 252–253, 255,
260, 312–313 Iatridou, Sabine 94, 112, 114
Fernández Soriano, Olga 265 Ibarra, Orreaga 248
Fleischman, Suzanne 60 Irurtzun, Aritz 46–47
Folli, Rafaella 140, 168, 292–293 Izvorski, Roumyana 112
Franco, Jon 8, 221, 223, 225–229, 231–233, Jackendoff, Ray 178–179, 189, 198
239, 243, 253 Jaeggli, Osvaldo 4, 221, 228, 230
Jaxontov, Sergej 93
García Fernández, Luis 91 Jespersen, Otto 178
García Murga, Fernando 182, 189, 198, 200, Jiménez Fernández, Ángel 44, 46
203–204 Jouitteau, Mélanie 4, 222, 228, 230, 233
Gawron, Jean Marc 106, 108
Gehrke, Berit 5, 7, 83, 84, 112, 120, 128– Kayne, Richard S. 18, 24, 139, 155
131 Kearns, Kate 178–179
Geis, Michael 335 Keller, Frank 139, 155
Geuder, Wilhelm 178, 181 Kennedy, Christopher 97
Giry-Schneider, Jacqueline 185, 193, 201 Keyser, Jay S. 140, 145, 176, 179, 198, 203–204,
Goenaga, Patxi 33, 87 206
Goethals, Patrick 317, 335 Kiss, Katalin É. 46
Goldberg, Adele 194 Koizumi, Masatoshi 262
Gómez, Rikardo 59, 78 Koontz-Garboden, Andrew 106
Grafmiller, Jason 49 Kortman, Bernard 319, 330, 338
Greenberg, Joseph 17–18 Krajewska, Dorota 5, 73, 83–84, 86, 112, 114–
Grimshaw, Jane 178–190, 193–194, 198 117, 133, 325
name index 349
Kratzer, Angelika 7, 82–84, 88, 97, 105, 107, Mohanan, Tara 178
111–113, 119, 124–125, 141, 143 Mounole, Céline 2, 6, 15, 63–66, 68, 71–73,
75, 93, 243, 246–247, 249–250
Labelle, Marie 165, 168
Lafitte, Pierre 40, 47, 204 Nedjalkov, Vladimir 93
Lafon, René 2, 71
Lahiri, Aditi 178, 180 Odria, Ane 4, 8, 82, 222–223, 229, 232, 234,
Laka, Itziar 43, 145–146, 148, 204, 213, 278, 240, 243–244, 246–247, 249, 251, 261–
299 263, 266–267, 269, 310, 312–313
Lakarra, Joseba 72, 75 Ordóñez, Francisco 41
Landa, Alazne 8, 243, 252–255, 312 Ormazabal, Javier 251, 261, 303
Landa, Josu 261–262 OEH 20, 34, 257, 276–280, 287
Larson, Richard K. 176 Ortega, Iván 44–45, 47
Lazard, Gilbert 3 Ortiz de Urbina, Jon 3–5, 6, 8, 40, 42, 47, 73,
Ledgeway, Adam 60 76, 83, 86, 92, 99, 126, 139, 142, 204, 208,
Legrende, Géraldine 168 223, 232, 244, 255–257, 261–262, 269,
Lehman, Winfred P. 17 291, 300, 304–305, 322, 332, 339
Leonetti, Manuel 44–46, 243, 251, 259 Oyharçabal, Bernard [also Beñat Oihartza-
Levin, Beth 4, 7, 95, 105, 139, 141–143, 146, bal] 4, 29, 31, 87, 145–148, 197, 198, 204,
152, 160, 179, 194, 204, 278, 299 206, 207, 209, 210–213, 227, 261–262,
Levinson, Stephen 335 285, 287, 291, 293, 305, 325
Lin, Tzong Hong 176, 178
López, Luis 204 Palay, Simin 319
Luján, Marta 5, 83 Pancheva, Roumyana 112, 114
Parsons, Terence 113
MacKenzie, Ian 139, 155 Pastor, Alberto 38
Maienborn, Claudia 106 Perez Gaztelu, Elixabete 28
Maldonado, Ricardo 221, 223, 225, 228, 233, Pérez Saldanya, Manuel 3, 9, 317, 319, 326,
237 330, 332, 335–338
Manterola, Julen 65 Pensado, Carmen 243, 251, 258
Marantz, Alec 5, 7, 82, 189, 191 Perlmutter, David 7, 139, 143, 230, 234, 283,
Marco, Cristina 5, 7, 83–84, 128–130 291, 301
Marín, Rafael 5, 83, 90–91, 106 Pesetsky, David 205
Martinez, Arantzazu 178, 182, 185, 188, 194, Piera, Carlos 200
196–200, 203–206, 208, 213–214 Pineda, Anna 7, 152, 154, 194, 244, 258, 260–
Masullo, Pascual 198, 262, 264 261, 330
Mateu, Jaume 140, 148–149, 157–158, 168, Pinkster, Harm 93
183 Potts, Chris 222, 235
Mathieu, Éric 204 Pylkkänen, Liina 4, 187, 194, 223, 232, 297,
Matushansky, Ora 204 305–306
McClure, William 152
McFadden, Thomas 119, 187, 262 Quer, Josep 44
McIntyre, Andrew 95, 124–125
McNally, Louise 97 Rafel, Joan 192, 193
Merlan, Francesca 139 Ramchand, Gillian 91, 105, 140, 148, 178,
Mester, Armin 178, 179–190, 198 180–181, 189, 192, 214
Michelena, Luis [also Koldo Mitxelena] 15, Rapp, Irene 120
19, 42, 59, 65, 257 Rappaport Hovav, Malka 7, 95, 105, 139, 141–
Mithun, Marianne 139, 182, 204, 206 143, 146, 152, 160, 179, 194
350 name index
Rebuschi, George 86 Tenny, Carol 143

Rezac, Milan 4, 222, 228, 230, 233, 243, 246– Thibault, Pierrette 155
247, 249, 251, 266, 267, 269, 303, 306, Torrego, Esther 4, 154, 205, 243–244, 251,
310 259–260, 294
Rodríguez, Sonia 182, 189, 198, 200, 203 Trask, Larry 26–27, 34, 59–60, 71
Rodríguez Mondoñedo, Miguel 251 Traugott, Elizabeth 179, 181, 335
Rodríguez Ordóñez, Itsaso 243, 246–247, Treviño, Esthela 41, 284, 294
251, 255 Tubino, Mercedes 284, 292, 294–295
Rohlfs, Gerhard 76, 78
Romero, Juan 251, 261–262, 303 Urgell, Blanca 63
Rosen, Carol 139 Uribe-Etxebarria, Myriam 5, 73, 83, 86, 92,
Rotaetxe, Karmele 40 99, 126, 145–146, 204
Rothmayr, Antonia 106 Urrutia-Cárdenas, Hernán 252
Rothstein, Susan 262
Varela, Soledad 200
Sadler, Louisa 101–102 Villasante, Luis 40, 73, 325, 338
Sainz, Koldo 59, 78 Vincent, Nigel 60
Sainz-Maza, Lorena 45 von Stechow, Arnim 120
Salaburu, Pello 142
Salles-Loustau, Jean 69, 77–78 Wasow, Thomas 5, 7, 82, 95, 119
Saltarelli, Mario 37 Webelhuth, Gert 226
Samioti, Yota 5, 82, 107, 111 Wichmann, Søren 139
Sankoff, Gillian 155 Wierzbicka, Anna 178
Santos Río, Luis 317, 336 Williams, Edwin S. 262
Sapir, Edward 139
Sarasola, Ibon 24, 145, 204, 257, 262, 280 Yamada, Masahiro 4, 8, 221–223, 225, 232,
Schäfer, Florian 5 235, 237–240
Shih, Stephanie 49 Yrigarai, Ángel 19
Siewierska, Anna 40
Sleeman, Petra 82, 95, 99 Zabala, Igone 73, 89, 145, 182, 188, 190, 192,
Smolensky, Paul 168 194, 197–200, 203–205, 208–209, 263
Sorace, Antonella 7, 139, 155–157, 159, 167 Zamparelli, Roberto 131
Strozer, Judith 226 Zuazo, Koldo 340
Svenonius, Peter 102, 204 Zubizarreta, María Luisa 9, 44, 279, 282–
Szantyr, Anton 330 283, 290
Zufferey, Sandrine 337
Zwicky, Arnold 335
Subject Index
Boldface numbers indicate (sub)sections where the subject is specifically addressed.
absolutive 52, 114 and secondary predication 268

adjectival participle 82, 154 as DP or PP 268
(un)bounded interpretation 127 Applicative head 222, 232, 256, 305
agent 124, 125, 126, 129, 132 appositions 18, 27, 31, 35
agreement 85 (non-)restrictive 31
and manner adverbs 121 Aragonese 2, 155
and temporal modification 120 Argument Fusion 180
and verbal participle 82, 84, 95, 116, 119, argument structure 180, 186, 190, 194, 224
124 arguments and participants 194
and Voice phrase 124 articles. See determiners
as complement of seem verbs 96 aspect 59
aspectual verb classes 105 (im)perfectivity in past 67
attributive and predicative 83, 86 Aspect head 88
compatibility with degree modifiers 97 aspectual particle 76
eventive vs. resultative 90 Bsq double-compound perfective 65
event-related modifiers 119 frequentative periphrases 75
experiential interpretation 114 habitual periphrases 75
modifying bare N or DP 99 lexical and grammatical 83, 103
modifying external arguments 103 old Bsq perfective forms 63
stative and resultative 107 perfective 282
temporal and locative modification 122 present perfect 64
vs. perfective adjective 91 progressive 72, 283
adjectives progressive and locative marking 66
Adjective Phrase 32 progressive grammaticalization 72
attributive 26 resultative periphrasis 75
classifying 29 auxiliary 180, 287
complement and adjunct 33 aux alternation 119, 154, 155, 224
complex attributive AdjP 35 aux alternation hierarchy 155
extraposition 32 aux selection 59, 63, 66–68, 78, 139, 140
perfective 90, 91 preposing 43, 47
pre- and postnominal 101–102 semiauxiliary 54. See also modal parti-
preposed 26, 27 cles
qualifying 29 Aymara 246
relational 26
stage-level 89 bare nominals 146, 148, 182, 195, 200–201
adverbs Basque Spanish 132, 243, 252, 260, 268, 312–
activity adverbial 194 313
modal 323 benefactive 62, 65
stative 89 Bengali 180
affixation 17 Biscayan 21, 54, 68, 72
Albanian 235, 237
allocutives 227, 239 Case 148, 181, 204–205, 213
animacy 244, 246, 252 Catalan 74, 76, 78, 140, 155, 158, 183, 185, 187,
a-objects in Sp. 251, 258, 260, 313 202, 330, 333, 338
352 subject index
causal clauses 317 causal 333

and modal 337, 340 -en 322, 325, 333, 337–338, 340
causal subordinators 318 -la 69, 321, 323, 325
content causals 317, 331 que 323, 326
epistemic causals 317, 323 completive clauses 325
external 317, 323, 324, 331–332, 335 compounds 14–16, 26–28, 279
internal 317, 319, 320, 323, 330 conditional clauses 49, 69, 334–335
rhematic 320, 325–326, 336 conjunctions 324
speech-act causals 317, 331, 336 adversative 328
thematic 320, 325–326, 332, 335 and wh-words 323, 328–330, 338–340
thematic and rhematic 317 causal 328
causatives 95, 208, 212, 263, 276 of modal origin 338
affix as a verb 277 of temporal origin 333
and infinitive adjacency 281 coordinating 319, 324, 327
animate causee 311 copulative 87
archaic affix 277 disjunctive 328
aspect and passive complement 282 modal 337–338
causee and ethical datives 305 consecutive clauses 325
causes and agents 292 converbs 86
clausal complement 279 copular verbs 92, 116–118, 126, 158
direct and indirect 284 copular verb vs. auxiliary 92, 117
impersonal 283, 289, 301, 304, 309 transitive copulas 93
indirect causation 279 with adjectival vs. passive participle 95
mono- or biclausal 283 with resultative participle 73, 83
negated complement 282
nominalized complement 280 datives 4, 24–25, 187, 191–192, 194, 208, 221,
of abs/dat verbs 301 243
of bivalent unergative 300 (non)argumental 231
of dativeless verbs 298 adnominals 24, 262
of ditransitive verbs 306 affected 4
of light verbs 278 affected experiencer 221, 223–224, 235,
of monotransitive 299 237–238
of simplex unergatives 299 agreement dropping 232, 245, 261–262,
of unaccusative 312 301
passivization 283, 293 alternation with absolutive 256
tensed complement 285 and secondary predicates 262
Chinese 178 as DP or PP 244, 261
clause union 282, 283 aspectual 226, 229
clitics 225, 312 attitude holder 235
aspectual 221 benefactive 221–223, 229–230, 232–233,
climbing 282, 301 235, 238, 239
Clitic Left Dislocation 41 binding a variable 236
cluster 226, 229 causee 244, 262–263, 268, 276, 297, 309
doubling 222, 227, 230, 232, 253 combinations 227, 229, 233
encliticization 327, 328 double dative 234, 262, 300, 302, 305–
non volitional and dative clitic 221 306, 311
comparatives 288 ethical 4, 221, 305, 308, 311
complementizers 77, 319, 324, 328, 337, 341 negation 237
bait- 30, 325, 338 person restrictions 221, 227–229
subject index 353
experiencer 177, 222, 230, 232, 240, 244, event 180, 189, 198, 237, 297
261–262, 265, 284, 288, 301 bare event noun 180
external possessor 221–223, 228, 230, configuration and theta-role 150, 167
232–233, 235, 239, 240, 244, 261, 262, configurations 140, 148
264 event structure 186
goal 233–234, 244, 262, 301, 303, events and states 148
309 flavor 149, 153, 169
high and low 223 in light verb constructions 176
in bivalent unergative 269 kinds and tokens 129, 131
in unergative and ditransitive 256, 260 exclamatives 45
malefactive 225 experiencer subject 177
subject co-referential 235, 239 extraposition 34, 37
typology 239
degree words 32, 33–34, 36, 38, 98 final clauses 323, 333
degree adverbs and PP s 34 focalization 42–44, 207, 209, 322, 330
extraction 39 contrastive 45, 48
demonstratives. See determiners corrective 42, 47, 52–53
derivation 17, 28 focal stress 44, 46
destinative 62 Focus Fronting 44
determiners 21, 37, 198, 213 Mirative Focus 44–46
and predicative agreement 65 presentational 52
definiteness 38–39 Quantifier Fronting 44, 46
demonstratives 21, 198 verb focalization 48
indefinite 32, 36 Verum Focus 44, 46
disjoint reference 125
DOM 243, 299, 306, 310 Galician 155
and dative marking 246 Gascon 2, 64, 66, 69, 76–78, 319, 334
and secondary predication 266 genitive 21–24, 329
and tense 251 and future tense 62, 65
animacy 250, 257 German 63, 84, 119–120, 122–124, 129–132,
human objects 247 155, 231, 235, 237, 262, 328
in causatives 310 Gipuzkoan 34, 38
marking 244 goal 24, 256, 261
objects and goal IO 256 grammaticalization 59–60, 63–65, 68, 72,
person 247, 252 73–76, 116–117, 179–180, 317, 318–319,
position of DOM objects 246 326–327, 333, 335, 337–338
specificity 249, 257 Greek 84, 94, 99, 107, 123, 124, 131–132
stigmatization 247, 310, 312 Guarani 246
vs. dative in bivalent unergatives 255
Dutch 99, 152, 155, 291 heaviness 14, 49
Hebrew 222, 235
English 60, 63, 76, 88, 94, 96, 99, 101, 102, Hindi 178, 245
107, 119, 124, 178–180, 185, 206, 226, 229, Hungarian 21, 278
262, 283, 289, 302, 310, 318, 323, 328,
330 imperatives 69
ergative 126, 131, 142, 190 impersonal 296
ergativity incorporation 146, 202, 204, 213
and telicity 152 and phonology 210
semantic split 139, 151 inference 317, 326–327, 335, 338
354 subject index
instrumental 333–334 modification with manner adverbs 88

interrogatives 42, 209, 235 resultant vs. target state 83
multiple wh-Q 40 resultative vs. stative 83
question words 319 stative and resultative 87
wh-extractions 235 stative participle vs. adjective 88
wh-words 43 target state 113
yes/no 50 temporal and spatial modification 83
intonation 49, 317 noun complement clauses 322
inversion predicates 232, 288, 301 noun modifiers 23
Italian 19, 140, 154–158, 281, 283, 292–
293 Occitan 65, 69, 154, 319, 330, 333, 338
Japanese 178, 190, 198, 262 palatalization 327

participles 61
Korean 178 adjectival. See adjectival part.
agreement 64, 86
Latin 14, 59, 60–61, 65, 68, 72, 330, 333, aspectual interpretation 111
338 attributive and predicative 117
leísmo 243, 252, 260, 312 attributive vs. predicative 99
Lexical Conceptual Structure 179 ending attached to V, not Root 110
locational nouns 17 experiential 104, 112
low vowel raising 327 imperfective 2, 61, 63, 66
inflected experiential perfective 73
mediopassives 283, 289, 295–297 non-verbal. See non-verbal participle
mood 59, 64 perfective 2, 26, 61, 64, 85
Basque indicative forms 63 possessive resultative 93, 116
exhortative and imperative 69 prospective 62, 65, 74
modality resultative 5, 60, 64–65, 73, 85, 111
need, must, want 73 copula with result. participle 84
potential 70, 279 suffixes 86
subjunctive 67, 279, 280, 284 temporally (un)bounded 106–107,
and (im)perfective 68 113
diachrony in Basque 68 resultative vs. continuous interpretation
use in Bsq and Rom 68 112
morphological case 245 particles 76
affirmative 78
Navarro-Aragonese 319 epistemic modality 77
negation 43 hearsay information 77
new speakers 53, 314 in Gascon 77
nominalization 323 interrogative 77
nonverbal participles 84 modal 76
adjectival. See adjectival participle partitive 142, 147, 200, 207–208
adjectival and adpositional 83 passives 92, 283–284, 294
adpositional 87 adjectival and verbal 119
adverbial 86 in causatives 282
and verbal projections 83 lexical and phrasal adjectival pass 119
as complement of creation verbs 88 passive agent 99, 289, 291. See also
continuous interpretation 112 adjectival part.
expressed agent 83 passive small clause 94
subject index 355
pattern convergence 17, 19, 25–26, 35, 42, Romance diachrony 60

46, 53, 59, 64–65, 71, 77, 159, 162, 164, simple past 64
167, 170, 185, 212, 243, 255, 276, 287, 305, theta-roles 150, 179, 187, 292
310, 312–313, 324, 337, 340–341 topic 41
Persian 180 contrastive 41
pluperfect 64–65 topicalization 48
polarity 33, 37, 45, 48 truth-conditional meaning 222–223, 227,
Portuguese 76, 155, 333, 338 235
predicative agreement 34, 38
prefixation 17 Urdu 176, 178
prosody 327
Punjabi 246 verb classes
purpose clauses 125 appearance and existence 158
causative alternation 142
quantifiers 21, 22, 34, 38, 54, 198 change 169
Quechua 278 change of location 142, 155
change of state 92, 112, 115–116, 141–142,
reanalysis 28, 206, 334 152, 156, 158, 162, 168
relational nouns 16 existence and continuation of state
relational suffix 23–24, 62, 87, 321, 323 157–158
relatives 29, 40, 325, 326 internal causation 143, 146, 152, 160, 162
appositive 31 motion 168
extraposition 31–32 psych verbs 183, 222, 232, 284
non-restrictive 30 stative 159
participial 87 verbs
pre and postnominal 29 bivalent unergative 243–244, 255, 259,
pronouns 30 260, 300
reduced 28 ditransitive 51, 224, 234, 237, 244, 246,
Rumanian 78, 338 260–262, 270, 300, 303
infinitive 63
scrambling 48–49 inversion. See inversion predicates
secondary predication 95, 244, 261 light 169, 176. See also complex
Semitic 246 unergative
Siberian Koryak 204 and argument structure 177, 179
specificity 244, 246, 253 and heavy counterparts 180, 188–189
and definiteness 249 basic repertoire 185
State phrase 149, 166, 169 causative 278
stimulus 178, 192–194 different verb, same NVE 186
subordinator. See complementizer fossilized 178, 214
Syntactic Argument Structure 179 list of egin LVC s in Basque 182
list with verbs other than egin 182
TAM non-separable units 184
Spanish and Basque chart 66–67 non-verbal element 194
telic verbs 91, 156, 168 noun as DO 205, 207
entailments 153 quantification 198, 201–202, 210
tense 59 reanalysis 206
compound tenses 61 referentiality of nominal 197
future and epistemic modality 77 V N adjacency 209
future periphrasis 61, 76 V N cohesion continuum 213
356 subject index
vs. synthetic 178, 200, 203 unergative 104, 139, 203, 212, 290–291,
vs. transitive constructions 204 298
vs. Vague Action Verbs 179, 199 and telicity 143
weather 190, 197–198 and Voice phrase 141
without faire and hacer 184 built on light verbs 144
periphrases 75, 250. See also aspect case and auxiliary 141
radical of Bsq verbs 62–63, 67–68, 277, complex 145
279 causative 278
stative 140, 167, 285 semantic classes 145
stative predicative constructions 163 dialectal variation 150, 152, 160
statives and theta-role 167 initial arguments 143
synthetic 51, 71, 78 stative 162
vs. compound forms 71 verb classes 143
unaccusative 41, 103, 139, 142, 166, 290– weather verbs 146, 286
291, 295, 297–298, 308 Voice phrase 141, 149, 166
argument role 141
case and auxiliary 140 word order 14, 233, 307–309
Event and State phrases 168 neutral 40
telicity 168
verb classes 141

Aligning Romance Grammar

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Aligning Romance Grammar

Загружено:

Авторское право:

Доступные форматы

Basque and Romance

Grammars and Language Sketches

The titles published in this series are listed at brill.com/gswl

Ane Berro Urrizelki

Copyright 2019 by Koninklijke Brill NV, Leiden, The Netherlands.

This book is printed on acid-free paper and produced in a sustainable manner.

3 Tense, Aspect and Mood 59

4 Non-verbal Participles in Basque and Spanish 82

5 (In)transitive Verbs: Unergatives and Unaccusatives 139

6 Light Verb Constructions in Basque and Romance 176

7 On Non-selected Datives: Ethical Datives in Basque and Spanish 221

8 Differential Object Marking in Basque and Spanish Dialects 243

9 Complex Causative Verbs and Causes in Basque (and Romance) 276

10 Grammaticalization Processes in Causal Subordination 317

Name Index 347

The book Basque and Romance: Aligning Grammars is a collection of articles

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004395398_002

2 Basque and Romance Grammars in Contact

Basque is a genetically isolated ergative language spoken by around 715,000

3 Topics Covered in This Book

ent typological distance created by ergative marking and complex agreement

4 Overview of the Chapters

As mentioned in previous sections, Basque and the surrounding Romance lan-

rise to synonymous LVC s in these languages. Regarding the morphosyntactic

present perhaps not-so-striking parallelisms. The Basque causative morpheme

Lazard, Gilbert. 2001. Le marquage différential de l’object. In Martin Haspelmath, Ekke-

Basque and Spanish differ rather drastically in the organization of constituents

Disregarding dvandva cocompounds, very productive in Basque but where

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004395398_003

type of compounding is also extremely productive in Basque, while it is much

3 Itself from (e)gi(n) ‘do’ and -le ‘-er’.

(1) Agirre Jauna/ El señor Aguirre ‘Mr. Aguirre’

(2) Kantauri Itsasoa/Mar Cantábrico ‘Bay of Biscay’

(3) Ibaizabal ibaia/El río Ibaizabal ‘the river Ibaizabal’

(4) 2000 urtea/el año 2000 ‘the year 2000’

(5) Gernikako Arbola Plaza/ Plaza Gernikako Arbola ‘Gernikako Arbola

9 Kayne (2017) for an antisymmetric approach to this aspect of morphology.

(6) El País Egunkaria/Diario El País ‘the newspaper El País’

(7) purpura kolorea/el color púrpura ‘the color purple’

As usual, it is possible to find exceptions, more in Basque than in Spanish,11

2.2 Order in Major Phrases

2.1.1 Determiners and Quantifiers

2.2.1.2 PP Dependents of Nouns

(10) etxearen jabea / el propietario de la casa

(11) etxeko jauna / el señor de la casa

(12) mugarik gabeko mundua / un mundo sin fronteras

(13) bihotzeko mina / una pena en el corazón

(14) bihotz oneko pertsona / una persona de buen corazón

(15) a. entzulegoari eskaini / eskertu

b. * entzulegoariko eskaintza / eskerrak

The dative verbal complement in (a) cannot be found as the complement to

(16) a. omenaldia Xalbadorri / Xalbadorri omenaldia

Basque and its mirror image head-complement-adjunct in Spanish. Thus, in the

2.2.1.3 Attributive Adjectives

Notice that it is (some) mostly classifying ‘adjectives’, as opposed to quali-

2.2.1.4 Relative Clauses

(17) a. [Zerbait dakite-n] eskualdun zaharr-ak alde dituzte.

b. Eskualdun zahar [zerbait dakite-n-ak] alde dituzte.

(18) Iaincoa, ceinec iustificaturen baitu circoncisionea fedetic

(19) Zuk ez al duzu lehengusu bat [filarmonika jotzen du-en-a]?

The appositive relative (Oyharçabal 2003: 802–807, de Rijk 2006: 487–490)

(20) Eta ez dut uste baden emazterik ezetz ihardetsiko

Oyharçabal (2003: 804) states appositive relatives are necessarily non-restric-

extraposed, we would expect restrictive relative clauses in general to undergo