Вы находитесь на странице: 1из 25

16

Logophoricity
Felix K. Ameka

16.1 Introduction

An important and pervasive feature of language use in daily life is that


speakers represent what other people are saying, thinking or feeling in
a form of language. Goffman (1974) observed that reported speech is an
integral part of everyday human interaction. In language studies reported
speech has been seen as ‘speech within speech, a message within message’
(Jakobson 1971: 130) or as an ‘utterance within an utterance’ (Voloshinov
1973). Languages provide diverse mechanisms to their speakers for indi-
cating what their own utterance is and what the utterance of others is (see
Coulmas 1986; Güldemann and von Roncador 2002; Holt and Clift 2006;
Lucy 1993; Banfield [1982] 2014). These and many other studies provide
accounts of the different forms and characteristics of reported speech –
direct, semi-direct and indirect – and its functions as manifested in diverse
languages and in various contexts of daily life. One of the issues that has to
be resolved with respect to represented speech is how to distinguish
between the words of the source or author that is being reported and
those of the reporter or animator, to use the term introduced by
Goffman (see Goffman 1981; Levinson 1988).
Many African languages in the Macro-Sudanic Belt1 (Güldemann 2003,
2008a, 2008b) have a paradigm of person markers which indicate corefer-
ence with the real or imagined author or source of a secondary discourse.
The term logophoric2 pronoun (or logophoricity) was introduced by Hagège
(1974) in his description of Tupuri and Mundang to characterize such
a paradigm of pronouns. This phenomenon, however, was noted in earlier
grammatical descriptions of the languages in the area. For instance, in his
grammars of Ewe, Westermann observed (1930: 60–1; see also 1907: 57):
‘In an objective clause the main sentence of which contains a verb of
saying, thinking, meaning, desiring etc. should the subject of the main
sentence occur in any kind of pronominal form (subjective, objective or

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
514 FELIX K. AMEKA

possessive) of the second or third person, this is expressed by ye, yeĎ in


the singular, yewo, yewóƒé in the plural. These pronouns would therefore
appear to introduce indirect speech, although in actual fact they introduce
direct speech.’
As Clements (1975) pointed out, the antecedent of the logophoric pro-
noun need not be the matrix subject. Also the form Ď is the possessive
linker and is not an integral part of the pronoun, although Westermann’s
characterization is an indication that the logophoric pronoun (in Ewe) can
stand for the possessor in a possessive phrase. The plural is formed by
cliticizing the plural marker onto it.3
As an initial example, consider the following dialogue taken from an
Ewe written play based on the Ewe migration story:4

(1) a. Mı́á-ƒé Fia-ga´ ̃ Agɔkɔli . . . dó gbe ná mi.


1p l - p o s s chief-big n a m e send greetings d a t 2p l
‘Our supreme chief Agokoli . . . sends you his greetings.’
b. É=bé gaƒoƒo ene sɔŋ=é nyé é=si
3s g =q t hour four i n t =af o c cop 3s g =this
ye=le mia-ƒé ası́-nu dzɔ=ḿ
l o g =be.at:p r e s 2p l - p o s s hand-mouth wait=p r o g
ko ye=mé-se nya áɖéké o
only l o g = n e g - hear word indef neg
‘Hei says hei has been waiting for signs from you for four whole hours
now and hei has not heard anything.’
c. Éya ta é-bé mı́a kplé vi=wò=wó
3s g because 3s g -q t 1p l and child=2s g =p l
kplé ʋɔnudrɔ´ ̃lá=wó mı́a-zɔ afɔ ɖeká
and judge=p l 1p l -move leg one
á-vá ye ƒéme fı́fı́ laa
p o t -come log house now right
‘Therefore hei says that we and your children and the judges should
come together to hisi house right now.’ (Kwamuar 1997: 15)

This excerpt is a text of a chief’s spokesperson. The spokesperson is the


animator reporting what the Principal, who is absent from the scene of
conversation, said to the intended addressees. The spokesperson first
introduces the chief as the one who sent him by bringing greetings from
him. In the immediately following sentence (1b), he reports the chief’s
message by framing it with the quotative verb or marker bé, which creates
a context for represented speech. The subject of the quotative verb is the
chief who was mentioned in the previous clause. The coreference to the
chief as the subject of the quotative verb is signalled by the ordinary third
singular pronoun. However, in the complement clause representing the
content attributed to him, he (the chief) is referred to with the logophoric
pronoun. It has become customary to refer to the one whose thought or
speech, etc. is being reported as the logophoric trigger (i.e. the chief or the
pronoun that refers to him in the main clause) and to refer to the stretch of

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 515

discourse attributed to the trigger as the logophoric domain. In sentence


(1c) the spokesperson continues to report the chief’s message and uses the
quotative verb frame. Note that the subject of this verb, the chief, is here
not referred to with the logophoric but the ordinary third singular pro-
noun and he is the logophoric trigger. In the report (the logophoric
domain), however, he is referred to with the logophoric in a possessive
phrase. Some would say that the logophoric pronoun is the target of the
marking triggered by the source in the matrix clause. In terms of gramma-
tical functions, the logophoric pronouns in (1b) are in subject function in
the dependent clauses and their antecedent is in subject function in the
matrix clause. In (1c), however, the NP in which the logophoric pronoun
occurs as possessor is the grammatical object of the verb vá ‘come’.
The logophoric pronoun in possessor role in this case refers back to the
subject argument in the matrix clause (the logophoric trigger).
The instances of the logophoric pronoun in example (1) illustrate some
of its other canonical features. First, that it is used to refer to the person/
entity ‘whose speech, thoughts feelings or general state of consciousness
are reported’ (Clements 1975: 141), i.e. the logophoric trigger. Secondly, as
(1b) partly shows, once the source is identified, the domain of the logo-
phoric can go over a stretch of discourse. Third, the logophoric pronoun is
a reference-tracking mechanism where it marks coreference with the
source, and disjoint reference is marked by the ordinary pronoun.
The logophoric pronouns, since the early studies, have been noted to share
this coreference marking function with other devices. In particular, both
Hagѐge (1974) and Clements (1975) draw attention to the so-called ‘indirect
reflexives’ known from classical languages like Greek and Latin and also to
the long-distance reflexives in other languages like Icelandic and Asian
languages like Chinese and Japanese. Clements compared the logophoric
pronoun in Ewe to these indirect reflexives thus: ‘The indirect reflexive,
like the logophoric pronoun of Ewe permits a speaker to avoid ambiguity
of reference’ (Clements 1975: 142). He called this avoidance of ambiguity of
reference a logophoric use. Clements cites various examples to show that the
reflexive form is used instead of the ordinary pronoun in Latin to avoid
ambiguity. Examples in (2) illustrate:

(2) a. A Caesare (. . .) invitor (. . .) sibi ut sim legatus


‘I am invited by Ceasari to be lieutenant to himi (sibi)’
(Cicero Epistulae ad Atticum ii 18, 3) (Clements 1975: 143)
b. Vos ex M. Favonio audistis Clodium sibi dixisse (. . .) periturum
Milonem
‘You heard from M. Favoniusi that Clodius has said to himi (sibi) that
Milo was to die.’ (Cicero Oratio pro Milone 44) (Clements 1975: 144)

Because of this use of the indirect and long-distance reflexives to disam-


biguate reference just as the logophoric markers, many theoretical
models of grammar as well as Discourse Representation Theory have

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
516 FELIX K. AMEKA

tried to account for these forms in a unified way (see, e.g., Sells 1987;
Stirling 1993; Huang 2000; Speas 2004). But Culy (1997: 846) has drawn
attention to the fact that logophoric pronouns and indirect reflexives
have different primary uses: the former for indirect discourse and the
latter for clausal coreference. Culy indicates that they both have
a common secondary use, namely signalling point of view. He acknowl-
edges that uses of logophoric markers pertaining to point of view involve
rather marginal constructions. In fact, the authenticity of the examples
from Anlo Ewe used by Clements to argue for this use of the Ewe logo-
phoric pronoun has been called into doubt by other native speakers
(cf. Essegbey 1994). Moreover, some other languages that have canonical
logophoric markers such as Eleme and other Ogonoid languages are said
not to have any point of view usages of the logophoric forms (Bond 2006).
The logophoric markers are involved in reference tracking like other
reference-tracking devices such as indirect reflexives, switch-reference
systems, etc. However, the logophoric marking systems, as Comrie (2004)
points out, have a constellation of properties that makes them
distinctive.
In this chapter, I use the terms logophoric and logophoricity in the
original sense of linguistic markers (pronouns, verbal markers) dedi-
cated to referring to the source of information that is being reported.
The aim is to describe the various properties that make these forms
distinctive and cast them in a cross-linguistic typological perspective.
The chapter is structured as follows: section 16.2 proposes a typology of
logophoric marking systems, while section 16.3 details the variation
that is found across languages in logophoric marking with respect to
person, number and grammatical function features. Section 16.4 dis-
cusses the ways in which languages create logophoric domains using
verbal and non-verbal linguistic signs. In section 16.5, the motivations
for logophoric systems and their functional load are explored. It is
argued that the key to understanding these systems is to pay attention
to cultural practices of communication in everyday discourse in lingua-
cultures in which the logophoric forms occur. Some concluding remarks
are offered in section 16.6.

16.2 A Typology of Logophoric Marking Systems

There are different dimensions of the typology of languages with respect


to the marking of coreference in the logophoric domain. Culy (1994a,
1997; see also Huang 2000) makes a distinction between ‘pure’ logophoric
languages and ‘mixed’ logophoric languages. The former have distinct
morphosyntactic forms for indicating coreference in the logophoric
domain. The ‘mixed’ logophoric languages, according to this typology,
are those languages that use reflexives in logophoric domains, or those

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 517

languages where the pronouns are not solely for marking logophoricity.
Various languages such as Igbo, Yoruba and Engemi are ‘mixed’ logophoric
languages. They have logophoric marking pronouns, but the pronouns
have functions beyond the logophoric such as being object pronouns in
non-logophoric domains (see discussions in von Roncador 1992 and Culy
1994a). Languages mark logophoricity using one or more of the following
means: (a) reported speaker pronouns or logophoric pronouns; (b) addres-
see logophoric pronouns; (c) verbal logophoricity markers-logophoric sub-
ject cross-reference markers, first person markers on verbs and logophoric
verbal affixes; (d) anti-logophoric pronouns.

16.2.1 Reported Speaker Logophoric Pronouns


The first group are those languages that have logophoric pronouns to
refer to the Source whose speech or thoughts are being reported. Some
of these languages are Aghem (Grassfields Bantu, Cameroon); the
Chadic languages of Nigeria – Angas and Goemai; the Dogon languages
of Mali and Burkina Faso – Donno Sɔ and Jamsay; the Songhay language
Koyra Chiini of Timbuktu, Mali; the Gbe languages – Ewe, Gen, Aja, Fon
of the Gulf of Guinea; the Adamawa languages – Mundang and Tupuri;
Wan, a Mande language of Cote d’Ivoire; and Maale, an Omotic lan-
guage of Ethiopia. See example (1) above for illustration of the Ewe
logophoric pronoun.

16.2.2 Addressee Logophoric Pronouns


The second group are those which have pronouns used to refer to the
addressee of the matrix clause in the logophoric context. Languages with
this kind of pronoun are the Chadic languages, such as Angas, Mupun,
Goemai, Pero, and Tikar (see Frajzyngier 1985). Many of the languages with
logophoric addressee pronouns also have logophoric speaker pronouns as
well. Tikar is the one language so far that has addressee logophoric pro-
nouns but no reported speaker logophoric pronouns. Consider the follow-
ing examples from Goemai (Hellwig 2011, p.c.; Ameka 2004) showing the
feminine, masculine and plural forms of the addressee logophoric pro-
nouns respectively. The addressee logophorics are glossed as LogB and the
speaker logophorics as LogA.

(3) S’a muk b’am d’i sek masha hok. (. . .)


hand 3Sg.Poss stick l o c . a n a p h b o d y friend(fem) def
Yin pa goe nyet s’a ji.
saying Sgf.LogB o b l leave hand Sgm.LogA. Poss
‘His1 hand stuck there to the girl-friend2. (. . .) (He1 said) saying, she2 should
leave his hand1.’ (FUAN)

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
518 FELIX K. AMEKA

(4) Ba doe de goesha muk (. . .)


return(sg) come(sg) d i r e c t i o n friend(masc) 3Sg.Poss
Yin gwa dok kat ndoe bi d’i
saying Sgm.LogB p a s t . r e m findi n d e f thing
l o c . a n a p h ndûûn bang
i n s i d e gourd
hok a?
def inter
Mûai yin hai ji kat ndoe bi ba
friend saying i n t e r j Sgm.LogA find i n d e f n e g thing
‘He1 returned (and) came to his friend2. (. . .) (He1 said) saying, did
he2 find anything inside the bag? The friend2 (said) saying, he2 did
not find anything.’ (MOESHA)

(5) Sai k’ur ba yong mûep d’oot,


then tortoise return(sg) call 3Pl softly
yin nwa goe doe ntyem
saying Pl.LogB obl come front
‘Then the tortoise1 returned (and) called them2 softly, saying, they2
should come first.’ (KUR)

Hellwig (2011: 228) points out that the logophoric pronouns, both addres-
see and speaker, are obligatory in reported speech contexts, but optional
in possessive contexts.

16.2.3 Verbal Logophoricity Markers


In a number of languages, logophoricity is signalled by a marking on
the verb. In some languages such verbal marking is used together
with a logophoric pronominal marking system. There are different
forms of signalling logophoricity on the verb. Curnow (2002) identi-
fies three types of the marking of logophoricity on verbs in African
languages. These are logophoric subject cross-reference marking, first
person markers on verbs and logophoric verbal affixes. Each of these
is described in turn.

16.2.3.1 Subject Cross-Referencing on the Verb in the Logophoric


Domain
There are languages which employ cross-reference marking on the verb.
For instance, in Akɔɔse a head-marking Bantu language of Cameroon,
a distinct subject cross-reference marker is used in the logophoric domain
to indicate coreference to the logophoric trigger in the matrix clause.
In example (6) the ordinary subject cross-reference marker is used in the
dependent clause to refer to the subject in the matrix clause. In (7),

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 519

however, a dedicated logophoric cross-reference marker is used to refer to


the reported speaker in the matrix clause.

(6) à-hɔ´bé ǎ á-kàg


he-said rp he-should.go
‘He said that he (someone else) should go.’
(7) à-hɔ́bé ǎ mǝ́-kàg
he-said rp l o g -should.go
‘He said that he (himself) should go.’ (Hedinger 1984: 95)
The logophoric cross-reference marker in Akɔɔse is used for second and
third persons as opposed to first persons, and it is a distinct subject
cross-reference marker different from all other person cross-reference
markers in the language. Curnow notes that such a system can be used
on its own as in Akɔɔse or it can be used in tandem with logophoric
pronouns, as is the case in the Nilo-Saharan languages Logo, Kaliko and
Moru.

16.2.3.2 First Person Logophoric Verbal Affix


Some languages have a first person verbal affix that cross-references the
logophoric trigger in the logophoric domain. In Donno Sɔ, a Dogon lan-
guage of Mali and Burkina Faso, there are agreement affixes on the verb
that are controlled by the subject and/or object both in matrix and in
dependent clauses. In logophoric domains a first person verbal affix is
used to refer to the Source, the logophoric trigger, in the main clause, as
in example (8).

(8) Oumar [inyemɛ jembɔ paza bolum] min tagi


Oumar l o g sack:d f drop left:1 s g 1s g : o b j informed
‘Oumarj told me that hej had left without the sack.’ (Culy
1994b: 123)
In this example, there is an overt subject in the logophoric domain, which
is represented by the logophoric pronoun inyemɛ ‘l o g ’ that refers back to
Oumar, the reported speaker in the matrix clause. Crucially there is a first
singular verbal affix on the verb ‘left’. This marker refers back to the
subject of the logophoric domain. This form is used only in logophoric
domains and only for third person subjects. However, Donno Sɔ allows
subject ellipsis. In such cases as in example (9), the only indicator of
logophoricity is the first person affix on the verb, referring to the logo-
phoric trigger. Note that this trigger is the third person and it is elided in
the dependent clause.

(9) Oumar [minɛ inyemɛ mɔ gɛndɛzɛm] gi


Oumar field log p o s s regard:p r o g :1s g said
‘Oumari said that hei will look at hisi field.’ (Culy 1994b:123)

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
520 FELIX K. AMEKA

Interestingly, this first person logophoric marker is not used to mark


reference to a first person subject in the subordinate clause. Such a verb
is zero-marked. This is the way in which a third person non-logophoric
subject is also marked. Consider example (10). Note the bare verb in the
subordinate clause.

(10) Oumar [ma jembɔ paza bali] min tagi


Oumar 1s g : s u b j sack:d f drop left 1s g : o b j informed
‘Oumar told me that I had left without the sack.’ (Culy 1994b:123)
As logophoric discourse is a report of a first person discourse, it seems
reasonable that a first person marker can be used as the expression of the
logophoric trigger in a logophoric domain. This form of marking is con-
sistent with a view that ‘logophorics may be thought of as embedded 1sg
and 1pl pronouns within quotations attributed to a third party speaker’
(Heath 2008: 648). Besides it is not an accident that logophoric pronouns in
some of the languages may have grammaticalized from first person pro-
nouns and other deictic markers or shifters (von Roncador 1992;
Dimmendaal 2001). For instance, the Ewe logophoric marker ye ‘l o g ’
may have evolved from the independent first person pronoun nye ‘1s g .
i n d e p ’ (cf. Heine and Reh 1984).
As we have seen, the first person marking, like the cross-referencing,
can occur in a language which has logophoric pronouns. But like the
subject cross-referencing system, it is not every language that employs
the first person marking system that has logophoric pronouns. The Nilo-
Saharan languages Karimjong and Lotuko do not deploy logophoric pro-
nouns, but they use first person verbal affixes (see Curnow 2002 for further
discussion). In general, first person marking occurs with both singular and
plural antecedents.

16.2.3.3 Logophoric Verbal Affixes


A group of languages, especially the Ogonoid languages of Nigeria such as
Gokana (Hyman and Comrie 1981), Kana (Ikoro 1996), Eleme (Bond 2006)
and other Cross-River languages like Efik and Ibibio (see Curnow 2002) use
verbal affixes to signal logophoricity. Compare the following pairs of
examples in (11), (12) and (13) taken from Eleme, Gokana and Kana respec-
tively in which the (a) sentences have a logophoric verbal affix while the (b)
sentences are not marked for logophoricity (Bond 2006: 236).
The distinction between the (a) sentences and the (b) sentences is that
there is coreference between the subject participant in the dependent
clause and that of the matrix clause.

(11) Eleme (Bond 2006: 236)


a. ɛ̀ -kɔ ɛ̀ -dɔ-ɛ̀
3-say 3-fall-l o g
‘Hei said that hei fell.’

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 521

b. ɛ̀ -kɔ ɛ̀ -dɔ
3-say 3-fall
‘Hei said that hek fell.’
(12) Gokana (Hyman and Comrie 1981: 20)
a. aé kɔ aé dɔ-ɛ̀
He said he fell-l o g
‘Hei said that hei fell.’
b. aé kɔ aé dɔ
he said he fell
‘Hei said that hek fell.’
(13) Kana (Ikoro 1996: 283)
a. à-kɔ̄ɔ̀ é-kı̄i-ѐ
he- c o n n he.d f -go-l o g
‘Hei said that hei would leave.’
b. à-kɔ̄ɔ̀ é-kı̄i
he-c o n n he.d f -go
‘Hei said that hek would leave.’
The categorial status of these markers in some of the languages is not
entirely clear. Hyman and Comrie (1981) analyse the Gokana marker as an
affix, but Ikoro (1996) claims that it is a clitic. Similarly, the Kana form here
is called an affix, but Ikoro (1996) argues that it is a clitic, as it can also dock
on predicate nominal complements as in example (14).

(14) Kana embedded nominal p r e d i c a t e l o g marked (Ikoro


1996: 285)
kɔ̄ɔ̀ ɛ̀ ɛ̀ -ɛ̀ nà kà-ѐ
c o n n he:c o p : p r e d -l o g my mother-l o g
‘Shei claims that shei is my mother.’
Irrespective of the structural type, these verbal markers share several prop-
erties with other logophoric marking systems. They are special markers
used in a logophoric domain typically licensed by logophoric predicates and
quotative markers. The languages differ with respect to the clarity of which
participant they are referring to in the matrix clause. Gokana seems vague
on this, but both Kana and Eleme provide means of clarifying who the
antecedent of the verbal logophoric marker is. In Eleme, the verbal marker
refers to a third person, while in Kana and Gokana it can refer to a second or
third person. In fact in Gokana it can even refer to a first person, although
this is less preferred (Hyman and Comrie 1981).
With respect to number, Eleme provides good parallels to other logo-
horic languages. Thus, it makes a distinction between a singular verbal
logophoric suffix -e ‘l o g ’ as in (11a) and a plural one -ba ‘l o g . p l ’ , as
in (15).

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
522 FELIX K. AMEKA

(15) ɛ̀ -kɔ-ri m-ɛ̀ -dɔ-ba


3-say-p l c o m p -3-fall-l o g . p l
‘Theyi said that theyi fell.’
Furthermore, just as a singular reported speaker can be referred to with
a plural logophoric pronoun interpreted as an associative plural (see
§16.3.2 below), the plural verbal logophoric suffix can have a singular
antecedent, as illustrated in example (16).

(16) àŋɛ̀ lama-mi kɔ̀ ɛ̀ -ba-dɔ̀ -ba


3s g tell-1s g c o m p 3-3p l . a s s . p -fall-l o g
‘Hei told me that theyi+k fell.’
All the languages with verbal logophoricity marking are, not surprisingly,
languages that are head marking at the clause level.

16.2.3.4 Anti-Logophoric Markers


In all the instances of logophoricity marking discussed so far, there is
a dedicated form for referring to a source whose speech and thoughts,
etc. are being reported. In other languages the reported speaker or source
is referred to with an ordinary pronoun, while there is a special pronoun
for use in logophoric domains which refer neither to the reported speaker
nor the current speaker nor the addressee. As this is the reverse of the
logophoric system, such pronouns have been called anti-logophoric. Anti-
logophoric pronouns signal the one whose thoughts and speech are not
being reported. Such forms are found in the Kwa languages Attie and
Adioukrou from Côte d’Ivoire and in the Nilo- Saharan language Mabaan
(Andersen 1999). As H. Hill (1995: 95) notes with respect to the Adioukrou
forms:
To clarify the reference of third-person pronouns, many languages use
a logophoric pronoun in the Q[uote] C[ontent] that has a special form and
marks coreference with the speaker of the Q[uote] M[argin]. Adioukrou,
however, solves this problem of third person reference in a different way.
It has a special form to mark DISJOINT reference with the speaker of
the QM.
Thus, the anti-logophoric markers serve similar function in reference
tracking. Consider this example from Adioukrou, adapted by Culy (1997)
from H. Hill (1995):

(17) a. lii dad eke ini/k im Dabu (Reporting s u b j pronoun)


3sf p said that 3sr p went n a m e
‘Hei said that hei/k went to Dabou’
b. lii dad eke ow’nk im Dabu (Antilogophoric)
3sf p said that d sr p went n a m e
‘Hei said that heki went to Dabou’

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 523

As indicated along the way, the systems of logophoric marking including


anti-logophoric marking can co-exist in a language. The variation in the
forms of logophoric marking in terms of features will be explored in the
next section.

16.3 Diversity of Logophoric Marking

In the preceding section, I outlined the different logophoric systems found


across the languages. In this section we turn now to show the variations
that occur with respect to person and number features of logophoric
markers in the various languages.

16.3.1 Person Features of Logophoric Pronouns


Languages differ as to the person features of the logophoric pro-
nouns. In all logophoric languages, the logophoric marker can be
used to refer to third persons. In many languages it is only the
third person that can be an antecedent. In Donno Sɔ, a Dogon lan-
guage of Burkina Faso and Mali, the logophoric pronoun is only used
for third persons. Similarly the logophoric pronoun in Sango is also
only used for third person antecedents. In yet other languages there
is syncretism between second and third person logophoric pronouns.
The Kwa languages Ewe and Ikposso as well as Wan, a Mande lan-
guage, are examples. Yagi Dii, an Adamawa language, is the only
language reported in the literature without controversy where the
logophoric pronoun can be used for the three persons (see Bohnhoff
1986 as reported in von Roncador 1992, and Huang 2000).
The variation among languages with regard to person features has
led to a proposal of an implicational person hierarchy for logophoric
pronouns, first articulated by Hyman and Comrie (1981):

(18) Implicational person hierarchy for logophoric pronouns


3>2>1

The implication says that if a language has first person logophoric


pronoun use, it would also have second or third person uses. In other
words, third persons are more easily accessible than second persons to
logophoric marking and a second person is more accessible than first
person.
As von Roncador (1992) points out, this implicational hierarchy con-
flates a number of possibilities of marking: (i) It is possible for a third
person logophoric as opposed to first and second person to be marked. For
instance, Donno Sɔ (Culy 1994b). (ii) There is syncretic third and second
person marking by the same form as opposed to a non-marked first person.
For example Ewe, Fon, Ikposso and Efik. (iii) The same form is used for all

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
524 FELIX K. AMEKA

three persons. This is suggested to be the case for the Gokana verbal affix
(Hyman and Comrie 1981).5
Given the nature of the person feature of the logophoric markers, one
can raise the question of where they fit in the distinction between Speech
Act Participants, i.e. first and second persons, vs third person (other or
substitutive). They seem to always substitute for first or second persons in
reported contexts. It goes to reinforce their separate nature and explains
the marking patterns that are found across languages.

16.3.2 Number Features of Logophoric Pronouns


There is variation across the languages with logophoric markers in terms
of the number features of the markers. In Ewe for instance, a plural
logophoric pronoun can be created by cliticizing the regular plural marker
to the logophoric form: ye vs. ye=wó. In a sense the plural clitic marker
functions in this instance like a suffix as nothing can come between the
logophoric and the plural marker. Other languages may have distinct
forms for singular vs. plural. Ikposso (a Kwa language of Togo and
Ghana) has distinct forms for the singular used for second and third person
singular, namely ntà ~ ntè ; and a set for second and third person plural,
namely, nātà ~ nātè. The allomorphy is partly controlled by aspectual
distinctions where one set occurs in perfective contexts and the other in
imperfective contexts (Soubrier 2013: 364).
In yet other languages there is only a singular form. Avatime, another
Kwa language of Ghana, is a case in point. In Avatime there is only
a singular logophoric pronoun used for the third person singular.
The form is yi, and its allomorphs vary according to the vowel of the
noun class of the referent (Rebecca Defina and Saskia van Putten p.c.)
This peculiarity might be due to the fact that the logophoric pronoun is
borrowed from the Inland Ewe dialects. Mundang, an Adamawa language
as well as Igbo, a Benue-Congo language of Nigeria, also have only singular
logophoric pronouns.
The generalization seems to be that if a language has logophoric pro-
nouns it will have a singular pronoun, but not all logophoric pronoun
languages have plural pronouns. This generalization has also been cap-
tured in an implicational hierarchy for number features of logophoric
pronouns (Huang 2000: 179):

(19) Number hierarchy for logophoric pronouns


Singular > plural
If a language has a plural logophoric pronoun it will also have
a singular one.

The interesting thing about the use of plural logophorics is that they can
have a singular antecedent in the linguistic context. If the reported

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 525

situation, however, implies that the logophoric trigger is included in


a group, the marking in the logophoric domain can be in the plural. For
instance, take a situation where a parent visited another family with the
children and announces their departure. Her announcement can be
reported in Ewe as in (20a) or (20b). Compare the forms in (20):

(20) a. Dadá bé ye=wó=dzó


mother qt l o g =p l =leave
‘Motheri says that theyi+j are leaving.’
b. Dadá bé ye=dzó
mother qt l o g =leave
‘Motheri says that shej is leaving.’
c. Dadá bé wó=dzó
mother qt 3p l =leave
‘Motheri says that theyj are leaving.’
If (20a) is used, the plural logophoric pronoun is interpreted as the mother
and others with her, the children. That is, it is understood as an associative
plural. For (20b) where the singuar logophoric form is used, the mother,
the reported speaker is not associated with the children. (20c) is the form
used where the reported speaker, mother, is not one of those who left.
The use of a plural logophoric pronoun to refer to a singular reported
speaker is a way of creating associative plural logophoric pronouns. But
such a context raises some questions as to whether this is restricted to
logophoric contexts or just a more general use of plural forms for associa-
tive plural readings. For instance, the plural in Ewe has associative plural
readings in certain contexts, e.g. Kof ı́=wó ‘Kofis’ has two readings: a regular
plural, i.e. two or more people called Kofi, and Kofi together with other
people. It is important for further research in logophoric languages to see
the extent to which the associative plural use is more general or just
restricted to logophoric domains.

16.3.3 Grammatical Functions of Logophoric Triggers and


Logophoric Pronouns
Languages also vary with respect to the grammatical roles that logophoric
triggers can serve in a matrix clause and those that a logophoric pronoun
can serve in the report content. In pure logophoric languages such as Ewe,
the logophoric pronoun in the logophoric domain can function in intran-
sitive subject (20a,b) as well as transitive subject roles, as in the two
instances in (1b) above. However, it can also function as a transitive object
(21a) or dative object (21b) or an oblique object (21c).

(21) a. nyɔ́nu=ɔ bé ɖevı́=á dzu ye


woman=d e f q t child=d e f insult l o g
‘The woman said the child insulted her.’

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
526 FELIX K. AMEKA

b. vi=nye tsi ná=m bé amedzró=á ná ga ye


child=1s g tell d a t =1s g q t visitor=d e f give money log
‘My child told me that the visitor gave her some money.’
c. é=bé zã dó ɖé ye=wó
3s g =q t night put.on a l l l o g =p l
‘He said that night fell on them.’
In a ‘mixed’ logophoric language like Igbo which has only singular logo-
phoric pronoun usage, it seems to occur only in subject function in
a logophoric domain.
Similarly, languages vary with respect to the grammatical function of
the logophoric trigger or reported source in the matrix clause or quote
margin. Frequently the logophoric trigger fulfils a subject function in the
matrix clause, as many of the examples we have seen so far illustrate (see
especially examples in (21)). However, when the predication in the matrix
clause relates to psychological states, the reported experiencer can have
other grammatical functions apart from the subject (see Huang 2000: 182,
and references there). To continue with Ewe examples, in this case also the
logophoric trigger can fulfil transitive object (22a) as well as dative object
(22b) functions.

(22) a. dzi dzɔ Áma bé ye=tu aƒé=á vɔ


Heart straighten n a m e q t l o g =build house=d e f p f v
‘Ama was happy that she finished building the house.’
(lit. Heart straightened Amai that shei finished building the
house.)
b. é=vé dɔme ná Áma srɔ̃ bé
3s g : i m p e r s =pain stomach dat n a m e spouse q t
ta.tɔ́=wó mé-kpé ye=wó o
leader=p l n e g -meet l o g =p l n e g
‘It angered Ama’s spouse that the leaders did not invite
them (l o g ).’
A matrix clause with a predicate like ‘hear’ can have two sources: one of
the reported perceiver and the other of the source of the perception of the
reported perceiver, typically coded as an ablative argument in the matrix
clause. When a logophoric marker is used in the logophoric domain of
such a proposition, the interpretation can be ambiguous, as illustrated in
(23) from Ewe:

(23) Áma se tsó Akú gbɔ´ bé Akós dzu ye


n a m e hear from n a m e place q t n a m e insult log
‘Amai heard from Akuj that Akos insulted her(l o g )i/j.’
The ambiguity can be resolved in context, but it shows that the logophoric
pronoun in the logophoric domain can refer to either the matrix subject or
the dependent argument in a postpositional phrase functioning as object

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 527

in the matrix clause. The variation among languages as to the accessi-


bility, as it were, of logophoric pronouns and triggers to various gram-
matical functions has led to proposals of two hierarchies, one for the
grammatical functions and the other for the associated semantic roles
(see note 6 for the hierarchies).6 In my view these hierarchies do not
explain much, as the critical thing about logophoric marking is not
about syntax or even the syntax–semantics interface but about dis-
course. In the next section I explore the various ways in which logopho-
ric domains are created cross-linguistically. I argue that even though
a lot has been said about predicate types and so-called logocentric pre-
dicates, they are less important. What is crucial is a ‘report opening’
marker (Stirling 1993: 260). I will argue that apart from complementizers
or quotatives, attention should be paid to non-verbal signals like intona-
tion and gestures.

16.4 Creating Logophoric Domains

A speech event involving the report of the speech, thoughts, wishes etc. of
others can be usefully characterized in structural terms as consisting of:

• a QUOTE MARGIN made up of a report verb and its participants, at least


a reported source and in some cases an addressee, and typically a report
opening marker or quotative, and
• a QUOTE CONTENT which is the report proposition.

(24) Structure of a speech report event


[Report Report [Reported
predicate <source, addressee> opener]Quote Margin Proposition(s)]Quote Content
This characterization applies to the three main speech report styles that
have been identified in the literature: direct, semi-direct and indirect (e.g.
Aikhenvald 2008). In the context of logophoric languages of Africa, there is
usually a prosodic break between the Quote Margin and the Quote Content
(Wiesemann 1990; H. Hill 1995) and the logophoric marker typically
occurs in the Quote Content proposition. Using such a framework, one
can identify certain elements that may be involved in the creation of
logophoric domains, i.e. the Quote Content.
In discussions of logophoricity, a lot of attention has been paid to what
have been called logophoric verbs or logocentric verbs or predicates (see,
e.g., Stirling 1993; Culy 1994a; Huang 2000). It has been observed that
there seems to be a hierarchy of which verbs are more accessible than
others to ‘license’ the use of logophoric markers:

(25) Implicational hierarchy for ‘logocentric’ predicates


Speech predicates > epistemic predicates > psychological predi-
cates > knowledge predicates > perceptive predicates

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
528 FELIX K. AMEKA

The idea is that if a language allows predicates at a point on the scale to


be part of a logophoric domain-creating margin, then it would also use
verbs higher on the scale for the same purpose. Thus, if a perception
predicate can license the use of a logophoric marker in the report content
in a language, then one expects that predicates of knowing and thinking,
wanting, believing and communication would do the same. A language
like Sango restricts its logophoric pronoun to propositions dependent on
the verb of saying in the language.
Some descriptions of logophoricity in individual languages go as far as
providing a list of the so-called logophoric predicates. Lefebvre and
Brousseau (2002: 79), for example, cite the following as logophoric verbs
in Fon (Gbe, Kwa):

(26) Logophoric verbs in Fon (Lefebvre and Brousseau 2002: 79)


ɖɔ ‘to say’ kɛ´n ‘to bet’
kàn byɔ́ ‘to ask’ gbɛ´ ‘to refuse’
yı́ gbѐ ‘to answer’ tù n ‘to know’
flin ‘to remember’ ɖì ‘to believe’
lin ‘to think’ ɖó nù kún ‘to hope’
wɔ̀ n ‘to forget’ jló ‘to want’
dó àkpá ‘to promise’ gblɔ̀ n adàn ‘to threaten’
sѐ ‘to hear’ mɔ̀ ‘to see’
zɔ̀ n ‘to command’ kú dlɔ̀ ‘to dream’
mɔ´ ‘to deny’ xwlé ‘to swear’
One can assign these verbs to the various semantic classes suggested in the
implicational hierarchy.
Even though a lot has been made of the predicates and their classes,
reflecting a syntactic bias in the approach to the study of logophoricity, it
has been noted that there are some occurrences of logophoric markers
that do not require these verbs. Rather, the main element that is required
for creating a logophoric domain within which a logophoric marker can be
used is a report opener. Thus for Fon, Lefebvre and Brousseau (2002: 78)
assert that the Fon logophoric pronoun émì ‘is permitted only in a clause
selected by ɖɔ̀ which can be used either as a verb meaning “to say”, or as
a complementiser introducing tensed complements of verbs of the SAY-
class’. So in a sense it is not about the verbs but rather about the report
opener ɖɔ̀ , which can be thought of as a quotative that creates a logophoric
domain.
A similar statement can be made with respect to the Ewe quotative
marker bé that we have seen in several examples. The generalization for
Ewe is that all multi-clausal constructions in which the quotative com-
plementizer is involved is a logophoric domain, and hence if there is
coreference between a participant in the dependent clauses and the
protagonist in the main clause, the logophoric pronoun is used. Thus,

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 529

the quotative bé is used to create a logophoric domain in purpose con-


structions as illustrated in (27).

(27) Esi yi asi.me bé ye-a-ƒle avɔ


n a m e go market q t l o g -s u b j v -buy cloth
‘Esii went to the market so that shei would buy cloth.’
In Kana, a language that uses logophoric verbal affixes, such use of the
logophoric occurs in a purpose clause where the matrix verb is not
a reporting verb. Crucially, in Kana also the quotative complementizer,
the report opener, glossed as ‘c o n n ’ in (28) below, is the marker that
creates the logophoric domain. As in Fon and many other languages this
quotative marker is evolved from the ‘say’ verb in Kana. See example (28)
from Ikoro (1996: 286).

(28) bà-lú kɔ̄ɔ̀ ba-é-tō-ѐ


they-come:f a c t c o n n they-d f -cry-l o g
‘They came to cry.’
Typically, examples such as (27) and (28) are explained as constructions in
which logophoric markers are used in non-reportive contexts. However, if
one understands purpose constructions as reporting the wants and wishes
of a source (i.e. that the utterance is reporting that the participant in
subject role in the matrix has something in mind that he or she wants to
do), then the use of the report opener and the logophoric marker fall in
place.
Another construction in which a report opener of the quotative com-
plementizer type occurs is the proximative aspectual construction. This
construction has not been noted in the literature as an area of the extended
use of logophoric markers. However, in Ewe the logophoric pronoun is
used to mark cross-reference between the argument in the dependent
clause and the subject participant in the matrix clause in such
a construction. Consider the following examples:

(29) a. Tsi bé ye-a-dza


Water q t l o g -s u b j v -drip
‘It is about to rain.’
b. ɖevı́=á bé ye-a-trú
child=d e f q t l o g -s u b j v -vomit
‘The child is about to vomit.’
The proximative is a grammaticalized construction related to the wants,
real or imagined, of the participant functioning as subject in the matrix
clause. As argued in Ameka (2008), the verb dı́ ‘want’ can be overtly
expressed in the matrix clause of a proximative construction. It can thus
be viewed as a report about the want or intention of the subject participant
in similar fashion to purpose clauses.

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
530 FELIX K. AMEKA

Other languages have dedicated report openers which are complementi-


zers. Engenni, an Edoid language of Nigeria, is such a language. It has two
complementizers, ga and na. The former is used to create logophoric
domains where the speaker wants to attribute discourse to someone else,
and the latter is used in situations based on the speaker’s direct perception
(see Thomas 1978). Thus, there is a dedicated complementizer that intro-
duces contexts where logophoric markers have to be used should the Source
whose speech, thought, etc. are being reported be referred to in that domain.
There seem to be languages in which the report opener that creates
logophoric domains is not a verbal element but rather a non-verbal linguis-
tic sign. Nikitina (2012) reports that in Wan, a Mande language of Côte
d’Ivoire, a logophoric domain is created neither by the so-called logocentric
predicates nor by a quotative complementizer. In discussing a specific exam-
ple from a narrative text, she observes that ‘the speech [of a character] is not
introduced by any overt verb of speaking; instead the narrator signals the
beginning of a character’s speech by intonation and gesturing’ (Nikitina
2012: 286). I conclude that the report opener in Wan that creates logophoric
domains is non-verbal, involving intonation and gesture.
A non-verbal report opener could be a stage in the grammaticalization
chain of such markers. In West African languages in general, as noted
above, there is always a prosodic break between the Quote Margin and the
Quote Content, and the last item in the margin is the quotative. It is
conceivable that, over time, the quotative could be eroded but the prosodic
mark remains. This could be the intonation. Such a suprasegmental could
be reinforced by a gesture.
Be that as it may, the only form necessary for creating a logophoric
domain is a report opener: a quotative which may be a verb or
a complementizer or non-verbal signals of intonation and gesture. In the
next section we review the functions, uses and motivations of logophoric
marking across languages.

16.5 Functions and Motivations of Logophoricity

The main function of logophoric marking is to avoid ambiguity of refer-


ence between the source of information (the reported speaker) and
a protagonist in a logophoric domain (Kibrik 2011). In this scenario, the
logophoric markers are in opposition to other markers that signal disjoint
reference, usually the ordinary third person pronouns. The reverse of this
is what happens in antilogophoricity.
Logophoric markers can also be seen as a device that allows the attribu-
tion of authorial responsibility in discourse (Hill and Irvine 1992). Thus,
a speaker indicates that they are only reporting someone else’s message
and the logophoric marker is used to refer to that person. Because of this

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 531

a speaker can distance themselves from the reported content and not
assume responsibility for it.
As such logophoricity can be viewed as an evidential strategy
(Aikhenvald 2004; Dimmendaal 2001, 2014). Even though in studies of
evidentiality, logophoricity is not usually considered a subarea of that
domain, if evidentiality is defined as being primarily concerned with the
source of information, then logophoric markers are in a sense devices for
evidentiality, even if only indirectly. In fact Stirling (1993) explicitly links
her analysis of logophoricity to evidentiality, and Sells (1987) states that
logophoricity is used to express evidentiality. Culy (1994a: 1084) also
suggests that if evidentiality is concerned with attitudes to knowledge,
then logophoricity can be seen as an evidential device. Thus, in terms of
indicating source of information and also in terms of authorial responsi-
bility and knowledge aspects of logophoric markers, logophoricity can be
considered an evidential-like device in the languages in which they occur.
The avoidance of referential ambiguity and the evidential strategy-
marking functions can be viewed as cognitive motivations for logophoric
systems. These functions can be shared with other linguistic categories.
The intriguing question about logophoricity is why it is predominantly
found in the Macro-Sudanic belt of Africa.7 To begin with, the presence of
logophoric markers in some of the languages can be attributed to areal
contact. As noted above, a language like Avatime has borrowed the form
from Ewe. Dimmendaal (2001) suggests that logophoricity is an archaic
feature in Niger-Congo and Nilo-Saharan languages of the area and sur-
mises that if it is not an inherited feature in Chadic (and Omotic) languages
from Proto-Afroasiatic, then these languages could have acquired it ‘as
a result of areal contact with these Niger-Congo and Nilo-Saharan lan-
guages’ (p. 155). Thus, the presence of logophoric markers in the area
can be explained in each case as being due to genetic inheritance or due
to areal diffusion. This still does not explain its sustainability in the belt.
Ameka (2004) argues that logophoricity in these African languages is an
elaboration of a pervasive cultural practice in their grammars of triadic
communication, i.e. ‘the art of communicating with another through
a third party’ (Yankah 1995: 2; see also Tarr 1979; Ameka and Breedveld
2004). In other words, the grammatical device of logophoricity is a usage
effect of the use of intermediaries in communication. H. Hill (1995) in
describing reported speech in Adioukrou, a Kwa language of West Africa, is
one author who relates the cultural practice and the grammatical marking
of reference in speech events of reported speech. She explains that
a speaker uses reported pronouns when they want to ‘encode someone
else’s speech as a reporter transmitting a message . . . When a speaker is
simply a reporter he is assuming the role of the neutral third party. Using
third parties is in accord with the preferred method of interaction in Adioukrou
culture’” (H. Hill (1995: 93, emphasis added).

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
532 FELIX K. AMEKA

Another aspect of the use of logophoric pronouns, especially in lan-


guages with a syncretic second and third person form as opposed to a non-
logophoric first person, is that the logophoric pronoun can be used for
a first person, the current speaker as opposed to the reported speaker. This
usage can be compared to how other evidentiality strategies are employed
as a distancing mechanism and also linked to cultural practices. Bamgbose
˙
(1986: 83) describes this use of the logophoric pronoun in first person
reference as ‘taboo use’. He argues that such a usage is motivated in
Yoruba culture consistent with the ‘deliberate avoidance of attributing to
oneself unpleasant things such as death, illness and other types of mis-
fortune’. He illustrates this point with a situation where someone could
say to you ‘you are sick’, and this can be reported in Yoruba as in (30).

(30) ó nı́ ara ò un ò dá


3s g say body l o g n e g well
‘He says that I (l o g ) am sick.’
A more general point can be made of the use of logophoric reference for
first person reference as situations involving referring to s e l f as o t h e r ,
as if there was a split personality. This is related to epistemology where one
presents knowledge of oneself as if it was someone else’s (see, e.g., Kuno
1987 for explorations in epistemology in grammar). It is also related not
just to situations of avoidance of misfortune but to the more general
cultural imperative of observing ambiguity in communication in African
cultures.

16.6 Closing Statement

Logophoricity is used for at least two distinct phenomena in linguistics.


One use with which I have been only tangentially concerned in this
chapter is that of indirect and long-distance reflexives to signal cross-
reference with a reported speaker to avoid ambiguity. The other use of
logophoricity, and the one that I have been concerned with in this chapter,
is in reference to special grammatical devices that are used predominantly
in a number of African languages to refer to the source of the reported
content in a reported speech event.
In these languages, it is an obligatory grammatical category where the
languages force the speakers to signal the source of the information in
a reported content. In this way they are comparable to evidentiality stra-
tegies. It is a mechanism for attributing authorial responsibility in dis-
course, thereby allowing speakers to distance themselves from the
reported content.
It has been argued here that the critical ingredient for the creation of
logophoric domains are report openers rather than so-called logocentric
predicates. The report openers are either quotative verbs or

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 533

complementizers, or they can be non-verbal signals like intonation or


gestures, as is the case in Wan (Mande, Cote d’Ivoire). More attention
should be paid to such non-verbal signals.
The motivations for the use and the occurrence of logophoric markers in
Africa are cognitive and cultural. Logophoricity is an instantiation of
a cultural preoccupation with third-party communication in grammar.
Logophoricity as a grammatical category shows that we cannot understand
grammar and cross-linguistic typology without paying attention to form,
history in terms of genetic inheritance and areal contact, meaning and
above all cultural practices. Through this chapter linguists should gain
more understanding of the fascinating phenomenon of logophoricity
which links to other areas such as evidentiality strategies and attitudes
to information source and reference tracking in discourse.

Notes

1. The Macro-Sudanic belt is ‘a broad belt south of the Sahara . . . sand-


wiched between the Atlantic Ocean and the Congo Basin in the south
and the Sahara and Sahel in the north, and spans the continent from the
Atlantic Ocean in the west to the escarpment of the Ethiopian Plateau in
the east’ (Güldemann 2008a: 151–2).
2. The term is based on analogy with other terms like anaphoric = back-
ward reference, cataphoric = forward reference and logophoric = word
reference.
3. In some Inland Ewe dialects the form of the pronoun is yi.
4. In the examples, and throughout the text boldface is used to draw
attention to particular items under discussion. Also, subscripts are
used to indicate co- and/or disjoint reference between participants.
5. When it comes to first person marking, there are conflicting reports in
the literature about Lele a Chadic language of Cameroon. Wiesemann
(1986) suggests that first persons can be triggers for logophoric marking
in this language. However, Frajzyngier (2001), which is a reference
grammar of the language, does not indicate that this is possible
(cf. Siewierska 2004; see §16.5 for the use of non-first person logophoric
pronouns for oneself).
6. The representation of these hierarchies is the following taken from
Huang (2000: 167):

Hierarchy for logophoric triggers


Surface structure: subject > object > others
Semantic role: agent > experiencer > benefactor > others

Both the surface structure or grammatical relations and the semantic


roles hierarchies are reminiscent of other hierarchies for NP accessibil-
ity to all sorts of syntactic structures such as relativization and case

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
534 FELIX K. AMEKA

marking. Their relevance for cross-reference marking in logophoric


contexts is epiphenomenal in my view.
7. ‘Pure’ logophoricity as a grammatical category has only been reported
in very few cases outside Africa. Bugaeva (2008) describes the logopho-
ric system in Ainu. There are reports of their occurrence also in the
Amazon (Aikhenvald 2004).

References

Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford University Press.


2008. Semi-direct speech: Manambu and beyond. Language Sciences 30:
383–422.
Ameka, Felix K. 2004. Grammar and cultural practices: The grammaticali-
zation of triadic communication in West African languages. The Journal
of West African Languages 30(3/4): 5–28.
2008. Aspect and modality in Ewe: A survey. In Felix K. Ameka and
M. E. Kropp Dakubu (eds.), Aspect and modality in Kwa languages, pp.
135–94. Amsterdam: John Benjamins
Ameka, Felix K. and Anneke Breedveld. 2004. Areal cultural scripts for
social interaction in West African communities. Intercultural pragmatics
1(2): 167–87.
Andersen, Torben. 1999. Anti-logophoricity and indirect mode in Mabaan.
Studies in Language 23(3): 499–530.
Banfield, Ann. [1982] 2014. Unspeakable sentences: Narration and representation
in the language of fiction. (Routledge Revivals). London: Routledge.
Bamgbose, Ayo. 1986. Reported speech in Yoruba. In Florian Coulmas (ed.),
˙
Direct and indirect speech, pp. 77–98. New York: Mouton de Gruyter.
Bond, Oliver. 2006. A broader perspective on point of view: Logophoricity
in Ogonoid languages. In John Mugane, John P. Hutchinson and Dee
A. Worman (eds.), Selected proceedings of the 35th Annual Conference on
African Linguistics: African languages and linguistics in broad perspective,
pp. 234–44. Somerville, MA: Cascadilla Proceedings Project. Available
online at: www.lingref.com, document #1313.
Bohnhoff, Lee E. 1986. ya.g Dii (Duro) pronouns. In Ursula Wiesemann
(ed.), Pronominal systems, pp. 103–29. (Continuum 5). Tübingen: Gunter
Narr.
Bugaeva, Anna. 2008. Reported discourse and logophoricity in Southern
Hokkaido dialects of Ainu. Gengo Kenkyuu 133: 31–75.
Clements, George N. 1975. The logophoric pronoun in Ewe: Its role in
discourse. The Journal of West African Languages 10: 141–77.
Comrie, Bernard. 2004. West African logophorics and the typology of
reference-tracking. The Journal of West African Languages 30(3/4):
94–103.

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 535

Coulmas, Florian (ed.). 1986. Direct and indirect speech. Berlin: Walter de
Gruyter.
Culy, Christopher. 1994a. Aspects of logophoric marking. Linguistics 32(6):
1055–94.
1994b. A note on logophoricity in Dogon. Journal of African Languages and
Linguistics 15(2): 113–25.
1997. Logophoric pronouns and point of view. Linguistics 35(5): 845–59.
Curnow, Timothy Jowan. 2002. Three types of verbal logophoricity in
African languages. Studies in African Linguistics 31(1/2): 1–26.
Dimmendaal, Gerrit J. 2001. Logophoric marking and represented speech
in African languages as evidential hedging strategies. Australian Journal
of Linguistics 21(1): 131–57.
2014. The grammar of knowledge in Tima. In Alexandra Y. Aikhenvald
and R. M. W. Dixon (eds.), The grammar of knowledge: A cross-linguistic
typology, pp. 245–61. Oxford University Press.
Essegbey, James. 1994. Anaphoric phenomena in Ewe. MPhil. thesis.
University of Trondheim.
Frajzyngier, Zygmunt. 1985. Logophoric systems in Chadic. Journal of
African Languages and Linguistics 7(1): 23–37.
2001. A grammar of Lele. Stanford, CA: CSLI Publications.
Goffman, Erving. 1974. Frame analysis: An essay on the organization of experi-
ence. Cambridge, MA: Harvard University Press.
1981. Forms of talk. Oxford: Basil Blackwell.
Güldemann, Tom. 2003. Logophoricity in Africa: An attempt to explain and
evaluate the significance of its modern distribution. STUF-Language
Typology and Universals 56(4): 366–87.
2008a. The Macro-Sudan belt: Towards identifying a linguistic area in
northern sub-Saharan Africa. In Bernd Heine and Derek Nurse (eds.),
A linguistic geography of Africa, pp. 151–85. New York: Cambridge
University Press.
2008b. Quotative indexes in African languages: A synchronic and diachronic
survey. Berlin: Mouton de Gruyter.
Güldemann, Tom and Manfred von Roncador (eds.). 2002. Reported dis-
course: A meeting ground for different linguistic domains. Amsterdam: John
Benjamins.
Hagège, Claude. 1974. Les pronoms logophoriques. Bulletin de la Société
Linguistique de Paris 69: 287–310.
Heath, Jeffery. 2008. A grammar of Jamsay, a Dogon language. Berlin: de
Gruyter Mouton.
Hedinger, Robert. 1984. Reported speech in Akɔɔse. The Journal of West
African Languages 14(1): 81–102.
Hellwig, Birgit. 2011. A grammar of Goemai. Berlin: De Gruyter Mouton.
Heine, Bernd and Mechthild Reh. 1984. Grammaticalisation and reanalysis in
African languages. Hamburg: Helmut Buske.

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
536 FELIX K. AMEKA

Hill, Harriet. 1995. Pronouns and reported speech in Adioukrou. The Journal
of West African Languages 25(1): 87–106.
Hill, Jane H. and Judith T. Irvine. 1992. Introduction. In Jane H. Hill and
Judith T. Irvine (eds.), Responsibility and evidence in oral discourse, pp. 1–23.
Cambridge University Press.
Holt, Elizabeth and Rebecca Clift (eds.). 2006. Reporting talk: Reported speech
in interaction. Cambridge University Press.
Huang, Yan. 2000 Anaphora: A cross-linguistic approach. Oxford University
Press.
Hyman, Larry M. and Bernard Comrie. 1981. Logophoric reference in
Gokana. Journal of African Languages and Linguistics 3: 19–37.
Ikoro, Suanu M. 1996. The Kana language. (CNWS Publications 40). Leiden:
Research School CNWS.
Jakobson, Roman. [1971] 1990. Shifters, categories, and the Russian verb.
In Selected writings: Word and language, pp. 130–53. The Hague and Paris:
Mouton.
Kibrik, Andrej A. 2011. Reference in discourse. Oxford University Press.
Kwuamuar, Sebastian. 1997. Ewɔ moya na Tɔgbi Agɔkɔli [It surprised Chief
Agokoli]. Accra: Bureau of Ghana Languages.
Kuno, Susumu. 1987. Functional syntax: Anaphora, discourse and empathy.
University of Chicago Press.
Lefebvre, Claire and Anne-Marie Brousseau. 2002. A grammar of Fongbe.
Berlin: Mouton de Gruyter.
Levinson, Stephen C. 1988. Putting linguistics on a proper footing: explora-
tions in Goffman’s concepts of participation. In Paul Drew and
A. Wootton (eds.), Erving Goffman: Exploring the interaction order, pp.
161–227. Cambridge: Polity Press.
Lucy, John A. 1993. Reflexive language: Reported speech and metapragmatics.
Cambridge University Press.
Nikitina, Tatiana. 2012. Logophoric discourse and first person reporting in
Wan (West Africa). Anthropological Linguistics 54(3): 280–301.
Sells, Peter. 1987. Aspects of logophoricity. Linguistic inquiry 18(3): 445–79.
Siewierska, Anna. 2004. Person. Cambridge University Press.
Soubrier, Aude. 2013. L’ikposso uwi Phonologie, grammaire, textes, lex-
ique. PhD thesis, Université Lumière Lyon 2, Lyon.
Speas, Margaret 2004. Evidentiality, logophoricity and the syntactic repre-
sentation of pragmatic features. Lingua 114: 255–76.
Stirling, Lesley. 1993 Switch-reference and discourse representation. Cambridge
University Press.
Tarr, Delbert Howard Jr. 1979. Indirection and ambiguity as a mode of
communication in West Africa: a descriptive survey. PhD thesis.
University of Minnesota.
Thomas, Elaine. 1978. A grammatical description of the Engenni language.
(Summer Institute of Linguistics Publications in Linguistics 60).

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016
Logophoricity 537

Dallas, TX: Summer Institute of Linguistics; University of Texas at


Arlington.
von Roncador, Manfred. 1992. Types of logophoric marking in African
languages. Journal of African Languages and Linguistics 13: 163–82.
Voloshinov, V. N. 1973. Marxism and the philosophy of language. New York:
Seminar Press.
Westermann, Diedrich. 1907. Grammatik der Ewe-Sprache. Berlin: Reimer.
1930. A study of the Ewe language. London: Oxford University Press.
Wiesemann, Ursula (ed.). 1986. Pronominal systems. Tübingen: Gunter Narr.
1990. A model for the study of reported speech in African languages. The
Journal of West African Languages 20(2): 75–80.
Yankah, Kwesi. 1995. Speaking for the chief: Okyeame and the politics of Akan
royal oratory. Bloomington: Indiana University Press.

Downloaded from https:/www.cambridge.org/core. University of Florida, on 05 May 2017 at 23:26:38, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/9781316135716.016

Вам также может понравиться