The Phonemes of an Araucanian Dialect

Author(s): Jorge A. Surez

Source: International Journal of American Linguistics, Vol. 25, No. 3 (Jul., 1959), pp. 177-
181
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/1263792
Accessed: 09-06-2017 18:27 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to
International Journal of American Linguistics

This content downloaded from 152.74.16.171 on Fri, 09 Jun 2017 18:27:06 UTC
All use subject to http://about.jstor.org/terms
 THE PHONEMES OF AN ARAUCANIAN DIALECT1

JORGE A. SUkREZ
UNIVERSITY OF BUENOS AIRES

0. Introduction 0.2. The corpus consists of 345 rather

1. Stress brief utterances in conversational form, 37
2. Vowels traditional tales, 7 descriptive narratives,
3. Consonants and 15 songs. The transcription is phonetic.
4. Phoneme alternations Each symbol is in general accurately and
5. Phoneme distribution unambiguously explained, except perhaps a
couple of them that need some interpreta-
0.1. This paper is based on a corpus col-tion as to their exact phonetic value.4 Lenz
lected by Rodolfo Lenz in the Upper Per- was consciously aware, in most cases, of
quenco near Pua (Chili, province of Malleco, free and conditioned variants-physiologi-
department of Mariluan) during the yearscal, and phonetically conditioned variants,
1896-1897.2 The materials were elicited from in his own terms-so that an approximate
Calvuin, a native informant about twenty- phonemic analysis of segmental phonemes is
five years old, who spoke Spanish quite de- already implicit in his remarks. Neverthe-
ficiently. He considered himself Moluche,less, the phonemic status of certain sounds
but since he lived among Pehuenches, andremains somewhat dubious owing chiefly, I
believe, to two circumstances: (1) sounds of
his dialect did not differ overtly from theirs,
Lenz names it, with some caution, Pe- low frequency, and morphemes of few occur-
huenche.3 rences seem not to have been sufficiently re-
checked; (2) the transcription reflects, rather
1 Araucanian is a native Southamerican lan-
guage spoken by about 200,000 people in Chili, and
undiscriminately, incidental forms of care-
lessnum-
by an only roughly calculated but far lesser speech, and divergencies in pronuncia-
ber of speakers in Argentina (more or lesstion probably due exclusively to differences
8,000).
They are concentrated chiefly in the provinces of The latter will probably have to
in tempo.
Bio-Bio, Malleco, and Cautin (Chile), and in the
be treated as morphophonemic alternations,
province of Neuqu6n in Argentina.
2 Rodolfo Lenz, Estudios Araucanos, Santiago
so that ultimately they cannot affect the
de Chile, 1895-1897 (reprint with different paging phonemic analysis. More regrettable is the
from Anales de la Universidad de Chile, vols. 90- fact that features of stress, intonation, and
98, 1895-1897. I quote from the reprint). I have juncture are either entirely disregarded or
made use of Lenz's other studies on the Araucan-
not satisfactorily described.
ian language, especially the one appeared in Bi-
0.3. The present paper attempts to supply
blioteca de Dialectologia Hispanoamericana 6.234-
an explicit and more consistent phonemic
244 (Buenos Aires, 1940). Lenz was a distinguished
formulation of Lenz's materials. Given the
phonetician; for data on his scholarship see Amado
Alonso, Biblioteca de Dialectologia Hispanoame-
nature and the deficiencies of the data, the
ricana 6.269-278.
Estudios Araucanos, 313-314, 447-448). For the
3 Lenz recognizes among the Araucanians from
Chili four linguistic groups: Picunche, Huiliche,
difficulties involved in the identification and clas-
Pehuenche, and Moluche. This division agrees sification of Araucanian tribes, see Cooper, ibid.
with the tribal classification by John M. Cooper in 692-693 and 755, with special reference to Pehuen-
Handbook of Southamerican Indians 2.690 (1946), che. For other classifications see John A. Mason,
where the Moluche is called Mapuche. But since ibid. 6.307-308 (1950), and P. Rivet-C. Loukotka
the differences between Pehuenche and Moluche in Les langues du monde, 1101-1102 (Paris, 1952).
are small, and only lexical, Lenz seems to consider 4 Most of Lenz's symbols have been replaced
them as two subdivisions of the same dialect in (cf.this paper by others of current usage.
177

This content downloaded from 152.74.16.171 on Fri, 09 Jun 2017 18:27:06 UTC
All use subject to http://about.jstor.org/terms
178 INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS VOL. XXV

analysis is not complete, nor consonant,

exempt of
[i] is in free variation with [ie],
doubts. But a tentative solution[I],isand [ii]; example:
given in [fiSa . fie6a -- fiE6a
all cases, except when-as in suprasegmental
. fii6a] old, great. All these segments may
phonemes-no conclusion at all can be drawn
be grouped in one phoneme /i/.
from the corpus. The main interest of[k]the
[o] after is in free variation with [uo],
phonemic system lies, I think, as inin the
[kom fact
kuom] all. Both variants may
that it exhibits five phonemically distinct
be grouped in one phoneme /o/.
positions in the nasal series.5 Also [i] remark-
after consonant (especially [w] and the
able are the frequent alternations
nasals)of pho-
is in free variation with ["i] and [Ii]
as in [&oji -N 6Dui] it dies out; [puwi -
nemes, perhaps of an affective character.
puwqi] he came. The variants maybe grouped
1. Lenz makes the following points con-/i/.9
in one phoneme
cerning stress, (appliable to all theThedialects):
remaining vowels are in contrast with
each other
(1) it is not very strong; (2) it moves and with the above listed. Con-
from
one syllable to another depending sequently, the
on there are six vowel phonemes:
rhythm of the utterance; (3) it /i u e o a/.from
moves
one syllable to another when particles are
The two semivowels [w] and [il are classi-
joined to the word; (4) the words
fied as bear in on account of their dis-
consonants
general penultimate stress whentribution.
ending in a
vowel, and ultimate stress when ending in a
consonant. An example from Huiliche is
3.1. The following consonants occur:
[pl, [t], It], [],, [k]: voiceless stops, at bila-
given of the mobility of stress: [toafmu mali
fii Aikal here is my house, but [fii Auk&
bial, dentalmo
or interdental, alveolar, prevelar,
kapan] I come from my house.6and From (4) it
velar positions. Sometimes they are very
seems that stress might be predictable with
strongly articulated, especially after [a].
reference to juncture; (2) points to an[6]:
[tr], utter-
voiceless prepalatal affricates; [t']
ance-stress; and (3) to a stress which is apicalmight
(more or less like English [tr] in
be stated in terms of classes of morphemes.7 tree); [6] is dorsal. Both are sometimes pre-
Since accents are only exceptionally ceded bymarked
a brief homosyllabic friction.
in the Pehuenche texts, no conclusion [A], [v], can
[E], [y], [y]: voiced fricatives, at
be drawn from the corpus. bilabial, labiodental, dental or interdental,
palatal, and velar positions. [,] is a very open
2. The following vowels occur:consonant, [i i u e a o little friction. Fricatives in
with
al. general are frequently produced without full
[i] and [a] are in complementary distribu- voice.
tion: [i] is always stressed, [a] occurs only in [W], [f], [Y]: voiceless fricatives, at bilabial,
unaccented syllables.8 Before a prepalatal labiodental, and palatal positions.
6 This fact seems not to have been observed up
[s], [s]: voiceless sibilants, apico-dental
to now in any other language; cf. Charles F. Hock- and lamino-alveolar respectively.
ett, A Manual of Phonology, 119 (1955). [s]: voiceless alveolar fricative.
B [tafamu] here; [mal-] (to) be; [-i] he; [ni] my; [r]: voiced frictionless alveolar continuant,
[suka} house; [-mo] from; [kopa-] (to) come; [-n] I.
perhaps retroflex.
7 The situation described by Lenz reminds us
considerably of the phonemic status of stress in [11], [1] , [:, [ voiced lateral continuants,
Wishram Chinook, as described by W. Dyk and at dental, alveolar, prepalatal, and palatal
D. H. Hymes, IJAL 22.238-241 (1956). positions. [1] is apical, but does not seem to
8 This distribution is explicitly stated by Lenz, be retroflex.
and the alternation corroborated by examples in
the corpus; examples: [kila - kola] three; [ripi ! 9 The raised vowels represent in all eases very
ropi] road. brief glides.

This content downloaded from 152.74.16.171 on Fri, 09 Jun 2017 18:27:06 UTC
All use subject to http://about.jstor.org/terms
NO. 3 THE PHONEMES OF AN ARAUCANIAN DIALECT 179

ofnasal
[m], [ n], [n], [ii], [n], : voiced parallelism
con-between the allophones of //
and those of the other fricatives seems to re-
tinuants, at bilabial, dental or interdental,
alveolar, prepalatal, palatal, andflect the posi-
velar introduction of Spanish loanwords
tions. [n] is apical, but does notwithseem[s] which
to beoccasioned the redistribution
retroflex. of the allophone. Lenz's testimony (see
[w]: voiced bilabial rounded semivowel. 3.32) corroborates this hypothesis, and,
[i]: voiced velar unrounded semivowel. besides, I believe the analysis of the other
3.2. Of the above sounds, the following dialects and of the works of the ancient
do not contrast: grammarians will also confirm it.10
[l] occurs only in contact with the front 3.32. In Araucanian words not of Spanish
vowels; in this position it varies freely with origin, both [s] and [s] vary freely with /?/.
[k], as in [eeli - keli] red. Although in the data this alternation actu-
L13], v , [v], [f] are in free variation, as in ally takes place in only a small number of
(to/a --- tava --- toaa - tafa] here. words, Lenz explicitly states that [?] "upon
[Y] occurs only initially after pause; in losing its voice seems to change easily to [s]
this position it varies freely with [y], as in and less frequently to [s]",11 but he leaves
[Yoka- y yaka-] (to) be afraid. out of consideration Spanish loanwords. If
[7] and [i] are in complementary distribu- we were to exclude them too, we would have
tion: [y] occurs before [a], [i] before vowels to assign [s] and [s] to /I/. But since loan-
other than [a]. (In syllabic final, it is not clear
words with [s] or [s] are very numerous, but
which of the two occur.) Examples are never in free variation with [1], and since
[WSein] volcano; [-ayeyal-] for laughing. loans with [b] never show [s] or [s] either, I
have assigned the two sibilants to one pho-
[s] and [s] are in free variation, as in [kansa-
kansa-] (to) be tired. neme /s/, and postulated an alternation of
[l] occurs only in juxtaposition with a back phonemes in the rest of the words. Undoubt-
edly, these complex facts in the data reflect
vowel; in this position it varies freely with
[1], as in [wile - wile] morning. a moment of phonemic transition and in-
[n] also occurs only in juxtaposition withstability, due to the introduction of Spanish
a back vowel, varying freely with [n], as inloanwords with dorso-alveolar [s].
[wine - wine] first. 3.33. [s] is a sound of extremely low fre-
Each of the above sets of sounds can be quency. Apart an onomatopoeic word ['iil]
joined in one phoneme. The remaining(the cry of the widgeon), and a few Spanish
sounds, except [s] (see 3.33), are in contrast
loans, in which it varies freely with /s/, it
with those listed above and with each other,occurs in the following morphemes (none of
so that they must be considered allophones which occurs more than twice): [Aiwal-] (to)
of distinct phonemes; there are then twenty-tuck up (one's clothes); [we?a] bad, in which
one consonant phonemes: /p t t k tr 6 ff y 10 The voiceless variants of fricatives are the re-
y r s 1 X m n n fi w/. sult of a devoicing process which is hardly re-
corded in the dialect of the North, Picunche, that
3.3. Some points in the phonemic analysis
need further remarks. has only [v], [#], and [r]; but which is already ac-
complished in Huiliche, the dialect of the South,
3.31. It may seem queer to find a partially
that possesses instead [f], [s], and [s] respectively.
devoiced [t] (represented by Lenz withThe
a free variation of voiced and voiceless frica-
special symbol), together with the non-occur-
tives in Pehuenche represents an intermediate
rence of the totally unvoiced [q]. To assumestep between the voiced sounds of Picunche and
those of Huiliche which are almost all voiceless.
that Lenz's symbol covers both the devoiced
11 Estudios Araucanos, 135. Furthermore, Arau-
and the voiceless forms, would be arbitrary:
canian words (not loans) only with [?] or (s] in the
Lenz distinguishes accurately elsewhere one
text are usually identified in footnotes by Lenz
type of sound from the other one. This lack
with the same words with [j].

This content downloaded from 152.74.16.171 on Fri, 09 Jun 2017 18:27:06 UTC
All use subject to http://about.jstor.org/terms
180 INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS VOL. XXV

it varies freely with /r/, /s/, andwell

/I/;as[kuse]
some early grammarians, attribute
old, in which it varies freely withthem an emotive
/s/ and value. But Lenz also ad-
/8/ but contrasts with /r/ ([kure] mits wife).
that perhaps the formal changes may
Neither these data nor Lenz's description
correspond to some modifications of meaning
which he didits
allow of any definite decision concerning not notice and unified therefore
status; I consider it here, rather in the translation. The following phonemes
arbitrarily,
an allophone of /s/, imputing the alternate:
non-occur-/t/, /t/, /6/: fita, fit3a, ffia old;
rence of [siwal-] to mere chance.12 /t/, /t/: 3ente, ,ente so much; /t/, /t/, /tr/:
3.34. The assignment to [y], [i], and wente,
wente, [w] wentre above; /t/, /s/: ki6u,
of a different phonemic status kisu than that/Z/, /s/, /r/: we6a, wesa, wera
alone;
assigned to the glides [i], [i] and [u]bad; /n/, /n/: mawin, mawin to rain; /m/,
mentioned
/n/: moXin,
in 2, is not arbitrary. There is contrast be-noXin to get drunk; /i/, /u/:
tween the first three, which have the leu3u
leu#fi, same river; /i/, /e/: kori, kore broth.
distribution as every other continuant (see
5), and the glides, which occur 5. Syllables have the following structures:
always as
V, CV,
part of syllable-nuclei, and can be in every
VC, CVC (the last two only before
case replaced by the vowel alone; consonant
besides,or pause); in Spanish loanwords
Lenz distinguishes consistently between also CCV, [y]the second consonant being al-
ways /r/ or /1/, as in /pofre/ < Sp./pobre/
and [i], [i] and ['], [w] and [u]. On the other
hand, [y], [w], and [i] cannot bepoor, /flanko/
joined in < Sp./blanko/ white. Some
monosyllabic consonant clusters do occur in
one phoneme with [i], [u], and [i] respectively,
words that are not loans, but alternating in
because they contrast in identical positions.
every case with CaC; examples: [tfa - tafa
Examples: [feyika&i] then; [feiirke] that is
here; [puwrkei - [puwarkei] he really comes.
right; [wiywin] I am thirsty; [piifia] bit (for
These clusters must be interpreted phonem-
horses); [uikai] he opened; [witr en] I am
ically as /CVC/.13
cold; [wul-] (to) sell; [puwkei] he is coming;
Distributionally, the consonants fall into
[puuyu] deer; [walwe] farm; [eluayu] I shall
two intersecting classes, according to their
give you; [Tifiim] bird; [wenii] friend; [kui']
hand; [reile] seven. The facts might be met
position in the syllable: stops occur only
initially; the remaining ones-which may be
with differently if we posited some phonemic
labeled continuants-occur both initially
feature of syllabification, which does not
and finally.14 The consonants occur before all
seem otherwise necessary.
vowels, except /w/ that does not occur be-
4. A characteristic of the dialect is the fore /o/; /y/ before /o/, /u/; /t/ before
/i/, /o/; /!/ before /i/, /e/; /n/ before /i/,
frequent alternations of phonemes. Lenz, as
/e/, /o/, /u/. The occurrences of /t/, /1/,
12 The marginal status of [?] in Pehuenche isand not/n/ are furthermore restricted to a few
what we would have expected after the compari-
morphemes, but all these restrictions may
son with the other dialects, for [s] is in Huiliche
the voiceless sound corresponding to [r] in Picun- be due to inaccuracy in the recording: Lenz
che. We could have expected therefore [s] to himselfbe in points out to the fact that only at
Pehuenche a free variant of [r], and Lenz states the end
in of his work with the informant he
the Introduction to all the dialects (p. XXIII) discovered those three sounds were distinc-
that this is precisely the case. But among the pre-
tively different from /t/, /1/, and /n/ respec-
liminary remarks on Pehuenche he does not men-
tively.15
tion the fact again, while, on the contrary, he
calls attention to the contrast between [kuse] and 13 It is mainly for this reason that I have re-
[kure] (p. 131). I have prefered to be guided onlyjected the phonemization of [tr] as /t/ + /r/.
by the evidence in the text, in which case no other 14 In two interjections this rule is infringed:
analysis seems to meet the facts better than the /ap/ hello!, and /8ot/ great Scott!, gee!
one I have proposed. 15 It is worth noting that the relevance of the

This content downloaded from 152.74.16.171 on Fri, 09 Jun 2017 18:27:06 UTC
All use subject to http://about.jstor.org/terms
NO. 3 THE PHONEMES OF AN ARAUCANIAN DIALECT 181

Bi-syllabic clusters of consonants

Xw; mp, mt,that
mt, kmtr, , me, m, my,
have been observed are the following: pp,mn, mnf, mi, mw; nt, nt,
nmy, mr, ml, mm,
t3, Ftt, /k, t , r, B3 1, /r , Onl, nk,
Omntr, nw; B,
, /n, np, /Bw;
nt, nk, ntr, n6, n,B, n3, ny,
bp, bk, Str, ~y, r, ; 7yp, yt, yk,
nr, ns, y7,
nl, nX, yl,
nm, nn, nfi, no, nw; nip, fit,
ym, yD; y#, yw; rp, rk, rt, rO, ilk,rl,
iinc,r),
nf,rw;
fir, sp,
nl, nm, fi, nw; ok, D6, D6,
st, sk; lp, 1k, jD; lp, It, 1k, ltr,Dr, Dl, 1l,
1c, om, ly,
on; wt,
ly,wk, wtr, wl, wm.
lr, im, In, ifi, ID; Xp, Xk, Xc, X#, As Xr,
manyXm, as five
Xi,vowels may appear in
succession, each belonging to a different
syllable. Not all the possibilities of occur-
dental sounds in front of the alveolar ones was al-
ready clearly recognized by Father Valdivia at the
rence have been observed; the most notice-
beginning of the XVIIth. century (Luis de Val-
able restrictions are (1) no more than two
divia, Arte, Vocabulario y Confessionario de la
Lengua de Chile, Lima, 1606; ed. facs. J. Platz-like vowels occur in contact; (2) the follow-
mann, Leipzig, 1887, p. 9). ing pairs do not occur: io, ei, ai, oi, oo.

This content downloaded from 152.74.16.171 on Fri, 09 Jun 2017 18:27:06 UTC
All use subject to http://about.jstor.org/terms