Adelaar Willem, Muysken Pieter. The languages of the Andes
Подождите немного. Документ загружается.
364 3 The Inca Sphere
Table 3.26 Chipaya consonant inventory (based on Cerr´on-Palomino, MS)
Labial Alveolar +affr
*
Apical Palatal Retroflex Velar +lab
**
Postvelar +lab
Plain p t c ˇcˇc
.
kk
w
qq
w
obstruents
Glottalised p’ t’ c’ ˇc’ ˇc
.
’k’ q’
obstruents
Aspirated p
h
t
h
c
h
ˇc
h
ˇc
.
h
k
h
q
h
obstruents
Fricatives s ¸sˇsˇs
.
hh
w
x
***
x
w
Nasals m n n
y
ŋ
Laterals l l
y
Lateral l
fricative
Vibrant r
Glides w y
*
affr. = affricate
**
lab. = labialised
***
We follow Cerr´on-Palomino and Olson in writing the postvelar or fortis velar fricative (and its
labialised counterpart) as x (x
w
), not x
.
(x
.
w
).
Muysken has not provided a better view of it. A comparison of the Vellard materials
with the data collected by Uhle is probably the best basis for further analysis.
There have been several attempts to study and document the Chipaya language. In
the 1960s, Olson collected extensive lexical material. Part of it is included in two ar-
ticles designed to present evidence for a putative genetic relationship of Uru–Chipaya
with the Mayan languages in Mesoamerica (Olson 1964, 1965). (For the reception of
Olson’s views on this see section 1.7.) A third article (Olson 1967) deals with Chipaya
syllable structure and contains a discussion of its phoneme inventory. In the 1980s,
Porterie-Guti´errez also brought together a substantial amount of Chipaya material, but
the publication of it was prevented by her untimely death. An annotated and translated text
belonging to this material was published posthumously by Howard (Porterie-Guti´errez
1990). In 2001, Cerr´on-Palomino began a new research effort geared at the documenta-
tion of Chipaya. The Chipaya data presented in the following pages are largely based on
his results obtained so far (Cerr´on-Palomino MS). A comparison with Uchumataqu will
be made when relevant and possible. The two languages are clearly related, although
probably not mutually intelligible (Torero 1992: 181). See Olson (1965: 37–8) for a list
of 87 obvious cognates, in which most pairs show only minor phonetic differences.
The consonant inventory of the Chipaya language is presented in table 3.26.
The consonant inventory in Olson (1967) differs from the one presented in table 3.26
by the fact that the aspirated consonants and the lateral fricative (lh) are interpreted as
3.6 The Uru–Chipaya languages 365
sequences of consonants, rather than as unit phonemes. He characterises the fricatives
in the postvelar column (x, x
w
)as‘fortis velars’. The palatal and the apical fricatives are
treated as allophones of a phoneme /s
.
/which occurs as palatal [ˇs] after high vowels and
as a backed alveolar [s
.
] elsewhere. Olson (1967: 300), furthermore, noted the use of a
glottal stop in women’s speech; for instance, in oqa
ʔ
a ‘I am going’ (as against oqaˇc
.
a
in men’s speech). This usage did not seem to be continued by the younger generation,
however.
Cerr´on-Palomino notes an absence of restrictions upon the use of aspirated and glot-
talised consonants within a root. They occur initially in a syllable but this need not be
the first syllable in the root (e.g. in tot
h
i ‘cow’s horn’, kunt’´ı¸s ‘to be sure’). Furthermore,
glottalised and aspirated consonants can co-occur in the same root (e.g. in p
h
anˇc’u ‘soft’,
ˇc
.
’ik
h
a ‘equal’). Considering their great frequency in native roots, there appears to be no
reason to assume that glottalisation and aspiration are borrowed features in Chipaya.
Note that glottalisation and aspiration are not contrastive in the labiovelar and
labiopostvelar stop series. According to Cerr´on-Palomino, it may be due to a recent
simplification of the system because free variation between labialised and non-labialised
aspirated consonants is still observed in roots such as q
h
a¸s ∼ q
hw
a¸s ‘water’. In addition,
the same variation can be found in the alveodental stop series (e.g. in t
h
at¸s ∼ t
hw
at¸s
‘to pile up’). These examples suggest that the labial element may have had the status
of a segmental phoneme, rather than that of a feature of the consonant with which it is
associated. At present, the loss of labialisation seems to have become a general tendency
in the language.
Nasal contrasts are maintained in syllable-final position, even that between alveoden-
tal n and velar
ŋ
,asinlan¸s ‘to touch’ and la
ŋ
¸s ‘to work’. The glide consonant w is
often realised as a fricative [β], especially when in contact with front vowels (e.g. siwi
[siwi ∼ siβi] ‘winter’). Both elements are reminiscent of the situation in the Aymara of
northern Chile (cf. section 3.3.3).
In comparison to Aymara and Quechua, the articulation point of the postvelars in
Chipaya is less retracted and does not normally reach the uvular range. This can make it
difficult to recognise the distinction between velars and postvelars. Nevertheless, there
is an ample choice of minimal pairs illustrating the contrast:
(330) kara ‘wide’ qara ‘comb’
(Cerr´on-Palomino, MS)
(331) k’oru ‘bowl’ q’oru ‘a type of hat’ (Cerr´on-Palomino, MS)
The consonant inventory of Uchumataqu, as established by Muysken (MS), is similar
to Cerr´on-Palomino’s Chipaya inventory but for the absence in the former of the apical
and retroflex series (¸s; ˇc
.
, ˇc
.
’, ˇc
.
h
, ˇs
.
). In addition, the alveolar affricate series is only
represented by plain c; the velar nasal (
ŋ
) and the labialised postvelars (q
w
, x
w
) are
lacking; and glottalised p’ and t’ occur very seldom. Possibly, the situation of near
366 3 The Inca Sphere
extinction in which the Uchumataqu language has found itself for decades may be
responsible for a reduction of the sound system, but it could also be the result of earlier
developments.
Both Chipaya and Uchumataqu have a five-vowel system with distinctive vowel length
(a, e, i, o, u; a:, e:, i:, o:, u:).
157
The contrast of high and mid vowels in Chipaya is
illustrated in the following examples:
(332) uwa ‘food’ owa ‘knee’
(Cerr´on-Palomino, MS)
(333) hik¸s ‘road’ hek¸s ‘to appear’ (Cerr´on-Palomino, MS)
The contrast is also retained in the environment of a postvelar consonant. In con-
tradistinction to Aymara and Quechua, the automatic lowering of high vowels is either
minimal, or does not occur at all in that environment:
(334) q
h
uˇca ‘potsherd’ q
h
oˇca ‘foot’ (Cerr´on-Palomino, MS)
(335) q
h
in
y
a ‘crippled’ q
h
en
y
a ‘slow’, ‘dull’ (Cerr´on-Palomino, MS)
The difference between short and long vowels is illustrated in (336) and (337):
(336) uˇs
.
a ‘north’ u:ˇsa ‘sheep’ [Spanish oveja]
(Cerr´on-Palomino, MS)
(337) qa¸s ‘to talk’ qa:¸s ‘to cry’ (Cerr´on-Palomino, MS)
Vowel length in Chipaya is often the result of contractions, which seem to occur
frequently in the language; for instance, we:n ‘night’ corresponds to Uchumataqu wiyani
(cf. Olson 1965: 37–8). It may provide an explanation for the relatively low functional
load of the length distinction.
Cerr´on-Palomino notes that after lateral and velar fricatives, as well as after aspirated
consonants, vowels tend to be voiceless when the following consonant is also voiceless,
e.g. l
oki [l
˚
oki] ‘mud’, t
h
ut¸s [t
h
˚
ut¸s] ‘to spit’. In Olson (1967) such voicelessness is
generally interpreted as the manifestation of a syllable-final fricative.
Chipaya accepts clusters of two consonants in word-initial and word-final position.
Most characteristic are the word-initial clusters consisting of a sibilant which is either
followed by a bilabial stop or nasal, or by a velar or postvelar stop. Their occurrence
is reminiscent of Puquina (cf. section 3.5). In these clusters the opposition between the
sibilants s, ¸s and ˇs
.
remains intact:
(338) skara ‘hat’ ¸skara ‘toad’ ˇs
.
kati ‘nearby’
(Cerr´on-Palomino, MS)
(339) smoya ‘mosquito’ ¸smali ‘strong’ (Cerr´on-Palomino, MS)
157
For Uchumataqu Uhle mentions a sixth vowel, underlined e, reported to sound like the vowel
in German ¨a. Its occurrence still needs to be studied in the word lists he collected.
3.6 The Uru–Chipaya languages 367
Word-final clusters can be part of a truncated root (for instance, in hik¸s ‘road’, from
*hik¸sa; compare hik¸sa-l
y
a ‘little road’). More often they are the result of suffixation, for
instance, when the infinitive marker -¸s is added to a verb root. Before -¸s root-final vowels
are suppressed (except when preceded by a sequence of more than one consonant).
(340) qulu ‘salt block’ qul-¸s ‘to salt’
(Cerr´on-Palomino, MS)
(341) wiya ‘dream’ wiy-¸s ‘to dream’ (Cerr´on-Palomino, MS)
As can be deduced from the example of the infinitive marker, vowel suppression
is a common procedure in Chipaya, in which respect it resembles Aymara. Although
in Chipaya word stress is characteristically located on the penultimate syllable, it may
remain on the last syllable when the vowel of a word-final syllable is suppressed. A
comparison with Uhle’s data shows that the infinitive marker was originally syllabic
(-ˇca), hence an example such as (342) in which stress is on the final syllable:
(342) u¸s´ın-¸s [<*u¸s´ın-ˇca] ‘to play’
(Cerr´on-Palomino, MS)
Within noun phrases the final vowel of the modifying element, which precedes the
head, is also subject to truncation; compare the situation in Aymara:
(343) o¸sp
h
eta [o¸s(a) p
h
eta]
nose opening
‘nostril’
(Cerr´on-Palomino, MS)
In Uchumataqu, Muysken (MS) recorded consonant clusters of even greater com-
plexity than those found in Chipaya but notes that these tend to become simplified in
the speech of the present generation of semi-speakers.
The principal morphological device of Chipaya is suffixation. The overall picture,
however, is rather different from that in the neighbouring languages. Nominal morphol-
ogyisrelatively elaborate and comprises a distinction between feminine and masculine
gender, as well as a fair number of case markers. Subject and object are not marked for
case. Personal possession is indicated by means of free pronouns marked for genitive
case. In a verb phrase person, gender and number of the subject can be specified by
means of clitic elements. These elements are normally attached to a non-case-marked
expression that accompanies the verb (preferably the object or an adverb, occasionally
the subject). Verbal suffixes indicate tense, aspect, mood and evidentiality. The shape
of tense/aspect suffixes is to a certain extent again determined by person, gender and
number of the subject. The preferred word order in Chipaya is subject–object–verb and
within the noun phrase modifier–head.
158
158
In Uchumataqu there are a few instances of the reversed order (head–modifier), e.g. in xa´us
w´ıri ‘anus’ (xa´us ‘hole’, w´ıri ‘buttocks’); see Muysken (2000: 102).
368 3 The Inca Sphere
The main strategy to indicate feminine gender in nouns is by the addition of a suffix
-i. This suffix replaces an original stem-final vowel, if present (344):
(344) a¸snu ‘donkey’ a¸sn-i ‘female donkey’
[Spanish asno]
(Cerr´on-Palomino, personal communication)
Final -i is often elided when preceded by a single consonant (345), except when such an
elision would lead to the coincidence of a feminine form with its masculine counterpart.
This is the case, for instance, in Spanish proper names ending in a consonant (346).
(345) u:ˇsa ‘sheep’ u:ˇs-i ∼ u:ˇs ‘ewe’
[Spanish oveja]
(Cerr´on-Palomino, personal communication)
(346) huwan ‘John’ huwan-i ‘Joan’ [Spanish Juan]
(Cerr´on-Palomino, personal communication)
The demonstratives ti: ‘this’ and ni: ‘that/he’, as well as the numeral ci: ‘one’, have
special feminine forms: ta:, na: and ca:, respectively.
The markers of the genitive case with nouns are also sensitive to gender. When used as
modifiers, masculine nouns take ˇs
.
,whereas feminine nouns replace their final -i with -a.
(347) kuˇc-ˇs
.
k
h
un
y
ikuˇc-a k
h
un
y
i [kuˇci ‘pig’, Old Spanish coche]
pig-G.MS ear pig-G.FE ear
‘pig ears’ ‘sow ears’
(Cerr´on-Palomino, personal communication)
According to Cerr´on-Palomino (personal communication), pronouns (both mascu-
line and feminine) have a special genitive case marker -(i)ˇs
.
,which has the function of
indicating the agent in a passive construction. (The long allomorph -iˇs
.
is found after
the personal pronoun stems wer ‘I’ and am ‘you’, which end in a consonant.) With
demonstrative pronouns a genitive of possession is also formed in this way.
When used independently or predicatively, all genitives of possession (whether nom-
inal or pronominal) are followed by an element -ta,which is sensitive to the gender
distinction; for instance, in (348):
(348) ni:-ˇs
.
-ta ni:-ˇs
.
-t-i
that-G-N that-G-N-FE
‘his one’ ‘his female one’
(Cerr´on-Palomino, MS)
The personal pronouns wer ‘I’ and am ‘you’ have special genitive forms we-t(a) and
am-ta, respectively, for both attributive and predicative possessive use. The short form
we-t is used as a prenominal modifier (349). The final element -ta is again subject to the
gender distinction (350).
3.6 The Uru–Chipaya languages 369
(349) we-t ˇcuwa
I-G plate
(Cerr´on-Palomino, personal communication)
‘my plate’
(350) we-ta we-t-i
I-G.N I-G.N-FE
‘(it is) mine’ ‘(it is) my female one’
(Cerr´on-Palomino, MS)
In a verb phrase with a third-person subject, gender is reflected in the choice of the
clitic that accompanies the verb. In the following examples the clitic elements -(i)ˇs
.
for
masculine, and -(i)l for feminine, are attached to the adverb maxn
y
a. The element -ki-,
which accompanies the future-tense marker -a-,isused with all persons except first
singular and plural (exclusive). It is not sensitive to gender.
(351) ni:-nak-ki maxn
y
a-ˇs
.
t
h
ax-a-ki-ˇc
.
a
159
that.MS-PL-TO early-MS sleep-F-F.NS-DV
‘They (those men) will sleep early.’
(Cerr´on-Palomino, MS)
(352) na:-nak-ki maxn
y
a-l t
h
ax-a-ki-ˇc
.
a
that.FE-PL-TO early-FE sleep-F-F. NS-DV
‘They (those women) will sleep early.’
(Cerr´on-Palomino, MS)
Although in the context illustrated in (351) and (352) the sole function of -(i)l and
-(i)ˇs
.
is to distinguish gender, they can have other functions as well. The element -(i)l
is used with the first person singular and plural exclusive, and with the third person
feminine singular and plural; -(i)ˇs
.
is used with the first person plural inclusive, with
the second person plural, and with the third person masculine singular and plural. A
special clitic -(i)m, akin to the second-person-singular pronoun am,isused in verb
phrases with a second-person-singular subject. (The long forms -il, -im and -iˇs
.
occur
after consonant-final personal pronouns.)
The use of the first and second persons singular is illustrated in (353) and (354).
Note that the required enclitics are attached to the subject of these sentences as no other
constituents are available.
(353) wer-il sat-a-ˇc
.
a
I-1 run-F-DV
‘I shall run.’
(Cerr´on-Palomino, MS)
(354) am-im sat-a-ki-ˇc
.
a
you-2.SG run-F-F.NS-DV
‘You (singular) shall run.’
(Cerr´on-Palomino, MS)
159
Cerr´on-Palomino reports that the ending -ki-ˇc
.
a is pronounced [kˇs
.
]innon-careful speech, in
which case stress is located on the vowel immediately preceding it.
370 3 The Inca Sphere
If a sentence contains a direct object, the latter often carries the clitic element:
(355) Luwisitu-ki t’anta-ˇs
.
lul-n
y
i-ˇc
.
a
Luisito-TO bread-3.MS eat-CU.3.MS-DV
‘Luisito is in the habit of eating bread.’
(Cerr´on-Palomino, MS)
Although the preferred word order is subject–object–verb, other order options are al-
lowed as well. As a result, a sentence such as (356), in which the object carrying the
clitic follows the verb, is acceptable.
(356) Luwisitu-ki lul-n
y
i-ˇc
.
a t’anta-ˇs
.
Luisito-TO eat-CU.3.MS-DV bread-3.MS
‘Luisito is in the habit of eating bread.’
(Cerr´on-Palomino, MS)
If there is no other stem to which the clitic element can be attached, it may also
occur as a proclitic with the verb itself. Example (357) illustrates the use of a proclitic
corresponding to a first person plural inclusive.
160
(357) ˇs
.
-lik-la
4-drink-4.IM
‘Let us drink!’
(Cerr´on-Palomino, MS)
Cerr´on-Palomino notes that the use of clitics indicating subject reference could be
entering a phase of obsolescence. Present-day speakers tend to omit them without af-
fecting the acceptability of the utterance. He observes, furthermore, that the location of
the clitic in the sentence can have a focalising function.
As we anticipated, the shape of the tense–aspect markers is also partly determined by
the person and the number of the subject, thus generating complicated patterns of con-
cord. These concord patterns may differ in accordance with the tense–aspect paradigm
involved. Consider the use of the element -ki- in the future-tense paradigm (see above).
As an additional illustration, the paradigm of the unmarked present tense is reproduced in
table 3.27. It contains a suffix -u- in forms with a first-person-singular subject, whereas
in all other person–number combinations the corresponding slot remains empty. The
exemplified phrase contains an adverb ¸sina ‘alone’ followed by the delimitative suffix
-l
y
a ‘only’ (a suffix presumably borrowed from Quechua, where it has the same shape
and meaning). This adverb also carries the enclitic that indicates person, gender and
number. The exemplified verb is t
h
ax- ‘to sleep’.
Table 3.27 suggests that there is no formal difference between singular and plural
in the third person. However, other tense–aspect paradigms have different forms for
singular and plural in the third person feminine. Such, for instance, is the case in the
160
Muysken reports that Uchumataqu also allows proclitic subject markers.
3.6 The Uru–Chipaya languages 371
Table 3.27 Unmarked present tense in Chipaya (from Cerr´on-Palomino, MS).
1 pers. sing. ¸sina-l
y
a-l t
h
ax-u-ˇc
.
a ‘I sleep alone.’
plur. (excl.) ¸sina-l
y
a-l t
h
ax-ˇc
.
a ‘We (excl.) sleep alone.’
2 pers. sing. ¸sina-l
y
a-m t
h
ax-ˇc
.
a ‘You (sing.) sleep alone.
plur. ¸sina-l
y
a-ˇs
.
t
h
ax-ˇc
.
a ‘You (plur.) sleep alone.’
3 pers. masc. sing. ¸sina-l
y
a-ˇs
.
t
h
ax-ˇc
.
a ‘He sleeps alone.’
masc. plur. ¸sina-l
y
a-ˇs
.
t
h
ax-ˇc
.
a ‘They (men) sleep alone.’
3 pers. fem. sing. ¸sina-l
y
a-l t
h
ax-ˇc
.
a ‘She sleeps alone.’
fem. plur. ¸sina-l
y
a-l t
h
ax-ˇc
.
a ‘They (women) sleep alone.’
4 pers. plur. (incl.) ¸sina-l
y
a-ˇs
.
t
h
ax-ˇc
.
a ‘We (incl.) sleep alone.’
past-tense (completive) paradigm illustrated below:
(358) na:-ki maxn
y
a-l t
h
ax-ˇcin-ˇc
.
a
that.FE-TO early-3.FE sleep-PA.3.FE.SG-DV
‘She slept early.’
(Cerr´on-Palomino, MS)
(359) na:-nak-ki maxn
y
a-l t
h
ax-ˇci-ˇc
.
a
that.FE-PL-TO early-3.FE sleep-PA.3.FE.PL-DV
‘They (those women) slept early.’
(Cerr´on-Palomino, MS)
The element -ˇc
.
a,which we have glossed as a declarative marker, is also found after
a noun or adjective in verbless equational clauses, such as (360):
(360) am q
h
oya q
h
ay- ˇcuka ew-ˇc
.
a
you house buy-FN new-DV
‘The house that you are going to buy is new.’
(Cerr´on-Palomino, MS)
The declarative marker -ˇc
.
a is almost certainly related to the very characteristic ending
-ˇcay of Uchumataqu. Muysken (2000) discusses its function in that language and finds
that it is used in declarative and exhortative sentences, as well as in yes/no questions and
in some wh-questions. It is not used in imperatives and in infinitival complements. Like
its Chipaya equivalent, the element -ˇcay is also found affixed to nouns and adjectives.
An example of a declarative sentence in Uchumataqu is:
(361) w´ır-il x´ala k’´ay˘ap´e
.
k
u-ˇcay
161
I-1 llama buy want-PN.1.SG-DV
‘I want to buy a llama.’
(Vellard 1967: 14)
161
In examples taken from Vellard we have tried to respect his phonetic notation. The apostrophe
(as in pˇe
.
k’-u-ˇcay ‘I want’) is reproduced in a slightly different way in order to avoid confusion
with the glottalisation sign.
372 3 The Inca Sphere
Vellard (1951: 37–9) presents some partial Uchumataqu verb paradigms. They sug-
gest a structure comparable to that of Chipaya (see also Muysken 2000). The Chipaya
examples given in Olson (1967) contain evidential elements referring to hearsay and
probability that appear to fill the slot of the tense–aspect markers (e.g. ap-ˇs
.
-ki-ˇc
.
a ‘he
is following, they say’; e:kt-qal-ˇc
.
a ‘he surely had hungered’).
162
Olson (1965: 33) also
notes causative formations such as t-xaw-un- ‘to cause to shout’ from q
h
aw- ‘to shout’.
A reflexive marker -s- is recorded by Cerr´on-Palomino in the form peka-s- ‘to love each
other’ (cf. Aymara -si-). These examples make it clear that the verbal morphology of
Chipaya still holds a number of areas to investigate.
Cerr´on-Palomino mentions several verbal affixes that are used as nominalisers. Verbs
such as pek(a)- ‘to want’ can be used with infinitive complements containing the affix -¸s
(362). Note the absence of an infinitive marker in a parallel construction in Uchumataqu
(361).
(362) ni:-ki ˇc
h
iˇswi-ˇs
.
lul-¸s pek-ˇc
.
a
he-TO meat-3.MS eat-IF want-DV
‘He wants to eat meat.’
(Cerr´on-Palomino, MS)
There are several other nominalising affixes, -ˇcuka ‘potential’, -n
y
i ‘agentive’, -ˇci and
-ta ‘resultative’, which can be used to form relative clauses and complement clauses.
Relative clauses and complement clauses precede the noun or the verb, respectively,
on which they depend. Their internal order is subject–object–verb. Note, however, that
the (human) subject of an embedded complement clause can be extracted to a position
preceding the (non-human) subject of the main clause, as illustrated in (360); a further
example is (363):
(363) am kula lul-ˇci ana wali-ˇc
.
a [Spanish vale ‘it is worth’; Aymara
you quinoa eat-SN not good-DV wali ‘good’]
‘The quinoa you have eaten is not good.’
(Cerr´on-Palomino, MS)
A complement clause without subject extraction is illustrated in (364):
(364) am-ki wer ˇc
h
iˇswi lul-ˇci ¸si¸s-ˇc
.
a
you-TO I meat eat-SN know-DV
‘You know that I have eaten meat.’
(Cerr´on-Palomino, MS)
Relative or complement clauses containing the nominaliser -ˇci,asexemplified in
Cerr´on-Palomino (MS), follow the pattern of main clauses in that their subjects are not
162
The source does not offer any morpheme glosses for these forms.
3.6 The Uru–Chipaya languages 373
marked for case. If the nominaliser is -ta (compare Aymara -ta) the agent of the clause
can take (possessive) genitive case marking:
(365) we-t qam-ta q
h
oya qac-ˇci-ˇc
.
a
I-G live-SN house be.lost-PA.3.MS-DV
‘The house where I used to live got lost.’
(Cerr´on-Palomino, MS)
Chipaya has a special nominaliser -i expressing the action goal of verbs of motion (a
function which in Aymara and Quechua is fulfilled by the agentive nominalisation). Less
immediate goals can be expressed by the combination -¸s-xapa (infinitive + benefactive
case).
(366) wer-naka lul-i-l oq-a-ˇc
.
a
I-PL eat-GO-1 go-F-DV
‘We will go and eat.’
(Cerr´on-Palomino, MS)
(367) wer-ki laŋŋ-¸s-xapa t
h
on-ˇcin-ˇc
.
a
I-TO work-IF-B come-PA.1-DV
‘I have come in order to work.’
(Cerr´on-Palomino, MS)
Chipaya has an active switch-reference system operating in subordinate clauses. The
endings used for this purpose have a double function. They specify the identity of
the subject of a subordinate verb in relation to the verb to which it is subordinated
(same or different), and they contain an indication of the temporal relationship of the
two verbs (simultaneous or consecutive). The following examples illustrate the use of
switch-reference with simultaneous events; the endings are -kan for same, and -nan for
different subjects. It has not been determined yet whether or not the element -¸s- in (368)
is to be seen as a part of the switch-reference marker. We assume that in (369) the element
-ˇs
.
corefers with the third-person subject of the main clause.
(368) ni:-ki ˇc
h
iˇswi lul-¸s-kan tik-ˇci-ˇc
.
a
he-TO meat eat-?-SM.SS die-PA.3.MS-DV
‘He died while he was eating meat.’
(Cerr´on-Palomino, MS)
(369) wer-ˇs
.
u:ˇsa kon-nan we-t hila-ki t
h
on-ˇci-ˇc
.
a [Aymara hila ‘brother]
I-3.MS sheep kill-SM.DS I-G brother-TO come-PA.3.MS-DV
‘When I was killing the sheep, my brother arrived.’
(Cerr´on-Palomino, MS)
Example (370) illustrates the use of switch-reference (different subjects) between
verbs referring to consecutive events. The ending -tan indicates that the subordinate
event is previous to the one expressed by the main verb.