Adelaar Willem, Muysken Pieter. The languages of the Andes

Подождите немного. Документ загружается.

424 4 Languages of the eastern slopes

Table 4.2 Yanesha



(Amuesha) phoneme inventory (based on Fast 1953 and Duff-Tripp

1998)

∗

Labial Palatal labial Dental Palatal Retroﬂex Velar Palatal velar

Voiceless stops p p

Voiced stops b b

Affricates c ˇcˇc

Voiceless fricatives s ˇsx

Voiced fricatives ˇz

Nasals m m

Lateral l

Vibrant r

Glides w y

Front Central Back

Mid e (ə)o

Low a

∗

Yanesha



vowels can be plain, long (a:, e:, o:), aspirated (a

, e

, o

), or glottalised (a

, e

, o

stresses the highly verb-centred character of the Arawakan clause, and surmises that

the original word order of the family was SOV, while now various orders are found,

including for the area under study SOV, VSO and OVS.

However, the special status of Yanesha



, also called Amuesha (Wise 1976), as a lan-

guage strongly inﬂuenced by Quechua, warrants a more detailed presentation. This case

shows how far-reaching Quechua inﬂuence has been on some neighbouring monta˜na

languages. Not only has Yanesha



adopted a great many Quechua loans, including some

core vocabulary, and the numerals from ‘six’ to ‘nine’, but it has also developed, for all

practical purposes, a three-vowel system like Quechua instead of the four-vowel system

of related Arawak varieties such as Campa. Furthermore, it has adopted a retroﬂex palatal

affricate ˇc

, according to Wise, from central Peruvian Quechua varieties such as Jun´ın,

Pasco or Hu´anuco. Finally, it does not allow the vowel sequences common in related

Arawak varieties but impossible in Quechua. The phonological changes in Yanesha



had

made it difﬁcult for earlier researchers to classify the language as Arawak (Wise 1976).

4.2.1 Yanesha



phonology

The consonant inventory of Yanesha



is given in table 4.2. It is based on Duff-Tripp

(1998), which builds on Fast (1953) and Duff (1957). These two sources present a three-

vowel system for the language, which is expanded to four vowels in Duff-Tripp (1998).

The Yanesha



texts available indicate that the overwhelming majority of words involve

4.2 The Arawakan languages 425

the three vowels a, e, o. These vowels occur in different phonation types: they can be

plain, long, aspirated, or glottalised.

A central feature of the consonant inventory is the

systematic opposition of palatal and non-palatal stops.

Aword series as in (9) illustrates vowel length and glottalisation, as well as the

opposition between palatal, alveolar and retroﬂex affricates:

(9) ˇcop ‘it grows’

ˇc

o:p ‘corn’

o:p ‘an insect’

ʔʔ

‘ﬁre’

ʔʔ

‘grandfather’ (Fast 1953: 192)

4.2.2 The principal grammatical features of Yanesha



Nominal elements

There is no grammaticalised gender in Yanesha



,but a number of kinship terms are

indeed differentiated for natural gender. Words for brother and sister and son and

daughter are also differentiated for the gender of Ego (po- ∼ p

e- is a possessive

preﬁx; the variation is dialectal; some nouns take pa

- or po

- instead; cf. Duff-Tripp

1997: 144):

(10) a. pa

ʔʔ

-mo

ʔʔ

naˇsen

‘his brother’

p-o

ʔʔ

ˇc

‘his sister’

ʔʔ

-se ‘her brother’

ʔʔ

-mo

ʔʔ

neˇz

‘her sister’

b. po-ˇcemer, p

e-ˇcemer ‘his son, daughter’

po-ˇcoyor, p

e-ˇcoyor ‘her son, daughter’

ʔʔ

-sen

‘his daughter’, ‘his/her niece (by sibling of

same sex)’

(Duff-Tripp 1997: 25, 31, 33)

There are compounds with the nucleus at the end:

(11) a. ˇson

ker-(e)po

worm-time

‘season of the worms’

b. aˇso

ʔʔ

ˇs-esme

ʔʔ

armadillo-tail

‘armadillo tail’

(Duff-Tripp 1997: 25)

Vowel length and aspiration are not systematically indicated in Duff-Tripp (1997, 1998).

426 4 Languages of the eastern slopes

Some compounds have a non-nominal initial element:

(12) a. ataˇz

-po

ʔʔ

big-house

‘big house’

b. anoˇz

-p

sit-place

‘a place to sit’

(Duff-Tripp 1997: 38)

Personal reference

Preﬁxal person marking in Yanesha



can refer to possessors, objects, and subjects. The

paradigm is as in (13) (Duff-Tripp 1997: 29):

(13) ne- 1.SG ye- 1.PL

pe- 2.SG se- 2.PL

po/p

e- 3.SG po-/p

e- .. -et 3.PL

Notice that this system has been considerably simpliﬁed when compared to the

Proto-Arawakan system reconstructed by David Payne:

(14)

∗

nu- 1.SG

∗

wa- 1.PL

∗

p- 2.SG

∗

hi- 2.PL

∗

l- 3.SG.MS

∗

na- 3.PL

∗

u- 3.SG.FE (David Payne 1991a: 376)

When the preﬁx is a possessor, there may be sufﬁxal person marking on the enclitic

element ‘to be’.

(15) no-pakl

-o

ʔʔ

-ce:n-en

1P.SG-house-L-be-1S.SG

‘I am/was in my house.’

(Wise 1986: 571)

The paradigm for this sufﬁxal marking is as follows:

(16) -Vn 1.SG -Vy 1.PL

-Vp

/-Vp 2.SG -Vs 2.PL

∅∅ 3.SG -et 3.PL

(Duff-Tripp 1997: 41)

The nature of the vowel in these sufﬁxes is determined by phonological characteristics of

the base word. Notice the close link between these sufﬁxes and the preﬁx paradigm

in (13).

When the possessor is not the original possessor of a noun, double preﬁxal marking

can occur:

4.2 The Arawakan languages 427

(17) no-pa

ʔʔ

-sme

ʔʔ

-o:r

1P.SG-3P.SG-tail-RL

‘my tail (which I have obtained from some animal)’

(Wise 1986: 575)

With respect to possession, four classes of nouns must be distinguished (Duff-Tripp

1997: 30–5):

(18) a. possessive marking obligatory (basic kinship terms such as ‘sister’

and ‘daughter’ and elements that always form part of a whole):

e-setma

ʔʔ

3P.SG-horizontal.beam

‘the horizontal beam of a house’

(Duff-Tripp 1997: 31)

b. possessive marker normally present (body parts, kinship terms

such as ‘husband’ and ‘wife’, abstract nouns, normally possessed

elements). These elements carry the privative marker -c/-Vc when

the possessor is not speciﬁed:

p-on

-ec

3P.SG-head head-PI

‘her/his head’ ‘head’

(Duff-Tripp 1997: 31)

ʔʔ

-n

-ec

3P.SG-word word-PI

‘her/his word’ ‘word’

(Duff-Tripp 1998: 603)

c. possessive marker optional (many objects, animals, kinship terms

such as ‘mother’ and ‘father’):

p-oˇck-ar oˇcek

3P.SG-dog-RL dog

‘her/his dog’ ‘dog’

(Duff-Tripp 1998: 254)

d. possessive marker impossible (a small class including ˇceˇsa

‘child’

and wokˇcaneˇsa

‘orphan’)

This classiﬁcation of four classes of nouns intersects with one concerning the form

of the possessive marking. Most inanimate objects, all body parts and some of the

kinship terms for children and siblings just receive a possessive preﬁx. However, an-

imate beings, most kinship terms, words for food and some nominalisations receive

the possessive preﬁx as well as a sufﬁx -Vr, illustrated with p-oˇck-ar ‘her/his dog’

in (18c).

With verbs, the person preﬁx normally marks the subject (cf. the examples below).

In certain circumstances, e.g. in relative clauses in which the subject is unspeciﬁed, the

428 4 Languages of the eastern slopes

preﬁxal person marking can refer to the object:

(19) n

ye-co:y-o

ʔʔ

t-amp-e:n-e

ʔʔ

which 1O.PL-light-EU-DA-PR-UN

‘which lights us (the sun)’

(Wise 1986: 571)

Marking of grammatical relations and adpositions

Objects and subjects are not marked overtly, but there is a generalised relational (locative,

genitive, instrumental) case marked -o,which can be further speciﬁed by an additional

sufﬁx or sufﬁx combination:

(20) non

-o locative ‘in/to the canoe’

non

-o-t

ablative ‘from the canoe’

non

-o-t

-en

‘from the canoe onward’

non

-o-t

-eye

ʔʔ

‘a bit beyond the canoe’

non

-o-net

allative ‘towards the canoe’

non

-o-wa delimitative ‘until the canoe’

non

-o-

ʔʔ

mar approximate location ‘near the canoe’

(Duff-Tripp 1997: 35–6)

There is no set of postpositions, with the exception of a marker -Vkop for benefactive

and purpose:

(21) ˇceˇsa-neˇsa-t

-eˇc

n-okop

child-group-DI-various-B

‘for all various classes of small children’

(Duff-Tripp 1997: 48)

Many spatial distinctions are marked with (possibly internally complex) adverbs,

such as:

(22) al

‘there’ aˇz

‘here’

ʔʔ

yo ‘there (distant)’ t

-aˇz

-o ‘there (not so distant)’

yom-ta

ʔʔ

n ‘on the other side’ an

-ta

ʔʔ

n ‘on this side’

-ap

ar ‘far’ (Duff-Tripp 1997: 126)

Verbs

An example of a highly complex verb in Yanesha



is given in (23):

(23) ∅∅-omaˇz

-am

-e

ʔʔ

t-amp

-es-y-e:s-n-e:n-a

3S-go.downriver-DB-EU-DA-EU-PL-EU-late-PR-RF

‘They are going downriver by canoe in the late afternoon stopping

often along the way.’

(Wise 1986: 582)

4.2 The Arawakan languages 429

Striking is the frequent occurrence of euphonic markers, which also appear to have a

grammatical function. Note also the use of a tense marker indicating ‘late afternoon’; it

contrasts with a marker indicating ‘early morning’.

The element -amp

-, glossed ‘dative’ by Wise, deserves special mention. It can add a

slot for an affected oblique object to an intransitive:

(24) a. ne-ma

t-a

1S.SG-run-RF

‘I run.’

(Wise 1986: 592)

b. ne-ma

t-amp

-s-ap

-a

1S.SG-run-DA-EU-2O.SG-RF

‘I run away from you.’

(Wise 1986: 592)

Sometimes this function of -amp

- leads to a passive-like structure, as in (25b):

(25) a. ∅∅-wa

t-a

3S-rain-RF

‘It rained.’

(Wise 1986: 593)

b. no-wa

t-amp

-s-a

1S.SG-rain-DA-EU-RF

‘I was caught in the rain.’

(Wise 1986: 593)

The rich sufﬁxal verbal morphology expresses a wide range of concepts. One example

is subjunctive mood:

(26) pe-ˇso:r-a

ʔʔ

n-m

-e:pa

ʔʔ

2S.SG-fall-AB-CM-SJ

‘Be careful lest you fall.’

(Wise 1986: 602)

Another one involves reciprocal marking:

(27) ∅∅-y-a

ʔʔ

n-a

ʔʔ

t-ann-a

ʔʔ

t-a

3S-cry-RR-EU-RC-EU-RF

‘They cry for one another.’

(Wise 1986: 579)

Wo rd order

Basic word order in main clauses is VSO:

(28) o

ʔʔ

ˇz

-at-eˇz

-ey aˇc

kaˇs

y-aˇc-or

already eat-CA-RP-1S.PL dwarfs 1P.PL-mother-RL

‘The dwarfs have already fed our mother.’

(Duff-Tripp 1997: 177)

Possibly a borrowing from neighbouring Quechua I dialects, where aˇc

kaˇs means ‘lamb’.

430 4 Languages of the eastern slopes

However, often a topic marked with -pa

(either subject or object) occurs preverbally:

(29) pa

ʔʔ

-tak-pa

ʔʔ

awo

ʔʔ

pamp

-en

-et-a

ʔʔ

-ˇcoy-o

3P-foot-TO HS bury-PR-3S.PL-AN 3P-ﬁeld-L

‘One foot (of his) they buried in the ﬁeld.’

(Duff-Tripp 1997: 179)

Auxiliary elements precede the verb, while oblique complements tend to follow it:

(30) aw-o

ʔʔ

∅∅-aw-an-m

-e:t ent-o

AX-HS 3S-go-AB-CM-3S.PL.RF sky-L

‘Then they went up to the sky.’

(Wise 1986: 605)

Modiﬁers such as possessors precede the head noun:

(31) a:ˇcp

-e

ʔʔ

m-a:r

mother 3P-baby-RL

‘mother’s baby’

(Wise 1986: 607)

4.2.3 Complex sentences in Yanesha



Adverbial clauses generally precede the main clause and are marked with -pa

(32) ama ye-ˇz

-en-a

ʔʔ

-pa

ʔʔ

ˇc y-e

ʔʔ

pota

ʔʔ

t-a [cf. Quechua ama ‘don’t’]

not 1S.PL-eat-PR-AN-TO F 1S.PL-wash.hands-RF

‘Before eating we wash our hands.’

(Duff-Tripp 1997: 191)

In relative clauses a relativiser n

appears at the beginning of the clause, before the

verb, and the clause is marked again with the topic marker -pa

(33) an

-n

apa

ʔʔ

-tak, n

∅∅-pamp

-en-et-a

ʔʔ

-ˇcoy-o-pa

ʔʔ

,awo

ʔʔ

anap

this-SQ 3P-foot, R 3S-bury-PR-3.PL-AN 3P-ﬁeld-L-TO,

HS 3S.answer.3O

‘The foot they buried in the ﬁeld answered.’

(Duff-Tripp 1997: 206)

Notice that the clause follows the antecedent.

4.3 Tupi–Guaran´ı

The very important Tupi–Guaran´ı language family is mostly spoken in the area south

of the Amazon, in Bolivia, Brazil, Paraguay and Peru. In Bolivia, it is represented by

Chiriguano–Ava, Chiriguano–Tapyi or Izoce˜no, Guarayo, Jor´a, Pauserna–Guarasugw´e,

Sirion´o and Yuqui. It also includes the Chan´e, a former Arawakan group which now

speaks Chiriguano, and the Tapiet´e. The Tapiet´e are claimed by Nordenski¨old (1912: 310)

to be originally a Mataco-related group. Some of the other Tupi–Guaran´ıgroups are small

(the nomadic Sirion´o) or nearing extinction (the Jor´a and the Pauserna–Guarasugw´e).

4.3 Tupi–Guaran´ı 431

Table 4.3 The relationship among the pre-Andean

members of the Tupi–Guaran´ı language family

(based on Rodrigues 1984–5)

I Chiriguano–Ava (including Tapiet´e)

Chiriguano–Tapyi or Izoce˜no

Guaran´ı

II Guarayo

Pauserna–Guarasugw´e

∗

Sirion´o, Jor´a

Yuqui

∗

III Cocama, Cocamilla

Omagua

∗

Pauserna–Guarasugw´e and Yuqui are not classiﬁed in

Rodrigues (1984–5).

The Chiriguano have always been a militarily powerful group; they invaded the Andean

territories proper both during the Inca period and in colonial times.

Tupi–Guaran´ı languages are found in Peru along the Amazon and Ucayali rivers.

There are two languages, Cocama, including Cocamilla, and Omagua. A few speakers

of Cocama can be found in Brazil as well. Early explorers of the Amazon such as Orellana

marvelled at the riches of the Omagua city-like settlements along the rivers. Cocama is

giving way to Spanish at the present moment, and the people are now often acculturated

and live dispersed. The Cocama were ﬁrst contacted in 1559, and the Cocamilla much

later, in 1651. In 1666 the two groups rebelled together against the Spaniards (Chirif

and Mora 1977).

The relationship among the Bolivian and Peruvian members of this extensive lan-

guage family (found also in Brazil and Paraguay) has been the subject of some disagree-

ment, due to different classiﬁcatory techniques. Firestone (1965) has applied lexico-

statistic techniques, and Lemle (1971) has traced some phonological innovations, while

Rodrigues (1984/5) has combined several types of diagnostic criteria, and Dietrich

(1990) has done a sophisticated analysis of the number of phonological and morpholog-

ical features separating these languages. The groupings proposed by Dietrich (1990), as

a consequence, not only reﬂect genetic relationships but also degree of innovation. Here

we follow Jensen (1998), a comparative reconstruction of Tupi–Guaran´ı morphosyntax,

in taking Rodrigues (1984–5) as our point of departure. The relevant branches of the

Tupi–Guaran´ıfamily are given in table 4.3.

While the family as a whole has preserved many of its original features relatively

intact, a number of languages have undergone considerable changes. Sirion´o, Jor´a and,

presumably, Yuqui have been much simpliﬁed, possibly because these languages are the

432 4 Languages of the eastern slopes

result of the adoption of Tupi–Guaran´ıasasecond language by people who originally

spoke something else (Rodrigues 1984–5: 43). The nature of these simpliﬁcation pro-

cesses remains to be studied, however. In any case, the present-day differences between

Guarayo and the other members of its group are much larger than their joint classiﬁcation

would suggest (Dietrich 1990: 111).

The case of Cocama also is of great interest from the perspective of language history

and language contact. While phonologically and lexically it is undoubtedly a Tup´ıan

language, closely related to Tupinamb´a from Brazil, structurally it is not (Rodrigues

1984–5: 43–4; Cabral 1995, 2000). It has different pronouns when spoken by the two

genders, and in several cases the feminine form is of Tupi origin (bold here), while the

masculine is not (the reverse never happens):

(34) 1.SG FE = ´ece MS = ta

2.SG ´ene

3.SG FE = ´ay MS = ´uri

1.IN ´ıni

1.EX FE = p´enu MS = t´ana

2.PL ´epe

(Cabral 1995: 132)

Rodrigues (personal communication) suggests that the language resulted from a fairly

recent migration of a group speaking a Tupinamb´a-related language from the lower

to the upper Amazon, perhaps in the immediate preconquest era, and its subsequent

incorporation in a different language group.

4.4 The Jivaroan languages

Jivaroan is represented by four languages: Aguaruna, Huambisa, Achuar (or Achual)

and J´ıvaro or Shuar. All four languages are closely related. Jivaroan speakers show a

particularly strong ethnic consciousness. The Aguaruna territory in Peru is quite ex-

tensive and covers parts of the departments of Loreto and Amazonas. The Huambisa

are mainly in Loreto. The Shuar (also known as the J´ıvaro) occupy the eastern slopes

of the Ecuadorian Andes, mainly in the province of Morona–Santiago, up to the very

limits of the Andean ridge proper. The Sangay volcano is their sacred mountain. To the

east of the Shuar live the Achuar (partly also in Loreto, Peru). Thus the southern part

of the Ecuadorian oriente is dominated by a single family of closely related languages,

Jivaroan. It may be that previously the Jivaroan area extended into the highlands of the

Ecuadorian province of Loja. Like their Peruvian relatives the Aguaruna, the Ecuadorian

Shuar are known for their high level of political organisation.

See, however, Schleicher (1998) for an explanation of the Cocama ﬁrst person-singular preﬁx -ta

within the context of Tupi–Guaran´ı historical development.

4.4 The Jivaroan languages 433

The Jivaroan groups have a long and complex history of interactions with the highland

cultures. The Shuar are renowned for having raided Andean territories as late as the

colonial period. They were ﬁrst contacted by the Spanish early on, in 1534, and rebelled

succesfully in 1599. In 1816 a new attempt was made to subdue them. The Aguaruna were

attacked by the Incas upon several occasions, and after 1549 the Spaniards attempted to

bring them together in reducciones.In1886 they rebelled against the Jesuits (Chirif and

Mora 1977). The Huambisa have managed to stay further apart from the Spanish, due

to their location further into the interior.

As mentioned above, Shuar is one of the most important languages of the Ecuado-

rian lowlands. Part of the Jivaroan language family, it is closely related to and mutually

intelligible with Achuar and slightly more distantly related to and partly intelligible with

Aguaruna and Huambisa. While the Shuar are very well known, under the name J´ıvaro,

and have been the subject of extensive ethnographic studies, particularly by Karsten

(1935) and Harner (1972), there is surprisingly little in terms of full linguistic descrip-

tions. The sketch here takes the brief account presented by Karsten (1935) as its basis,

and includes some of the data presented in Beuchat and Rivet (1909, 1910a), Alvarez

([c.1915] 1983), Flornoy (1938), Turner (1958), Pellizzaro (1969), Juank (1982), Rouby

and Riedmayer (1983) and Gnerre (1986). It has not been possible to completely ho-

mogenise the spelling from the different sources, since it is not always clear whether the

differences arise from orthographic conventions, phonetic versus phonological spelling,

or from regional or diachronic variation. Finally, there is the problem that in Macas a

pidgin version of Shuar was in use by and with mestizo settlers (Gnerre 1986: 340),

and it may be that Karsten’s account is inﬂuenced by this pidgin Shuar. Gnerre (1986:

309–11) draws attention to the wide variety of registers that were available in traditional

Shuar, including ceremonial visiting dialogues, ceremonial war party dialogues and tra-

ditional narrative styles. In the course of the twentieth century, these styles have largely

disappeared; it would be interesting to study the features of recently developed styles,

such as the Shuar used in radio transmissions.

4.4.1 Shuar phonology

The major source on Shuar phonology is Turner (1958). He presents the phoneme in-

ventory given in table 4.4. There is a nasal–oral contrast in the vowel system, but most

of the consulted sources indicate nasality inconsistently or not at all. Sequences of same

vowels occur, some of which may be due to the loss of an intervocalic consonant (e.g.

in waakis ‘Gualaquiza’). Karsten (1935) sometimes indicates vowel length, where other

sources have plain vowels. Most sources use the symbols e or i for the central vowel



established by Turner (1958). There is also a vowel [e] which in most contexts is best

analysed as underlyingly /a/. Thus we have: