Adelaar Willem, Muysken Pieter. The languages of the Andes

Подождите немного. Документ загружается.

204 3 The Inca Sphere

The palatal resonants *l

and *n

have been subject to depalatalisation (to l and n,

respectively) in a number of Quechua I dialects. Surprisingly, the two changes do not

always coincide. Ancash Quechua, for instance, maintains palatal l

but has changed *n

to n,whereas the opposite obtains in Pacaraos Quechua. Other dialects, such as Tarma

Quechua, have undergone both depalatalisations. As in the case of the sibilants, there

appears to be a close relationship between the historical process of depalatalisation and

the rise of sound-symbolic mechanisms. In Tarma Quechua, the palatal resonants are

now used as the hypochoristic counterparts of the non-palatals; an interesting example

is (27):

(27) *l

u > l

anu ‘thin’

> l

u ‘very thin’

If the sound changes were automatic, both consonants should have been de-

palatalised. In this case, however, one depalatalisation sufﬁces to open the possibility

of a hypochoristic back-formation (l

u ‘very thin’). Whether or not *l

remained

unmodiﬁed for this reason remains a matter of speculation, but the example certainly

illustrates the fact that depalatalisation is far from being a change without exceptions.

Furthermore, the impact of depalatalisation must not be overestimated, because minimal

pairs involving a palatal and a non-palatal nasal were rare in Proto-Quechua, whereas

non-palatal *l was conﬁned to just a few lexical items, if it occurred at all.

In the Argentinian Santiago del Estero dialect the palatal lateral *l

has become

avoiced alveopalatal fricative [ˇz]. The same has occurred in many Ecuadorian and

northern Peruvian dialects, where either the fricative or an affricate [d

ˇz

]isfound. Both

in Argentina and in Ecuador these changes illustrate an incipient sprachbund as they

seem related to parallel developments in the local Spanish.

(28) *l

aki ‘sorrow’ > (Imbabura) ˇzaki

> (Cajamarca) d

ˇz

aki

(29) *al

qu ‘dog’ > (Cotopaxi) aˇsku

> (Cajamarca) ad

ˇz

A rather unusual development of *l

is attested near Concepci´on in the depart-

ment of Jun´ın. There the palatal lateral became a voiceless alveopalatal affricate ˇc

(Cerr´on-Palomino 1989a: 42–3, 60–1). A similar development was observed in Salasaca,

in Ecuador (see section 3.2.8).

(30) *qil

ay ‘money’ > (Concepci´on) iˇcay

(Huancayo) il

All Quechua dialects have an alveodental ﬂap r.Inaddition, most dialects have a

slightly affricated retroﬂex vibrant ˇr,which is regularly used to represent the Spanish

3.2 The Quechuan language family 205

rolled r or rr in loan words. This sound may also function as a distributional variant

of the ﬂap r in native words. This is the case in Ayacucho Quechua word-ﬁnally (e.g.

in yawar [yawaˇr] ‘blood’), and in San Pedro de Cajas (northern Jun´ın) Quechua word-

initially, e.g. in rumi [ˇrumi] ‘stone’. Bolivian Quechua, like several varieties of Bolivian

Spanish, has a voiced alveodental fricative [z] in word-initial position, instead of the

retroﬂex, e.g. in rumi [zumi] ‘stone’. The distribution of ˇr in Ayacucho Quechua has led

to a marginal phonemic opposition between r and ˇr in cases like arpas (Spanish arpa)

‘harp’ and yawaˇr-pa ‘of the blood’ (Parker 1969a). The dialect of Pacaraos exhibits the

unusual case of a phonemic opposition between a ﬂap r and a trilled rr in word-initial

position (see section 3.2.9).

The most radical change affecting the vibrants is the lateralisation found in Quechua I

Huanca (Concepci´on, Huancayo and Jauja) and in some of the neighbouring Yauyos

dialects (Cacra, Hongos).

(31) *rimay ‘to speak’ > (Chongos Bajo) limay

(32) *qunqur ‘knee’ > (Chongos Bajo) un?ul

All Huanca dialects have acquired new r-like sounds, e.g. in yawaˇr ‘blood’. They are

probably due to borrowing from one of the neighbouring non-lateralising dialects, such

as Ayacucho. The *r > l change may have had a wider distribution in early colonial

times.

The nasal consonants *n and *m have been preserved without major modiﬁcations in

the modern Quechua dialects, except for Cuzco Quechua and Bolivian Quechua where

they are no longer distinguished in syllable-ﬁnal position. Traditionally, all nasal allo-

phones in syllable-ﬁnal position that are not pronounced [m] are assigned to a phoneme

/n/ and are written n regardless of the differences in pronunciation. Before a word bound-

ary, before glottal stop (in the Huanca dialects), before resonants, and, in some dialects,

before fricatives as well, a velar allophone [

ŋ] occurs. Elsewhere, the articulation point

of a syllable-ﬁnal nasal is assimilated to that of the following consonant. The treatment

of [m] before a labial stop is not uniform. Some authors write m, e.g. ˇcampa ‘piece of

grass-cover’, wasi-m-pa ‘of his house’. Others write m within a root and n at morpheme

boundaries; still others write n in both cases. In most dialects, the labial character of m is

lost before a labial resonant (m, w), e.g. Ayacucho Quechua qam [x

am] ‘you’, but qan-

man [x

ŋmaŋ] ‘to you’. The dialects of the Tarma region (northern Jun´ın) have developed

The name of Lima, the Peruvian capital, is a case in point. It was derived from the expression

*rimaq ‘the one who speaks’, ‘an oracle’. The name of the river Rimac, which ﬂows through the

city, reﬂects the pronunciation of a more conservative dialect. Although the old town of Lima

was situated on the coastal plain, its name may have become known in the pronunciation of the

Jauja dialect, where it would have been approximately [limax].

206 3 The Inca Sphere

a marginal distinction between alveodental [n] and velar [

ŋ] before the glides y and w,

e.g. [wanyay] ‘to use guano’ (Andean Spanish guanear), [wa

ŋyay] ‘to strike a blow’.

Proto-Quechua and most modern dialects exhibit a straightforward syllable-structure

of the CV(C) type, with the sole exception that word-initial syllables are (C)V(C). Long

vowels are structurally equivalent to a VC sequence. Quechua has few constraints on

the occurrence of consonant clusters at syllable boundaries. This situation is particularly

evident in the conservative central Peruvian Quechua I dialects, where over a hundred

combinations are permitted. As a result of consonant lenition in syllable-ﬁnal position,

the number of combinations has been reduced substantially in Cuzco Quechua and in

Bolivian Quechua.

Word-ﬁnal consonant clusters and medial clusters of more than two consonants are not

allowed. A connective element -ni- serves the purpose of avoiding impossible consonant

sequences in nominal morphology (see the discussion of the possessive markers in

section 3.2.6 for an example). Its use, however, is not entirely restricted to nouns, and it

appears to have played a role within the history of the verbal personal reference system

as well. The connective -ni-isalso used after long vowels, as in the following example

from Chup´an in northern Jun´ın, where it separates recurrent instances of the ﬁrst-person

subject marker -



: in a verb form:

(33) tarpu-pa:ku-:-n´ı-:

sow-PL-1S-EU-1S

‘We (exclusive) sow.’

In most Quechua dialects stress is assigned to the penultimate vowel of each polysyl-

labic word form and thus provides an easy phonology-based criterion for word delimita-

tion. Exceptionally, in interjections and emphatic expressions, stress can be word-ﬁnal,

e.g. in alal´aw ‘how cold!’. In Tarma and the dialect of northern Jun´ın word-ﬁnal stress

can occur as a result of the elision of a ﬁnal syllable.

(34) aywa-mu-ra-ygi ∼ aywa-mu-r´a-y

go-H-PA-2S

‘You came.’

In the Quechua I dialects stress is maintained in word-ﬁnal position when the long

vowel morpheme referring to ﬁrst-person subject or possessor is immediately followed

by aword boundary. (Length oppositions not connected with the ﬁrst-person morpheme

are normally neutralised in word-ﬁnal position, in which case stress is regularly assigned

to the penultimate syllable.) Example (35) illustrates the use of a ﬁrst-person marker in

Huanca; cf. also (33).

(35) w´asi ‘house’

was´ı-: ‘my house’

3.2 The Quechuan language family 207

In Pacaraos Quechua (see section 3.2.9) stress is phonemic. It may lie either on the

penultimate or on the ﬁnal syllable. Word-ﬁnal stress in Pacaraos is either due to the

elision of a ﬁnal syllable, for instance, in akˇs´u-k, from akˇs´u-kta (potato-AC), or to a

word-ﬁnal occurrence of the ﬁrst-person marker -´y. Otherwise, penultimate stress is

predominant.

More complex, non-penultimate stress patterns, involving word-initial stress and

stress dependent on syllable structure, are found in QI Ancash and in QI Cajatambo

(cf. Torero 1964: 461; Parker 1976). The northern Peruvian dialects QIIA Ferre˜nafe

(Escribens 1977) and QIIB Amazonas (Taylor 1975; Chaparro 1985) also have deviant

stress patterns, as had the sixteenth-century coastal dialect described by Santo Tom´as.

For the dialect of Huaraz (department of Ancash) the following stress pattern is de-

scribed (Torero 1964). Syllables can be either long (CVC, CV:) or short (CV). If a word

contains a long syllable which is not word-ﬁnal, the latter is stressed. If there are several

non-ﬁnal long syllables, either the last one, or the one which has a long vowel can be

stressed, or stress can be distributed over several long syllables. If there are no non-ﬁnal

long syllables, the ﬁrst syllable is stressed or the ﬁnal syllable when it contains a long

vowel.

3.2.6 Grammar

Quechua is essentially agglutinative. Its morphological structure is almost entirely based

on the use of sufﬁxes and is extremely regular. Vocalic alternations, including length

and quality alternations, occur to a limited extent in part of the dialects. There are no

preﬁxes

and compounds are exceptional.

Syntactic constructions in Quechua are basically head-ﬁnal (except for relative clauses

which may follow their antecedents). Constituent order in main clauses is basically free,

but there is a certain preference for the order subject–object–verb, which is required

in dependent clauses. Possessive relations are indicated both on the head and on the

dependent nominal. Case relations are marked by special sufﬁxes which are attached at

the end of the noun phrase. Quechua is a nominative–accusative language. Nominative

case is not marked.

Equations and other constructions involving a nominal predicate contain the copula

verb ka- ‘to be’, which in other contexts has the meaning ‘to exist’. In most dialects,

the third-person copula ka-n ‘he/she/it is’ is left out so long as it is not needed for

Whether or not an Ancash-type accentuation pattern reﬂects that of Proto-Quechua is a matter

of debate. Against this view, advocated by Torero (1964), we may argue that the essentially

penultimate stress of Pacaraos Quechua provides a plausible explanation for the development

of the ﬁrst-person markers in both branches of the Quechua linguistic family, hence it is likely

to have been inherited from Proto-Quechua.

Ecuadorian Quechua has an element ila- ‘step-’, which could be interpreted as a preﬁx.

208 3 The Inca Sphere

carrying other sufﬁxes. There is no straightforward passive, nor is there an overt distinc-

tion between transitive and intransitive verbs. Special verbal endings, nominalisation,

case marking and the so-called independent or sentential sufﬁxes (see below) take the

place of conjunctions, which are virtually non-existent in the language, unless they are

borrowed from Spanish. However, some deictic expressions may function as sentence

organisers in conditional and correlative constructions. They are also used to indicate the

relation between sentences in a discourse, e.g. ˇcay-qa (that-TO) ‘but’; ˇcay-mi (that-AF)

‘therefore’, ‘so’.

Adjectives are similar to nouns in their syntactic behaviour. It is not always easy to

distinguish between the two categories. In an example such as (Ayacucho Quechua) rumi

wasi ‘stone house’, rumi can be interpreted either as a noun ‘stone’, or as an adjective

‘made of stone’. The main criterion for establishing the difference is that a noun can func-

tion by itself as the subject in a sentence, whereas real adjectives can only act as subjects

when followed by an element that indicates their status as an independent item; an ele-

ment frequently so used is ka-q ‘(the one) that is’, e.g. in hatun ka-q ‘the (a) big one’.

The original morphological complexity of Quechua has been preserved remarkably

well in most of the dialects (including dialects which are on the verge of extinction,

such as Pacaraos Quechua). As an exception to this, the Ecuadorian and Colombian

dialects (including the Ecuadorian and northern Peruvian jungle dialects) exhibit a rather

simpliﬁed morphology. For instance, the possessive markers on nouns are no longer used

there. Compare (36) from Ecuador to (37) from Ayacucho, Cuzco or Huancayo.

(36) kam-pax wasi

you-G house

‘your house’

(37) wasi-ki

house-2P

‘your house’

From a formal point of view, most Quechua sufﬁxes are easy to identify. There exist

afew portmanteau sufﬁxes occurring at the end of a verb form, which refer to speciﬁc

combinations of tense, mood and person. Personal reference marking in verbs involves

two speech-act participants, the subject and a direct or indirect object. The object par-

ticipant must be human. Some combinations of subject and object reference in verbs

involve the use of sufﬁxes that have different meanings from those which they convey

when considered independently from those combinations.

Apart from the somewhat complex cases just mentioned, each instance of a sufﬁx

corresponds to a particular meaning. However, the exact semantic nature of the verbal

derivational sufﬁxes and their pragmatics are often difﬁcult to determine. Some sufﬁxes

have different meanings depending on the contexts in which they occur. Combinations

3.2 The Quechuan language family 209

of sufﬁxes may convey special meanings even in cases where the sufﬁxes involved are

not contiguous. Some sufﬁx meanings are clear and straightforward, but others are quite

subtle and can be mastered by a non-native speaker only with difﬁculty. The use of the

verbal derivational sufﬁxes exhibits a remarkable variety throughout the dialects. This

circumstance constitutes an obstacle for communication among Quechua speakers of

distinct areas and is one of the main reasons why the different Quechuan varieties are

usually treated as linguistic entities in their own right. The morphological transparency

which emanates from most descriptions of Quechua is no more than a matter of ap-

pearance. Short grammatical labels or one-line characterisations of the meaning of a

sufﬁx generally hide a complex reality. The order in which sufﬁxes occur in a verb

form is essentially ﬁxed, although more than one option may be available in some

parts of the sufﬁx inventory. Descriptive studies differ in the extent to which they suc-

ceed in capturing the rules that govern sufﬁx order in Quechua. For an admirably thor-

ough analysis of the different order options in a Quechua dialect (Ayacucho) see Parker

(1969a).

Morphology plays a very dominant role in Quechua. Several functions which in other

languages are assigned to intonation, to word order or to lexical expressions (function

words) are indicated by means of morphological markers in Quechua. There exists a

special set of afﬁxes that can be attached to verbal, as well as to nominal expressions.

They also occur with expressions which are neither verbal nor nominal. In the litera-

ture the members of this set are referred to as independent sufﬁxes, sentential sufﬁxes,

class-free sufﬁxes or enclitics. Independent sufﬁxes cannot occur in every position in

the sentence. With a few exceptions, subordinate clauses and noun phrases with a hi-

erarchically organised inner structure function in their totality as bases of attachment

for the independent sufﬁxes. The combination of a nominal predicate and the copula

verb ka- ‘to be’ can contain no more than a single locus for the attachment of in-

dependent sufﬁxes. In such constructions independent sufﬁxes are normally attached

to the nominal predicate even though the latter precedes the copula, as is very often

the case.

The functions of the independent sufﬁxes include data source, polar question marking

(-ˇcu)

, topic or contrast (-qa), notions such as ‘still’ or ‘ﬁrst’ (-raq), ‘already’ (-n

a),

‘even’, ‘too’ (-pas, -pis, -si) and ‘on the other hand’, ‘as you know well’ (-taq). In com-

bination with interrogative expressions (wh-words), the independent sufﬁxes acquire

special meanings. Negation is indicated by the combination of an independent sufﬁx

following the negated element and a lexical expression (mana in statements, ama in

The shape of the sufﬁxes listed here is a reconstructed form, which may emerge differently in

some dialects. Henceforth, this will be the case whenever there is no reference to a particular

dialect.

210 3 The Inca Sphere

exhortations) that occurs before it. In most dialects the independent sufﬁx used in nega-

tions (-ˇcu)isidentical to that which marks a polar question; in Tarma, for instance, we

ﬁnd:

(38) ali-ˇcu

good-IR

‘Is it good?’

(39) mana-m ali-ˇcu

not-AF good-NE

‘It is not good.’

In Huanca the independent sufﬁx marking a polar question has a special form, different

from the one used in negations (40).

(40) tanta-kta-ˇcun apa-mu-nki

bread-AC-IR carry-H-2S

‘Will you bring bread?’

(Cerr´on-Palomino 1976a: 232–3)

(41) mana-m tanta-kta-ˇcu

no-AF bread-AC-NE

‘No, not bread.’

(Cerr´on-Palomino 1976a: 232–3)

A similar situation obtains in Ancash Quechua where the interrogative sufﬁx is -ku,

as opposed to -cu (< *-ˇcu)innegative expressions (Parker 1976: 148–9).

The independent sufﬁxes that indicate data source are usually referred to in the lit-

erature as evidentials or validators. The data source system is primarily based on a

three-way distinction. The validity of the source from which the information was drawn,

either through personal witness, hearsay, or conjecture is consistently marked in most

declarative sentences. The contrast is illustrated in (42) in examples from northern

Jun´ın/Tarma.

(42) mana-m ali-ˇcu mana-ˇs ali-ˇcu mana-ˇc

ali-ˇcu

not-AF good-IR not-HS good-IR not-DU good-IR

‘It is not good (I know).’ ‘It is not good (I heard).’ ‘It is not good

(I guess).’

In most Quechua dialects the validators have syllabic, as well as non-syllabic allo-

morphs. The syllabic allomorphs (-mi;-ˇsi, -si; -ˇc

i, -ˇci,-ˇca

) occur after a consonant, a

diphthong, or a long vowel; the non-syllabic allomorphs (-m; -ˇs,-s; -ˇc

, -ˇc) occur after a

short vowel.

The forms with ﬁnal i occur in the Quechua I dialects except in Huanca. The a variant occurs in

the other dialects.

3.2 The Quechuan language family 211

Table 3.3 The Quechua four-person

system

Speaker Addressee

1 +−

2 −+

3 −−

4 ++

Validators tend to select the ﬁrst available unit of attachment in a sentence. However,

the presence of another independent sufﬁx, such as the topic marker -qa or the additive

marker -pas,-pis,-si

(‘too’, ‘even’), causes a validator to move to the next available

unit. It is also possible to mark out a particular constituent of the sentence as the answer to

a question by attaching a validator to it. For that purpose, the validator can move further

to the right. This combination of facts may lead to the impression that the conﬁguration

of a topic marker and a validator represents some sort of a topic–comment structure. This

impression is invalidated by examples such as the following sentence (from Ayacucho),

in which there is no topic marker; the validator remains attached to the ﬁrst constituent,

in this case, the topic:

(43) kay-mi qiqa

this-AF chalk

‘This is a (piece of) chalk.’

Nominal morphology comprises devices for case, number and personal reference

marking. Personal reference markers identify the possessor of a noun in relations of

genuine possession or in any other relation between two nominal entities envisaged as

such. A four-term personal reference system, similar to that which obtains in Aymara

(Hardman et al. 1988: 18; cf. also section 3.3.4), is found in the conservative central

Peruvian Quechua I dialects. It can be described in terms of the presence (or absence) of

a speaker and an addressee, the two main participants in the speech act (see table 3.3).

The fourth person refers to a group of people including both the speaker and the

addressee. In practice, it takes the place of a ﬁrst-person-plural inclusive as it is found in

other native American languages. By contrast, Quechua I ﬁrst-person markers may refer

either to the speaker, or to a group of people including the speaker but not the addressee.

In most Quechua I dialects, the ﬁrst-person-plural exclusive possessor with nouns is not

morphologically distinct from the ﬁrst-person-singular possessor. The distinction can be

The afﬁx -si is found in a number of Quechua I dialects, along with one or both of the other

forms which occur throughout the main branches of the Quechua family.

212 3 The Inca Sphere

made explicit by adding a lexical expression of the possessor, as in (44) and (45) from

northern Jun´ın:

(44) wayi-:

house-1P

‘my/our (exclusive) house’

(45) nuqa:-kuna-pa wayi-:

I-PL-G house-1P

‘our (exclusive) house’

In verbs the difference between a ﬁrst-person-plural exclusive and a ﬁrst-person sin-

gular is indicated by means of a verbal plural marker (see below).

The proto-forms underlying the Quechua second, third and fourth nominal person

markers are -(y)ki, -n and -nˇcik, respectively. They or their reﬂexes are found in all

present-day dialects, except those that no longer mark personal reference on nouns. By

contrast, the ﬁrst-person ending is variable (see section 3.2.3). Consonant-ﬁnal stems

are followed by the connective element -ni-inorder to permit the attachment of personal

reference markers, as in (46) from Ayacucho or northern Jun´ın.

(46) mikuy-ni-ki

food-EU-2P

‘your food’

In virtually all Quechua dialects noun plurality can be indicated by means of the

plurality marker -kuna. Although the marking of plurality is not an obligatory procedure

in Quechua, the sufﬁx -kuna can be used with any noun referring to a set of individualised

items. There is no distinction between count and non-count nouns.

(47) mikuy-kuna

food-PL

‘foodstuffs’, ‘sorts of food’, ‘livestock’

Along with the straightforward process of noun referent pluralisation here described,

most Quechua II dialects have developed (or maintained) a system of number marking

that is accessory to personal reference marking. The afﬁxes in question pluralise a

possessor, a verbal subject, or a verbal object. Both types of plurality (noun referent

plurality and personal reference plurality) are kept apart consistently.

The southern Quechua IIC branch exhibits the most elaborate system of personal

reference number marking. There are separate endings for second-person plurality

(-ˇcik,-ˇcis

) and for non-second-person plurality (ﬁrst- and third-person -ku). These

The endings in -s (-ˇcis and -nˇcis) are found in Quechua IIC, except Ayacucho. They may either

co-occur with reﬂexes of -ˇcik and -nˇcik,orbethe only option as is the case in Cuzco Quechua.

3.2 The Quechuan language family 213

sufﬁxes do not take the place of the regular person markers but follow them.

Two

plural markers may occur in a sequence, but combinations of -ku and -kuna are avoided.

Consider (48), (49), (50) from Ayacucho:

(48) wasi-ki-ˇcik-kuna

house-2P-PL.2-PL

‘your (plural) houses’

(49) wasi-y-ku

house-1P-PL.1/3

‘our (exclusive) house(s)’

(50) wasi-y-kuna

house-1P-PL

‘my houses’, ‘our (exclusive) houses’

In the northern Quechua II dialects (Cajamarca, Amazonas) personal reference num-

ber is marked differently. Reﬂexes of *-l

apa ‘all’ and *-sapa ‘each’ are found instead

of -ku and -ˇcik.InEcuadorian Highland Quechua the marker -kuna corresponds to both

-ku and -kuna of the southern Quechua II dialects.

Case marking plays an important role in Quechua nominal morphology. Case is

marked on the ﬁnal word of a noun phrase or a relative clause when preceded by its an-

tecedent. Only the subject, a nominal predicate and some adverbial expressions (mostly

referring to time) occur without a case marker. Furthermore, the accusative case marker

is optionally absent in the context of a nominalised verb. Case markers are limited in

number, more speciﬁc spatial relations being paraphrased by means of body-part nouns

or special positional nouns.

Most case-marking sufﬁxes vary in shape across the dialects, albeit not all in an

equally radical way. The accusative case marker indicates a direct object, an indirect

object or a deliberately selected geographic goal. Its shape is -ta,but some dialects

(Huanca, Pacaraos, colonial Cuzco Quechua) provide evidence of an allomorph -kta

after (short) vowels, which undoubtedly goes back to Proto-Quechua. In a similar way,

the genitive case marker is -pa in most dialects but is subject to allomorphic variation in

some conservative dialects (Huanca, Cuzco). Huanca -p and Cuzco Quechua -q,aswell

The fourth-person ending -nˇcik (-nˇcis) could be interpreted as a compound ending due to its

partial similarity with the pluraliser -ˇcik (-ˇcis). From a synchronic point of view, such an in-

terpretation is defendable for Quechua IIC (and for some Quechua IIB dialects). Historically,

however, -nˇcik goes back to Proto-Quechua, whereas there is no evidence of such antiquity

for the pluraliser -ˇcik,which does not occur in Quechua I. It seems plausible to assume that

both endings took their origin from the lexical element *ˇcika ‘size’, albeit at different times

(cf. Cerr´on-Palomino 1987a: 205).