Adelaar Willem, Muysken Pieter. The languages of the Andes
Подождите немного. Документ загружается.
224 3 The Inca Sphere
especially since it was borrowed into Andean Spanish. It has been conveniently labelled
sudden discovery tense.
35
With regard to Ayacucho Quechua, it is also referred to as
the narrative past because it often denotes a plain past tense in narratives (ending
-sqa). However, sudden discovery is not necessarily confined to the past. It may refer to
the present, and even to the future, when the unpredictable outcome of an experiment is
involved. Due to its range of meanings, the status of sudden discovery as a tense is some-
what debatable. An English translation for the sudden discovery category is ‘it/he/she
turned out (to be)’.
Other tenses found in some of the dialects are a recent past (in Ancash Quechua)
and a perfect comparable to the present perfect tense of English (e.g. in Pacaraos, see
section 3.2.9).
The two tenses of the optative refer to the possibility of an event in the near future,
and to an event that has failed to take place in the past, respectively. Formally, they are
alike, except for the fact that the latter is followed by the third-person past-tense form
of the verb ‘to be’, kar(q)a,which in this case functions as a lexical past-tense marker.
(The examples (87)–(88) are from Ayacucho Quechua.)
(87) maqa-nki-man
hit-2S-PO
‘You could hit (him/her).’
(88) maqa-nki-man ka-r(q)a
hit-2S-PO be-3S.PA
‘You could have hit him/her (but you didn’t).’
Subordinate verb forms are morphologically marked as such in Quechua. They nor-
mally refer to events which are either prior to, or simultaneous with the main event in
the sentence. Although the exact nature of the relation between the two verbs (cause,
condition, temporal background) is left undetermined, a further specification can be
obtained through the addition of aspect markers and independent suffixes. The addition
of the independent suffix -pas (-pis,-si)‘even’, ‘also’ adds a concessive meaning to the
subordinate verb. A progressive aspect marker indicates simultaneousness; a negation in
combination with the independent suffix -raq ‘still’, ‘first’ denotes previousness of the
main event (‘before ...ing’). In some dialects, borrowed Spanish conjunctions (espe-
cially si ‘if’) may precede a subordinate verb form.
36
The exact semantic interpretation
35
The term ‘sudden discovery tense’ is from Adelaar (1977: 94). In the linguistic literature similar
categories, as in Turkish for instance, have been referred to as the ‘inferential past’ (Comrie,
personal communication).
36
Subordinate verb forms have often been interpreted as cases of nominalisation. There is a good
argument for not following this line. Although direct objects preceding a nominalised verb form
may occur without the accusative case marker -(k)ta, this is never the case with subordinate verb
forms.
3.2 The Quechuan language family 225
of a subordinate verb is to a high degree dependent on the meaning of the main verb
which it accompanies. The following sentence, taken from a popular legend concerning
the town of Tapo, near Tarma, illustrates this fairly well:
(89) mana ˇciwaku pla:nu-nˇci-ta aspi-pti-n-qa tapu-m limaq ka-n-man ka-ra
not blackbird map-4P-AC erase-DS-3S-TO Tapo-AF Lima be-3S-PO be-3S.PA
‘If the blackbird had not erased our map, Tapo would have been Lima
(viz. the nation’s capital).’
The main point of interest in relation to subordinate verbs in Quechua is the existence
of an explicit switch-reference mechanism. Most Quechua dialects have a subordination
marker, such as -pti-,which indicates a change of subject. The use of this marker must
be accompanied by an indication of the grammatical person of the subject. This is
understandable because the subjects of the two verbs involved are not the same. In
Ecuadorian Quechua, however, different subject marking is not accompanied by any
personal reference morphology (cf. the case of Tsafiki in section 2.17). The verb form
shows that the subjects are different but not to what grammatical person they belong. The
Ecuadorian different subject marker -(x)pi may have been derived from a combination
of the agentive nominalisation in *-q with the locative case marker -pi, rather than from
*-pti- (in spite of the resemblance).
The shape of the same subject marker differs according to the dialects (cf. sec-
tion 3.2.3). The most common markers are -r (-l)inQuechua I and -ˇspa-(-spa)in
Quechua II.
Switch-reference is of utmost importance for the structure of Quechua discourse.
After a first introduction, the main protagonists in a narrative are not again explicitly
mentioned but are expected to be identified through a correct interpretation of the switch-
reference forms. In order to help this process of interpretation, a subordinate verb which
resumes the content of the previous sentence is often used to introduce a new sentence.
The presence of switch-reference markers permits the hearer to establish whether or not
the subject of the previous sentence is still in focus. Consider the following passage in
a narrative from San Pedro de Cajas (Adelaar 1977: 408–9):
(90) ˇcawra-q’ na:-ˇsi ˇcaka-ra-ya-n.
then-TO already-HS become.dark-CN-PR-3S.
ˇcaka-ru-pti-n-ˇsi na: laso:ˇcu-nuy yapay ˇc
.
a-ru-n alqu-q’ kurba:ta-ku-ˇs’.
become.dark-PF-DS-HS already at.eight-CP again arrive-PF-3S
dog-TO necktie-V-SN
‘At that time, they say, it was already getting dark. When it had
become dark, at about eight o’clock, the dog arrived again, wearing
a necktie.’
226 3 The Inca Sphere
Additional subordination markers occur in some dialects. The ending -ˇstin (or -stin)
indicates simultaneousness with same subjects. Some Quechua I dialects have a negative
subordinator, -nni or -:ni; cf. also Pacaraos Quechua miku:ni ‘without eating’.
Nominalisation has a place at the core of Quechua grammar.
37
Nominalised verbs
in Quechua exhibit the syntactic and morphological characteristics of nouns, while re-
taining a substantial part of the complex verbal morphology as well. Externally, they
can be marked for case (as nouns are), whereas internally they can take case-marked
complements (as verbs do). Personal reference markers are either verbal (when an ob-
ject is involved) or nominal (when a possessive or pseudo-possessive relation to the
event is expressed). With nominalised verbs the subject and possessive markers are for-
mally the same and cannot always be distinguished on the basis of their function and
meaning.
The flexibility of the Quechua nominalisation system is enhanced by its faculty to
combine with case affixes and, to a lesser extent, aspect markers. Potentially, each
combination of a nominalisation with a particular case category yields a specific class of
complement clauses. These configurations complement the rather limited possibilities in
terms of semantic specification that are provided by the subordinative (switch-reference)
categories.
Quechua dialects differ in the number of nominalisers they use, but a minimal system
of four nominalisations is found throughout the family. One category of nominalisation
(ending -y) refers to the event in abstracto and, occasionally, to its result or a generic ob-
ject in the widest sense. It has the characteristics of an infinitive.Personal reference with
infinitives is typically nominal (pseudo-possessive), and there are no tense distinctions,
as in (Ayacucho):
(91) miku-y kawsa-y
eat-IF live-IF
‘to eat’, ‘food’ ‘to live’, ‘agricultural products’
The infinitive is frequently used with the accusative case marker -(k)ta and an auxiliary
verb. Verbs such as ati(pa)-‘to be able’, muna-‘to want’, yaˇc
.
a-/yaˇca- ‘to know how to’
and (the reflexes of) qal
y
a.yku- (QI)/qal
y
a.ri- (QII) ‘to begin’ are frequently used as
auxiliaries, as illustrated by (92)–(93) from Tarma:
(92) rima-y-ta xala.yu-ru-n
speak-IF-AC begin-PF-3S
‘He began to speak.’
37
Foradetailed study in a generative context see Lefebvre and Muysken (1988).
3.2 The Quechuan language family 227
Although grammatically the infinitive functions as the object of the auxiliary verb,
the resulting construction as a whole has several characteristics of a compound verb.
Subject and object markers may be distributed over both components of the con-
struction. For instance, the ventive (hither) marker -mu-may appear on the auxiliary
verb, although from a semantic point of view it belongs to the event denoted by the
infinitive.
(93) yarba-y-ta xala.ya-mu-ra
38
descend-IF-AC begin-H-3S.PA
‘He began to descend’. ‘He began to come down’.
In Ecuador, the -y infinitive was replaced by the instrumental nominalisation in -na
(see below). It has been retained, however, for the exclusive purpose of the auxiliary
construction just outlined.
Forms that result from other nominalisation strategies refer to participants or items
involved in the event denoted by the base verb. They can also refer to a place, time or
means, or to the fact of the event itself. The nominalisation in -q is usually referred
to as the agentive. Agentive forms are subject-centred. They refer to the subject of an
event denoted by the base verb. Tense distinctions are not relevant. The agentive can
be used to construct relative clauses in which the subject is identical to the antecedent.
With verbs of motion it may indicate a goal. Agentive nominalisation also constitutes the
basis of the habitual past tense (see above). (The examples (94)–(95) are from Ayacucho
Quechua.)
(94) pukl
y
a-q
play-AG
‘one who plays’, ‘a player’
(95) pukl
y
a-q ri-saq
play-AG go-1S.F
‘I shall go and play.’
Forms resulting from instrumental nominalisation in -na are everything but subject-
centred. They refer to events not yet realised. As such, they may either denote a means,
a place, a time in the future, an object to be affected, the necessity of an event, or plainly
the fact that it will occur. Instrumentals are frequently used in relative clauses in which
the subject is not identical to the antecedent. When used by itself, the instrumental
38
Actually, the root ‘to descend’ is yarbu-. The sentence is a rare example of a case in which a
vowel modification rule (u > a)iscopied from an auxiliary verb onto the infinitive which it
accompanies. The alternative option in which the infinitive retains its original vowel u is more
common (cf. Adelaar 1977: 117).
228 3 The Inca Sphere
nominalisation may refer to an obligation. Examples (96)–(99) from Tarma Quechua
illustrate different uses of instrumental (future) nominalisation.
(96) xu-x miku-na-n-guna-ta
give-HB eat-FN-3P-PL-AC
‘They used to give (them) their victuals.’ (lit. ‘their things to eat’)
(97) mana-m musya-ra-ˇcu yanu-ku-na ga-na-ta
not-AF guess-3S.PA-NE cook-CU-FN be-FN-AC
‘He did not suspect that there would be a way (means) of cooking them.’
(98) ˇcay xiˇsya-y-xa asta
39
wanu-na-n-gama nuna-ˇc
.
u:-xa ga-n
that be.ill-IF-TO until die-FN-3S-LI man-L-TO be-3S
‘That illness remains in a man until (the moment) he dies.’
(99) aywa-na-:
go-FN-1S
‘I must go.’
One of the most important applications of the instrumental nominalisation is its com-
bination with the benefactive case suffix -paq. This combination refers to a purpose
clause, as in (Ayacucho):
(100) yanapa-wa-na-yki-paq
help-1O-FN-2S-B
‘...so that you might help me’
Alongside -na,anarchaic form -nqa was preserved in colonial Cuzco Quechua. In
Ecuadorian Quechua, a reflex of *-nqa is still used today in the local equivalent of the
-na + -paq construction just described (for an example see section 3.2.8 on Salasaca
Quechua). Interestingly, this construction has become involved in an extension of the
switch-reference system. It is now used for purpose clauses in which the subject is the
same as that of the main clause. The different subject function has been taken over by
the original third-person imperative ending -ˇcun.Aswementioned before, Ecuadorian
subordinate verbs lack personal reference marking. (For a parallel situation in Tsafiki,
see section 2.17.)
The applications of the stative nominalisation in -ˇsqa (-sqa, -ˇska,-ˇsa) are largely
the same as those of its instrumental counterpart, the difference being the fact that
statives refer to events that are realised either previously, or simultaneously with re-
gard to the main event. As a rule, forms resulting from nominalisation in -ˇsqa are not
39
The particle asta (from Spanish hasta ‘until’) is often found in combination with the limitative
case marker -gama (-kama in most other dialects), which has essentially the same meaning.
3.2 The Quechuan language family 229
Table 3.5 Valency-changing suffixes in Quechua
-ˇci- ‘causative, permissive transitiviser’ (adds a new
agent argument)
-ku- ‘reflexive’
-na(ku)- ‘reciprocal’
-ka(:)- (QI) ‘medio-passive’, ‘no control’
-na(:)- (QI), -naya- (QII) ‘experiencer’ (transfers a need or desire to an object)
-pa(:)- (QI), -pa- (QII) ‘applicative transitiviser’ (adds a new object)
-ra(:)- (QI), -raya- (QII) ‘state or prolonged event’
-wˇsi- (QI), -ysi- (QII) ‘accompanied action transitiviser’ (to help someone
perform an action)
subject-centred. (Nevertheless, subject-centred statives do occur sporadically with in-
transitive bases.) Example (101) from Tarma illustrates the use of a stative participle.
(101) lapa-n waxta-ˇc
.
u xampi-nˇci-guna-m kanan-xa xunxa-ˇsa
all-3P slope-L medicinal.plant-4P-PL-AF now-TO forget-SN
‘On all the slopes our medicinal plants are now forgotten.’
Stative participles are frequently found in compound tenses, where they occasionally
acquire an active (subject-centred) meaning. This is the case of the experiential past in
Pacaraos Quechua, illustrated in (102).
(102) rika-pu-ˇsq´a-s(u) ka-nki
see-LS-SN-IR be-2S
‘Have you ever seen it?’ ‘Did you ever get to see it?’
The derivational or modal affixes that operate as extensions of the verb root together
constitute the richest and most complex part of Quechua morphology. As we anticipated,
they make up a very heterogeneous set from a semantic and a functional point of view.
Some can function as valency-changing affixes; see table 3.5.
The following examples illustrate the use of these suffixes. The examples (110a, b),
illustrating the suffix -ysi-, are from the Ayacucho dialect; the others exemplify the
dialect of San Pedro de Cajas in northern Jun´ın.
(103) a. wanu- ‘to die’ wanu-ˇci- ‘to kill’, ‘to cause to die’
b. maqa- ‘to beat’ maqa-ˇci- ‘to make (someone) beat’
‘to have (someone) beaten’
(104) a. mayla- ‘to wash’ mayla-ku- ‘to wash oneself’
b. wanu- ‘to die’ wanu-ˇci-ku- ‘to kill oneself’,
‘to commit suicide’
230 3 The Inca Sphere
(105) a. maqa- ‘to beat’ maqa-naku- ‘to beat each other’
b. maqa-ˇci-naku- ‘to have each other beaten’
(106) a. miku- ‘to eat’ mika-na(:)- ‘to provoke hunger’
(impersonal subject)
b. mika-na-y ‘hunger’
(107) a. punu- ‘to sleep’ punu-ka(:)- ‘to fall asleep’
b. laqa- ‘to stick’ laqa-ka(:)- ‘to get stuck to someone’
(108) a. ayku- ‘to laugh’ ayku-pa(:)- ‘to laugh at’, ‘to smile at’
b. muna- ‘to want’ muna-pa(:)- ‘to desire’, ‘to long for’
(109) a. wila- ‘to warn’ wila-ra(:)- ‘to be in a state of having
been warned’
‘to be in a state of having
announced’
b. rirka- ‘to look’ rirka-ra(:)- ‘to observe’, ‘to watch’
c. hita- ‘to throw’ hita-ra(:)- ‘to lie down’
(110) a. apa- ‘to carry’ apa-ysi- ‘to help someone carry’
b. puri- ‘to walk’ puri-ysi- ‘to accompany’
Two derivational extensions are closely linked with the personal reference system.
The ventive or cislocative suffix -mu- (‘hither’) either denotes a psychological approach
experienced by the speaker, or any motion directed towards the location of the speaker.
In this function -mu-isobligatory to the extent that it cannot be left out without arousing
the suggestion of a motion away from or neutral with regard to the speaker’s location.
In narratives the notion of ‘speaker’s location’ may acquire a derived interpretation, as
it can represent a location which the speaker has in mind, viz. a place focused in the
narrated event, as in (Ayacucho):
(111) apa-mu-y apa-y
carry-H-2S.IM carry-2S.IM
‘Bring it (here)!’ ‘Take it away!’
If -mu-isfound in a verb which cannot be interpreted as a verb of motion, the
action is supposed to be carried out at an indicated place and to have consequences for
the speaker. The suggestion may be that of a circular effect which emanates from the
speaker; example (112) is from Tarma Quechua:
(112) ˇcay-ˇc
.
u:-mi ulxu ga-y-ta yaˇc
.
a-ka-mu-nxa
that-L-AF man be-IF-AC know-RF-H-3S.F
‘(I will send my son to the Army.) There he will learn how to be a man.’
The benefactive suffix -pu-isused in many Quechua dialects, in combination with
personal reference markers otherwise referring to object, for the purpose of encoding a
3.2 The Quechuan language family 231
beneficiary, as in Ayacucho (113):
(113) ˇcura-pu-sayki
keep-BN-1S.2O.F
‘I shall keep it for you.’
In Cuzco Quechua, -pu- has acquired the additional function of a reversive (back to
original state) or an itive (motion away from the scene). In the latter case, it functions as
the semantic opposite of -mu-.
Aspect systems are more or less well developed in virtually all Quechua dialects. In
Quechua IIC, the progressive aspect marker is reconstructed as *-ˇcka- (possibly from
*-ˇc
.
ka-). It is still found as such in Ayacucho Quechua, as well as (very seldom) in the
seventeenth-century variety of Quechua used in the manuscript of Huarochir´ı (akin to
Quechua IIB). Modern reflexes of the progressive aspect marker are Cuzco -sya-or-ˇsa-,
Bolivian Quechua -sa- and Santiago del Estero -ˇska-. The progressive aspect marker
in Quechua I dialects is either -yka(:)- or its reflex -ya(:)-.InEcuadorian Quechua,
progressive aspect meaning is conveyed by -ku- (also -xu-), originally a marker of the
reflexive category.
More complex aspect systems are found in Quechua I, in particular, in the Huanca
dialects and in northern Jun´ın (including Tarma). In these dialects, reflexes of the suffix
*-rqu- (-ru-, -
ʔ
(lu)-, -:(lu)-) operate as a perfective
40
counterpart of the progressive
marker -ya(:)-. Both aspect categories are mutually exclusive and stand in opposition
to the unmarked forms which have a habitual or general truth value. In the dialects
of the province of Tarma, aspect marking is mutually exclusive with negation at the
sentence level. In some northern Quechua I dialects (Hu´anuco, eastern Ancash) a suffix
-ski- occupies a place in the aspect system comparable to that of *-rqu-insouthern
Quechua I.
Directional affixes referring to the semantic categories ‘outward’ (*-rqu-), ‘inward’
(*-yku-), ‘upward’ (*-rku-) and ‘downward’ (*-rpu-) can be reconstructed, at least for
Quechua I, on the basis of such word sets as yarqu- ‘to leave’, yayku- ‘to enter’, yarku-
‘to climb’ and yarpu- ‘to descend’. The suffixes -rku- and -rpu- are still used productively
to express direction in many varieties of Quechua I. The suffixes -r(q)u- and -y(k)u- are
used in most Quechua dialects but have acquired new functions.
The remaining derivational extensions mainly bring about semantic additions to the
verb root, often resulting in lexicalised verb–suffix combinations. Some semantic addi-
tions are straightforward, for instance, that of -(y)kaˇc
.
a(:)- (QI)/ -(y)kaˇca- (QII), which
conveys the meaning ‘up and down’, ‘back and forth’ or ‘hesitatingly’. But the semantic
40
The use of the perfective aspect in southern Quechua I is reminiscent of the use of perfective
verbs in Slavic languages such as Russian.
232 3 The Inca Sphere
additions of other verbal extensions are elusive and difficult to define. The meaning of
-ri- ‘inchoative’ is a case in point. Additionally, some of the suffixes with straightforward
semantic applications have derived meanings, which belong to the pragmatic level and
are often hard to translate.
The order in which the verbal extension suffixes appear when combined is basi-
cally fixed. Some combinations of suffixes, however, are so infrequent that it would
seem rather artificial to speak of fixed-order classes. Certain valency-changing suf-
fixes (particularly, causative -ˇci- and the reflexive and reciprocal markers) tend to dis-
play different order options with corresponding differences in interpretation. In other
cases, the existence of different order options may be of no semantic consequence
(see, for instance, the discussion of Ayacucho Quechua verbal suffix order in Parker
1969a). When the order position of a suffix is fixed, as is the case of pluralisers and
aspect markers in Quechua I, for instance, it need not reflect the logical semantic
build-up of the verb form as such. In what follows, the Tarma Quechua forms (77)–
(79) are expanded with the causative marker -ˇci-. The ordering of the suffixes is quite
unexpected.
(114) wata-ˇci-ba:ku-n
tie-CA-PL-3S
‘They have it tied.’
(115) wata-rga-ya:-ˇci-n
tie-PL-PR-CA-3S
‘They are having it tied.’
(116) wata-ra-:ri-ˇci-n
tie-PF-PL-CA-3S
‘They (eventually) had it tied.’
In the examples (114)–(116) the presence of an aspect marker relocates the pluraliser
to the left of the causative suffix. The exact location of the pluraliser is furthermore
determined by the choice of the aspect marker. It follows the perfective but precedes
the progressive aspect marker, even though the two aspect markers make up a mutually
exclusive set (as we saw before).
An interesting but widely neglected aspect of Quechua morphology is the existence
of several types of root reduplication both in the verbal and in the nominal sphere. In
Tarma Quechua, different types of verbal reduplication refer to simulated action (‘to
act as if’), frustrated intention (‘to try without success’), eagerness (‘cannot wait to
do something’) and resulting condition. Nominal reduplication may denote dispersed
units, distributed qualities, pronoun plurality, ‘every’ and again resulting condition.
An example of nominal reduplication denoting a resulting condition based on the
3.2 The Quechuan language family 233
adjective rakta ‘thick’ is:
(117) rakta-y=rakta-y-ta-m xagu-ra-ya-:
thick-IF=thick-IF-AC-AF clothe-ST-PR-1S
‘I am dressed warmly.’
3.2.7 Characteristics of the Quechua lexicon
Most modern Quechua dialects have assimilated relatively large amounts of borrowed
vocabulary from Spanish. At the same time, however, a comparison with early colonial
texts shows a remarkable continuity in the native lexicon. Most lexical items that were in
use in the sixteenth century are still in use today. The loans have enriched the Quechua
lexicon, rather than substituted obsolete native terms. Only in certain culturally sensitive
domains, such as religion, social and political organisation, and kinship terms, has the
original native lexicon become much reduced.
An interesting peculiarity is the rather limited number of native roots in many do-
mains of Quechua vocabulary. Quechua roots can have a wide spectrum of semantic
applications, leaving the impression of a certain lack of semantic differentiation. This is
counterbalanced by the richness of the derivational morphology, which is available for
the expression of all sorts of semantic distinctions in an ad hoc way. Quechua narrators
have no stylistic prejudice against repetitions of same vocabulary items within a given
discourse. This fact may have slowed down the formation of new vocabulary in the
language.
The area in which the aforementioned scarcity of items is most conspicuous is that of
the verbs. Quechua verb roots are not formally categorised for the distinction transitive–
intransitive. Several verb roots, such as paki- ‘to break’ or tikra- ‘to turn’, can be used
both transitively and intransitively. Disambiguation, if necessary, is left to the derivational
morphology and the syntax.
A typical example of root economy in Quechua is the absence of a basic verb for the
notion ‘to kill’. In order to express this notion, there is no alternative but to use the regular
causative derivation of the verb ‘to die’, wan
y
u-ˇci-. That wan
y
u-ˇci-isstill synchronically
a derived base is corroborated by the potential insertion of additional affixes between
the root and the causative marker -ˇci-. Also from a semantic point of view, wan
y
u-ˇci-
preserves its original composite structure. It can be interpreted not only as a causative
‘to kill’, but also as a permissive ‘to let die’. On the other hand, Quechua has several
non-composite roots denoting specific ways of killing, for instance, (Ayacucho) sipi- ‘to
murder’, naka- ‘to butcher’.
Aremarkable case of low semantic differentiation is the verb of communication
n
y
i- (ni-inmost contemporary dialects). Although it is usually associated with the
meaning ‘to say’, it covers a wide range of interpretations varying from ways of saying