Adelaar Willem, Muysken Pieter. The languages of the Andes

Подождите немного. Документ загружается.

524 5 The Araucanian Sphere

of indicative and person markers may have occurred, and in the ﬁrst person singular there

is no indicative marker at all. The difference in pronunciation between the non-singular

ﬁrst-person conditional endings and those of the other two moods appear to be induced

by the phonological context, rather than by their being different endings.

The third-person endings can be put into dual or be pluralised by means of the elements

u (dual) and e

ŋ

n (plural), respectively, which are located after the verb (30). These

elements can apply to a third-person subject, as well as to a third-person object. When

they indicate plural or dual of the subject, they can be attached directly to the verb after

the loss of their initial vowel e (31).

(30) elu-e-y-ew eŋŋnpˇc

give-I-3.ID-3.OV 3.PL tobacco

‘He gave them tobacco.’

(Augusta 1966: 39)

(31) ˇc

ipa-ke-y-ŋŋnpu



go.out-CU-3.ID-3.PL night

‘They go out at night.’

(Smeets 1989: 461)

When the subject is either ﬁrst or second person, a third-person object can be marked

straightforwardly by means of the sufﬁx -ﬁ-. This marker is used with deﬁnite, known

objects; if the object is indeﬁnite it can be left out.

(32) n-n mapu

take-1.SG.ID land

‘I took land.’

(Smeets 1989: 19)

(33) n-ﬁ-n mapu

take-3O-1.SG.ID land

‘I took the land.’

(Smeets 1989: 19)

When a third-person actor is combined with a ﬁrst- or second-person patient, the

latter is indicated by means of the corresponding subject ending. At the same time,

two elements must be added: a sufﬁx -e-,which occupies a position to the left of the

subject ending, and an element -ew or mew,which appears immediately after the subject

ending. When the ﬁrst- and second-person markers are singular and in the indicative,

-ew is added directly to these markers; the second-person marker -y-mi loses its ﬁnal

vowel. When the markers are dual or plural the added element is -mew or -mu.

(34) leli-e-n-ew

watch-I-1.SG.ID-3.OV

‘He watched me.’

(Salas 1992a: 120)

(35) leli-e-y-m-ew

watch-I-ID-2.SG-3.OV

‘He watched you.’

(Salas 1992a: 121)

5.1 Araucanian or Mapuche 525

(36) kul

i-e-yin

-mu

pay-I-1.PL.ID-3.OV

‘He paid us.’

(Smeets 1989: 195)

(37) pe-e-y-mu-mew

see-I-ID-2.D-3.OV

‘He saw the two of you.’

(Salas 1992a: 125)

(38) leli-e-ˇci-mu

watch-I-1.SG.IM-3.OV

‘May he watch me!’

(Smeets 1989: 235)

As we have seen in the example of

ŋ

m-e-y-ew (29), a similar formation to the above

obtains when an obviative third-person actor is combined with a proximate third-person

patient. All the forms in -e ...(m)ew have in common that they involve a third-person

actor which is considered to be lower than the patient in a saliency hierarchy deﬁned by

Arnold (1996: 26) as 1 > 2 > 3 proximate > 3obviative (cf. also Salas 1992a: 125). The

sufﬁx -e-,which is present in all these forms, has been interpreted as an inverse marker

(Arnold 1996: 30).

The inverse marker -e- also plays a role in the combination of a second-person actor

with a ﬁrst-person patient. When the inverse marker -e- is accompanied by a ﬁrst-

person-singular subject ending (-n, -l-i), the resulting form refers to the transition of

a second-person actor to a ﬁrst-person patient with the restriction that both must be

singular. In this case the imperative and indicative forms are identical (Smeets 1989:

235; Salas 1992a: 127–8).

(39) leli-e-n

watch-I-1.SG.ID(IM)

‘You watched me.’ ‘Watch me!’

(Salas 1992a: 128)

When either the actor or the patient is non-singular, another inverse sufﬁx, -mu-, takes

the place of -e-. Number distinctions are indicated in the ﬁrst person (the patient), but

are left implicit in the second person (the actor); the imperative endings are used where

relevant (cf. Smeets 1989: 199).

(40) leli-mu-ˇci

watch-I-1.SG.IM

‘Watch (dual or plural) me!’

(Salas 1992a: 127)

Grimes (1985) is reported to have been the ﬁrst to treat the Mapuche personal reference

transitions as an inverse system.

526 5 The Araucanian Sphere

(41) leli-mu-yu

watch-I-1.D.ID

‘You (any number) watched the two of us.’

(Salas 1992a: 128)

In the combination of a ﬁrst-person actor and a second-person patient the inverse sufﬁx

plays a limited role, possibly because the requirement of the actor being hierarchically

lower in saliency than the patient is not met. Nevertheless, a form in -e-yu is used to

denote the combination of a ﬁrst-person actor with a second-person patient when the total

number of participants is no more than two; hence the presence of the ﬁrst-person-dual

ending -yu.Ifthe sum of the participants is more than two, the reﬂexive sufﬁx -(u)w-

is used in combination with the ﬁrst-person-plural ending -yin

(42) leli-e-yu

watch-I-1.D.ID

‘I watched you.’

(Salas 1992a: 128)

(43) leli-w-yin

watch-RF-1.PL.ID

‘I watched you (more than two)’. ‘We watched you’.

‘We (more than two) watched each other’.

(Salas 1992a: 128)

The situation outlined above is that of the Mapuche heartland (former Araucan´ıa).

Valdivia (1606) recorded a very different use of the marker -e- in the seventeenth-century

Santiago dialect. In that variety the inverse system was apparently not checked by any

considerations of hierarchy, the inverse sufﬁx -e- being freely combined with the second-

person subject endings in order to indicate any combination of a ﬁrst-person actor with

a second-person patient. There were number distinctions only for the (second-person)

patient, not for the (ﬁrst-person) actor: elueymi ‘I/we give to thee’, elueymu ‘I/we give

to you two’ and elueymn [elueymn] ‘I/we give to you all’.

A similar situation was

recorded for the ‘Indians of the South’, presumably the Huilliche, by Augusta (1990:

84); cf. also Salas (1992a:128).

The Mapuche language has an elaborate system of verbal nominalisations, which play

a central role in the formation of complex sentences, relative and temporal clauses, etc.

The nominalisation in -(



)n,asinlef-



n ‘run’, ‘running’ or aku-n ‘arrive’, ‘arrival’, has

many characteristics of an inﬁnitive. It is limited in its morphological possibilities, in

The marker -(u)w- in its transitional function is to be kept apart from -(u)w- in its truly reﬂexive

function, because the order and combinational possibilities of the two differ considerably (Smeets

1989: 385–6).

Strangely, with a singular second-person object, the imperative form (elueymi)was identical to

its indicative counterpart, but when the object was dual or plural, there was a difference: eluemu

‘let me/us give to you two’, eluemn [eluem

n] ‘let me/us give to you all’.

5.1 Araucanian or Mapuche 527

particular in relation to tense, but it is used in a wide range of syntactic constructions.

The actor corresponding to the event can be indicated by a possessive modiﬁer (44).

(44) ramtu-e-yu chew n

imle-n

ask-I-1.D.ID where 3P be-IF

‘I asked you where he lives.’

(Smeets 1989: 238)

Stative and agentive nominalisations can be formed by adding the afﬁxes -el and -lu,

respectively. Both types of nominalisation can be combined with tense (see later) and

have an extensive range of syntactic uses. Nominalised verbs in -el can refer either to

the event denoted by the base itself, or to any entity which is not the actor of that event.

It can be used in relative clauses or when referring to occurrences in general. The actor

can be indicated by a possessive modiﬁer.

(45) kim-nie-n ˇcem mi entu-el

know-CN-1.SG.ID what 2P.SG take.out-SN

‘I know what you took out.’

(Smeets 1989: 239)

(46) fali-y ta-mnkel

u-el [Spanish valer ‘to be worth’]

be.worth DC-2P.PL help-SN

‘It is good that you helped.’

(Smeets 1989: 257)

The agentive nominalisation in -lu refers to the actor or theme of the event denoted

by the base (‘the one who ...’), especially in relative clauses. It can also act as the verb

in a temporal clause. In one speciﬁc combination, with the future marker -a-,itcan be

used as a ﬁnite verb, thus competing with the future indicative.

(47)

a-lu el-el-ŋŋe-ke-y ko kin

e metawe mew

die-AG put-BN-PS-CU-3.ID water one jug OC

‘The deceased is provided with water in a jug.’

(Augusta 1990: 192)

(48) piˇci ˇce kim-nu-lu

small human know-NE-AG

‘a child that does not know’

(Smeets 1989: 282)

(49) pe-ﬁ-lu n

iˇcaw amu-tu-y

see-3O-AG 3P father go-RS-3.ID

‘When he saw his father, he went back.’

(Smeets 1989: 287)

(50) fey f

a-ŋŋe-a-lu

he/she husband-CV-F-AG

‘She will be married some day.’

(Smeets 1989: 290)

In the Mapuche of the Araucan´ıa an adjectivising sufﬁx -ˇci takes the place of -lu in

relative clauses of which the verb precedes the antecedent immediately. Interestingly, this

528 5 The Araucanian Sphere

must be the result of relatively recent restructuring. Valdivia (1606) describes relative

clauses in -lu followed by ˇci, and then by the antecedent. It appears that in such cases ˇci

wasadeictic element.

(51) el

apra-pa-ˇci kye



nmew

a.little rise-H-AJ moon OC

‘at the time when the moon had risen only a little

(was in its ﬁrst quarter)’

(Augusta 1990: 191)

(52) huya acu-tu-lu chi Patiru vey may ta inche

[wiya aku-tu-lu ˇci patiru βey may tain

ˇce]

yesterday arrive-RS-AG DC Father that then FO I/me

‘The Father who arrived yesterday, that was me.’

(Valdivia 1606: 47–8)

Although the syntactic uses of -el and -lu nominalisation are complementary, one

remarkable anomaly has to be mentioned. In temporal clauses, a form in -el can replace

the verb form in -lu when the subject of the verb is a ﬁrst person singular (cf. Smeets

1989: 262, 290). The replacement is optional. With any other person or number it would

be impossible. Example (53) illustrates the use of ﬁrst person singular -el (in amu-el).

It further contains an instance of -el, shortened to -l,asrequired after the future-tense

marker -a- (in umaw-tu-a-l); such forms in -a-l indicate a goal.

(53) fey-eŋŋnay-w-y-ŋŋnin

ˇce amu-el n

i umaw-tu-a-l

he/she-3.PL enjoy-RF-3.ID-PL I go-SN 1P.SG sleep-V-F-SN

‘They were glad when I went to sleep.’

(Smeets 1989: 262)

As in Quechua (but unlike Aymaran), the personal reference transitions of Mapuche

have been copied onto the nominalisations. Although not all the possible combinations

can be marked explicitly, the direct–inverse distinction is reﬂected by the use of different

forms (cf. Arnold 1996: 35). In the direct (non-inverse) transitions, as well as in the

(inverse) transition of a second-person actor with a ﬁrst-person patient, a special ending

-ﬁ-el takes the place of -el in the present-day dialect of the Araucan´ıa heartland. In this

ending the element -ﬁ-nolonger has its original function of a third-person object marker.

In the seventeenth-century Santiago dialect -ﬁ-el did not occur; instead of it, the ending

-bi-n [βin] was used.

(54) ta-mi elu-bi-n

[ta-mi elu-βi-n]

DC-2P.SG give-3O-IF

‘what I givetoyou’

(Valdivia 1606: 28)

This state of affairs still exists in the Argentinian Ranquelino dialect (Fern´andez Garay, personal

communication).

5.1 Araucanian or Mapuche 529

The possessive modiﬁer which must precede the -ﬁ-el nominalisation normally refers

to the subject (actor) of the nominalised verb. There is one notable exception, however.

In the interaction between a ﬁrst and a second person the possessive modiﬁer always

marks second person, even when it corefers to the object (1S.2O), rather than to the

subject of the nominalised verb. The example from Valdivia (54) illustrates this anomaly.

Disambiguation of the actor and patient roles is effectuated by the addition of personal

pronouns located on either side of the construction, following the SVO pattern (Smeets

1989: 272).

(55) in

ˇce mi pe-ﬁ-el eymi

I2P.SG see-3O-SN you

‘that I see you’

(modiﬁer refers to patient)

(56) eymi mi pe-ﬁ-el in

ˇce

you2P.SG see-3O-SN I/me

‘that you see me’

(modiﬁer refers to actor)

The inverse transitions that have a third-person actor are expressed in the nominalisa-

tion by means of the ending -e-t-ew,which, among other elements, contains the inverse

marker -e-.Inthis case it is the patient that is identiﬁed by the possessive marker.

(57) n

i pe-e-t-ew

3P see-I-N-3.OV

‘the fact that an (obviative) third person sees him/her/them’

(58) n

i pe-e-t-ew

1P.SG see-I-N-3.OV

‘the fact that (any) third person sees me’

(59) yu pe-e-t-ew

1P.D see-I-N-3.OV

‘the fact that (any) third person sees the two of us’

Another nominalisation strategy in Mapuche involves the sufﬁx -m,which refers to a

place, an instrument, or a tense-marked event. It is used for speciﬁc events with known

participants. It does not occur by itself but is found in combinations with temporal

elements, namely -a-m ‘future place or means’, ‘aim’, -mu-m ‘past place or means’,

‘past inﬁnitive’ and -pe-ye-m ‘usual place or means’ (cf. Smeets 1989: 263–71). The

sufﬁx -we indicates a place or instrument with general value (no speciﬁcation of tense or

participant).

The sufﬁx -fe (-voe in seventeenth-century Santiago Araucanian) indicates

an actor without any speciﬁcation of tense or situation.

The sufﬁx -we in Mapuche is reminiscent of Aymara -(:)wi,which has almost exactly the same

function. Both are frequently used in place names.

530 5 The Araucanian Sphere

(60) meta-we ‘jug’

[meta- ‘to carry in arms’]

(61) meta-we-fe

‘jug maker’ (Augusta 1966: 145)

Several other strategies are in use to form adjectives from verbs, e.g. -fal in ay



-fal

‘lovable’ from ay



- ‘to love’. A non-productive formation is the sufﬁx -en,which forms

both adjectives and nouns from intransitive verb stems ending in -



-, e.g. ray-en ‘ﬂower’,

from ray



- ‘to ﬂower’, and a

k-en (∼ a



n)‘dry’, from a



- ‘to dry’ (cf. werk-en in

section 5.1.5).

The verbal tenses of Mapuche are particular in that there is no real present tense. If the

verb represents an event, the unmarked tense is normally translated as a preterit. If, on

the other hand, it refers to a state or quality, a present-tense interpretation is preferred.

This state of affairs seems to be typical for the Mapuche of the Araucanian heartland.

There is no sign of it in Huilliche (Contreras and Alvarez-Santullano 1989), nor was it

common practice in the Mapocho dialect of Santiago. Valdivia (1606: 13) describes it as

a characteristic of the southern Indians (Beliches), but adds that they would preferably

include a sufﬁx -lle-[l

e] in the verb form (elullen ‘I gave it’). In modern Mapuche

-l

e-isanemphatic sufﬁx not related to tense. Valdivia also afﬁrms that the Indians of

Santiago were in the habit of ‘adorning’ their verb forms with an inﬁxed element -po-

and a postposed element -che [ˇce]; e.g. pe-po-n-che ‘I see’. Some basic examples of the

unmarked tense in Araucan´ıa Mapuche are:

(62) kpa-n

come-1.SG.ID

‘I came.’

(63) kim-n

know-1.SG.ID

‘I know.’

There are explicit markers which indicate past (-fu-) and future (-a-). These tense

markers can occur in most combinations, although not in the imperative, nor with certain

nominalisers. They can also be used cumulatively (-a-fu-)inorder to indicate a future of

the past. Since the unmarked tense refers to past actions as well, the use of -fu- implies

that an event is completed and that its results are no longer valid, nor relevant. Very

often it refers to actions that failed (cf. Smeets 1989: 300–3). After a vowel a the future

sufﬁx usually takes the shape -ya-. The sufﬁx -(



)wye- indicates previousness and can be

Null verbalisation is found with verbs of manufacturing; compare ilo- ‘to slaughter’ (ilo ‘meat’),

kofke- ‘to make bread’ (kofke ‘bread’).

Smeets (1989: 203–5) assigns perfective aspect meaning (implying successful completion) to

the unmarked tense form, but makes an exception for the verbs meke- ‘to be busy’ and

e- ‘to

be’, ‘to exist’. In this perspective, the verbs kim- ‘to learn’, ‘to know’ and nie- ‘to get’, ‘to have’

are ambivalent.

5.1 Araucanian or Mapuche 531

followed by other tense markers in anterior past and anterior future constructions. The

sufﬁx -(



)wma can indicate a perfect tense or previousness, but it does not take personal

reference markers. The resulting form can be interpreted both as a verb and as a noun

(Augusta 1903: 44; Smeets 1989: 291–2).

(64) aku-a-n

arrive-F-1.SG.ID

‘I will arrive.’

(Augusta 1990: 29)

(65) aku-fu-n

arrive-PA-1.SG.ID

‘I was arriving/did arrive (but that is no longer relevant).’

(Augusta 1990: 28)

(66) aku-la-ya-fu-n

arrive-NE-F-PA-1.SG.ID

‘I was not going to arrive (I did not have that intention).’

(Augusta 1990: 29)

(67)

a-wye-ya-n

die-PV-F-1.SG.ID

‘I will be dead by then.’

(Augusta 1990: 44)

(68) nˇc

am-ka-e-n-ew ˇcum-ŋŋe-wma ti walon

conversation-V-I-1.SG.ID-3.OV how-LS-PE DC war

‘He told me how the war had been.’

(Smeets 1989: 292)

In order to describe ongoing events explicitly linked to the present, the Mapuche

language has a series of options which are partly morphological and partly syntactic.

The sufﬁx -nie-ishomophonous and probably historically identical to the verb nie- ‘to

have’. It is used with transitive verbs, where it conveys the meaning of a continued or

ongoing transitive action (45), (69). The sufﬁx -(k



)le- (-k



le- after consonants, -le- after

vowels) is used either to denote a state resulting from an event, or, with intransitive

verbs, an ongoing event, depending on the (telic or atelic) semantics of the verb (21),

(70), (71); cf. Smeets (1989: 368–75). The sufﬁx or root -meke- can indicate an event in

progress with any verb of action (72). Finally, the adverb petu ‘still’ located before the

verb also has the effect of referring to present tense (73). Without further indication all

verbs carrying these elements are interpreted as present tense.

(69) ay-nie-e-yu

love-CN-I-1.D.ID

‘I love you’

(Smeets 1989: 388)

Walon:avariant of malon ‘raid’.

532 5 The Araucanian Sphere

(70) aku-le-y

arrive-PR-3.ID

‘He is arriving.’

(Smeets 1989: 370)

(71) an-m-kle-y

sit-T-ST-3.ID

‘It is planted.’

(Smeets 1989: 369)

(72) i-meke-n

eat-PR-1.SG.ID

‘I am eating.’

(73) petu i-n

PN eat-1.SG.ID

‘I am eating.’

The verbal sufﬁx -ke- indicates a customary action. It can be combined with other

elements, including tense and aspect markers. Without further-tense speciﬁcation it

refers to a habit in the present, as in (47), (74) and (75).

(74) kpa-ke-y-mi

come-CU-ID-2.SG

‘You always come.’

(Augusta 1990: 40)

(75) i-ke-fu-y-ŋŋn

eat-CU-PA-3.ID-PL

‘They used to eat.’

(Augusta 1990: 40)

Negation is indicated morphologically in Mapuche. There are three different markers:

-la-, -nu- (or -no-) and -ki-l-.Inmodern Mapuche, the distribution of the negative markers

is relatively straightforward. The marker -la-isused in the indicative mood, as in (16),

(66) and (76); -ki-l- (or -ki-nu-l-) is used in the imperative (77); and -nu-isused in the

conditional mood (78), in all nominalisations (48), and in nominal expressions such as

negative pronouns (79) and negative nominal predicates (80). In the latter case nu is a

free element or a clitic, rather than a sufﬁx.

(76) amu-la-yu

go-NE-1.D.ID

‘The two of us did not go.’

(Smeets 1989: 236)

(77) amu-ki.l-yu

go-NE-1.D.IM

‘Let us not go (the two of us)!’

(Smeets 1989: 236)

5.1 Araucanian or Mapuche 533

(78) amu-nu-l-iu

go-NE-CD-1.D

‘if the two of us do not go’

(Smeets 1989: 236)

(79) ˇcem nu rume

what NE ever

‘nothing’

(80) fey

θθomo nu

he/she/it woman NE

‘He is not a woman.’

The ending -ki-l- consists of two elements, which can be separated by the inverse

marker -e- and by the third-person object marker -ﬁ-. In some transitional combinations

requiring the inverse marker, the endings following -l- are those of the conditional mood,

rather than those of the imperative. This is the case in the transition of a second-person-

singular actor to a ﬁrst-person-singular patient:

(81) leli-ki-e-l-i (not:

∗

leli-ki-e-l-ˇci)

watch-NE-I-CD-1.SG

‘Do not watch me!’

(Smeets 1989: 237)

The formal relationship between the negative imperative marker and the condi-

tional mood marker is strongly conﬁrmed by Valdivia’s observations concerning the

seventeenth-century Santiago dialect. In that dialect the negative imperative marker

-qui- [ki] was entirely embedded in the conditional mood paradigm (82), although some

hybrid forms were also recorded (83).

(82) elu-qui-l-mi (not:

∗

elu-qui-l-¯ge)

[elu-ki-l-mi]

give-NE-CD-2.SG

‘Do not give!’

(Valdivia 1606: 24)

(83) elu-qui-l-e elu-qui-le-pe

[elu-ki-l-e] [elu-ki-l-e-pe]

give-NE-CD-3 give-NE-CD-3-3.IM

‘That he may not give!’ ‘That he may not give!’

(Valdivia 1606: 24)

Valdivia (1606: 24) also reports the use of -no-, in competition with -la-inindicative

forms such as elu-po-no-n-che ‘I do not give’.

The Mapuche verb can contain a large variety of derivational extensions. Among

them are valency-changing afﬁxes, as well as spatial, modal and aspectual afﬁxes. Some