Adelaar Willem, Muysken Pieter. The languages of the Andes

Подождите немного. Документ загружается.

504 5 The Araucanian Sphere

BOLIVIA

PARAGUAY

ARGENTINA

Córdoba

RANQUELCHE

(19th CENTURY)

GÜNÜNA

YAJICH(†)

Santiago

Concepción

Temuco

Osorno

HUILLICHE

Chiloé

TEHUELCHE

MAPUCHE

Buenos Aires

URUGUAY

PERU

BRAZIL

Map 12 The Araucanian Sphere: twentieth-century distribution of indigenous languages

5 The Araucanian Sphere 505

An overview of the Araucanian Sphere would not be conceivable without paying some

attention to the Argentinian pampas and Patagonia (including the provinces of Buenos

Aires, La Pampa, R´ıo Negro and Chubut), which for centuries constituted an outlet for

the expansion of the freedom-loving Araucanians. Among the native groups of these

sparsely peopled southern extensions were the nomadic Pehuenche,who inhabited the

eastern Andean slopes and foothills south of the Huarpeans, in the province of Neuqu´en.

These Pehuenche owed their name to their dependency on the collection of pine-nuts

of the Araucaria tree (pewen in Mapuche). Most early sources agree in afﬁrming that

the Pehuenche had a language and an identity of their own, distinct from those both

of the Araucanians and the Huarpeans. Unfortunately, nothing is known of the original

Pehuenche language. The present-day group called Pehuenche is located in Chile, on

the upper reaches of the Biob´ıo river (between Santa B´arbara and Lonquimay) and in

other pre-cordilleran areas further south, between the lakes Icalma and Panguipulli. They

speak a Mapuche dialect and are not necessarily related to the historical Argentinian

Pehuenche. The Querand´ı of Buenos Aires were the ﬁrst to become exposed to the

European colonisation of the R´ıo de La Plata in the seventeenth century. They soon

lost their ethnolinguistic identity. Only two sentences and a few words of their language

were recorded by French sailors around 1555. On the basis of these few data Viegas

Barros (1992) shows that Querand´ımay have been closely related to the language of the

G¨un¨una K¨une or Puelche.Ifthis conclusion is correct, it would identify the Querand´ıas

the northernmost representative of the Chon language family. The Tehuelche complex,

which includes G¨un¨una K¨une and various other representatives of the Chon family,

will be discussed in chapter 6. During the eighteenth and nineteenth centuries all these

peoples came into contact with the expanding Araucanians, who either subjugated them

by force, or assimilated them through alliance. As a result, a number of Araucanian-

speaking mixed groups came into existence, some of which are still aware of their

non-Araucanian origins. This, for instance, is the case with the Araucanised Tehuelche,

wholiveinElChal´ıa in southwestern Chubut. The Tehuelche origin of this group is

conﬁrmed by the use of Tehuelche vocabulary items and phonological features, such

as the use of a uvular stop and glottalisation, common in Tehuelche but otherwise

unknown in Araucanian (Fern´andez Garay 1997a). For some interesting examples of

Chubut Mapuche usage see also D´ıaz-Fern´andez (1992). Some of the mixed groups,

such as the Ranqueles,

for a time became powerful nations until they were annihilated

by the Argentinian army in 1879–81.

In spite of its geographical location, which could be considered peripheral within a

South American context, the Araucanian Sphere belongs to the oldest inhabited areas

The Ranqueles (or ra



l-ˇce ‘people of the reeds’) originally had their territory in southern

C´ordoba. In 1900 they were relocated in Colonia Emilio Mitre, in northwestern La Pampa

(Fern´andez Garay 1998a).

506 5 The Araucanian Sphere

of the American continent. Excavations conducted at Monte Verde near Puerto Montt

in the far south of continental Chile have seriously challenged the convictions of ar-

chaeologists that the human occupation of South America must have been posterior

to the Clovis horizon (9500–9000 BC), when Paleo-Indian big-game hunters roamed

on the North American plains (Fiedel 1992). The existence of a 12,000-year-old pre-

agricultural village in this remote area of South America suggests an earlier arrival of

the ﬁrst inhabitants than was held possible so far (Dillehay 1989–97).

Although the Araucanian Sphere region was neither an area of cultural prestige,

nor of great artistic production, it was certainly not unimportant from a political and a

demographic point of view. In spite of ﬁerce resistance of the Araucanians, the Inca rulers

succeeded in conquering the northern part of it, including central Chile and the Cuyo

region, but their control must have remained limited to the existence of military outposts

and probably some mitimaes (see chapter 3). The limit of the Inca military advance, to

either the Maule river (altitude of Talca) or to the Biob´ıo (altitude of Concepci´on), is

still a matter of debate. The Incas called the Araucanians Auca (from Quechua awqa

‘enemy’, ‘rebel’), a name which the Araucanians proudly adopted as a self-designation.

Soon after their ﬁrst attempts at colonisation, the Spanish conquerors became ac-

quainted with the stubborn and ﬁerce resistance of the Araucanians. The dramatic events

of the mid sixteenth century, when the Spanish conquistador Pedro de Valdivia tried to

establish a colony in central Chile at the cost of many Spanish and Indian lives, are

vividly narrated in Alonso de Ercilla’s epic poem La Araucana (1569–89). After these

turbulent beginnings, a status quo arose in which central Chile and the far south (in-

cluding the town of Valdivia and the island of Chilo´e) were under Spanish control,

whereas most of the intervening area (including the present-day provinces of Arauco,

Biob´ıo, Malleco and Caut´ın) remained independent and outside the domain of Christian

inﬂuence. The Araucanians strengthened their position thanks to an unusually effective

military organisation, great strategic experience and dexterity in the use of horses. In the

seventeenth century the Araucanians supported the Dutch in their short-lived attempt of

1642–3 to establish a colony in Valdivia under the leadership of Hendrik Brouwer. As

soon as it became clear that the newcomers could not live up to the Indians’ expectations

and started to behave like a colonial power themselves, the Araucanians withdrew their

support.

The dominant group among the unsubdued Araucanians were the Mapuche (‘people

of the land’). Consequently, the denomination Mapuche became almost a synonym of

Araucanian. The latter name was probably derived from that of the town of Arauco,

situated at the northern border of the independent Mapuche territory. This territory was

A more speciﬁc denomination used in the colonial period is that of promaucaes (from Quechua

purum awqa ‘uncivilised enemy’).

5 The Araucanian Sphere 507

known as La Araucan´ıa during the colonial period and in the nineteenth century. Outside

its borders the Araucanians were either completely assimilated (in the north), or gradually

dwindled in numbers also by assimilation (in the south). After the independence of Chile

(in 1818) the Mapuche in the independent territory retained their autonomy, although

it came gradually under the pressure of a consolidating Chilean state. In 1860 Antoine

Or´elie de Tounens, a notary-clerk from the P´erigord in France, came to the Araucan´ıa

and had himself crowned King of Araucania (and later also Patagonia) under the name of

Or´elie-Antoine I. Although many Mapuche recognised and supported him, the Chilean

and Argentinian authorities succeeded in evicting him twice in succession. Some of

Or´elie-Antoine’s relatives still claim rights to the Araucanian throne today.

The ﬁnal blow to the Araucanian independence in Chile came in the 1880s when

the Chilean military occupied the area in a campaign known as paciﬁcaci´on de la

Araucan´ıa (paciﬁcation of Araucania). The Mapuche were conﬁned in reserved areas

(reducciones), separated by pieces of land that were handed out to Chilean peasants

and newly arrived immigrants from Europe. This policy, aimed at destroying the unity and

organisational traditions of the Mapuche, created many antagonisms between Indians and

non-Indians. So far, it has proved effective. The Mapuche became a marginal and much

discriminated population, plagued by poverty and internal strife. During the Pinochet

dictatorship of 1973–89 the Mapuche suffered heavy oppression and adverse legislation

aimed at the dismantlement of the reducciones system. It is estimated that more than

a hundred thousand Mapuche have migrated to Santiago and other towns outside their

native territory.

Meanwhile, the provinces of Malleco and Caut´ın (in the ex-Araucan´ıa)

have remained predominantly Mapuche. The Araucanians who entered Patagonia and

the Argentinian pampas did so partly in order to establish colonies, and partly for

raiding expeditions. These raiding expeditions, known by the name of mal´on,were

conducted with great efﬁciency and violence. They had a terrifying effect on the white

settlers in these sparsely inhabited regions. Alliances between Indian groups sometimes

also involved adventurers of European extraction, who tried to escape the control of

the young Argentinian state. Some of these groups for some time acted as separate

political entities, a situation which is described by the Argentinian colonel Mansilla in

the account of his visit to the Ranqueles in 1870 (Mansilla 1947). Throughout most of the

nineteenth century southern Argentina remained dominated by Indians, Buenos Aires

being situated close to the frontier. The anti-Indian animosity among the authorities

and citizens of Buenos Aires grew such that the Indians were ﬁnally subdued in several

military campaigns, especially those of 1833 and 1879–81. The latter campaign led by

Most of the statistical information in this chapter has been obtained from the website Estado

actual de las lenguas abor´ıgenes de Chile (http://rehue.csociales.uchile.cl) of Gilberto S´anchez.

It contains information from the 1988 census.

508 5 The Araucanian Sphere

General Roca, during the government of Sarmiento, is known as conquista del desierto

(‘conquest of the desert’) and was geared towards the total destruction of the Indian

communities in the pampas and in northern Patagonia. The survivors of this campaign

were rounded up in reservations.

Considering that most of the other languages belonging to the domain of this chapter

are extinct and very poorly documented, our attention will mainly go to the Araucanian

language and its local varieties. At the end of the chapter some general features of a

Huarpean language (Allentiac) will be given.

5.1 Araucanian or Mapuche

At the arrival of the Spanish conquerors, Araucanian was the dominant, if not the only

language spoken in central Chile, the non-insular part of southern Chile and the island of

Chilo´e. Luis de Valdivia,

the author of the ﬁrst description of Araucanian (1606), speaks

of the ‘language which is used in the entire Kingdom of Chile’ (la lengua que corre

en todo el Reyno de Chile). In a later stage of the colonial period the name Chilidugu

(ˇcili

u ‘language of Chile’) came into use (cf. Havestadt 1777). The modern name

of the Araucanian language is Mapudungun, to be analysed as mapu-

u-n ‘language

of the (people of the) land’, by analogy with Mapuche (mapu-ˇce ‘people of the land’),

as the Araucanians nowadays prefer to call themselves. Nevertheless, the language too

is often referred to as Mapuche.

Early colonial observers (Bibar 1558) point at the linguistic homogeneity of central

and southern Chile when the area was ﬁrst visited by Spanish military expeditions.

There is a marked contrast with the situation of linguistic diversity and multilingualism

emerging from seventeenth-century descriptive accounts of almost any other area in

the Americas. Spanish religious authorities used the Araucanian language in order to

subdue and evangelise non-Araucanian peoples, such as the Chono of the islands south

of Chilo´e, to the Christian faith.

Although it appears that dialect differences did not stand in the way of mutual under-

standability, Valdivia gives a more subtle picture of the situation of the language. His

work is of particular interest due to his exposure to a variety of Araucanian that has long

been extinct, the dialect of the bishopric of Santiago. This variety was also known as

Mapocho or Mapuchu

after the name of the river that runs through the Chilean capital

today. Valdivia explicitly mentions several distinctive elements of the Santiago dialect,

which are not found in the varieties spoken today. When he mentions such characteristics,

he often contrasts them with linguistic habits that were prevalent among the Araucanians

The Jesuit grammarian Luis de Valdivia is not to be confused with the conquistador Pedro de

Valdivia, who died under Araucanian torture in 1552.

The river name Mapocho seems to have no relation with ‘Mapuche’, the present-day name of the

Araucanian people.

5.1 Araucanian or Mapuche 509

further south (‘arriba’), to which he often referred as the Beliche (from wil

i-ˇce ‘people

of the south’). Some of the characteristics that Valdivia considered typical for ‘arriba’

or Beliche speech apply nowadays to the language of the Mapuche heartland in Malleco

and Caut´ın, in southern Chile. An example is the absence of the pluralising morpheme

y`uca [yka], which, according to Valdivia, was used in the bishopric of Santiago as a

postposed element with non-animate nouns (e.g. ruca y`uca [ɹuka yka] ‘houses’), in con-

tradistinction to the preﬁx-like element pu which marked plurality with animate nouns,

as in pu che [pu ˇce] ‘people’. ‘The Beliches apply pu to every sort of thing’, says Valdivia

(1606: 10). The modern Mapuche language uses pu for human beings and only excep-

tionally for objects (Augusta 1903: 16), which is roughly in accordance with Valdivia’s

observation about the Beliche practice. In addition to such explicit mentions, Valdivia’s

description contains several lexical items and grammatical elements that are not found

in present-day varieties of the language. Examples are the verb cu- [ku] ‘to go’ (modern

Mapuche amu-) and the verbal derivational sufﬁx -clo-[klo] ‘to help’, which has no

equivalent in present-day Mapuche. In part, they may have been speciﬁc features of the

Santiago dialect, but we could also be dealing with expressions that became obsolete

during the last four centuries.

During the colonial period all of central Chile north of the Biob´ıo river became

thoroughly Hispanicised. The Mapocho dialect and maybe some other varieties became

extinct. The Araucanian dialects that survived are those of the former Araucan´ıa and

those of the territories further south, including Chilo´e. The varieties that spread into

Argentina were brought there mainly by refugees or migrants from the former Araucan´ıa,

which is conﬁrmed by the close linguistic similarity between the Chilean and Argentinian

Mapuche. The one considerably divergent variety of Araucanian still spoken today is

the dialect of the present-day Huilliche, who live in the southernmost provinces of

Valdivia, Osorno and Chilo´e, in the Chilean region of Los Lagos. Its local name in

San Juan de la Costa (Osorno), cesu

un ‘speech of the people’, reﬂects some of its

phonetic characteristics (Alvarez-Santullano Busch 1992).

Further characteristics of

the Huilliche dialect are discussed in Salas (1992a: 86–92).

In the past several attempts were made at subdividing the Chilean Araucanians into

ethnolinguistic subgroups. Best known is the classiﬁcation of Lenz (1895–7) who, in

addition to the Huilliche, distinguished the Picunche (pikum-ˇce ‘people of the north’) in

the northern part of the former Araucan´ıa (province of Malleco), Moluche or Ngoluche

(

ulu-ˇce ‘the Chilean Mapuche as they are called by their (eastern) Argentinian kinsmen’,

probably related to

-ant



‘evening’) in the southern part of Araucan´ıa (province of

Caut´ın), and the Pehuenche (‘people of the Araucaria pine’), who live near the cordillera.

Croese (1980, 1985) distinguishes eight dialect areas in Chile, two of which correspond

The equivalent expression in Mapuche would be ˇce-

u-n.

510 5 The Araucanian Sphere

to Lenz’s Picunche and one of which corresponds to Huilliche. There appears to be

a consensus that, with the exception of Huilliche, all the Chilean dialects are mutually

intelligible. The Picunche dialect stands out by the fact that its interdental and labiodental

fricatives are voiced ([ð], [v]), whereas they are voiceless ([

θ], [f]) in the Mapuche

dialects further south (Salas 1992a: 92). In this respect, the Picunche dialect coincides

with the extinct northern variety described by Valdivia. Valdivia’s transcription (<d>,

<b> ∼ <v>) suggests that the fricatives in that variety were also voiced.

The Argentinian dialects are only slightly different from their Chilean relatives, with

the exception of occasional substratum inﬂuence from Tehuelche or other local lan-

guages. They may share speciﬁc features with one of the Chilean dialects, thus betraying

the homeland from which they once migrated. For instance, the Ranquelino dialect of

La Pampa province shares the use of voiced fricatives with Picunche, which may be its

closest relative in Chile (cf. Fern´andez Garay 1991).

Approximately 400,000 Mapuche people live in the Chilean region of Araucan´ıa

(provinces of Caut´ın and Malleco), where they are the majority of the rural population.

There seem to be no reliable ﬁgures concerning the number of speakers of the language,

but it may be assumed that a substantial part of the Mapuche population in this area

continue to speak it. Some smaller groups of Mapuche are located in Arauco and Biob´ıo

(region of El Biob´ıo) and in Valdivia (region of Los Lagos). The Huilliche people inhabit

a discontinuous area in the region of Los Lagos, including the coast of Osorno (San Juan

de la Costa), the interior of Valdivia (Isla Huapi in Lake Ranco) and several locations near

Quell´on at the southern tip of Chilo´e (e.g. the community of Huequetrumao). Referring

to Isla Huapi and San Juan de la Costa, Contreras and Alvarez-Santullano (1989) insist

that only people over sixty can speak the language to a satisfactory degree. Given the

critical situation of the Huilliche language, it is unlikely that there remain more than

afew thousand speakers. Estimates concerning the number of Mapuche in Argentina

ﬂuctuate between 27,000 and 60,000 (Mart´ınez Sarasola 1992: 493). The largest group

of Argentinian Mapuche live in Neuqu´en, but there are also communities in the provinces

of Buenos Aires, Chubut, La Pampa and R´ıo Negro. Very little can be said about the

actual number of speakers of Mapuche in Argentina, as in many locations the language

is being replaced by Spanish (Fern´andez Garay 1996).

5.1.1 Mapuche studies

The Mapuche language has been studied more intensely than many other South American

Indian languages. A good overview of what has been done is Salas (1992b). The Mapuche

grammatical tradition begins with Valdivia (1606), whose work was followed in the

second half of the eighteenth century by the grammars of Febr´es (1764) and Havestadt

(1777), the latter one in Latin.

At the end of the nineteenth century Rodolfo Lenz (1895–7) published an extensive

series of studies, known as Estudios Araucanos (Araucanian Studies), including many

5.1 Araucanian or Mapuche 511

texts in different dialects, accompanied by ethnographic and linguistic observations.

Lenz also became known for his claim that Chilean Spanish, as pronounced by the lower

class, is essentially Spanish with Araucanian sounds (Lenz 1893: 208; 1940: 249). The

arguments upon which he based this assertion were successfully refuted by Alonso

(1953). Nevertheless, the idea of an Araucanian substratum in Chilean Spanish may

encounter a more favourable ear today.

Approximately in the same period, Bavarian missionaries renewed the Mapuche de-

scriptive tradition. Their most outstanding ﬁgure was F´elix Jos´edeAugusta, known for

his grammar (Augusta 1903) and dictionary of the language (Augusta 1916), by far the

best dictionary of the Mapuche language to date. Augusta also published a collection

of traditional text and ethnographic notes, Lecturas Araucanas (Araucanian Readings),

collected by himself and Sigifredo de Fraunhæusl (Augusta 1910). To the same grammat-

ical tradition belongs a grammar by Ernesto de Moesbach (1963). Doubtless, his most

valuable contribution is the edition of the autobiography and memoirs of Pascual Co˜na,

a Mapuche chief, who lived through the turbulent years of the last great Argentinian

raids and the paciﬁcation of the Araucan´ıa (Moesbach 1930). Both in content and in

form it is a monument of native American literature.

Among the more substantial contemporary contributions to the study of Chilean

Mapuche, the work of Salas (1979, 1992a, 1992b) occupies a central position. Salas

(1992a) also contains an extensive treatment of the different genres in Mapuche folk lit-

erature, including some annotated and translated texts. Smeets’s unpublished dissertation

of 1989 contains the most detailed grammatical description of the Mapuche language

so far, accompanied by analysed and translated texts.

Catrileo (1987) is a classroom

textbook for learning Mapuche. A recent grammatical description of Mapuche is Z´u˜niga

(2000).

Most of the publications concerning the Chilean Mapuche language are in the form of

articles in journals. We will mention just a few examples. Important articles on Mapuche

phonology are Su´arez (1959), Echeverr´ıa Weasson (1964), Echeverr´ıa and Contreras

(1965), Lagos Altamirano (1981) and Rivano (1990). The complex system of personal

reference marking in Mapuche is discussed, inter alia,inFontanella de Weinberg (1967),

Salas (1978, 1979), Grimes (1985), Rivano (1988, 1989) and Arnold (1996). Several

aspects of Mapuche grammar are discussed by Harmelink (1987, 1988, 1990, 1992).

The historical-comparative position of Araucanian is discussed in Key (1978), Stark

(1970) and Croese (1991). In relation to the Argentinian varieties of Mapuche we can

mention the publications of Fern´andez Garay on the Ranquel dialect (Fern´andez Garay

1989b, 1991, 1998b, 2001), as well as Golbert’s text edition Epu pe˜niwen (Two Brothers)

(Golbert de Goodbar 1975).

A published version of Smeets (1989) is in preparation.

512 5 The Araucanian Sphere

5.1.2 The sounds of Mapuche

The phonological structure of a Mapuche word is relatively straightforward. It is remi-

niscent of Quechua in that consonant sequences are limited to the intervocalic position

and consist of no more than two consonants.

However, in contradistinction to Quechua

and Aymara, vowel sequences are common, as in (1):

(1) leli-a-e-n-ew

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

‘He will watch me.’

The Mapuche vowel inventory consists of six vowels a, e, i, o, u,



. Their pronunciation

and analysis do not present any particular difﬁculties, except for



. There are no length

distinctions. The vowel



has a variable realisation, which can either be a schwa [ə], or

an unrounded high central vowel [¨ı] according to the environment in which it occurs. In

many cases the presence of the vowel



has not been perceived at all in the past. In fact, the

recognition of



as a unit phoneme is a relatively recent phenomenon. In the publications

of the Bavarian missionaries the different alleged realisations of



were distinguished in

writing, either as <¨u>,oras<

ə>, e.g. w¨un· ‘daybreak’, w

n· ‘mouth’ (Augusta 1916),

whereas in reality these two words do not differ in pronunciation [w



n]. Furthermore,

Valdivia (1606) and Augusta (1903) often write consonant sequences, where contem-

porary descriptions would assume the presence of an intervening



(2) Valdivia (1606: 10) modern

tva tfa ‘this’

(3) Augusta (1903/1990: 29) modern

konn kon- n ‘I entered.’

The non-recognition of the intervening vowel is not just an omission. Valdivia

(1606: 9) explicitly states that Araucanian had word-ﬁnal clusters.

It is the overall word

structure of the Mapuche language that has incited contemporary linguists to assume

the presence of a vowel not recognised so far. The morphophonological behaviour of



Occasionally, triconsonantal clusters involving the semi-vowel w occur, as in a

fwin ‘I had

him on the back of my horse’ (cf. Smeets 1989: 55). The element -fwi-isacontraction of -fu-

‘past tense’ and -ﬁ- ‘third-person object marker’.

Valdivia (1606: 9): ...ysesiguen despues dellas dos consonantes assi de las que en Espa˜na

solemos juntar, como de las que no solemos, y vna de las consonantes que suelen juntar es la

<g> –[

ŋ] [authors’addition] – que pusimos en el notable seg˜udo. Y no se deue pensar que entonces

ay nueua sylaba de mas de la vocal, porque no es mas de vna (‘...and after these (vowels) two

consonants may follow, like those we are accustomed to put together in Spain, as well as those

we would not put together. And one of the consonants they are accustomed to put together is

<g> [

ŋ], which we have treated in our second remark. And one must not think that there is a new

syllable then in addition to (the one of) the vowel, because there is no more than one’).

5.1 Araucanian or Mapuche 513

may have played a role in the way its status was perceived by different authors. As in the

case of the ‘impure’ vowels of Mochica (section 3.4), it can be used as a full root vowel

(as in w





n ‘mouth’), but also as a means to avoid non-licensed consonant sequences.

Example (3), kon-



n,isacase in point. The ﬁrst-person-singular subject ending is

-n; after bases ending in a consonant -



- is inserted in order to avoid sequences such

as *-nn.

An additional complication is the fact that the vowel



is often accompanied by a

voiced velar glide or semi-vowel [γ ]. Although there is some individual variation, the

presence of the glide is on the whole predictable, so that it need not be analysed as a

separate consonant phoneme. However, the rules governing the behaviour of [γ ] are so

speciﬁc and complicated that it turns out easier to include it with a separate symbol in

the orthography (Smeets 1989: 32–4). In some environments, i.e. between a full vowel

and a consonant, the semi-vowel is the only possible realisation, e.g. in reγle ‘seven’.

Elsewhere, γ is added either before or after the vowel



, according to the position in the

word.

(4) γy ‘name’

kuwγ ‘hand’

Valdivia (1606) is quite consistent in writing < ´u> (or any of its graphic variants < `u>,

<¨u>) both for the vowel, and for the semi-vowel (re ´ule, ´uj, cu´u). Augusta (1916) uses

the symbol <q> for some postvocalic realisations of [γ ]: word-internally (e.g. in reqle

[ɹeγle]); word-ﬁnally in (near) monosyllables (e.g. liq ∼ l¨uq [liγ ] ∼ [lγ ]‘white’, kuq

[kuwγ ] ‘hand’); and word-ﬁnally after vowels other than



(e.g. umaq [umaγ ] ‘sleep’).

Elsewhere, including in word-initial position, he does not write the semi-vowel (e.g. ¨u´ı

[γ y] ‘name’, ant¨u [ant(γ )] ‘day’).

The relation that holds between the semi-vowels y and w and the vowels i and u, respec-

tively, is somehow comparable to that between γ and the vowel



.However, the position

of y and w as consonant phonemes is much ﬁrmer due to their typically consonantal

behaviour. In Mapuche of Caut´ın the verbal indicative (or third-person subject) ending

-y is obligatorily preceded by a connective vowel



after a base ending in a consonant.

(Optionally, this sequence may be replaced by -iy; cf. Smeets 1989: 42, 227.)

(5) lef-y

run-3.ID

‘He/she ran.’

(6) lef-y-mi

run-ID-2.SG

‘You ran.’