Adelaar Willem, Muysken Pieter. The languages of the Andes
Подождите немного. Документ загружается.
384 3 The Inca Sphere
occurred after a vowel (384), whereas the latter was found after bases ending in a
consonant (385).
(384) k
.
ilap´anya a´ıta-k
.
alo
[cf. haita-tur ‘to drink’]
ckilapana drink-2S.IM
‘Drink ckilapana!’
178
(Mostny 1954: 142)
(385) can-alo [cf. ckan-tur ‘to give’]
give-2S.IM
‘Give!’
(Tschudi 1866–9, in Echeverr´ıa y Reyes 1967)
The negation of the imperative was expressed by means of a suffix -ˇca- (Mostny 1954:
153). A good example of this construction is again found in the Lord’s Prayer versions
(386). The base verb is the Spanish loan dejar ‘to leave’, ‘to abandon’.
(386) cum deja-cha-calo
1O.PL leave-NE-2S.IM
‘Don’t let us (fall into temptation)!’
(Tschudi 1866–9, in Echeverr´ıa y Reyes 1967)
Example (386) also illustrates the use of prefixes or preposed elements referring to
person of object, in this case, cum [kun] ‘us’. Va¨ısse et al. mention at least two other
elements that are used in this way: ack for first-person-singular object and tchencki
for second-person-singular object. One may venture the idea that the latter was in fact a
complex form referring to the transition of a first-person subject (cki-) to a second-person
object (tchem-).
The categories of deverbal nominalisation comprise the characteristic infinitive ending
in -tur, amply illustrated in the current section. Doubtless, there are other possibilities,
such as those illustrated by the forms l´alack-ma ‘daybreak’ and lalck-tchir ‘light’, both
derived from lalck-tur ‘to become light (of days)’ (Va¨ısse et al. 1896: 24). More verbal
morphology consists in the existence of a transitive or causative formation involving an
element -un-, as in (387):
(387) lalck-tur ‘to become light’
lalack-un-tur ‘to give/make light’
(Va¨ısse et al. 1896: 24)
The existential verb ‘to be’ in Atacame˜no is given as ttanzi [t’ansi] in San Rom´an
and as ttansi-r or ttans-tur in Va¨ısse et al. Its negation was expressed by sin(t)cha.
178
Ckilapana:afermented drink made from pods of the algarroba tree (Prosopis chilensis). The
ending -ya is not explained in this case.
3.8 The Lule–Tonocot´e language 385
Example (388) from Mostny illustrates this:
(388) t´e sinˇcaot’´ansi
tea not.be or be
‘Is there any tea, or not?’
(Mostny 1954: 170)
The SOV word order of Atacame˜no is illustrated in (389). Note the absence of a
subject prefix.
179
The element -ne- remains unexplained; it may be part of the present-
tense ending. There is no case marker for the goal.
(389) ´acca l´ıcan s´aj-ne-ma
[lickan ‘town’; sack-tur ‘to go’]
Itown go-?-PN
‘I am going to town.’
(San Rom´an 1967: 78)
Case markers and other elements referring to grammatical relations cannot easily be
recovered because they are not treated in the sources. Only a few case endings, such
as -ckol ‘instrumental’ (Va¨ısse et al. 1896: 19; Mostny 1954: 154) and -p(a)s ‘allative’
(Mostny 1954: 154), have been identified beyond doubt.
From a lexical point of view, it is interesting to note that the Atacame˜no language had
a full inventory of decimally organised numerals, which also included a non-borrowed
term for ‘hundred’: sem(m)a ‘one’, ppoya ‘two’, pp´alama ‘three’, tchalpa [ˇc’alpa] ‘four’,
m´utusma (San Rom´an: m´utsisma) ‘five’, m´ıtchala ‘six’, tch´oya [ˇc’oya] ‘seven’, tch´olama
[ˇc’´olama] ‘eight’, t´eckara ‘nine’, sutchi ‘ten’, haaras ‘hundred’ (San Rom´an: aras). The
connecting element was -ta,asinsuchi-ta ppoya ‘twelve’ (San Rom´an 1967: 80).
Atacame˜no had borrowed many lexical items from Quechua, notwithstanding von
Tschudi’s emphatic denial of such being the case (S´aez Godoy 1971: 20), e.g. ckausa-tur
‘to live’, from Quechua kawsa-, and tussu-tur ‘to dance’, from Quechua tusu- (for other
examples see the preceding pages). Borrowings from other languages are represented in
tchamma ‘force’, from Aymara ˇc’ama, and in horsa-tur ‘to rape’, from Spanish forzar.
3.8 The Lule–Tonocot´e language
The Lule were one of the semi-nomadic peoples who inhabited the Gran Chaco between
the Pilcomayo river and the Andean foothills of northwestern Argentina. Their language
is known through a grammar and vocabulary published by Machoni de Cerde˜na in 1732.
Machoni called his work Arte de la lengua Tonocot´eyLule (Grammar of the Tonocot´e
and Lule Language), suggesting that the Lule and Tonocot´e languages were one and
the same. This is confusing because in early colonial sources such as the Relaciones
geogr´aficas de Indias (Jim´enez de la Espada 1965, I: 390–6; II: 78–85) the two nations
179
According to San Rom´an (1967: 87), the Atacame˜nos ‘used few auxiliaries and abbreviated
everything they could’.
386 3 The Inca Sphere
and their languages are treated as separate. In the sixteenth century the Tonocot´ewere
part of the sedentary indigenous population of the provinces of Tucum´an and Santiago
del Estero, now occupied by speakers of the Quechua dialect of the same name (cf.
Mart´ınez Sarasola 1992: 549–52). They suffered constant attacks from the Lule. At
the beginning of the eighteenth century part of the Lule population had been brought
together in a mission in eastern Salta, which was subsequently transferred to a location
south of Tucum´an in 1737, where a town called Lules still exists (Furlong 1941).
Machoni justifies his procedure by affirming that the language which he describes
was spoken by five nations: Tonocot´e, Lule, Ysistin´e, Toquistin´e and Oristin´e. He claims
that the Tonocot´e, whom he located in the interior of the Gran Chaco, were the most
numerous of the five and had not yet been converted to Christianity in his time. Since
the languages of the tribes in question have long been extinct and only one variety has
been documented, the question whether Lule and Tonocot´ewere indeed identical, and
if not, whether Machoni’s grammar deals with the former or with the latter, may remain
open.
The view that Machoni’s Lule is genetically related to the moribund or extinct Vilela
language of the R´ıo Bermejo basin has received attention since the end of the nineteenth
century (Lafone Quevedo 1894a, b, 1895; Balmori 1967). It is probably correct, although
the relationship is certainly not a close one (Viegas Barros 2001). There is no easily
detectable genetic relationship between Lule–Vilela and the other two main linguistic
groups of the Argentinian Chaco, Guaicuruan and Matacoan. There are some similarities
with Matacoan, which may be due to contact.
Machoni describes a relatively small inventory of sounds consisting of five vowels, a,
e, i, o, u and a limited number of consonants. Stress is predominantly word-final.
180
The
symbol y is used as an equivalent of i, especially when stressed, but also for referring
to a semi-consonant. According to Machoni, the main difficulties of Lule pronunciation
consist in distinguishing between the sibilants c, ¸c, and s and in the existence of complex
consonant clusters involving sibilants in word-initial and word-final postion (e.g. quelp¸c
[kelpt
s
]‘Ispit’, slimst [slimst] ‘I make a sound with the nose’, oal´ecst [wal´ekst] ‘I know’,
stu¸c [stut
s
]‘Ithrow’).
Machoni’s description of the sibilants is not free of contradiction and is therefore
hard to interpret. His symbols c, ¸c, and apparently also z, seem to stand for a sound
which may be identified tentatively as an alveodental affricate [t
s
]; the language has
no ch [ˇc] as in Spanish. However, Machoni adds that in some words the c has to be
pronounced as a (Spanish) ¸c or ss (normally the pronunciation of his symbol s). An-
other inconsistency in the use of the symbol c is that it can both refer to [t
s
] and to
180
Word-final accent is usually written with an acute by Machoni. The examples given are in
Machoni’s original spelling.
3.8 The Lule–Tonocot´e language 387
[k]. This becomes evident from the different derivations of the imperative, as in uec-y
‘die!’ [wet
s
´ı] from uec-¸c [wet
s
t
s
]‘Idie’ against poqu-y [pok´ı] ‘dig!’ from poc-¸c [pokt
s
]
‘I dig’.
The simplicity of the Lule sound inventory is surprising when we consider the rather
complex inventories of the related Vilela language (which has postvelar stops, a voicing
distinction and a glottalisation distinction) and of almost all the surrounding languages,
including the modern Quechua dialect of Santiago del Estero. It is conceivable that
Machoni may have missed such distinctions. There are no indications to that effect,
except for the word tt´a ‘egg’, in which the double consonant symbol may point at the
existence of glottalisation. On the other hand, the highly informative vocabulary that
accompanies Machoni’s grammar contains evidence of voiceless laterals (quilh´a [kil
´a]
‘Indian girl’; cf. Vilela kil
e ‘woman’) and, possibly, of a voiceless nasal (nhal´a pul´u
‘sugar cane’). The existence of a glottal stop is suggested by se¸c ∼ s.he¸c [s
ʔ
et
s
]‘I
support’, in which s and the vowel must be pronounced separately, and se¸c [set
s
]‘Icry’,
in which they are pronounced without interruption.
From a superficial point of view, Lule appears to be a typically Andean language with
a suffixing structure, a moderately complex morphology and case markers. However, a
closer look at the verbs included in Machoni’s Spanish–Lule vocabulary shows that this
picture is not entirely correct. These verbs exhibit different sorts of derivational mor-
phology including prefixation, compounding and reduplication. There are, for instance,
verb sets characterised by a root and different prefixes referring to an instrument or body
part with which the action is performed. The root cannot occur without one of these
prefixes, nor are these prefixes formally related with corresponding nouns. For instance,
the root -calam- ‘to press’ is absent from the vocabulary, except when it occurs with one
of the instrumental elements illustrated in (390).
(390) ni-calam- ‘to press with the hands’ cf. ys ‘hand’
yacs-calam- ‘to press with the feet’ cf. ell´u ‘foot’
si-calam- ‘to press with knees or body’ cf. toi- ‘body’
ti-calam- ‘to press the earth (of rain)’ cf. auy´a ‘rain-water’
na-calam- ‘to press the earth (of running water)’ cf. t´o ‘water’
The root -moi- ‘to kill’ is found both with a generic and with an instrumental prefix
(391):
(391) tac-moi- ‘to kill’
s-moi- ‘to kill with an arrow or a spear’
nic-moi- ‘to kill (chicken) with the hands’
na-moi- ‘to kill with a string or lasso’
ap-moi- ‘to kill with a bullet’
388 3 The Inca Sphere
Compounding can be illustrated by the encoding of bodily positions in a verb such as
ui- [wi] ‘to sleep’ (392):
(392) lo-ui-
181
‘to sleep while sitting’ cf. lo-, loho-
182
‘to sit (down)’
qui-ui- ‘to sleep while standing’ cf. qui- ‘to stand’
ele-ui- ‘to sleep while lying’ cf. ele- ‘to lie (down)’
Further illustrations of compounding involving verb roots can be observed in (393)
and (394):
(393) ti-tuc- ‘to try’
ne-tuc- ‘to try and eat’, ‘to taste’
(394) tic-tun- ‘to finish’
ne-tun- ‘to finish eating’
Note that the common term for ‘to eat’ is not *ne-butcai-. Furthermore, Lule has a
wide array of terms for eating depending on what is eaten: e.g. pel- ‘to eat flour’, let-
‘to eat soft fruits’, ues- ‘to eat beans or cabbage’, anc- [ant
s
] ‘to eat soup’.
Lule verbs may differ lexically in function of the number (395) or the gender (396)
of the subject:
(395) u´os-¸c arrive-1S.SG ‘I arrive’
ual-cen arrive-1S.PL ‘we arrive’
(396) pac- [pat
s
] ‘to carry (of men)’
yes- ‘to carry (of women)’
Reduplication of the initial syllable is found in derived verbal substantives that refer
to an instrument, as in (397):
(397) ne-neyu-p´e ‘flute’ cf. neyu- ‘to play the flute’
Verbs are inflected for person and number. The endings which are used for the non-
future indicative tenses coincide roughly with the endings indicating person and number
of the possessor with nouns. Examples with the verb amaici- ‘to love’ and uy´a ‘house’
are given in table 3.32.
The ending of the verbal first person singular varies between -¸c,-s,-t, and zero. The lat-
ter two possibilities are used as alternatives after roots ending in a consonant cluster with
a final sibilant, as in oal´ecs-t (∼ oal´ecs-¸c, oal´ec-¸c)‘Iknow’ or stops-t (∼stops-¸c, stops)
181
In the examples the symbol v has been replaced by u, because the former is a mere notational
variant of the latter.
182
The interpretation of the symbol h between vowels is ambiguous. Apart from indicating a glottal
or velar fricative, it can also refer to a glottal stop.
3.8 The Lule–Tonocot´e language 389
Table 3.32 Lule personal endings for unmarked tense and nominal possession
amaici- ‘to love’ uy´a ‘house’
1 pers. sing. amaici-¸c ‘I love.’ uy´a-s ‘my house’
plur. amaici-c´en ‘We love.’ uya-cen ‘our house’
2 pers. sing. amaici-c´e ‘You (sing.) love.’ uya-c´e ‘your (sing.) house’
plur. amaici-lom ‘You (plur.) love.’ uya-lom ‘your (plur.) house’
3 pers. sing. amaici-p ‘He/she loves.’ uya-p ‘his/her house’
plur. amaici-p´an ‘They love.’ uya-p´an ‘their house’
Table 3.33 Lule future and imperative
verbal paradigms
Future Imperative
1 pers. sing. -n-s –
plur. -n-cen -pe, -cen
2 pers. sing. -psse + V, −y
plur. -n-(pe)lom -uan
3 pers. sing. -n-t -pep, -to
plur. -n-pan -(n-)pan
‘I pour (a liquid)’. The choice between -¸c and -s appears to be arbitrary. As for substan-
tives, Machoni affirms that they normally take -s,but that some take -¸c; e.g. y(h)´a-¸c ‘my
brother’.
The third-person-singular ending of the verbs has two forms: -p and -t. According to
Machoni, the verbs that take -t instead of -p are those that have roots ending in -a-, -e-,
-u-, -li- and -qui-, e.g. e-t ‘he sees’, taclu-t ‘he throws’.
The Lule verbal system comprises different tenses and moods, as well as devices for
subordination and nominalisation, all indicated by means of suffixation. There are no
indications of object encoding within the verb. Near and remote past are characterised
by the affixes -ni- and -(y)ate-, respectively, which are inserted between the root and the
person and number affixes. The imperative and, to a lesser extent, the future paradigm
have special endings for person and number. These are represented in table 3.33.
Apart from the second-person-singular ending all future endings contain an element
-n-. This same element is used to distinguish between an unmarked active participle in
-ton and a future active participle in -n-ton, e.g.:
(398) amaici-t`on ‘he who loves’
amaici-n-t`on ‘he who will love’
183
183
In this case Machoni uses a grave instead of an acute accent. The reason for it is not known.
390 3 The Inca Sphere
Both participles can be used as finite verbs replacing present-tense and future forms,
respectively. The second-person-singular imperative ending consists of a vowel -y after
roots ending in a consonant, and a lengthened final vowel after roots ending in a vowel.
This is indicated by doubling of the last vowel. It is not sure whether the added vowel
was pronounced separately, as vowel length, or that it was merely stressed. There are
some irregular imperatives, such as cai ‘go!’, ei ‘see!’ and nei ‘come!’ (from ca-, e- and
ne-, respectively).
(399) taclu- ´u ‘throw!’
taclup-y ‘turn upside down!’
The subordinate form of the verbs is formed by adding the element -l´e to the unmarked
tense of the indicative (e.g. amaici-¸c-l´e ‘when I love’). The element -l´e coincides in
form with a case marker that indicates location involving motion (‘into’, ‘from’). The
interpretation of the case marker depends on the meaning of the verb or the context. The
second l´e in (400) and (401) is described as a part´ıcula de ornato (ornamental particle)
by Machoni.
(400) uya-l´el´e ne-¸c
house-L – come-1S.SG
‘I come from the house.’
(401) uya-s-l´el´e ca-¸c
house-1P.SG-L – go-1S.SG
‘I go to my house’
Lule nouns referring to human beings can be marked for plural, although this is not
obligatory. The exact rules of plural formation are difficult to glean from Machoni’s
grammar.
(402) cu´e ‘son or daughter’ cu´e-l ‘children’
zal´a ‘young man’ zala-til ‘young people’
Personal possessive endings are added to the plural ending, if any, e.g.:
(403) Pedro cu´e-l-p uy´e
Pedro child-PL-3P.SG not.be
‘Pedro has no children.’
As can be seen in (403), the genitive construction is characterised by possession marked
on the head. The modifier (possessor) precedes the head and remains unmarked.
184
If
184
Adjectives follow the noun they modify, as in t´o ytyps ‘lukewarm water’.
3.9 Extinct and poorly documented languages 391
the head consists of the words p´e ‘father’ or umu´e ‘mother’, the possessive ending -p is
left out whenever the possessor is present. If the possessor is not present, the ending is
compulsory.
(404) Juan pe-p ‘the father of Juan (not present)’
Juan p´e ‘the father of Juan (present)’
The personal pronouns of Lule are not related to the verbal endings. Some of them,
however, can be related to verbal endings in the Vilela language: quis ‘I’, u´e ‘you’, u´a
‘we’, mi-l ‘you (plural)’. In Vilela -ki-s is a first-person-dual or plural verbal ending
(singular -ki), whereas -mi refers to a second person singular (Lozano 1977).
185
The
reflexive pronoun in Lule is formed on the basis of a root lo-towhich personal possessive
endings are added. Pronouns used in object function, including the reflexives, are located
before the verb.
(405) u´ey´o ne-¸c
you take come-1S.SG
‘I have come to take you.’
[y´one- ‘to take’, ‘to fetch’]
(406) lo-s tec-¸c [los tekt
s
]
self-1P.SG fight-1S.SG
‘I battle’, ‘I fight’
The numeral system of Lule is not typically Andean in that it consists of only four basic
items: alape´a ‘one’, tam´op ‘two’, tamlip ‘three’ and locu´ep ‘four’ (cf. lucu´e ‘man’). The
numeral locu´ep provides the basis for higher numbers, as in locu´ep moitl´e tamlip ‘seven’
(‘three added to four’) and in locu´ep moitl´e locu´ep alape´a ‘nine’ (‘four added to four
plus one’). The word for ‘ten’ ysyauomp is derived from ys ‘hand’ and yauo´o ‘together’.
It is used as a basis for further counting, as in ysav´omp moitl´e tam´op ‘twelve’ (‘two added
to ten’). Multipliers precede the number they multiply, as in tamlip ysav´omp ‘thirty’.
Lule contains several loan words from Quechua, such as yapa- ‘to add’ (Quechua
yapa-), tanta ‘bread’ (Quechua t’anta), tity ‘tin’ (Quechua titi), utc´u ‘cotton’ (Quechua
utku). Interesting are several loans from Spanish that have been modified almost beyond
recognition, e.g. capl´a ‘goat’ (Spanish cabra), ceual´a ‘millet’ (Spanish cebada), polot´o
‘bean’ (Spanish poroto) and telec´o ‘wheat’ (Spanish trigo).
3.9 Extinct and poorly documented languages of the Inca Sphere
As anticipated in section 3.1, the Middle Andean region was originally a linguistically
diverse area with numerous local languages. The decline of these languages was almost
185
It is interesting to observe that -l, -l
and -s also act as nominal pluralisers in Matacoan (Tovar
1981; Gerzenstein 1994).
392 3 The Inca Sphere
certainly triggered by the political unification of the area by the Incas, but many of them
were still widely spoken when the Spanish conquerors arrived. Much of the linguistic
diversity in the Middle Andean region was lost during colonial rule and in the first
century of independence. With the exception of the languages treated in the preceding
sections 3.4–3.8, all the languages that vanished during that period have remained un-
documented. Place names, personal names and, in the best case, short word lists must
be relied upon for a reconstruction of some of the characteristics of these languages.
Their existence is sometimes confirmed by mentions in the historical literature. For a
discussion of these languages we distinguish three relevant areas: Ecuador, northern
Peru and northwestern Argentina.
3.9.1 Ecuador
In the highlands of Ecuador it is possible to reconstruct much of the original language
situation, characterised by a series of distinct language areas that can be identified on the
basis of toponymy and other information. In the coastal area this is nearly impossible.
Coastal Ecuador was (and is) a densely populated area, but, apart from its northernmost
sector (see chapter 2) its linguistic past remains hidden. At the time of the conquest, the
ethnic group that inhabited the region of Guayaquil and the lower valley of the Daule
river was called Huancavilca (Cieza de Le´on 1553). The Huancavilca were in the habit
of removing part of their front teeth, and this is what their name in the native language
is reported to have referred to (Velasco 1789). There were intensive trade relations with
the islanders of Pun´a and with the inhabitants of the port of T´umbez in northern Peru.
Jij´on y Caama˜no (1941: 385–97) and Paz y Mi˜no (1961a) treat all these groups as a
linguistic unity, called Atall´an by the latter. Jij´on y Caama˜no adds the language of the
area of Manta in the province of Manab´ıtothis complex. He considers the Manab´ı–
Huancavilca–Pun´agroup to represent a single language with dialects, belonging to
afamily that also comprises Mochica and the Ca˜nar and Puruh´a languages in the
highlands further east. Although other classifications (Loukotka 1968) have followed
this suggestion, there seems to be no significant empirical evidence for it. Several other
groups, such as the Indians of the Car´aquez bay (Manab´ı), the Chono of the Guayas
river basin, the Campaces of the upper Daule valley and the Colima (location unclear),
have not been associated with the Huancavilca complex. Jij´on y Caama˜no proposes a
Barbacoan connection for the latter two, suggesting that the Campaces may have been
identical to the Tsafiki-speaking Colorado (cf. section 2.17 ; see also Hartmann 1980).
Newson (1995) proposes a similar connection for the Chono.
In 1593 the Diocesan Synod of Quito issued an instruction calling for catechisms and
confessionaries in the following languages: Quillacinga, Pasto, Puruh´a, Ca˜nar, ‘the
language of the plains’ (la de los Llanos) and the Atallana (Jij´on y Caama˜no 1941:
385). Assuming that the Atallana language mentioned in this document was in fact the
3.9 Extinct and poorly documented languages 393
Tall´an language of Piura in present-day Peru (see section 3.9.2), we may identify the
‘language of the plains’ with that of the Huancavilca complex, which could have been
predominant in southern coastal Ecuador. Another document of 1605, Descripci´on de
la Gobernaci´on de Guayaquil, refers to an extreme linguistic diversity in the Portoviejo
area in the province of Manab´ı. The same source reports that in about 1600 Spanish was
already becoming the contact language in the coastal area (Hartmann 1980).
In the highland part of Ecuador we find a succession of languages, which from north to
south are the following: Pasto, Cara, Panzaleo, Puruh´a, Ca˜nar, and a language complex
comprising Palta, Malacato, Rabona, Bolona and Xiroa. These languages are mentioned
in early sources such as Cieza de Le´on (1553) and the Relaciones geogr´aficas de Indias
(Jim´enez de la Espada 1965). They have been the object of studies by the Ecuadorian
scholars Jij´on y Caama˜no (1940, 1941) and Paz y Mi˜no (1940, 1941a, b, 1942, 1961b),
whoinventoried recurrent roots and endings, as well as other salient features of these
languages that can be deduced from place names, family names and historical accounts.
However important their work may be have been, a reassessment of the results following
modern techniques and insights is very much needed.
Pasto is usually classified with the Barbacoan family, together with its neighbour
Cuaiquer or Awa Pit (cf. section 2.17). The (extinct) variety of Muellamu´es (Nari˜no,
Colombia), of which a short word list was recorded during the nineteenth century, was
clearly related, though not identical, to Awa Pit and had a higher incidence of Quechua
loan words (cf. Jij´on y Caama˜no 1940: 197). It may have represented one of the last
surviving dialects of the Pasto language. Muellamu´es piar ‘maize’ can be recognised
in the place name Piartal, in Carchi province (Ecuador); the interpretation of -tal is
‘rock’ (Paz y Mi˜no 1940). The Pasto territory was situated around the border towns of
Tulc´an and Ipiales, striding the Colombian–Ecuadorian border. In Ecuador it included
the Andean part of Carchi. Typical elements in Pasto place names are -quer (Mayasquer,
Altaquer) and -es (Ipiales, Pupiales, T´uquerres). The language has been replaced by
Spanish.
The territory of the Cara or Caranqui language extended from the valley of the
Mira and Chota rivers to the city of Quito, covering the Andean region situated in the
province of Imbabura and in the north that of Pichincha. Important centres in this area are
Otavalo, Cayambe and Ibarra (Carangue). In a study of local place and family names,
Salomon and Grosboll (1986) show that the Cara linguistic area may have reached
the northern outskirts of Quito, where it bordered on the Panzaleo language further
south. The area is today inhabited by a dynamic indigenous population speaking a
variety of Quechua known as Imbabura Quechua. In particular the people of Otavalo
and surrounding villages are known for their rich folklore and economic success. Cara
may have been used until well into the eighteenth century (Caillavet 2000: 103), before it
became fully replaced by Quechua. A possible substratum element from Cara preserved