Dialectal variation in the Aymara language of Bolivia and Peru

Material Information

Dialectal variation in the Aymara language of Bolivia and Peru
Briggs, L. T. ( Lucy T )
Place of Publication:
Gainesville, Fla.
Lucy Therina Briggs
Publication Date:
Copyright Date:
Physical Description:
2 v. (xvi, 848 leaves) : map ; 28 cm.


Subjects / Keywords:
Aymara language ( lcsh )
Aymara language -- Dialects ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )


General Note:
Dissertation--University of Florida.
Includes bibliographical references (leaves 835-846).
General Note:
General Note:
Statement of Responsibility:
by Lucy Therina Briggs.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator ( with any additional information they can provide.
Resource Identifier:
03125344 ( oclc )
000178680 ( alephbibnum )


This item has the following downloads:

Full Text







This description of dialectal variation in the
Aymara language of Bolivia and Peru is dedicated
to all the Aymara speakers who helped make it
possible, and to the Aymara linguists of the
future who will improve upon it.


This study is based on research conducted from

1970 through 1975 at the University of Florida and in

Bolivia and Peru, under the auspices of (1) a graduate

teaching assistantship (1970-71) in the Aymara Language

Materials Project funded by Title VI of the National

Defense Education Act, (2) a National Science Foundation

graduate fellowship (1971-74), and (3) a University of

Florida College of Arts and Sciences graduate fellowship

(1974-75). To the sources of that support I wish to

express my deep appreciation.

Special thanks are due also to my parents,

Ellis Ormsbee Briggs and Lucy Barnard Briggs, who

gave me financial and moral support throughout my doc-

toral studies.

My field work in Bolivia was authorized by the

Institute Nacional de Estudios LingiUsticos (INEL) and

facilitated by the Instituto de Lengua y Cultura Aymara

(ILCA). My field work in Peru was authorized by the

Institute Nacional de Investigaci6n y Desarrollo de la

Educaci6n (INIDE). Copies of this study are being made

available to the three named entities, for whose coopera-

tion I am very grateful.

The list of persons who assisted me in the

research is long. I wish here to single out three per-

sons whose contributions were crucial to my undertaking

the task and bringing it to a conclusion. They are

Dr. M. J. Hardman, director of my doctoral dissertation,

and two native speakers of Aymara who were my teachers at

the University of Florida: Ms. Juana Vasquez, writer and

artist, and Mr. Juan de Dios Yapita, founder and director

of the Instituto de Lengua y Cultura Aymara of La Paz and

professor at the Universidad Nacional de San Agustin in

La Paz. Whatever insights I have gained concerning Aymara

language and culture are due in large measure to their

knowledge and patient guidance. Without the training I

received from them and their considerable help in the

analysis, this study could not have been completed.

Specifically, Ms. Vasquez helped me review the extensive

literature on Aymara, commenting on the Aymara examples

contained therein. Mr. Yapita reviewed a near-final

draft of the whole manuscript of this study, sometimes

listening to tapes to check the accuracy of the tran-

scriptions of the Aymara examples. Both worked with me

in the analysis of translation dialects culminating in

Chapter 9.

In the various stages of the work Dr. Hardman

was my constant mentor, challenger, and support. The

final draft also benefitted from the suggestions of the

other members of my doctoral committee, as well as those

of Professor Bohdan Saciuk of the Program in Linguistics

of the University of Florida, who kindly read and com-

mented on Chapters 3 and 4. I also wish to thank Dr.

Charles Palmer for preparing the final versions of the

maps, and Ms. Patricia Whitehurst for typing the final

draft of the manuscript.

In acknowledging the help I have received, I do

not wish to imply that this study is free of errors. For

them I take full responsibility, trusting that researchers

who follow me will correct my mistakes with the same zeal

that I have applied to correcting my predecessors, in

pursuit of the objective we all share--ever more accurate

descriptions of the Aymara language.



ACKNOWLEDGEMENTS ----------------------------------- iv

ABSTRACT ------------------------------------------- xii


1 INTRODUCTION ------------------------------- 1

1-1 Demography -------------------------- 1
1-1.1 Number and location of Aymara
speakers ---------------------------- 1
1-1.2 Status ------------------------------ 6
1-1.3 Monolingualism, bilingualism,
and multilingualisim ----------------- 7

1-2 History ----------------------------- 8
1-2.1 Language family --------------------- 8
1-2.2 Dialects ---------------------------- 11
1-2.3 Summary description of La Paz
Aymara ------------------------------ 15

1-3 The Present Study ------------------- 18
1-3.1 Theoretical bases ------------------- 18
1-3.2 Purposes and scope ------------------ 22
1-3.3 Methodology and data ---------------- 24
1-3.4 Sites and sources ------------------- 29
1-3.5 Organization of the study ----------- 35
1-3.6 Conventions and terminology --------- 36

2 A SURVEY OF THE LITERATURE ----------------- 43

2-1 Introduction ------------------------ 43

2-2 Colonial Period --------------------- 43

2-3 Prelinguistic Studies--19th and 20th
Centuries --------------------------- 52



2-4 Linguistic Studies ------------------ 62
2-4.1 Synchronic studies ------------------ 62
2-4.2 Historical studies ------------------ 78

2-5 Summary and Projection -------------- 78

SHAPE OF MORPHEMES ------------------------- 80

3-1 Introduction ------------------------ 80

3-2 Phonemes ---------------------------- 80
3-2.1 Phonemic inventory ------------------ 80
3-2.2 Allophones -------------------------- 83
3-2.3 Canonical forms --------------------- 86
3-2.4 Restrictions on phoneme occurrence -- 87

3-3 Nonphonemic Phenomena --------------- 89
3-3.1 Stress ------------------------------ 89
3-3.2 Intonation -------------------------- 90
3-3.3 Spanish loans ----------------------- 90

3-4 Phonological Correspondences Within
and Across Dialects ----------------- 91
3-4.1 Vowel correspondences --------------- 92
3-4.2 Consonant correspondences ----------- 100
3-4.3 Correspondences of vowels and/or
vowel length and nonstop consonants -- 127
3-4.4 Metathesis --------------------------- 132
3-4.5 Correspondence of final (C)CV
sequence with another CV sequence
or zero ------------------------------ 136
3-4.6 Correspondence of final /n(V)/ and
zero --------------------------------- 137
3-4.7 Correspondence of initial /ja/ and
vowel length or zero, and corre-
spondence of final /ma/ and zero ----- 139
3-4.8 Combinations of correspondences in
one word ----------------------------- 139

3-5 Conclusion --------------------------- 140


4-1 Introduction ------------------------- 145



4-2 Morphologically Determined Vowel-
Deleting and -Retaining Rules
(Morphophonemics of Suffixes) ------- 146
4-2.1 Phonological and morphological
conditioning ------------------------ 147
4-2.2 Morphological conditioning ---------- 148
4-2.3 Phonological and syntactic
conditioning ------------------------ 149
4-2.4 Morphological and syntactic
conditioning ------------------------ 149

4-3 Phonotactically Conditioned Rules
(Canonical Form Conditions) --------- 149
4-3.1 Word-initial position --------------- 150
4-3.2 Medial position --------------------- 154
4-3.3 Final position in morphological
word -------------------------------- 185
4-3.2 Final position in syntactical word -- 192

4-4 General and Dialect-Specific Rules -- 193
4-4.1 Variation in morphophonemics of
suffixes ---------------------------- 194
4-4.2 Variation in other morphophonemic
rules ------------------------------- 195

4-5 Conclusion -------------------------- 197

5 VARIATION IN THE NOUN SYSTEM --------------- 202

5-1 Introduction ------------------------ 202

5-2 Closed Classes of Noun Roots -------- 203
5-2.1 Interrogatives ---------------------- 203
5-2.2 Demonstratives ---------------------- 210
5-2.3 Personal pronouns ------------------- 215
5-2.4 Numbers ----------------------------- 216
5-2.5 Positionals ------------------------- 222
5-2.6 Temporals --------------------------- 226
5-2.7 Ambiguous noun/verb roots ----------- 243

5-3 Noun Suffixes ----------------------- 245
5-3.1 Class of limited occurrence --------- 245
5-3.2 Class 1 suffixes -------------------- 259
5-3.3 Class 2 suffixes -------------------- 274
5-3.4 Class 3 suffixes (verbalizers) ------ 300


5-4 Summary and Conclusions -------------- 309
5-4.1 Types of variation in the noun
system ------------------------------- 309
5-4.2 Dialectal patterning ----------------- 311

6 VARIATION IN THE VERB SYSTEM ---------------- 317

6-1 Introduction ------------------------- 317

6-2 Verbal Derivational Suffixes --------- 318
6-2.1 Class 1 suffixes --------------------- 321
6-2.2 Class 2 suffixes --------------------- 345
6-2.3 Class 3 suffixes --------------------- 374

6-3 Verbal Inflectional Suffixes --------- 384
6-3.1 Introduction ------------------------- 384
6-3.2 Verbal inflectional distinctive
features ----------------------------- 386
6-3.3 Tenses ------------------------------- 389

6-4 The Verb sa.da 'to say' -------------- 445
6-4.1 sa.ha with Simple tense -------------- 447
6-4.2 with Future tense -------------- 450
6-4.3 sa.ha with other tenses -------------- 453
6-4.4 Dialectal patterning ----------------- 454

6-5 Summary and Conclusions -------------- 455
6-5.1 Types of variation in the verb
system ------------------------------- 455
6-5.2 Dialectal patterning ----------------- 459


7-1 Introduction ------------------------- 468

7-2 Particles and Syntactic Suffixes ----- 469
7-2.1 Particles ---------------------------- 469
7-2.2 Syntactic suffixes ------------------- 473

7-3 Basic Sentence Types ----------------- 508

7-4 Morphosyntactic Processes ------------ 509
7-4.1 Reduplication ------------------------ 509
7-4.2 Subordination ------------------------ 517
7-4.3 uka linker and summarizer ------------ 562
7-4.4 sa.ha embedding ---------------------- 564
7-4.5 Negation ----------------------------- 576

7-5 Conclusion --------------------------- 587


8 VARIATION IN SEMANTICS --------------------- 592

8-1 Introduction ------------------------ 592

8-2 Linguistic Postulates --------------- 593
8-2.1 Four-person system ------------------ 594
8-2.2 Human/Nonhuman ---------------------- 597
8-2.3 Directly/Indirectly acquired
knowledge (data source) ------------- 631
8-2.4 A nonpostulate: Singular/Plural ---- 634

8-3 Semantic Variation in Roots and
Suffixes ---------------------------- 643
8-3.1 Noun system ------------------------- 643
8-3.2 Verb system ------------------------- 654

8-4 Metaphor ---------------------------- 667

8-5 Summary and Conclusion -------------- 672


9-1 Introduction ------------------------ 675

9-2 Phonology --------------------------- 679

9-3 Morphophonemics --------------------- 680

9-4 Morphology -------------------------- 681

9-5 Morphosyntax and Syntax ------------- 683

9-6 Semantics --------------------------- 693
9-6.1 Linguistic Postulates --------------- 693
9-6.2 Other semantic peculiarities -------- 706

9-7 Summary and Conclusion -------------- 712

10 CONCLUSION --------------------------------- 716

10-1 Dialectal Variation in Aymara ------- 716

10-2 Regional Dialect Groups and
Features ---------------------------- 718
10-2.1 Northern group: La Paz, Juli,
Socca, Huancand --------------------- 722
10-2.2 Southern group: Jopoqueri (and/or
Corque), Salinas, Morocomarca (and/or
Calacala) --------------------------- 724


10-2.3 Intermediate dialects: Calacoa
and Sitajara ----------------------- 727
10-2.4 Peripheral (as distinguished from
central) dialects ------------------ 728
10-2.5 Cross-regional features ------------ 733
10-2.6 Cross-dialectal perceptions -------- 735
10-2.7 Attitudes toward Aymara language
and culture ------------------------ 736

10-3 Interpretation of Research Results
and Their Implications ------------- 739

10-4 Directions for Future Research ----- 746


SENTENCES ---------------------------------- 752

B ONOMATOPOEIC PARTICLES --------------------- 782

COMMON EXPRESSIONS ------------------------- 784

RIDDLE ------------------------------------- 803

E INDEX OF SUFFIXES -------------------------- 811

REFERENCES ------------------------------------------ 835

BIOGRAPHICAL SKETCH --------------------------------- 847


Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Lucy Therina Briggs

August, 1976

Chairman: M. J. Hardman-de-Bautista
Major Department: Linguistics

The Aymara language is spoken on the high Andean

plains of Peru and Bolivia from the northern tip of Lake

Titicaca to the salt flats south of Lake Poop6. South-

west of Lake Titicaca it is spoken in the upper reaches

of some of the river valleys that descend to the Pacific

coast and to the east it extends into the subtropical

Yungas valleys, but its domain is primarily the altiplano.

The total number of speakers approaches two million, of

whom about one and a half million live in Bolivia and

the rest in Peru. There are also a few speakers in

northern Chile. Predominantly farmers or herders, the

Aymara have traditionally traded over a wide area, Aymara

women playing a major role in regional marketing of agri-

cultural produce.


Aymara belongs to the Jaqi family of languages

whose other extant members are Jaqaru and Kawki, spoken

in the department of Lima, Peru. The relationship of

Aymara to Quechua, the other major Andean language, is

undetermined. Dialects of Aymara have not hitherto

been systematically studied, although dialectal variation

has been known to exist since colonial times. The present

study was conceived to begin the task of determining the

extent and character of dialectal variation in Aymara.

Based on research in ten Aymara communities and

incorporating data from a survey of the literature from

colonial times to the present, this study examines

regional variation in phonology, morphophonemics, mor-

phology, syntax and morphosyntax, and semantics, and

three translation dialects not specific to any one region:

Missionary, Patr6n, and Radio Aymara. The appendices

include the elicitation list used in fieldwork, a list

of onomatopoeic particles, regional versions of greetings,

a brief dialogue, a saying, and a riddle, and an index

of suffixes.

The study confirms that all dialects share the

basic structures attributed to two La Paz dialects in

earlier studies by M. J. Hardman and associates at the

University of Florida. Aymara is a polysynthetic language

in which suffixes play not only morphological but also


syntactic roles and retain or lose their own or preceding

vowels according to complex morphophonemic rules. All

dialects also have certain linguistic postulates: a

system of four grammatical persons, a distinction of human

and nonhuman reference, and a distinction of direct and

indirect knowledge. All dialects are mutually (though

not equally) intelligible.

Regional differences occur primarily in phonology

and morphophonemics. Two dialects have a phoneme lacking

in the others, and there is considerable variety in phono-

logical shapes in morphemes stemming from phonemic insta-

bility and morphophonemic variation. Regional patterning

involves two overlapping distinctions: (1) a division

into northern and southern dialects (with two intermediate

dialects sharing some features of both), and (2) a division

into central and peripheral dialects reflecting the spread

of La Paz influence toward outlying areas that retain

certain features La Paz has lost. The dialectal picture

is further complicated by the existence of certain features

shared by a few dialects without regard to regional


While many La Paz innovations are attributable

to Spanish influence and all dialects of Aymara appear

to be adopting Spanish loans at an accelerating rate,

Aymara is a vigorous language that will survive due to

natural population increases for at least several genera-

tions. In the long run the future of the language will

depend on many factors, not least of which will be the

extent to which its speakers succeed in fostering its

use as a vehicle of literature and education.




1-1 Demography

1-1.1 Number and location of speakers

Aymara is spoken on the high Andean plains of

Peru and Bolivia from the northern tip of Lake Titicaca

to the Uyuni salt flats south of Lake Poop6 (see Figure

1-1). Southwest of Lake Titicaca it is spoken in the

upper reaches of some of the many river valleys that

descend to the Pacific coast, and to the east it extends

into the subtropical Yungas valleys, but its domain is

primarily the altiplano.

The majority of native speakers of Aymara today

are Bolivians and constitute approximately a third of

the Bolivian population (Hardman et al. 1975:3.2). As

the total estimated population of Bolivia in 1973 was

5.3 million (U. S. Department of State 1974:1), Bolivian

Aymara alone may account for well over a million and

a half speakers.

In Peru, according to the national census of 1961,

persons for whom Aymara was the first (maternal) language

'%. C
Cachuy (Kawki)
4. QTupe (Jaqaru) ..
I ".. / L A /

\ / / /.

'- \ IHuancane/ Vitocota/ \

Moquegua, Jesis e Macca /
SAPurea Where Aymara Is Spoken
0 200 AREQUIPA 10 200 L

Tarata/ uaraco
tud Tacna 'I, a anac
Moquegua / Jesis de Machaca6 /
Area Where Aymara Is Spoken I Sta n ar n c h
./Tarata/*V/ O/urv0
SPrimary Aymara Sites For This Study TacnD /
Corquc 0
Secondary Aymara Sites For This Study Arica /qP/

0 Other Sites Where Aymara Or '/9
Another Jaqi Language Is Spoken L.Coipasa OSalinas\ d
S/jGarcia M
l Certain Other Towns Or Cities Lic // //
Mentioned In Text salt Fl

International Boundaries \ 0
- *- Departmental Boundaries \ '

'%~.. ~


- .~ .-*.


iarca .- \



) 1 ^e

Figure 1-1. Area Where Aymara is Spoken




made up only 3.5% of the population five years old and

older, or 290,125 out of a total of 8,235,220 (Repdblica

del Perd 1966:4-45). According to the 1972 Peruvian

census, their number had grown to 332,593, although this

then constituted only 2.9% of the total population five

years old and older (Republica del PerO 1974:2.646).

Allowing for the inclusion of Peruvian Aymara children

under five, the total of Aymara speakers in Bolivia and

Peru today may be estimated as nearing two million.

The 1961 Peruvian census gave breakdowns of

Aymara speakers by department and province. Unfortu-

nately, such figures are not yet available for 1972.

The 1961 census indicated that of the total 290,125,

83.9% were in the department of Puno, in the provinces

of Puno, Chucuito, and Huancand. Of the rest, 8.4%

were in the interior highlands of the departments of

Moquegua (province of Mariscal Nieto) and Tacna

(province of Tarata), with the remaining 8.5% scattered

in the departments of Arequipa (provinces of Arequipa

and Islay), Puno (provinces of Sandia and San Roman),

Lima, and Cuzco.

For Bolivia, reliable statistics on numbers

and location of Aymara speakers are not available. The

majority of speakers are generally considered to be in

the departments of La Paz and Oruro. There are also

Aymara in the northern and western parts of the department

of Potosi and (Javier Alb6 and Walter Penaranda, personal

communications) along the western border of the department

of Cochabamba. The presence of Aymara throughout the

department of La Paz is well known although the north-

eastern provinces beyond the Cordillera Real (Larecajas,

Mubecas, Bautista Saavedra, and part of Camacho) are

shared with Quechua speakers, some villages being pre-

dominantly Quechua, others Aymara. The situation in

eastern Oruro is similar, with a preponderance of Quechua

as one approaches the Potosi border.

In northwestern Potosi between the departments

of Oruro and Cochabamba the linguistic situation is

complex. The mining centers just east of the Oruro-

Potosi border are Quechua speaking, but surrounding

towns, such as Calacala and Morocomarca (see Figure 1-1)

are often Aymara. In some of these, as in Calacala, the

younger generation is bilingual in Spanish and Quechua

rather than Aymara. Although persons over 15 are capable

of telling stories in Aymara they obviously prefer to

use Quechua; children under 12 do not understand Aymara.

The situation is like that noted by 0. Harris (1974)

in some other communities in the province of Bustillos

and in the province of Charcas, where Aymara is spoken

only at home; its use in public is met with embarrassment,

if not shame, and Quechua is used primarily in public or

with strangers. This situation suggests a kind of

diglossia, perhaps a relic of an earlier time when general

languages coexisted with the particular languages used in

each locale. Harris has also noted, however, that many

areas of PotosT traditionally considered to be Quechua

speaking are inhabited by Aymara-speaking groups for

whom Quechua does not appear to be becoming the dominant

language. For example, the valleys of San Pedro de

Buena Vista in northern PotosT and the area of Llica in

western Potost near the Uyuni salt flats are Aymara


In some cases, according to Harris, the designation

of a given ayllu (clan group) as Aymara or Quechua speak-

ing is inappropriate; language cuts across ayllu lines.

For example, Harris found that the Machas, who live near

the border of Chuquisaca department and are generally

considered to be Quechua speaking, speak Aymara in the

most remote part of their valley, near Carasi, province

of Charcas. The situation of the Laymis, on the other

hand, is the reverse: in the high puna near Uncla they

all know Aymara while in the remote parts of their

valley they speak Quechua.

These are examples of the complexities that need

further study to determine the exact areas where Aymara

is spoken today. The mobility of the Aymara must also

be taken into account. Predominantly farmers or herders,

the Aymara have traditionally traded over a wide area.

Aymara women play a major role in regional marketing

of agricultural produce. Aymara families that move to

the cities maintain close ties with their villages and

frequently own agricultural property at several ecologi-

cal levels, a system of vertical archipelagos that has

existed since prehistoric times (Murra 1968 and 1972).

Aymara is also spoken in the environs of Arica,

Chile, and is taught at the Universidad del Norte in

that city (Juan de Dios Yapita, personal communication).

It may also be spoken along the Chilean border of the

Bolivian department of Oruro. Whether the Aymara popula-

tion of Chile is native or predominantly of recent Bolivian

or Peruvian Aymara settlers also needs further clarifi-


1-1.2 Status

Both Bolivia (in 1970) and Peru (in 1971) have

recognized Aymara as a national language, together with

Quechua and Spanish, but this action has failed to alter

the social fact that monolingual Aymara speakers are

effectively barred from active participation in national

life (Hardman et al. 1975:3.2). This situation is offset

by the efforts of small but active groups of Bolivian

Aymara speakers to educate other Aymara and the public

at large on Aymara language and culture and to stimulate

the production of written literature in the language.

These efforts received considerable impetus from the

participation of members of the Bolivian Aymara community

in the Aymara Language Materials Project at the Univer-

sity of Florida (see 2-4.12) and have continued under

the leadership of the Instituto de Lengua y Cultura

Aymara (ILCA) and the Instituto Nacional de Estudios

LingUFsticos (INEL) in La Paz. The Centro Pedag6gico

y Cultural de Portales in Cochabamba has promoted the

development of teaching materials in Quechua and Aymara

and is involved with ILCA and INEL in sociolinguistic

surveys to determine speaker attitudes toward education

in the two languages.

The government of Peru has recently embarked on

an educational development plan which includes primary

education in vernacular languages for those who do not

speak Spanish. Programs in Aymara have yet to be

developed but the government is financing the translation

into Spanish of the Aymara teaching and reference grammar

produced at the University of Florida (Hardman et al.

1975) for use in training teachers of Aymara children.

1-1.3 Monolingualism, bilingualism, and multilingualism

Figures for Aymara monolingualism, Spanish-Aymara

bilingualism, or multilingualism of other types are

either lacking or untrustworthy. The 1961 Peruvian

census indicated that of the total of 290,125 speakers

of Aymara five years or older, 162,175 said they did

not speak Spanish when asked if they did and that 96%

of these were in Puno (Repbblica del Peru 1966:2-3).

Other official Peruvian sources differentiate coordinate

and subordinate bilingualism but the application of these

terms to actual cases varies. Redefinitions and refine-

ments of the terms monolingual and bilingual are needed

for the Andean situation, which includes such complexi-

ties as those of northern Potosi (1-1.1). Some persons

encountered in this research who were designated by

other Aymara speakers as monolingual appeared to have a

receptive if not productive competence in the Spanish

language within a narrowly defined set of topics. Other

persons who on first acquaintance appeared to be fairly

fluent in Spanish later proved to have many difficulties

in comprehension and production. The role of cultural

and social factors must also be taken into account.

(See the remarks for bilingualism of sources, 1-3.3.

Present-day Aymara dialects that show heavy Spanish

influence are discussed in Chapter 9.)

1-2 History

1-2.1 Language family

Aymara is a member of the Jaqi language family

(Hardman 1975) which Torero (1972b) prefers to call the

Aru family. Other extant members of the family are

Jaqaru and Kawki, spoken in Tupe and Cachuy, respectively,

in the highland province of Yauyos about 150 miles south

of Lima, Peru. Jaqaru is still vigorous but Kawki is

dying out.

Citing historical and toponymical evidence,

Torero (1972b) has established the probable extension

of this language family in the 16th century as from the

area of two present provinces of Lima, Huarochirf and

Yauyos, south to what is now southern Bolivia. Aymara

occupied the most extensive area, south and southeast

of the River Pampas basin in southern Huancavelica and

northern Ayacucho departments, while the other Jaqi

languages were spoken in a more restricted area to the

north. Hardman (1966:15) has reported evidence for the

existence of a Jaqi language in the valley of Canta

north of Lima in the early 1900's. According to Torero

the language family entered its expansive phase with

the rise of Huari (Ayacucho) and Aymara has since moved

south, taking over territories of other languages such

as Puquina. Torero has cited his and Hardman's

glottochronological calculations as indicating approxi-

mately 1,490 years of minimal divergence between Jaqaru

and Aymara beginning in approximately A. D. 480. By

this reckoning Kawki and Aymara are 1,130 years apart,

having diverged around A. D. 840.

On the basis of these dates and linguistic evi-

dence from Hardman of a closer linguistic relationship

between Kawki and Aymara than between Jaqaru and Aymara,

Torero has posited the following phases in the expansion

of the language family: (1) a first split in the fifth

century A. D. or before, (2) a second split in the

ninth century, and (3) a proto-Aymara period a few cen-

turies before the establishment of the Inca empire.

Tying these in with archaeological evidence, Torero has

identified the first phase with the beginning of Nazca

influence in the region of Ayacucho and the second phase

as occurring during the Vinaque culture centered in the

city of Huari, which controlled the area from Yauyos in

Lima to southern Cuzco and Arequipa between A. D. 500

and 1000. The third phase coincided with the third

stage of the Middle Horizon after the decline of the

important Vinaque centers (Torero 1972b:92,94,97).

With respect to the expansion of Aymara on the

altiplano, Torero has calculated the date of divergence

between the dialect of Moho in the province of Huancand,

department of Puno, and that spoken near La Paz, Bolivia

as about A. D. 1550. While holding that a comparison

of these dialects with one from southern Bolivia would

show a longer period of separation, he has tentatively

suggested that Aymara penetrated the area around Lake

Titicaca during the 13th century A. D. in the latter part

of his third phase (Torero 1972b:62-63).

1-2.2 Dialects

According to colonial and later sources cited

by Tschopik (including Cieza de Leon, Bertonio, Rivet,

and Markham), the following were independent Aymara

states existing before the Inca conquests and in

Tschopik's view probably also dialect groups (Tschopik









Pacasa or Pacaje

Caranga or Caranca



Vilcanota valley between Com-
bapata and Tinta (department
of Cuzco, Peru)

Between Tinta and Ayaviri
(department of Puno, Peru)

On the plains of the Pucara
and Ramis rivers as far as the
city of Puno, Peru

On the southwestern shore of
Lake Titicaca between Puno and
the Desaguadero River

North of Arequipa (Peru) on the
upper course of the Colca River

East of Arequipa in the upper
drainage of the Tambo River
(department of Moquegua, Peru)

South of Lake Titicaca along
both banks of the Desaguadero
River (Bolivia)

South of the Desaguadero River
to Lake Coipasa (Bolivia)

Northeast of Lake Poop6 near
Chuquisaca (Bolivia)

Quillaca or Quillagua Southeast of Lake Poop6

Omasuyo East of Lake Titicaca (Bolivia)

Collahuaya Provinces of Muiecas and
Caupolic6n (Bolivia)

Aymara is spoken today in the areas attributed

to the Lupaca, Pacasa or Pacaje, and Caranga or Caranca,

and in parts of the areas attributed to Omasuyo and

Collahuaya, other parts of which, like the other regions

cited above, are now largely Quechua speaking. Other

areas where Aymara is spoken today are not included above,

e.g. the present region of Mariscal Nieto province of

Moquegua and of Tarata province of Tacna, Peru. In any

case, determining where the early Aymara groups were

located is complicated by the ancient Andean system of

maintaining vertical archipelagos of colonies at differ-

ent ecological levels (Murra 1968 and 1972). For example,

the report of the administrative inspection visita

general) by Garci Diez de San Miguel of the Lupaca-

controlled Chucuito province in 1567 cited the names

of towns subject to Chucuito but located in the valleys

sloping westward to the Pacific and eastward to the Yungas

and the valley of Cochabamba. Among these towns were

the following, most of which have modern counterparts:

Moquegua and Torata (department of Moquegua), Sama and

Tarata (department of Tacna), Larecaja (department of

La Paz, east of the Cordillera Real), Chicanoma

(Chicaloma is a modern town in Sur Yungas province,

La Paz department), and Capinota (western Cochabamba

department) (Diez de San Miguel 1964:14,17,27,203;

modern departments and correspondences supplied).

According to the Relaciones geograficas de Indias

(colonial geographic reports), the Pacaxe (sic) also had

colonies interspersed among those of the Lupaca near

the Pacific coast (Jiminez de la Espada 1881:1.338).

The only detailed information so far available

as to the size of any Aymara-speaking population during

the colonial period is also for the Lupaca. The earliest

figures date from the visit general of 1549 which gave

a total of 18,032 heads of household in Chucuito province

(Diez de San Miguel 1964:202-203). The 1567 visit

found a total of 63,012 persons, children and adults,

of whom 15,047 were Urus; the figure of 63,012 was said

to include the population of the Chucuito colonies

mentioned above. Of the total, 15,404 were identified

as taxpayers. The principal cacique of the province,

Martin Cari, claimed an additional 5,000 taxpayers,

but Diez de San Miguel disputed this claim,saying that

the original figure of 15,404 already included 'many

Indians that the said caciques and heads of ayllus

declared they had in PotosF and La Paz and the province

of Charcas and other parts of these Kingdoms' (Diez de

San Miguel 1964:204-206). It may be hoped that as addi-

tional visits from the colonial period become available

to scholars (Murra 1970), more details for the populations

of the other Aymara nations may come to light.

With respect to the linguistic situation in

Chucuito and its colonies,Diez de San Miguel gave

little information. He used the term aymaraes to refer

to the people but not their language, recommending that

priests sent to the area remain long enough to learn

'la lengua colla' (Diez de San Miguel 1964:227).

Originally the name of one Aymara nation, Colla acquired

a wider connotation under the Incas after they designated

their southernmost province Collasuyu. According to

Tschopik (1946:503), Cieza de Le6n's Cr6nica, written

about 1550, used the terms Colla and Collao indiscrimi-

nately, and the use of the term Aymara to designate a

language first appeared in a relation of Polo de

Ondegardo of 1559. The term apparently did not come

into general use until the 17th century, however.

It is not clear from the visit of 1567 whether

the Urus, who were considered a separate ethnic-cultural

group from the Aymara, were nevertheless native Aymara-

speakers. Torero believes the Urus spoke a language

related to Chipaya but also spoke one or more of the

general languages of the area (Torero 1972b:60). Urus

living among the Aymara would speak Aymara but to what

extent or degree of native fluency is unknown. There is

some evidence that the Urus constituted a servant class

(Hardman, personal communication).

Differences among Aymara dialects have always

been considered minor from colonial times to the present.

The priests who went to the mines in Potosi to preach

and hear confessions had no trouble understanding Aymara

speakers from different provinces, according to the Jesuit

missionary Ludovico Bertonio (1612, A 2). Bertonio

(1603b and 1612) occasionally identified certain forms

as preferred by the Lupaca but not until recent times

have compilers of Aymara word-lists or grammars some-

times indicated the geographical origins of the forms

cited. The published literature gives no indication

whatever of social differentiation of dialects as dis-

tinguished from regional variation.

1-2.3 Summary description of La Paz Aymara

The most complete and accurate ethnographic

and grammatical description of Aymara to date, based

on that spoken in Compi and Tiahuanaco, two communities

near La Paz, Bolivia, is contained in Outline of Aymara

phonological and grammatical structure by Hardman et al.

(1975:3). The Outline describes Aymara as a polysynthetic

language in which suffixes and retention or loss of

vowels perform almost all grammatical functions.

Suffixes have complex but usually regular morphophonemics.

Some suffixes require a preceding consonant, others

require a preceding vowel, and others allow either;

some suffixes also determine the retention or loss of

their own final vowel. Syntactic units are signalled

by final suffixes and by morphophonemic vowel loss or

retention. Suffix order is usually fixed, as is word

order within the noun phrase; otherwise word order is

fairly free although some orders are preferred.

The phonemic inventory consists of three vowels,

vowel length, and 27 consonants, including plain,

aspirated, and glottalized stops and affricates,

as well as fricatives, nasals, laterals, glides, and

a flap or trill.

Morpheme form classes are roots and suffixes

which together form stems. Root classes are nouns,

verbs, particles, and a class of interrogatives cutting

across the others. Suffix classes are noun derivationall),

verb derivationall and inflectional), independent nonfinal

(occurring before inflection on verbs and before final

suffixes on nouns and particles), and final suffixes

(occurring on any word, after all other suffixes).

Verb roots are bound; nouns and particles are free.

Class change through verbalization and nominalization

(a special kind of derivation) is extensive and recursive,

creating verb and noun themes. Inflection, defined as

closing a root, stem, or theme to further derivation,

is limited to verbs and to one noun 'suffix', zero

complement vowel drop; noun case suffixes permit further

derivation. Accumulation of suffixes on one stem is common.

Morphosyntactic subordination is accomplished

by nominalization and use of final suffixes to mark

dependent clauses. Syntactic processes include use of

the demonstrative uka 'that' as linker and summarizer,

and sentence embedding with the reportive verb

'to say'.

Apart from these features, Aymara shares with

the other Jaqi languages certain linguistic postulates.

Hardman (in press a) has defined linguistic postulates


.. recurrent categorizations in [a]
language . the most tightly tied to
the perceptions of the speakers. . .
The most powerful . are those involved
in the obligatory grammatical system . .
Typically, a postulate is realized at
several levels . morphologically,
syntactically, and in the semantic struc-

The principal Jaqi linguistic postulates according to

Hardman (1972a) are a four-person system, a distinction

of human and nonhuman, and a distinction of direct and

indirect knowledge with respect to data source. These

three postulates are marked throughout Aymara structure

in morphology, syntax, and semantics.

1-3 The Present Study

1-3.1 Theoretical bases

Theoretical bases for this study are two: one

concerning the nature of language, and the other concern-

ing scientific description of a language.

A language or dialect is a system of interlocking

contrasts, or rules, used in social and cultural inter-

action by a given community. Like all natural phenomena

language is always changing. At any moment certain con-

trasts are being neutralized; certain rules are being

suspended temporarily or broken; new rules are being

created as people adapt language to their own needs.

Some rules are more resistant to change than others;

these include the linguistic postulates. But language

variation is a major fact of the nature of language:

variation within one idiolect, within one dialect, or

in a group of dialects. The extreme of language varia-

tion is the proliferation of languages that are mutually

unintelligible although perhaps still related and sharing

a number of rules. The point at which dialects become

separate languages is arbitrary, usually determined by

political rather than strictly linguistic considerations.

To investigate variation within a language is to

seek a more complete description of that language. In

a praiseworthy attempt to go beyond one-dimensional,

single-dialect descriptions of English, Labov and his

followers have developed the concept of variable rules

and a methodology to collect and analyze data reflecting

them, using statistical measurements. There is an

obvious need for such studies of other languages con-

ducted by native speakers trained in linguistic field

methods. In early stages of research, however, what is

needed are structural descriptions upon which to base

later studies in greater depth. Such descriptions,

while limited in accuracy and completeness and relatively

informal in presentation, may be considered scientific

if they meet certain criteria. These are (1) use of

sound field methods for gathering and recording not only

linguistic, but also relevant social and cultural data;

(2) collection of sufficient data to insure the identi-

fication of significant features; and (3) adherence to

analytical methods that respect the structure of the

target language or dialect and that base the description

on that structure.

The last criterion implies a willingness to

experiment with different models and to select models

that best fit the data, while avoiding the temptation

to choose for description only those aspects of a

language or dialect that lend themselves to description

in terms of currently popular models, or worse, to force

the data into a distorting mold. Ideally, this approach

requires the analyst to know how to use a variety of

models. Traditional phonemic, generative, or stratifi-

cational models might be used for phonology and morpho-

phonemics; structuralism or tagmemics might be appropriate

for morphology; generative or interpretive semantics,

case grammar, or Chafian models might be used for

syntax and semantics.

Unfortunately, questions of time and expediency,

as well as personal taste, limit the linguist's access

to different models. Moreover, theories and models have

a way of evolving into dogmas with schools of more or

less fanatic leaders and disciples who demand commitment

to one model, one terminology, one faith. The linguis-

tics scholar wishing to try different models must learn

to switch philosophies and metalanguages with the skill

of a simultaneous interpreter. Even then, he or she

often finds that communication across theoretical

boundaries is difficult, if not impossible. Theoretical

schools tend to draw circles to shut each other out;

few draw circles to take each other in (Edwin Markham,

Outwitted). But different languages and different parts

of a given language may call for different theoretical


For the best results different models should

be kept in mind at every stage of analysis. For early

stages of the investigation the discovery procedures

developed by Pike and Nida are still unsurpassed;

Hockett's item and arrangement and item and process

models are still relevant. Generative grammar has

refined the item and process model, enabling it to

account more adequately for relationships among rules.

The generative phonology model, focusing on distinctive

features that underlie or compose phonemes, can illumi-

nate and show in an easily grasped notation aspects of

phonology and morphophonemics that structural phonemics

is not so well prepared to handle; case grammar or

Chafian semantics may offer similar advantages lacking

in earlier models. Having a repertory of models or

conventions to choose from in presenting the rules

discovered enhances the likelihood that the linguist

will choose the model best reflecting the rules'


This point of view does not reject the

existence or the importance of language universals

or the need to search for ultimate truth. It merely

holds that at present, linguistic diversity, whether

among languages or within one language, is more interest-

ing than linguistic uniformity. By the same token, it

favors the encouragement of theoretical diversity as

ultimately leading to more accurate and therefore more

scientific language description.

1-3.2 Purposes and scope

As noted above, the existence of dialectal vari-

ation in Aymara was known in the 17th century, although

differences were dismissed as insignificant. The Aymara

examples cited by Hardman et al. (1975) include some

cases of dialectal variation between Compi and Tiahuanaco

(La Paz) but other contemporary published references to

Aymara-dialectal variation are extremely rare. This

study was conceived to carry forward the task of describ-

ing such variation and thereby to increase knowledge of

the language as a whole.

Using the description by Hardman et al. (1975)

as a basis for comparison, I decided to sample selectively

the Aymara spoken over a wide area (approximately

that shown by the shaded area of Figure 1-1), investigating

phonological, morphological, morphophonemic, syntactic,

and semantic variation with a view toward seeking answers

to the following questions: What is the extent of dia-

lectal variation in Aymara? Is it indeed minor, or does

it affect intelligibility? In what parts of the grammar

does variation occur? Does it occur within as well as

across dialects? What kinds of variation are most preva-

lent and/or significant? What are the major features

distinguishing dialects and the major dialect groups

identifiable on the basis of them? What of interdialectal

attitudes: are some dialects more prestigious than

others? What of the effects of the dominant language,

Spanish: is Aymara an 'oppressed language' showing

signs of decline, as suggested by Alb6 (1973a), or is

it vigorous and growing?

Apart from their intrinsic interest for linguis-

tics, answers to such questions would have a number of

practical applications. A full description of variation

in Aymara is needed for reconstruction of proto-Aymara

and of proto-Jaqi, a task already begun by Hardman.

Descriptions of areal features in conjunction with

information contained in colonial documents may enable

historians to reconstruct past population movements and

relationships among areas (Murra 1970:20). And now that

programs of bilingual education are being undertaken or

considered in Peru and Bolivia, there is a growing demand

for detailed descriptions of Aymara and other indigenous

languages to be used for developing teaching materials.

The field work for this study was conducted from

July to September 1972 and from March 1973 to January

1974. After returning to the University of Florida,

I reviewed the literature on Aymara from the 17th century

to the present, incorporating into the analysis of field

materials Aymara texts and grammatical information of

relevance to the present study. However, the focus of

this study is on Aymara as presently spoken in Peru and


1-3.3 Methodology and data

Methodology for the present study was based on

that of Pike (1947), Nida (1965), and Samarin (1967)

as interpreted and applied by Hardman. Two complementary

kinds of data were sought: (1) free texts recorded on

tape, and (2) materials obtained through an elicitation

list of words, phrases, and sentences presented orally

in Spanish to Aymara-Spanish bilinguals for translation

into Aymara.

Free texts included messages of greeting; tradi-

tional folk tales, riddles, songs, and sayings; and con-

versations among native speakers, or monologues, on such

topics as life in the community, festivals, local agri-

culture, education, illness, and other personal experi-

ences. Recordings were also made of a Baptist sermon

in Aymara, of Baptist and Catholic hymns, and of several

Aymara radio broadcasts in La Paz.

The elicitation list in Spanish was developed

to obtain a body of data readily comparable from one

site to another. Based on the longer Swadesh list, it

included words, phrases, and sentences originally

elicited in Aymara or one of the other Jaqi languages

in earlier research by Hardman and associates and sub-

sequently translated back into Spanish as spoken in the

Andean area. In the course of the field work for this

study, these materials were modified in order to elicit

already-identified Aymara grammatical categories and

syntactic structures, and individual lexical items show-

ing dialectal variation. Eventually the list was refined

to eliminate material not showing variation and to focus

on areas of differences. In areas having the velar

nasal phoneme,Jaqaru words containing it were added to

the list in an effort to elicit Aymara cognates.

Although the use of Spanish in the elicitation

list sometimes resulted in Aymara translations that

reflected Spanish syntactic patterns, this drawback was

minimized by deliberate inclusion of Andean Spanish

examples already paralleling Aymara structure, and

balanced by analysis of free texts recorded in Aymara.

Grammatical structures were also elicited directly in

Aymara. For example, verb tense paradigms were

elicited by changing subject and complement pronouns

once an example of a verb with person/tense suffix had

occurred. Because the purpose of the investigation was

to elicit variation, it was important to encourage use

of local forms which sources might tend to suppress if

La Paz dialect forms were used to elicit them; the use

of Spanish avoided this problem. For example, using

Spanish made it possible to elicit in each site a brief

selection of common remarks (see Appendix C). A somewhat

abbreviated version of the elicitation list is given in

Appendix A.

The free texts provided examples of previously

unattested forms or variations which were then sought

elsewhere. Although no attempt was made to obtain the

same folktales in different places, a few occurred more

frequently than others, providing readily comparable

data complementary to that obtained through the elicita-

tion list. All free texts were tape-recorded; in most

cases translations of the elicitation list were not.

A total of approximately five hours of tape-recordings,

plus another five hours of data transcribed directly

without recording, constitutes the basic corpus of this

study. (Included in the total are approximately 30

minutes kindly recorded on my behalf in Torata, Moquegua

by Francisco Gangotena and Carlos Saavedra.) Another

five hours of tape-recorded texts were used as background

only. (Included were some recordings made in Oruro and

northern Potosi by Javier Alb6 and generously made

available to me.) Several hundred hours were spent

in reviewing original transcriptions with one or more

native speakers, whenever possible the source of

the text or another resident of the same community.

(A text recorded in a trilingual area of Potosi was

checked with a Quechua speaker from Llallagua to see

whether a certain unusual form might be a Quechua loan,

but no formal attempt was made to compare Aymara texts

with Quechua.) Some texts were later checked with

speakers from other areas, revealing further dialect

similarities or differences. Transcriptions were then

exhaustively analyzed with regard to phonology, morphology,

morphophonemics, syntax, and semantics, including linguis-

tic postulates.

The key factor making possible my entry into

and acceptance in the several Aymara communities where

I conducted research was my previous training in field

methods and study of Aymara language and culture under

Hardman and the Bolivian Aymara linguists Juana Vasquez

and Juan de Dios Yapita. Carter (1972) has noted that

ethnographic research in a given community can succeed

only if it is desired by a leading member of the community;

this is true of linguistic research also. As the first

native speaker of Aymara to receive formal linguistics

training and to teach Aymara at universities in Bolivia

and in the United States, Yapita is such a leader. In

1972 he founded the Instituto de Lengua y Cultura Aymara

(ILCA) in La Paz to encourage the development of scholarly

research conducted by and for members of the Aymara

community. My initial field work in Bolivia was under-

taken with his approval and support and with the help

of persons who had been his students or were otherwise

associated with him; subsequently my work contacts ex-

tended to persons who knew him only by reputation.

In order to facilitate my research in Peru,

where his work was not known, Yapita provided me with

a two-minute tape-recorded message in Aymara conveying

his greetings on behalf of ILCA and the Bolivian Aymara

community. This message, supplemented by my explanation

of the purposes of my research and by my assurances that

its results would be made available to ILCA to advance

the study of Aymara language and culture, served to

create a very favorable climate for cooperation among

the Peruvian Aymara I met.

Although my command of spoken Aymara was rudi-

mentary, making it necessary for me to rely heavily on

Spanish as a contact language, my familiarity with

Aymara grammar and culture and my association with

community leadership enabled me to accomplish most of

my research goals.

Questions of ethics and the sources' rights to

privacy were considered during the research. In most

cases recordings were made with the participants' knowl-

edge; in the few cases when participants were not aware

that they were being recorded, the recordings were later

played for them with the offer to erase anything unac-

ceptable; this was never requested. Typed transcripts

of some of the recorded stories were later provided to

the tellers, and some have been published through ILCA

or the Aymara Newsletter (2-3.12) with due credit given

the authors. In order to respect sources' privacy,

recorded texts or transcriptions containing information

of a personal nature will not be made public without

the source's permission.

1-3.4 Sites and sources

Communities mentioned in this section may be

found in Figure 1-1, except as noted.

In Peru I collected data in the following com-


Huancand (province of Huancane, department

of Puno)

Juli (province of Chucuito, department of


Calacoa (province of Mariscal Nieto, depart-

ment of Moquegua)

Tarata and Sitajara (province of Tarata,

department of Tacna)

In addition, I recorded in Puno examples of the Aymara

of Socca (province of Puno).

In Bolivia I collected data in the following


Corque (province of Carangas, department of


Salinas de Garci Mendoza (province of Ladislao

Cabrera, department of Oruro)

Calacala (province of Bustillos, department of


Compi (province of Omasuyos, department of

La Paz)

Achocalla (province of Murillos, department

of La Paz; just south of the capital, and

not on Figure 1-1)

In addition, in Bolivia I obtained examples of Aymara

from the following communities although I did not visit


Jopoqueri (Carangas)

Morocomarca (Bustillos)

Jes6s de Machaca, San Andres de Machaca,

and Taraco (province of Ingavi, department

of La Paz)

Yapita and Vasquez furnished additional data from their

own dialects representative of the communities of Compi

and Tiahuanaco (Ingavi), respectively, as modified by

many years of residence in the city of La Paz. As

Aymara translator for the American Universities Field

Staff Film Project, Vasquez also provided examples from

the dialect of Vitocota (near Ayata in the province of

Munecas, department of La Paz) from the sound tracks

of four films made there in 1972.

Serendipity was largely responsible for choice

of sites; weather conditions, in particular as they

related to the state of the roads, determined the timing

of visits. In Peru I intended to visit areas of greatest

Aymara concentration as indicated in the 1961 census.

While attending a meeting of the International Linguistic

Association in Arequipa in March 1973, I met three young

Aymara men who responded to Yapita's recording and my

description of research interests. One was originally

from Juli and provided me with a letter of introduction

to his family there. Another was a school teacher in

Lima, originally from Huancand; I interviewed him in

Lima and visited his family in Huancand. The third

was from Socca, near Puno; although I was unable to

visit there, I did obtain several texts from him in


After the Arequipa meeting I went to Tacna

hoping to visit Aymara-speaking areas in Tarata province.

An Aymara taxi driver encountered by chance in Tacna

took me to the town of Tarata, where he helped me find

and interview Aymara speakers. Some months later, after

roads became passable, he took me to visit his mother

in the nearby town of Sitajara. Shortly thereafter

he drove me to Moquegua, where he helped me locate an

Aymara bus owner from Calacoa who arranged for me to

visit that community and stay with his wife's family.

In Bolivia I hoped to collect data in areas

where the Aymara was popularly thought to be different

from that of the city of La Paz and its environs. The

initial selection of areas to visit was made in consul-

tation with Yapita and with Javier Alb6, an anthro-

pologist with many years of residence and travel in

Bolivia, but as in the case of Peru, final choice of

sites was fortuitous. A former student of Yapita's

was my first source on the Aymara of Carangas. Through

him I met a teacher who arranged a visit to his brother's

family in Corque. Later, the teacher's wife invited me

to accompany her on a visit to her mother in Salinas

de Garci Mendoza (hereafter referred to as Salinas),

a trip postponed some months because of impassable roads.

With the help of a young woman related to the teacher,

I made contact with Aymara speakers in the Quechua mining

town of Llallagua, in northern Potosf, who invited me

to visit the nearby town of Calacala. In UncFa, near

Llallagua, I met through the local parish priest a young

Aymara man from the village of Morocomarca and inter-

viewed him in Uncia as time did not allow a visit to


As in the case of choice of sites, the selection

of sources (the term source is being used here in pref-

erence to the somewhat negatively loaded term informant)

was random. An attempt was made to obtain data from both

sexes, of different ages and educational levels ranging

from illiteracy to completion of several years of uni-

versity study. Occupations included market seller, certi-

fied school teacher, bus driver, farmer, student (ele-

mentary, secondary, normal school, or university), taxi

driver, housewife, university professor, administrator

of community development programs, and Baptist minister.

All sources 40 and under were bilingual in the sense of

being able to carry on an intelligible conversation in

Spanish, although in some cases their phonology was

heavily Aymara (see 1-1.3).

Below is a chart showing sex, age, and (for

sources over 40) knowledge of Spanish, the latter

determination based usually on a speaker's self-evaluation.

Ages shown are approximate as age was not obtained for

all sources. An average of six sources was consulted

for each major site, the numbers ranging from one to


0 19 20 39 40 59 60+ Totals

+Sp -Sp +Sp -Sp

M 3 16 5 1? 25

F 7 13 5 2 1 2 30

(The question mark refers to a speaker who was not

heard to speak Spanish and was not asked if he did.)

As may be seen, a preponderance of sources were

under 40 and hence by definition bilingual in Spanish.

(Sources from Calacala and Morocomarca were trilingual

in Spanish, Aymara, and Quechua.) Monolinguals were

relatively less accessible to me than were bilinguals

primarily because lack of time and difficulties of

travel prevented visits to remote communities where

monolingualism is reportedly widespread and secondarily

because most monolinguals encountered in the communi-

ties visited were elderly and infirm, often with missing

teeth and consequent faulty diction. Determination of

the location of communities with a high proportion of

monolinguals of different ages must await future

research, preferably with the participation of native

speakers trained in linguistic field methods.

With regard to training of native speaker

linguists, an informal attempt was made throughout

this research to instill in sources the basic concepts

of anthropological linguistics and field methods, by

example if not in formal classes. For example, two

young women from urban centers were taken on field trips

to act as interpreters and to learn the basics of

informant-investigator relations. One source who

already had a firm grasp of the Yapita phonemic alphabet

(3-2) was asked to transcribe a tape-recorded story from

a dialect other than his own, at whose telling he had

been present. His mistakes in transcription were sig-

nificant in showing points of difference between the

two dialects and, when brought to his attention, made

him aware of the ways that one's own language or dialect

grid may structure one's perception. Throughout the

research I maintained and encouraged in all persons with

whom I worked an attitude of respect or even enthusiasm

toward the diversity that became apparent, noting, however,

their occasional linguocentric comments (see 10-2.6).

1-3.5 Organization of the study

This study is organized into chapters on the

following topics: a survey of literature on Aymara

from the colonial period to the present; variation in

phonemics and in phonological shape of morphemes within

and across dialects; variation in morphophonemics, in-

cluding rules general to all dialects and rules limited

to certain dialects; variation in the noun system, in

the verb system, and in morphosyntactic and syntactic

structures and processes; variation in semantics, including

a section on the nonvarying linguistic postulates; three

translation dialects: Missionary, Patr6n, and Radio

Aymara; and a conclusion summarizing kinds of variation,

identifying regional dialect groups on the basis of sig-

nificant variation, and offering suggestions for future

research. Following the chapters there are Appendices


as follows: (A) the elicitation list used in field work;

(B) a list of onomatopoeic particles; (C) regional versions

of greetings and certain common expressions; (D) regional

versions of a saying and a riddle; and (E) an index of


1-3.6 Conventions and terminology

In this study the following conventions are

observed, conforming in most cases to those used by

Hardman et al. (1975).

Aymara examples are usually written in the

Yapita phonemic alphabet (see 3-2) modified by the use

of a colon (:) rather than a dieresis mark (") for vowel

length in order to permit separation of morphemic length

from the vowel it occurs on. Occasional examples are

given in phonetic transcription within square brackets

([]). Place names are spelled as they appear on maps

for ease of reference although users of the Yapita

alphabet prefer to spell them according to its rules,

e.g. Qumpi for Compi. Aymara examples from published

sources other than Hardman and associates are usually

converted to Yapita orthography. All examples not

attested by Hardman et al. (1975) or later reported

by Vgsquez, Yapita, or me are preceded by a raised

cross ( ); morphemic as well as phonemic analyses of

such forms are mine. An asterisk (*) before an example

means it is unattested; context will indicate whether it is

rejected by native speakers or presumed to exist on the

basis of other evidence.

Periods separate morphemes within a word, e.g.

uta.xa 'a/the/my house.' Unsuffixed bound roots (verbs)

are followed by a hyphen, e.g. juta- 'come.' In citation

form, suffixes which may close a construction are pre-

ceded by a hyphen, e.g. the final suffix -xa. Other

suffixes are preceded and followed by a hyphen, e.g.

the verbal derivational suffix -t'a-. Recurrent sub-

morphemic partial are placed within hyphens and within

slants, e.g. /-pa-/. A lowered v before a suffix indi-

cates it must be preceded by a vowel; a lowered v after

a suffix indicates it must retain its own final vowel

when followed by another suffix; a lowered c before a

suffix indicates it must be preceded by a consonant; a

lowered c after a suffix indicates it must drop its

final vowel before a following suffix and/or when it

occurs word-finally. For example, in most dialects

the possessive/locational suffix -vnac is preceded by

a vowel but drops its own vowel word-finally and before

following suffixes, e.g. uta.n, uta.n.xa 'in/of the

house.' When a suffix may be preceded by either a vowel

or a consonant the more common (or base) occurrence is

indicated above a tilde (~) and the less common below

the tilde, e.g. ja first person possessive suffix.

A similar notation after a suffix indicates it may in

some circumstances keep and in others lose its own final

vowel. If no subscript v or c follows a suffix, either

other factors determine the retention or loss of the

final vowel or the rule has not yet been determined.

Grammatical persons are indicated by the numbers

1, 2, 3, and 4, usually followed by a p, e.g. -ja 1p

possessive suffix, -ma 2p possessive suffix. Verbal

inflectional suffixes have the subject person on the

left and the complement person on the right of a right-

pointing arrow, e.g. 1+2 means Ip subject, 2p complement.

Verb tenses are sometimes abbreviated as S

(Simple), F (Future), I (Imperative), RDK (Remote

Direct Knowledge), RIK (Remote Indirect Knowledge),

D-l (Desiderative), D-2 (Remonstrator), IF (Inferential),

and NI (Non-Involver).

Aymara examples are underlined when treated as

base forms or words (that is, as morphemes or combina-

tions thereof):

juma.mpi 'with you' juma 2p pronoun

-mpi 'with'

S. the suffix -jama and variants .

Jupa.x wali suma jaqi.wa. 'He/she is a good

Aymara examples are placed between slant lines when

treated as allomorphs:

S. /-mpi/ occurs in La Paz, /-nti/ in Salinas

. The suffix -jama has the allomorphs

/-jama/, /-:ma/, /-ja/ . .

What may sometimes appear to be inconsistency with re-

spect to this notation will be due to the fact that a

morph that at the individual dialect level is a base

form (morpheme) may be considered an allomorph of a mor-

pheme at the supradialectal level and/or in another

dialect. In such cases context will dictate whether the

morph is to be treated as an allomorph or as a morpheme.

If there is only one invariant allomorph for all dialects,

it will always be underlined unless given in phonetic

transcription. If there is one allomorph that occurs

almost everywhere (even if some dialects have others),

it will be considered the base form of the morpheme in

question and underlined, e.g. the final suffix -xa

which has the allomorph /-:/ (vowel length) in some

areas. If a morpheme has several allomorphs, they may

all be referred to at once in alphabetical order and

underlined, e.g. the suffix -mpi ~ -nti, the suffix

-taki -tak"i ~ -tay.

Examples are glossed in one of two ways: (1)
If the example is short or for added clarity (for

example when a morpheme occurs without its final vowel)

individual morphemes are glossed to the right:

jani.w 'no' jani 'not' -wa final suffix

(2) If the example is long, it will be glossed

beneath and followed by a free translation of the whole


Kuna.r un.ta.t sar.naq.ta.xa,
anything look at go 2+3
around S

jaqi.tak p'inqa, anu.tak unra.
people shame dog honor

'How stupid you are, you are a shame to the

human race, an honor to dogs.' (La Paz/Compi)

(Yapita 1975:3)

As shown here, examples from Yapita are identified as

being from La Paz/Compi; similarly, material obtained

from V6squez is identified as being from La Paz/Tiahuanaco.

Thus is noted the fact that both have lived many years

in the city of La Paz although their dialects are

basically those of their communities of origin.

Noun suffixes and verbal derivational suffixes

are identified in either (or both) of two ways: (1)

by a term describing the function of the suffix, e. g.

-mpi agentive/instrumental, or (2) by a gloss, e. g.

-mpi 'with, by'. Function terms are not placed with-

in single quotes, while glosses are. The function term

may be a neologism like distance (used for the verbal

derivational suffix -waya-); such terms are those used

by Hardman et al. (1975).

Aymara sentences, defined by the presence or

absence or certain final suffixes, have the first word

capitalized and usually have a period, question mark,

or exclamation point at the end. Often the final suf-

fixes in Aymara convey semantic and emotional overtones

which are conveyed in spoken English by intonation and

in written English by punctuation. When a question mark

or exclamation point would be superfluous in Aymara

they are usually omitted even though the English gloss

may carry them. Aymara examples that are not sentences

in Aymara may in some cases be translated by sentences

in English but in such cases the Aymara punctuation will

usually be adhered to in the gloss as well as in the

Aymara example unless meaning would thereby be obscured.


kama.cha.ta.:.rak.pacha.:t"a 'what could have

happened to me'


Embedded quotes are shown within angled brackets:

s.i.way. '"They are crying,"
cry 3-3 sayThey are
S they say.'


Spanish words occurring in Aymara sentences

are written as Spanish if they were so pronounced, e. g.

content 'happy, content'. If they were phonologically

adapted to Aymara (Aymarized) they are written as Aymara,

e. g. kuntintu. In some cases decisions whether to treat

a given word as Spanish or Aymara were arbitrary, and a

few hybrids occurred, e. g. Pedru.



2-1 Introduction

This chapter discusses the contents and merits

of selected works relating to Aymara dating from the

colonial period to the present. The two major biblio-

graphical sources for works on or in Aymara are (1)

Bibliograffa de las lenguas quechua y aymara by Josd

Toribio Medina (1930), and (2) the monumental four-volume

Bibliographie des Langues Aymard et Kicua by Paul Rivet

and Georges de Crdqui-Montfort (1951-1956; henceforth

Rivet). To my knowledge no bibliographical work

specifically on Aymara has been published since Rivet's

fourth volume (1956).

2-2 Colonial Period

As is well known, the Spanish found no written

materials in the languages of the Inca Empire. In the

16th and early 17th centuries all works published in or

on Aymara were written for the purpose of spreading the

Christian faith by missionaries assisted by unnamed Aymara

converts bilingual in Aymara and Spanish. Such works

consisted of catechisms and other religious tracts and

of grammars to be used by missionaries wishing to learn

to speak and understand the language. The earliest work

known to contain Aymara is the anonymous Doctrina

christiana, y catecismo para la instrucci6n de los Indios,

published in Lima in 1584 (Rivet 1951:4-9).

According to Rivet (1956:631) a study of the

early Catholic evangelization of Peru from 1532 to 1600

and the use of Aymara and Quechua as languages of con-

version is Cristianizaci6n del Pert (1532-1600) by

Fernando de Armas Medina (1953). Two other publications

useful for information about Aymara society in the 16th

century are Visita hecha a la provincia de Chucuito . .

en el aNo 1567 by a colonial administrator, Garci Diez

de San Miguel (1567), reporting on his inspection of

Chucuito province (see 1-2.2), and an ethnological

appraisal of the Diez de San Miguel inspection, Una

apreciaci6n etnol6gica de la visit by John V. Murra


The first attempt at a complete grammar of Aymara

was written by Ludovico Bertonio in the early 17th

century. Born in 1552 in Italy, Bertonio joined the

Company of Jesus in 1575. Sent to Peru in 1581, he

remained there for 44 years, dying in Lima in 1625 or

1628 (Rivet 1951:26-27). Bertonio apparently spent most

of his time in Juli, capital of the Aymara-speaking

Lupaca kingdom, on the shores of Lake Titicaca. There

he wrote three grammars of Aymara, a Spanish-Aymara/

Aymara-Spanish dictionary, and several Aymara transla-

tions of religious texts. In 1603 he published two

grammars, an Arte breve de la lengua aymara (1603a) and

an Arte y grammatica muy copiosa de la lengua aymara

(1603b). A facsimile edition of the latter was published

in Leipzig in 1879 by Julio Platzmann (Rivet 1953:35).

Juan de Dios Yapita owns a volume containing the first

14 pages of the Arte breve (Bertonio 1603a) bound

together with pages 19 through 348 of the Arte y

grammatica muy copiosa (Bertonio 1603b); the latter is

missing the title pages, a section entitled Al lector,

and pages 207 and 208. A photocopy of the Yapita volume

is in the library of the University of Florida.

In his longer grammar of Aymara (1603b), Bertonio

gave a detailed description of the language in terms

of Spanish and Latin. Spanish spelling is adapted

(inconsistently and inaccurately) to Aymara sounds,

Spanish grammatical categories are translated into the

nearest Aymara equivalents, and Spanish sentences are

rendered into Aymara. The grammar is valuable not only

as an example of the Latinate grammars of American

languages written during the colonial period but also

for the wealth of material it provides on the Aymara

language of the Lupaca kingdom in the early 17th century.

These data and Bertonio's analyses must be carefully

reinterpreted, however, in the light of techniques of

contemporary linguistic scholarship and recent discoveries

concerning Aymara language and culture. A review of

individual forms attested by Bertonio (for example, verb

derivational and inflectional suffixes) shows many forms

identical with some in general use today, others in use

in only one or a few present-day Aymara dialects, and

still others not attested in modern Aymara but extant

in other Jaqi languages. In some cases the semantics

of a form have shifted since Bertonio's time, if his

translations may be taken as accurate. But that is a

problem: reviewing Bertonio's Aymara sentences with

Juana Vdsquez has revealed that most of them are unac-

ceptable. At best they sound translated; at worst they

are perceived as simply incorrect even when archaic or

unknown terms are replaced by contemporary terms. The

book contains no native Aymara texts--no sentences form-

ing a narrative that might have been spoken in the

language by native speakers--but only translations from

Spanish or Latin of isolated words, phrases, or sentences.

Bertonio's grammatical analysis of the language

missed many important features because of its focus on

Spanish categories. Nevertheless, Bertonio was a care-

ful observer and tireless organizer of his material.

On the morphological level his analysis is often accurate

in detail. For example, his grasp of the inclusive/

exclusive distinction in the Aymara person system is

essentially correct (Hardman 1972b). With respect to

syntax and cultural content, however, Bertonio's gram-

matical analysis must be characterized as distorted and


According to Rivet (1951:52-53), there exists a

third Aymara grammar by Bertonio, Arte de la lengua aymara

(1612), containing sentences in Aymara and Spanish and

a list of Aymara words; the only known copies are repor-

tedly owned by the Posnansky family in La Paz and by the

Biblioteca Nacional in Sucre, Bolivia.

In 1612 Bertonio published his Vocabulario de la

lengua aymara which has since appeared in several fac-

simile editions, most recently in La Paz, Bolivia in

1956. This lengthy book is a dictionary, the first part

(474 pages) Spanish-Aymara and the second (398 pages)

Aymara-Spanish, with approximately 50 entries to a page.

A thorough study of this book with native speakers is

long overdue to determine how many and which forms are

in use today and to correct errors evident in a sampling

of the entries.

LaBarre (1948) (see 2-3) performed the useful

service of culling out and repeating, with English trans-

lations, some categories of words in the Vocabulario such

as kinship terms, diseases, and sins to be reported in

confession. Checking these with VAsquez revealed that

many terms cited by Bertonio are perceived today as

awkward translations of Spanish terms into Aymara rather

than as native words or expressions. In this connection

it is interesting to note that in the introductory section

to the Vocabulario Bertonio indicated that he took the

entries (1) from Aymara translations of the lives of

Christ and the saints, sermons, comparisons of vices and

virtues, and so forth written by certain Aymara brought

up as Christians during the 35 years that the Jesuits

had been in Juli, and (2) from similar materials collected

by other priests. The entire dictionary, in other words,

was based on materials translated from Spanish into

Aymara, not the other way around.

An example with respect to kinship is illustrative.

In modern Aymara the nominalized verb 'to carry'

may also be used with the metaphorical meaning of 'con-

temporary', 'of the same age' (i.e., a person carried by

his/her mother at the same time as another person was

similarly carried). Bertonio translated it (or the

derived term 'having a contemporary') as

'relative' (Spanish pariente), and used it to translate

sentences like 'If that woman is your relative you may

not marry her.' The present-day meaning of the Aymara

sentence (see 6-3.34.13) is 'If that woman is your

contemporary you may not marry her.' It seems unlikely

that a semantic shift has occurred with this word since

the 17th century. Rather, the Bertonio gloss probably

reflects an initial difficulty in translation when the

Aymara interpreter, having no one term in his language

for 'relative', finally approximated it with apa.Ra.

The bafflement of the Aymara at hearing an injunction

to marry only someone older or younger may only be

imagined. Other such translation errors or distortions

have contributed to the development of the translation

dialects Missionary and Patr6n Aymara (Chapter 9).

Such errors may well have contributed to the

difficulties the missionaries encountered in their

efforts to convert the Aymara. Bertonio acknowledged

such difficulties in the Vocabulario in a section

addressed 'to the priests of the Aymara Nation'. Deny-

ing that Aymara was a difficult language (he said the

Jesuits in Juli learned to preach in the language in a

year), he conceded that students of the language tended

to become disheartened, discerning in the Aymara a low

capacity for learning and a strong resistance to con-


S. they are so given to bad customs,
their hearts are so full of spines and
thistles, that the seed of the divine
word planted there will not bear fruit
. (Bertonio 1612:unnumbered page
facing A 3; English translation mine)

The stubborn refusal of the Aymara to be con-

verted, in spite of the best efforts of gifted and ener-

getic priests like Bertonio, was attributed then and

later to incapacity coupled with sheer cussedness.

Recent discoveries with respect to Aymara linguistic

postulates (see 8-2) have put the Aymara recalcitrance

in a new light. In any event, negative stereotypes of

the Aymara character had by the end of 35 years become

fully accepted among the colonizers and were to persist

well into the 20th century (see 2-3).

Very similar to Bertonio's work although shorter

is a grammar by another Jesuit assigned to Juli, Diego

de Torres Rubio, whose Arte de la lengua aymara appeared

in 1616. A photocopy, the original of which belongs

to the Instituto de Lengua y Cultura Aymara in La Paz,

is in the University of Florida Library. The photocopy

and the original lack pages 65 through 68 and pages 72

through 77. The volume contains following the Arte

the complete Catecismo en la lengua espanola y aymara

del Piru originally published in Sevilla (1604) on the

basis of materials dating from a Provincial Council in

Lima in 1583. According to Rivet (1951:75), several

known copies of the Torres Rubio 1616 grammar are bound

with the Catecismo in this manner.

It is not known whether Torres Rubio and Bertonio
collaborated or worked independently. They were almost

the same age and had similar careers. Torres was born

in 1557 in Spain, joined the Jesuits in 1572, and arrived

in Peru in 1579. He died there in 1637 or 1638 (Rivet

1951:71). In 1967 Mario (to be distinguished from

Alejandro) Franco Inojosa published a version of the

Torres Rubio Arte in modern Spanish, giving the Aymara

in Torres' original spellings followed by transcriptions

in the official Peruvian alphabet for Aymara and Quechua

adopted in 1946.

After the middle of the 17th century the early

fervor of missionary activity subsided and for the next

hundred years little was published in Aymara except

occasional sermons, few of which have survived. As

described by Tovar (1961:186-194) the alternating lin-

guistic policies of the Spanish conquerors help explain

the relative dearth of materials published in Aymara

between the second half of the 17th century and the

late 18th century. In 1550 it was decided to teach

the Indians in Spanish. As this attempt proved unsuc-

cessful, in 1583 the policy of using native languages

was adopted, stimulating the production of grammars and

religious texts in those languages. By 1596 the earlier

policy was reinstated over the missionaries' objections.

The impasse was resolved in practice by the use of gen-

eral languages which at first included Aymara although

during the 17th century it gave way to Quechua. By the

late 18th century the Spanish crown had expelled the

Jesuits from Peru and shortly thereafter the crown re-

stored the Spanish-only policy. Nevertheless, the wealth

of material on New World languages which the Jesuits

had gathered soon began to be published in Europe, mostly

in the form of comparative vocabulary lists. From that

time on the amount of published Aymara material gradually


2-3 Prelinguistic Studies--19th and 20th Centuries

Prelinguistic studies are those written without

benefit of the theories or techniques of modern linguis-

tic scholarship or dealing primarily with other than

linguistic aspects of Aymara culture.

The great European philologists Hervas, Vater,

Adelung, Pott, and J6han, writing in the late 18th and

early 19th centuries, included in their encyclopedic works

references to Aymara taken from earlier sources and super-

ficial comparisons of Aymara words with those of other

languages. In the second decade of the 19th century

political speeches and documents relating to the inde-

pendence movements in South America were published in

some native languages including Aymara.

In 1826 appeared the first Protestant materials

in Aymara, translations of the New Testament. From then

on a series of such translations emanated first from the

British and Foreign Bible Society and subsequently from

the United States. Catholic materials (mostly by Bolivian

priests) began to appear in greater numbers also. Later

in the 19th century there began to appear accounts by

European scholar-adventurers of their travels on the

Bolivian and Pervuian altiplano. These usually included

grammatical sketches of Quechua and Aymara or word lists

of numbers, animals, plants, medicines and diseases, and

kinship terms.

The first detailed ethnographic account of the

Aymara was On the Aymara Indians of Bolivia and Peru

by David Forbes (1870), based on research done in Bolivia

and Peru from 1859 to 1863. A British mining engineer

of scholarly bent and the stamina necessary to remain for

long periods at altitudes up to 15,400 feet, Forbes was

best at concrete measurement and description of the

activities he witnessed. His account of the Aymara was

somewhat sympathetic, revealing the relentless physical

hardships and social injustices they suffered, but some

of his explanations for Aymara behavior suggest he may

have given too much credence to tales spread by whites

and mestizos based more on myth than on fact.

Forbes gave Aymara names for objects, activities,

and the like most of which, though deformed by an inade-

quate transcription, are recognizable today. His

grammatical analysis of Aymara is sketchy but accurate

so far as it goes. Appendix C of his book is a vocabu-

lary of Aymara words, including kinship terms, with

English translations. Forbes cast light on the status

of Aymara studies at the time in remarking on his fruit-

less efforts while in Bolivia to obtain a copy of a 17th

century Aymara grammar or dictionary even though he had

advertised in the papers that he would pay the 'high

sum of 50 dollars' (274, fn.) for it.

In 1891 the German physician-turned-philologist

Ernst Middendorf published Die Aimard-Sprache, the fifth

volume of his study of aboriginal languages of Peru

(Rivet 1952:558). The introduction to Middendorf's Aymara

grammar was translated into Spanish by the Bolivian scholar

Franz Tamayo in an article published in 1910 in La Paz

(Rivet 1952:558). Later, the Peruvian scholar Estuardo

Ndiez, working from an incomplete copy of the Tamayo

translation, revised and added some notes to it and

published it in a volume entitled Las lenguas aborigenes

del Perd (1959) prepared under the auspices of the Univer-

sidad Mayor de San Marcos in Lima to commemorate the 50th

anniversary of Middendorf's death. The following section

refers to that volume (1959:96-102).

Middendorf indicated that his grammar was based

on Bertonio's and on the dialect then spoken in La Paz,

which he visited on several occasions. He stated that at

that time both whites and mestizos spoke Aymara but in

most cases only as a lingua franca for communication with

Aymara servants or sellers in the marketplace. Middendorf

was able to find only a few persons with enough knowledge

to teach him the language. Like Forbes, he found no one

who possessed a Bertonio grammar, adding that no one had

even heard of such a book, not even the President of

Bolivia or members of his cabinet. Middendorf was finally

able to find some lawyers who had lived among rural

Aymara and claimed to know more of the language than the

city-dwellers. With them Middendorf reviewed a copy of

the Bertonio grammar in his possession, comparing forms

then in use with earlier ones, noting both, and using

them to draw up rules of sentence formation. In the

introduction to his Aymara grammar he devoted several

paragraphs to Aymara vowel-dropping, giving examples

of inflected verbs, and commented on Aymara verbs of

going and carrying. It is to be hoped that someday

Middendorf's grammar of Aymara may be translated into


In 1917 another Aymara grammar based largely

on Bertonio's appeared, by Juan Antonio Garcfa, a Bolivian

priest. Subsequently, etymologies and word lists for

such topics as kinship, place names, and musical instru-

ments proliferated, and a number of Aymara stories, poems,

and legends were written by self-styled Aymara scholars

(aymar61ogos). Novels on Indian themes, such as Alcides

Arguedas' Raza de bronce (1945), contained some Aymara

phrases. There was a continuous spate of dictionaries

or handbooks of Aymara, Quechua, and Spanish words and

phrases, as well as both Catholic and Protestant publica-


The matter of a standardized alphabet for Aymara

and Quechua has engaged the sporadic attention of scholars

and governments for years. In 1939 the Twenty-Seventh

International Congress of Americanists proposed an

alphabet for Aymara and Quechua which was adopted by offi-

cial Peruvian government decree in 1946 (Rivet 1956:265).

In 1954 the Bolivian government adopted a virtually

identical alphabet approved earlier that same year by

the Third Inter-American Indigenist Congress (Rivet 1956:


Catholic missionaries on the altiplano adopted

this alphabet. It represents an improvement over earlier

ones in that it shows phonemic vowel length; distinguishes

plain, aspirated, and glottalized stops and an affricate

in the proper articulatory positions; and distinguishes

the velar and postvelar fricatives. But it uses the

five Spanish vowels to represent the three phonemic Aymara

vowels and allophones of two of them which are not always

predictable from the environment, unnecessarily confusing

the transcription.

Meanwhile, Protestant missionaries developed a

variation which employs some Spanish letters, such as c

and qu to represent the Aymara velar stop, in the belief

that their use makes it easier for the Aymara to learn

Spanish. This alphabet, known as the CALA alphabet for

the first initials of the Comisi6n de Alfabetizaci6n y

Literature Aymara (Aymara Literacy and Literature Com-

mission), was adopted as official by Bolivian government

decree in 1968, apparently without rescinding support of

the earlier alphabet. Since then the two official Bolivian

alphabets have coexisted in uneasy competition.

Beginning in the 1930s American anthropologists

turned their attention to the study of Aymara society.

The Aymara (1946) and The Aymara of Chucuito, Peru

1. Magic (1951) by Harry Tschopik and The Aymara Indians

of the Lake Titicaca Plateau, Bolivia (1948) by Weston

LaBarre are generally considered classics, but research

in the last decade has shown them to be deficient in

important respects, based as they were on data obtained

through mestizos and whites. A more balanced account

is The Aymara of Chinchera, Peru: Persistence and change

in a bicultural context (1964) by John Marshall Hickman,

reporting on a Peruvian Aymara community near that studied

by Tschopik 20 years earlier; Another look at Aymara

personality (1966) by John S. Plummer questioned earlier

negative assessments of the Aymara character. William E.

Carter has conducted extensive studies among the Bolivian

Aymara in the department of La Paz. His Bolivia, a profile

(1971:89-91) brought together the various expressions of

the Aymara negative stereotype and put them in historical

perspective. The Bolivian Aymara (1971) by Hans C. and

Judith-Maria Buechler is a somewhat superficial network

analysis of the community of Compi on Lake Titicaca.

The languages of South American Indians (1950) by

John Alden Mason contains a short section on the Aymara

language, but it is full of inaccuracies, not only with

respect to the supposed relationship of Aymara to other

languages, but also to identification of Aymara-speaking

areas and dialects. Catflogo de las lenguas de America

del Sur (1961) by Antonio Tovar represents a slight

improvement in the information provided but the work is

still incomplete and inaccurate and the brief grammatical

description of Aymara is very weak. Other publications

on Aymara well into the 1960s testify to the sorry state

of scholarship with respect to the language.

Characteristic are the many virtually identical

handbooks or catalogues of common expressions in Aymara,

Quechua, and Spanish published in Bolivia, Peru, and

even Chile from the middle of the 19th century to the

present. (The latest to come to my attention is dated

1971, but new editions have probably appeared since then.)

These little books contain the kind of Aymara spoken by

white and mestizo landowners who learned to speak the

language imperfectly as children and whose attitude

toward the Aymara people and their language ranges

from kindly but patronizing to contemptuous. This kind

of Aymara is referred to as Patr6n Aymara (from the

Spanish patr6n 'owner') by Bolivian Aymara native speakers.

In the catalogues individual forms may be correctly trans-

lated but Spanish phrases are translated word for word

into Aymara that is usually discourteous if not insulting,

and often incorrect. Moreover, chaotic spelling reflects

a very inadequate grasp of Aymara phonology. (See Chapter

9 for examples.)

A variation on the catalogue is Gramatica del

kechua y del aymara (1942) by German G. Villamor. It

contains short grammatical descriptions of Quechua and

Aymara, a brief three-way dictionary of words from those

two languages and Spanish, and sections on history, myths,

and superstitions. Insofar as the Aymara is concerned,

the book is deficient in every respect, with incorrect

material poorly arranged. Another variation on the cata-

logue is Vocablos aymaras en el habla popular paceha

(1963) by Antonio Paredes Candfa, containing Aymara words

purported to occur in colloquial La Paz speech. Accord-

ing to Vdsquez, who reviewed the book with me, many of

the Aymara forms are incorrectly translated and in any

case are terms used by whites or mestizos in the city

rather than by rural Aymara. The book is useful primarily

to show what the white or mestizo understands by certain

Aymara words. Usually the context is not culturally

Aymara and the tone is patronizing when not actually


Two works which avoid the condescending tone of

the foregoing are a short Spanish-Aymara dictionary by

Mario Franco Inojosa, Breve vocabulario castellano aymara

(1965), and a more complete dictionary, Diccionario breve

castellano-aymara aymara-castellano (1970), by Pedro

Miranda. Mario Franco Inojosa, who updated the Torres

Rubio grammar (1616), uses the official Aymara alphabet

adopted by the Peruvian government in 1946. Most terms

he cites are the same as those used in La Paz, making it

useful for quick reference for that dialect; however,

the book is cheaply printed and has many typographical

errors. The Miranda dictionary (1970) is more complete

and better printed, and employs the official Bolivian

alphabet adopted in 1954.

Although it reflects patr6n and missionary

usages, by far the best of the prelinguistic grammars

of Aymara is GramAtica y diccionario aymara (1965) by

Juan Enrique Ebbing. This reference grammar was modeled

on Bertonio's longer grammar (1603b) and like it contains

a wealth of detail, although the geographical origins of

the forms attested are not given. The author's method

is to explain a Spanish grammatical category and then to

give its translation equivalent in Aymara. This makes

for a repetitious presentation as the same Aymara form

may translate several different Spanish forms, and the

method skews Aymara structure into a Spanish mold as in

the case of Bertonio's grammars. While some of the Aymara

examples are acceptable to native speakers, much of the

Aymara sounds translated and the book suffers from a

generally patronizing tone. The phonology is better than

most but confuses velars and postvelars. In spite of

its faults, however, this grammar shows an understanding

of certain aspects of Aymara usually overlooked such as

the fact that certain suffixes are essential to the Aymara

sentence, and although given to stating general rules to

which exceptions must then be made, Ebbing at least in-

cludes the exceptions, making up in accuracy of data for

loss of economy in presentation. As a handbook for study-

ing Aymara, his grammar is useful as a source of Aymara

glosses of individual Spanish forms. Translations of

Spanish sentences should be checked with Aymara native

speakers, however, as they are written in a style asso-

ciated with Catholic priests (for examples see Chapter 9).

The nadir in prelinguistic grammars of Aymara

was reached in Suma lajjra aymara parlaba (1969) by

Erasmo Tarifa Ascarrunz. Another example of Patr6n Aymara,

this book contains a wealth of material, but so badly

analyzed and presented as to be very difficult to use.

As usual in prelinguistic grammars,the Aymara sounds as

if it were translated from Spanish. On the other hand,

the Spanish translations of the Aymara (or Spanish

sentences from which the Aymara was translated) are in

the popular Spanish of the Andean area which reflects

Aymara structure to a considerable extent even in the

usage of monolingual Spanish speakers. In all, the book

is an interesting compendium of fact and misconception

which should be checked with native speakers before any

of its contents are accepted as valid.

2-4 Linguistic Studies

2-4.1 Synchronic studies

As far as I am aware, the first linguist to

note in print that Aymara has a three- rather than a

five-vowel system was Bertil Malmberg (1947-48). Kenneth

Pike, in his Phonemics (1947:153), included an Aymara

problem with data that clearly implied a three-vowel

system, although Pike left this conclusion to the reader.

Tschopik (1948) and LaBarre (1950) provided partially

phonemic renditions of Aymara folktales with English

translations but without grammatical analysis. Although

transcribed with five vowels and no indication of vowel

length or of syntactic final vowel dropping, the texts

appear to be otherwise phonologically accurate native

Aymara. (The informants are identified as monolingual.)

These texts are useful for the dialectal variants they

contain and for comparison with present-day dialects

from the same areas for the purpose of identifying changes

during the past 30 years.

The first morphological analysis of Aymara was

made by Thomas Sebeok (1951a). However, it was

based on an Aymara version of Little Red Ridinghood

translated from Spanish, rather than on a native Aymara

folktale, and the text is an example of Patr6n Aymara.

Sebeok (1951b) also collected data for an Aymara diction-

ary, using data from Tschopik, LaBarre, Villamor, Pike,

and Floyd Lounsbury as well as his own. Each entry

consists of a set of Aymara words sharing the same root

morpheme, with English (or in the case of Villamor's

data, Spanish) translations.

2-4.11 Missionary grammars and associated studies

The first attempt at a fairly complete grammatical

description of Aymara by a linguist using the methods of

modern scholarship was made by Ellen M. Ross, whose

Rudimentos de gramdtica aymara (1953; second edition

1963) was published by the Canadian Baptist Mission in

La Paz with an introduction by Eugene Nida. The preface

indicates that the Aymara of Huatajata (department of

La Paz) is the dialect on which the grammar is based and

that it is similar to that of the city of La Paz. Three

Aymara native speakers collaborated with Ross in producing

the grammar, a trilingual textbook for English-speaking

missionaries and Spanish speakers. Making use of aural/

oral language-teaching methods, the book presents graded

Aymara dialogues and drills with translations into Spanish

and grammatical explanations in Spanish and English. The

grammar includes cultural notes such as a comment on the

importance of greetings among the Aymara. While it has

an index of grammatical forms and topics (in Spanish),

it lacks a table of contents and thus cannot easily be

used as a reference grammar.

In any case, although it represents a tremendous

improvement over its predecessors, Rudimentos contains

frequently inaccurate grammatical analyses. More impor-

tant, the text still reflects, in the tradition of earlier

Aymara grammars,the usage of missionaries and their fol-

lowers. For this reason the Ross grammar should be used

with caution by persons not wishing to be identified with

or as missionaries. Also, the CALA writing system used

presents the learner with certain difficulties, especially

with respect to the postvelar fricative symbolized as jj

and reduplicated as the unwieldy and confusing cluster


A reference grammar for native speakers of Aymara

is Ross's Manual aymara para los aymaristas (n.d. [con-

siderably after Ross 1953]). Its purpose is to enable

Aymara speakers already bilingual and literate in Spanish

to learn to read and write Aymara and to become aware of

differences between Aymara and Spanish structure which

create difficulties for Aymara monolinguals wishing to

learn Spanish. As indicated earlier, the CALA writing

system used by Ross is designed to familiarize Aymara

speakers with Spanish spelling with a view to facilitating

their learning to read and write in that language.

Accordingly, Spanish loans, even those which entered

Aymara hundreds of years ago and are completely adapted

to Aymara phonology, are spelled as Spanish and the five

vowels of Spanish are used even though Ross recognized

that Aymara has a three-vowel system. Evidence that

the CALA alphabet does in fact accomplish the objective

of making it easier for Aymara monolinguals to learn

Spanish is lacking.

The Ross Manual is in effect a contrastive study

of Spanish and (Missionary) Aymara, often describing

Aymara in terms of Spanish, although this is warned

against (n.d.:65). The manual is also prescriptive,

for example in Lesson IX on punctuation. The grammatical

analysis is lacking in some important respects; for

instance, the four-person system is not completely

understood. The distinction of personal and nonpersonal

knowledge is recognized, however, for the first time.

The importance of morphemic vowel length and morphopho-

nemic vowel dropping is also understood and the reader

is urged to write as he speaks, although this injunction

is not always followed in the examples given in the text.

The role of sentence suffixes (called enclitics) is well

covered. But while the Manual has its strengths, never-

theless the message conveyed by the book is that learning

to read and write Aymara is merely a means toward learning

to be fully literate in Spanish and not a worthy end in

itself. This attitude is clear in a discussion of the

embedding of direct quotes in Aymara: the reader who

wishes to write a more involved style is urged to consult

a good Spanish grammar or to observe the style of writers

in that language (Ross n.d.:121).

Two subsequent teaching grammars of Aymara owe

much to Ross. Paul Wexler and his associates attempted

in Beginning Aymara: A course for English speakers (1967)

to write a linguistically sound pedagogical grammar of

Aymara specifically for English speakers. Intended for

Peace Corps volunteers, this grammar was based on research

carried out in Bolivia by three American field workers

who spent a short time there aided by three Aymara native

speakers from La Paz who were bilingual in Spanish. It

is of value primarily as an example of what happens when

linguistic researchers fail to take cultural as well as

linguistic factors into account in spite of their obvious

importance in a grammar designed for foreigners proposing

to live and work in an unfamiliar society. While care-

fully organized into graded dialogues and drills on topics

generally relevant to altiplano life, the Aymara sentences

in the book sound translated from Spanish, often using

missionary and/or patr6n terminology, and are therefore

both culturally and linguistically unacceptable to some

native speakers. Wexler recognized that the Aymara of

the informants probably showed heavy Spanish influence,

but he was evidently unaware of the social dimension of

their dialect--its evangelical cast--although he did

recommend further research with monolingual speakers.

The book also suffers from problems of translating Andean

Spanish into English. For example, wank'u (Wexler

wanc'u) is translated 'rabbit' instead of 'guinea pig',

probably because the Andean Spanish for guinea pig is

conejo (Peninsular Spanish 'rabbit').

The second Aymara grammar owing much to Ross,

and the best of the missionary grammars to date, is

Lecciones de Aymara (1971-72) by Joaqufn Herrero,

Daniel Cotari, and Jaime Mejfa, said to be based on a

dialect from roughly the same area as that of the Ross

grammars. Herrero is a native of Spain; Cotari and

Mejia are Bolivian Aymara speakers bilingual in Spanish.

Developed for use at the Maryknoll Language Institute

in Cochabamba, this grammar is superior to its predeces-

sors in grammatical analysis, but it has the same charac-

teristics perceived by some native speakers as non-Aymara

or substandard. An innovation useful for students of

Spanish dialects is the provision of two translations of

each Aymara dialogue, one in Andean Spanish and the other

in Peninsular Spanish.

The alphabet used by Herrero et al. is that

adopted by the Bolivian government in 1954. It differs

from the CALA alphabet only in its use of k and q for

the velar and postvelar stops, respectively, instead

of the CALA c and gu for velar and k for postvelar. The

phonology section includes numerous minimal triplets

illustrating plain, aspirated, and glottalized stops.

The importance of morphophonemic vowel dropping is clearly

grasped and suffixes are designated as weak (retaining

previous vowel) and strong (dropping previous vowel) when

they are first introduced, helping the learner to produce

correct forms from the beginning. The book is good on

the Aymara four-person system (while not calling it that),

avoiding Ross's error, repeated by Wexler, of designating

the inclusive fourth person as dual. Full verbal inflec-

tional paradigms with affirmative and negative examples

are presented in the body of the text.

A much shorter, less complete grammar is Mdtodo

de aymara (1973) by Marcelo Grondin, using the same alpha-

bet as Herrero. Published in Oruro, the book mentions

certain forms as different from those occurring in La Paz

but fails to include the distinctive allomorph of the first

person possessive suffix (with velar nasal) found in the

province of Carangas, Oruro. The Aymara four-person system

is clearly grasped, vowel-dropping is understood, and the

role of sentence suffixes noted, but the Aymara is pre-

sented in short dialogues that sound nonnative. The trans-

lations of the dialogues are in Andean Spanish.

The question arises why grammars of Aymara con-

tinued for so long to reflect only missionary and patr6n

usages. The answer lies in the fact that for many years

all linguists who undertook to study Aymara in depth were

missionaries who, however well prepared in linguistic

field methods, were more concerned with translating

Scripture from Spanish or English into Aymara than in

eliciting native texts on which to base a description of

the language. Their informants, being members of the same

religious community, were ready to accept the missionaries'

authority in matters of style and content. Many missionary

linguists, notably Nida (1957:58-60), are aware of the

linguistic pitfalls inherent in their approach and try

to avoid them; but it is unrealistic to expect missionary

grammars to be completely free of the distorting influence

of translation.

The few nonmissionary linguists who approached

Aymara did so either through missionaries or through

mestizos and whites. In such circumstances it is re-

markable that Tschopik and LaBarre were able to elicit

native Aymara folktales free of missionary or patr6n

influence. Sebeok was not so fortunate; the story on

which he based his morphological analysis is in Patr6n

Aymara. When the Wexler team sought to study Aymara they

proceeded through missionary contacts and thereby unwit-

tingly acquired informants trained in that tradition.

So long as all linguistic research was conducted

with sources speaking varieties of Missionary or Patr6n

Aymara, only data reflecting such dialects could be

obtained. A new point of entry into the Aymara community

was needed.

2-4.12 Aymara-centered studies

In 1965 M. J. Hardman arrived in Bolivia as a

Fulbright lecturer in linguistics. Together with Julia

Elena Fortin, Director of Anthropology in the Bolivian

Ministry of Education, Hardman founded the Instituto

Nacional de Estudios LingUisticos (INEL) in La Paz for

the purpose of training Bolivians in linguistics for

national development. Hardman had already investigated

Aymara's sister languages, Jaqaru and Kawki, and had

determined their relationship as members of the Jaqi

language family. Hardman's Jaqaru: Outline of phonological

and morphological structure (1966) is the first grammar of

a Jaqi language described in its own terms rather than

from the point of view of Spanish. A second edition in

Spanish translation is now in press in Peru. (Pre-Hardman

Peruvian sources for the study of Jaqaru and Kawki are

the writings of J. M. B. Farf6n and of Jose Matos Mar.)

One of Hardman's students at INEL in La Paz was

Juan de Dios Yapita, a native speaker of Aymara from the

community of Compi on the shores of Lake Titicaca. Yapita

had been educated in La Paz but maintained close ties with

monolingual friends and relatives in Compi. As the out-

come of an assignment, Yapita wrote the first phonemic

alphabet of Aymara ever produced by a native speaker of

the language and later, together with Herminia Martin and

others studying under Hardman's direction, did field work

in the provinces of Ingavi, Pacajes, Andes, Omasuyos, and

Manco Capac, department of La Paz. Hardman also did

field work in the province of Larecaja. The first pub-

lished result of this research was Martin's Bosquejo de

estructura de la lengua aymara (1969), a brief sketch of

the Aymara spoken in the town of Irpa Chico, province of

Ingavi. It is important as the first published description

of Aymara by a linguist for linguists, combining both

adequate theory and competent field investigation.

On the basis of Aymara research undertaken by

Hardman and associates in Bolivia, the Aymara Language

Materials Project began at the University of Florida in

1969 with support from the U. S. Office of Education of

the Department of Health, Education, and Welfare. The

goal of the project was to produce teaching and reference

grammars of Aymara reflecting linguistic and cultural

realities of the language from the point of view of

native speakers. The materials were prepared by a team

consisting of Hardman, two Bolivian native speakers of

Aymara trained in linguistics and anthropology (Yapita

and Juana Vasquez, who is from Tiahuanaco), and three

graduate students in anthropology and linguistics who

assisted with the analysis and tested the teaching

materials in Aymara classes. Their work was supplemented

by extensive help from a number of other native speakers

of Aymara as well as from other University of Florida

graduate students and staff.

The primary fruit of the project is a three-volume

work by Hardman, Vasquez, and Yapita entitled Aymar ar

yatiqabataki ('to learn Aymara') which appeared first in

1973 and in a revised edition in 1975. Volume 1, which

bears the title of the whole work, is a course in Aymara

for English and Spanish speakers, consisting of graded

dialogues based on rural Aymara life and drills based

on the dialogues, with translations into both Spanish and

English and accompanying tape recordings with English

translations. Volume 2, Teacher's manual to accompany

Aymar ar yatiqahataki, is keyed to the course and provides

cultural as well as grammatical explanations. Volume 3,

entitled Aymara grammatical sketch in the first edition

(1973) and Outline of Aymara phonological and grammatical

structure in the second (1975), is a detailed reference

grammar which may stand alone. It incorporates University

of Florida master's theses by Laura Martin-Barber on

phonology and by Nora C. England on verbal derivational

suffixes, and my term paper on the structure of the

substantive system, as Chapters 3, 6, and 8, respectively.

The project has also produced a computerized concordance

glossary of words, roots, and suffixes.

Secondary results of the project include numerous

student papers for graduate courses in anthropology and

linguistics at the University of Florida, for example

Norman Tate's ethnosemantic study of verbs of carrying

(1970) and a paper by Andrew Miracle and Juana V6squez

on ethnosemantic categories of feces in Aymara (1972).

Published articles related to the project include Hard-

man's on Aymara and Jaqi linguistic postulates (Hardman

1972a and in press a), Yapita's discussion of the role

of linguistics in Bolivian national development (1973b),

and Pedro Copana's recommendations concerning the educa-

tion of rural Aymara children (1973).

An increasing number of materials written in the

Aymara language have appeared as a result of the project.

The Aymara Newsletter has been published irregularly at

the University of Florida since 1970, originally under

the alternating editorships of Yapita and Vasquez, who

also collaborated on a correspondence course for Aymara

speakers (Vasquez & Yapita 1969). V6squez has written an

Aymara primer (1970) and is preparing another. Yapita has

edited several mimeographed Aymara literary journals,

among them Yatiiasawa (1970) and Literatura aymara

(1972-73). He has also produced materials for teaching

his phonemic alphabet in Bolivia (1973a)and a Spanish-

English-Aymara vocabulary (1974).

Former Yapita students who are members of the

Bolivian Aymara community have produced materials of their

own. Representative are articles by Vitaliano Wanka

Torres, describing results of the Aymara literacy pro-

gram he directs in Tiahuanaco (Wanka 1973a and 1973b);

an Aymara primer for adults (1974) by Francisco Calle P.,

of which a first edition of 17,000 was printed (Chaski

2:1974); and a bilingual manual on medicinal plants and

herbs (1974) by Gabino Kispi H. (Wanka and Kispi spell

their surnames, traditionally Huanca and Quispe, in

Yapita orthography.)

So far, lack of funds has precluded formal pub-

lication of more than a few of the Aymara-centered

materials that have begun to appear in growing numbers.

2-4.13 Sociolinguistic studies

In Peru the government has in recent years sought

the participation of missionary linguists associated with

the Summer Institute of Linguistics and of secular linguists

from academic institutions in the development of educa-

tional programs for speakers of indigenous languages

(loosely referred to as bilingual education programs).

While most publications on problems of multilingualism

in Peru focus on Quechua and the jungle languages rather

than on Aymara, several recent studies include references

to Aymara. As noted in 1-1, the two most recent Peruvian

national censuses (for 1961 and 1972) contain basic demo-

graphic data on the numbers and location of Aymara speakers

in Peru. The proceedings of a round table on problems

of Quechua and Aymara monolingualism held in 1963 have

been published in Mesa redonda sobre el monolingUismo

quechua y aymara y la educaci6n en el Perd (1966). One

of the participants in the round table was Alberto

Escobar, a Peruvian linguist who later founded the

government- and Ford Foundation-supported Plan de Fomento

LingUistico (linguistic development plan) at the National

University of San Marcos in Lima and who has written

several thoughtful essays on the language problems of

Peru such as an article on literacy programs (Escobar

1972a). Escobar edited a collection of articles entitled

El reto del multilingUismo en el Perd which appeared in

1972, the year the Peruvian government inaugurated a new

policy of bilingual education. The book includes articles

by Hardman on Aymara linguistic postulates (Hardman 1972a),

by Alfredo Torero on historical background (Torero 1972b),

and by Escobar on linguistics and politics (Escobar 1972b).

Domingo Llanque Chana, a Peruvian Aymara who is

a Maryknoll priest and at present (1976) vicar general

of the Prelature of Juli, has presented in Spanish trans-

lation an interview he conducted in Aymara with a 56-year-

old Aymara man from a rural community near Lake Titicaca

(D. Llanque Chana 1973). The topic is social interaction

among the Aymara, including the way they treat outsiders

as well as each other. To my knowledge this is the first

time the topic has been discussed in print by an Aymara.

The author observes that the basic element of Aymara

interaction is mutual respect expressed primarily through

courteous speech as exemplified in greetings.

A graduate of a normal school in Puno, Justino

Llanque Chana, has given an overview of the educational

situation of Peruvian Aymara based on the results of his

1973 survey of 85 high school students in the town of

Chucuito near Puno (J. Llanque Chana 1974). The survey

revealed negative attitudes toward Aymara language and

culture which the author interpreted as confirming the

alienating effects of an educational system stressing

acquisition of Spanish skills while banning (in theory if

not in strict practice) the use of vernacular languages.

Meanwhile in Bolivia, where the government has yet

to give formal support to bilingual education, only one

organization has so far as I know published materials

relevant to sociolinguistics: the Centro Pedag6gico y

Cultural de Portales in Cochabamba, which is supported by

the Patiho Foundation. In connection with a series of

educational conferences and seminars for Aymara and Quechua

speakers, Portales (as it is usually called) began in

1973 to publish in mimeographed form such materials as

articles by Javier Alb6 on the future of Aymara and

Quechua (which he considers to be 'oppressed languages';

Alb6 1973a) and on Aymara and Quechua educational radio

programs in Bolivia (Alb6 1973b). Also in 1973 Portales

published the Yapita phonemic alphabet and in 1974, my

article on the Aymara four-person system (Briggs 1974a)

and a summary of Hardman's article on Aymara linguistic

postulates (Hardman 1972a).

Portales has also assisted sociolinguistic sur-

veys. In 1973 and 1974, Yapita and Pedro Plaza, the

director of INEL, conducted with Portales and Ford Founda-

tion support sociolinguistic surveys of groups of Aymara

and Quechua speakers in Bolivia using methods developed

by Wolfgang Wdlck for Quechua in Peru (Wblck 1972 and


A valid contribution to knowledge of the Aymara-

speaking population of northern Potosi department is an

article by the British anthropologist Olivia Harris

(1974) on the Laymis and Machas (1-1.1).

2-4.2 Historical studies

Torero set forth well-grounded theories as to

the history of Aymara and its sister languages (1-2.1)

in an article entitled LingUistica e historic de los Andes del

Perd y de Bolivia (1972b). The relationship of Aymara and

Quechua, the other major language family of the Andean

area, is still a matter of debate. Mason (1950:196)

proposed 'Kechumaran' as a term 'to designate the yet

unproved but highly probable subphylum consisting of

Quechua and Aymara.' Also supporting a fairly close

relationship between Quechua and Aymara are Carolyn Orr

and Robert E. Longacre (1968) and Yolanda Lastra de

Suarez (1970). Hardman ascribes similarities in lexi-

con and phonology, where they exist, to geographic

proximity and overlap rather than to a genetic relation-

ship (Hardman, personal communication). Louisa Stark

(1970) has provided convincing data in support of Hard-

man's position.

2-5 Summary and Projection

The foregoing survey of representative litera-

ture on and in Aymara shows how scholarly and

not-so-scholarly treatment of the language has changed

in accordance with the focus of each period and the

development of more adequate method and theory. While

the production of written texts in Aymara is still meager,

it is increasing. Like the spoken language on which they

are based, these texts show dialectal variation, and they

will provide material for further dialect studies as well

as investigations of literary style.



3-1 Introduction

The basic phonology of Aymara has been described

by L. Martin-Barber (Hardma, et al. 1975:3, Chapter 3).

To the phonemic inventory therein described must now be

added the velar nasal, although it is of limited occur-

rence. Variations in phonological shape of morphemes

both within and across dialects are paralleled by morpho-

phonemic rules operating within dialects, to be discussed

in Chapter 4, to which this chapter is an introduction.

3-2 Phonemes

3-2.1 Phonemic inventory

Figure 3-1 shows Aymara phonemes in Yapita

phonemic orthography.

There are three vowels (front, central, and

back) and a phoneme of vowel length. The 27 consonants

are divided into voiceless and voiced. Voiceless conso-

nants are 12 stops, three affricates, and three fricatives.


Vowels: i a u Vowel length:


alveolar velar
palatal postvelar

Plain p t k q
Aspirated p" t" k" q"
Glottalized p' t' k' q'
Plain ch
Aspirated ch"
Glottalized ch'
Fricatives s j x

Laterals 1 11
Nasals m n n (nh)
Glides w y
Flap r

Figure 3-1. Aymara Phonemes (Yapita Phonemic Alphabet)

Voiced consonants are two laterals, four nasals, two

glides, and a flap. The stops occur in four positions

of articulation and three manners: plain, aspirated,

and glottalized. The three affricates are all palatal

and pattern with the stops. The three fricatives are

alveolar, velar, and postvelar. The laterals are alveo-

lar and palatal; the glides are bilabial and palatal;2

the flap, which may be realized as a trill, is alveolar.

Included in the total of four nasals is the

velar nasal /nh/ ([ ]), a phoneme in the related language

Jaqaru. The velar nasal has phonemic status in only two

Aymara dialects encountered to date, in the provinces

of Carangas (Oruro, Bolivia) and Tarata (Tacna, Peru),

specifically in the communities of Jopoqueri and Corque

(Carangas) and of Tarata and Sitajara (Tarata). Although

the phoneme occurs in few morphemes, these have a high

functional load. In both Carangas and Tarata the phoneme

occurs in homophonous allomorphs of two suffixes: /-nha/

first person possessive and /-nha/ verbal inflection of

first to third person, Future tense. In Tarata the

phoneme occurs in two more suffixes of the Future tense

(see Figure 6-3) and in at least two noun roots: anhanu

'cheek' and panhara 'grinder.'

The two areas where the velar nasal phoneme occurs

are separated by the province of Pacajes, department of

La Paz, whose dialects were not investigated directly for

this study. The phoneme was not found in the Pacajes

dialects investigated in the research for Hardman et al.

(1975). Late in the research for the present study,

evidence was obtained for the existence of a relic of

the /nh/ phoneme in La Paz/Compi dialect, in one word:

the noun ch'inhi 'nit' (Spanish liendre). This con-

trasts with intervocalic /n/ and /n/ (e. g. ch'ina 'human

posterior' and nunfu 'breast, teat'). Another apparent

relic of /nh/ is a velarized allophone of /i/ occurring

in Morocomarca (4-3.21.2). It may possibly also occur

in other Aymara dialects not yet investigated.3

3-2.2 Allophones

The allophones described by L. Martin-Barber

(Hardman et al. 1975:3) exist for the dialects of Aymara

investigated for this study. The following additional

comments may be made.

3-2.21 Vowels

In Spanish-influenced dialects the mid vowel /a/

may approximate Spanish /a/ but in monolingual Aymara

(and some nonmonolingual dialects, for example in northern

Potosi, a trilingual Quechua-Aymara-Spanish area) /a/ is

more closed, being realized frequently as [A] or [a].

/i/ and /u/, as noted by L. Martin-Barber, are lowered

in the environment of postvelar consonants /q/ and /x/

and raised in the environment of /h/ and /y/ and word-

initially. Elsewhere intermediate or high allophones

occur. Additional study will be needed to determine the


3-2.22 Consonants

Most allophonic variation of consonants in Aymara

is morphophonemically conditioned, and is therefore dis-

cussed in Chapter 4.

Friction attending the velar and postvelar frica-

tives /j/ and /x/ and the flap /r/ is variable but whether

the variation is dialectal, stylistic, or idiosyncratic

is yet to be determined. Impressionistically it was

noted that some Juli speakers pronounced initial /j/ with

heavy friction whereas in other dialects initial /j/ is

more often a glottal [h]. Dialects having /j/ (Salinas,

Jopoqueri) where La Paz and other dialects have /k/, for

example in the incompletive verbal derivational suffix

-ja- ~ -ka-, articulate a somewhat prevelar, palatalized

/j/ and a clearly postvelar /x/. The difference is quite

noticeable even to a nonnative in such pairs as

Chur.j.t.wa. 'I'm giving it to him/her/them.'

Chur.x.t.wa. 'I gave it to him/her/them.' (Jopoqueri)

(The second example has the verbal derivational suffix