NORTH AND SOUTH AMERICAN

LINGUISTIC CONNECTIONS1
M a ry R itc h ie K e y

i. A Working Hypothesis

In this paper a hypothesis is suggested for intercontinental

linguistic connections; that is, that the Uto-Aztecan (North
America), and Aztec languages (Mexico) have affinities with

I. This article contains parts o f a paper, The History and Distribution o f the Indigenous
Languages <f Bolivia, presented at the American Anthropological Association, November,
1978, Los Angeles. I gratefully acknowledge a Fulbright-Hays Research and Lectureship
in Comparative Linguistics and Indian Languages o f Chile, 1975. This gave me the
Mapuche material, and time to develop the files, which led to this present paper. Research
funds from the School of Humanities at my university made it possible to expand my
files by adding the Mosetene, Mapuche, and Quechua material. I acknowledge the
generosity o f the New Tribes Mission for sending me the unpublished vocabulary list
and grammar of the Tsimane (= Chimane - Mosetene) language. I also express my
appreciation to my professor, Winfred Lehmann, for reading the manuscript. Andrl
Martinet also made helpful comments. During my early work on fluctuation of phonemes,
Prof. Martinet recognized the theoretical implications for historical linguistics, and thus
I was encouraged to explore these matters further.
Students in my Historical Linguistics classes during the past few years used these
comparative data in their individual projects. This contributed substantially to the
work assembled here. Everyone contributed, but I particularly want to mention the
thorough and careful work in the term papers o f Linda Daetwyler, Ruth Cavender,
and Kim Richardson. Members of the classes also included: Gail Cameron, Jacquie
of. an, Erica Lansdown, Eva Litochleb, Mary Mastren, Eileen Matsumoto, Karel
Mundt, Sherry Rathsam, Cyndi Fann Reilly, Debbie Ross, and Martha White.
When this article was in its final stages, I received a letter from James Loriot, asking
about my Tacanan comparative studies. I was surprised and pleased to know that
Loriot, who has spoken Shipibo-Conibo since childhood, has been working on the
connections that are dealt with in this article. It is comforting when one’s work is corrob­
orated by so competent a linguist as Loriot. His material essentially points toward the
same conclusions as mine. In some areas he has worked out more details, and he attempts
proto forms at higher levels. His concerns go much farther than mine; he posits Trans­
pacific relationships. In his Eastern Trans-Pacific group he includes: Athabascan;
Uto-Aztecan; “ Hunikunean” , a group which includes Muskogee, Mayan, G i, Panoan,
Tacanan, Otomanguean, Jivaroan (?), Quechumaran, and Taras can. He has given me
permission to mention these interests o f his. W e look forward to his publications that
will give evidence o f these claims. I am grateful for the wealth o f comments which he
generously sent me about my own material, and which I will incorporate into my files
for future use.
La Linpnstiquet vol. 17, fasc. 1/1981
Quechua, Aymara (Bolivia and Chile)*, Mapuche (Araucanian, in
Chile), Mosetene ( = Chimane = Tsimane, in Bolivia), Tacanan
(Bolivia), Panoan (Bolivia and Peru), and the Fuegian languages.
M y hypothesis is not based merely on the resemblances
between the vocabulary, which, o f course, could be borrowings
or could be accidental. M y judgment is based on my previous
work with the sound correspondences within the families listed
above and their proto forms, the phonetic actualizations, the
fluctuations (defined below), and the distribution patterns of
individual languages.

2. South American Comparative Studies

Progress in comparative linguistics of South America has

moved very rapidly in the seventeen years since my dissertation
o f 1963 when I elaborated on the affinities o f Tacanan and
Panoan*. One o f the main reasons is the solid contribution of
Latin American linguists who have been caught up in the interests
o f historical migrations and classification of the some 600 languages
o f South America. The languages that have a very close genetic
relationship have probably already been identified. Substantial
work has been done for some large families, such as the Arawak,
Guarani, Chibchan, and Tucanoan. Some work has been done
bringing families together, for example the Tacanan and Panoan
o f Bolivia and Peru. Distant relationships are being studied
nowadays with vigorous attention.
Historical discoveries are often fortuitous. The South American
picture exemplifies this. The discoveries often have to do with
the linguist’s background rather than logical procedures. The
Chipaya-M ayan hypothesis was set forth in 1964, for example,
because Ronald Olsen had spent some time among the Mayans
in Mexico before going to Bolivia. Suarez’ important contributions
resulted from extensive work with colleagues in Argentina among
the Fuegian languages, which preceded his move to the north
where he gained access to Swadesh’ materials. The abundantly

a. Qjiechua and Aymara previously have not been proved to be related to the other
families listed in this article. I posit this relationship from the point of view o f Quechua
and Aymara in “ Quechumaran and Affinities” , in Script* EthnoUgica (Festschrift for
Prof. Marcelo Bduiroo), Buenos Aires, Argentina, in press.
3. Extensive references have been given in my recent publications, and for the sake
o f brevity I will give only the main sources here.
useful material published by Esther Matteson and colleagues
resulted from the interest of whoever-happened-to-be-present at
a workshop in Colombia. The Aztec hypothesis which I present
below is possible because I happened to have spent many years
in Mexico and published on the phonemes and compiled a dic­
tionary in Aztec. These discoveries are not made in ways that
one would expect in an organized, abundantly-funded effort
that the space age is well-acquainted with. In spite o f the piece­
meal effect, the situation of South American linguistics is, for
several reasons, one of the most exciting and rewarding areas of
research that exists today. Because the languages are non Western
they offer a rich laboratory of data for discussions of universals
and linguistic theory. Historical connections between North and
South America can be corroborated by linguistic evidence,
a powerful source of verification.

3. Considerations in Comparative Work

in Recently Written Languages

The phonological problems of comparative work in recently

written languages are probably no different from the phonological
phenomena in all languages of the world. But the attitude and
approach of the investigator is different. This has advantages
and disadvantages. One perceives an unwritten language in a
pristine atmosphere; one may actually “ hear” more in an innocent
and receptive attitude. When dealing with written languages,
one is apt to deal with the language in terms of the symbol (the
writing) rather than of what one hears. It is important to remember
this when the linguist approaches the matter of fluctuation or
variation of articulation. For example, when we “ hear” the
words, ,some fish’, we may “ hear” a labiodental [f] because the
symbol indicates such. But it is likely that a bilabial fricative [p]
is actually articulated. The possibility of observing more variation
in pronunciation, then, is an aid in trying to understand sound
change from language to language. O n the other hand, archaic
spellings and old documents are a great help in reconstructing
the past history of languages.
Fluctuation of phonemes in everyday articulation of natural
language is probably more common than educated/literate people
have realized. By fluctuation I mean those varieties of pronun­
ciation that actually occur in natural speech. These include the
whole range o f phonetic realizations as well as alternation of
phonemes tolerated in casual, natural speech. In Chama, for
example, the following fluctuations in pronunciation occur: /dopi-
kikwi ~ dobikikwi/ ‘enters’ ; /ekwi ato i~ ekwi adoi/ ‘my brother’ ;
/kaxa’sie ~ xaxa’sie/ ‘chonta palm’ ; /hawatemi ~ hawatewi/
‘piraha fish’. In Mapuche (Araucanian): /fita ~ flea/ ‘husband’ ;
/m upi~ mipi/ ‘feather’ . “ Learned” persons are so attached to
the symbolization o f the symbol that they no longer pay any
attention to the varying behavior that goes on out-of-awareness.
There is extensive fluctuation of phonemes in some South American
languages. Clairis has documented this for Qawasqar (Alacaluf).
I have documented it in publications since 1968. In the Chama
and Mapuche languages, the patterns o f fluctuation are related
to the patterns of reflexes and correspondences of the related
languages. A theoretical explanation of phoneme fluctuation is
suggested by the reflexes of sister languages in the form of bor­
rowings or retentions of a former state. This information is crucial
in recognizing relationships and distinguishing cognates from
loanwords.

4. Characteristics o f American Indian Languages

Morphemes in the agglutinative languages of South America

very often have a one-syllable shape. The one-syllable morphemes
can be joined to form stems in an amazing variety of combinations
found in cognates across family lines. The following examples
are from closely related languages, so there is no doubt about
their validity.
mover (move) Chama (Tacanan) wewa- / -nena-
Tacana (Tacanan) wana- (arreglar ‘arrange’)
Amahuaca (Panoan) wana-, nawa:- (arreglar)
murcillago (bat) Mosetene tini
Proto-Tacanan *bina
Proto-Panoan *kaii:
barba (beard) Mapuche ketre
Proto-Tacanan *kesa
Proto-Panoan *kwini
Cashibo (Panoan) kweSa

I am calling this potential for arrangement morpheme split,

i.e. stems can be formed by splitting morpheme combinations
into various other combinations. This kind of morpheme split
also occurs in the Uto-Aztecan languages:
despertar (awaken) U A *pusa
Zacapoaxtla (Aztec) ihsa

Apparently the morpheme split process can be carried across

family lines, for example in the following possible cognate groups:

aAo (year) Zacapoaxtla (Aztec) iiwit

Tetelcingo (Aztec) Sihpa
Mapuche (Araucanian) t'ipantu
Chacobo (Panoan) Sinipa
ropa (clothes) Zacapoaxtla (Aztec) tasa:l
♦Panoan ♦tari
Tacana (Tacanan) sa (shirt)
Reyesano (Tacanan) sapi

The hazards of morpheme identification o f cognates in very

distant relationships, of course, are horrendous.Even more
hazardous is the situation o f metathesis, whichoccursthroughout
the languages. Metathesis of phonemes is illustrated in the following
in closely related dialects of Quechua:

sesos (brains) Proto-Qjiechua ♦nutqhu

Cochabamba (Qjiechua) nuhtu
Cusco (Quechua) nutqhu / fiusqhu
Ayacucho (Quechua) hutxu
Riobamba (Quechua) fiutku
Tena (Quechua) fiuktu
brillar (to become bright)
Ayacucho (Qpechua) aiikya-
Huar&s (Qpechua) akci- / acki-

The wide range of possible morpheme combinations plus the

wide variety of vowel changes (such as vowel harmony) are
illustrated in the following possible cognates:

levantar (raise) Mapuche wit'ah-

Mosetene ieeme
Cavinena (Tacanan) wesa-
Chama (Tacanan) aha-
Tacana (Tacanan) tsewa
Amahuaca (Panoan) wani-
lavar (wash) Mapuche kiiattin
Mosetene iikon
Proto-Panoan ♦fioka-
Chama (Tacanan) iakwa-
 agujero (hole) Mapuche we£o 9
Amahuaca (Panoan) sowi:
gente (people) Mapuche kona (servant, soldier)
Capanahua (Panoan) n6kl

This type of morpheme split and vowel variation apparently

also occurs in distant relationships, as for example, in the following
possible cognate groups:
esquina (comer, also rincdn)
Proto-Quechua *k’uCu
Mosetene katsyeye
Cavinena (Tacanan) etsoko
caminar (walk, road) Proto-Qjiechua *puri-
Mapuche rip!
humo (smoke) Uto-Aztecan *kwici
Cashibo (Panoan) tsil kwi
comer (eat) Uto-Aztecan *kumi
Proto-Qjiechua *mikhu-
dia (day) Zacapoaxtla (Aztec) to:na-l
Amahuaca (Panoan) nata*
ropa (clothes) Zacapoaxtla (Aztec) koto:n (sarape)
Mapuche tukun
Amahuaca (Panoan) kotS:
barro (clay) Tacana (Tacanan) roto
Quechua turn

The historical processes involved in these languages may

cause a reshaping of the syllable, resulting in consonant clusters
in some o f the languages:
sol (sun) Quechua inti
Mapuche anti
Proto-Panoan *niti

5. Cognate Identification Versus Loanwords

Because of the hazards at the phonological level in trying to

match one-syllable morphemes, I have stayed within a very
close semantic range for identification. Most of the examples in
my comparative files have identical meanings. Loanwords are a
major stumbling block when it comes to certainty of cognate
relationship. There is a very large vocabulary of Amerindian
(and/or debatable) origin in the Spanish of Latin America, and
even in the English of the Americas, for example: arroyo, barranco,
buho (and other words for ‘owl’), cancha, cochino, copal, coyote,
chicha, chili, chocolate, hammock, jerk(y) (charqui), ocelot,
pampa, poncho, pulque, puma, serape/sarape, tomato, trapiche.
In many cases, the word has been taken over so completely that
it is no longer thought of as not-Spanish or not-English. Other
vocabulary items are less known on a wide scale, but deeply
embedded in the Spanish of a particular area. I compiled a non-
Spanish vocabulary of over 700 items which are used by mono­
lingual Spanish speakers in the Bolivian and Peruvian area
(Key, 1966). For example, cuscu/cuhcu/cuhcuh ‘fiebre (fever,
malaria)’ ; kurici/kurice ‘pantana (waterhole, swamp)’. This large
vocabulary cannot be ignored by South American linguists; often
the fluctuation in pronunciation follows the patterns of correspon­
dences in the Indian languages of the area, as happens in the
examples cited above. Sound changes do occur in loans of long
standing. Note the following examples in onomatopoeic words
and words that are thought to be obviously borrowed.
puerco (pig) Mapuche kuii
Cavinena (Tacanan) koii
Chama (Tacanan) kwefi
Tacana (Tacanan) koii
ovcja (sheep) Quechua uwiha
Aymara wesi
Tacana (Tacanan) wiia
Mapuche ufiia ~ ufisa
Selknam (Chon) wo: i5
Alacaluf wese
hipo (hiccough) Zacapoaxtla (Aztec) tsikno:li:s
Proto-Qjiechua *c’uku-
Proto-Qjiechua *kik’i
Cochabamba (Quechua) hik’u
Ayacucho (Quechua) hikfu
Chama (Tacanan) heki-so’ o
Proto-Tacanan *codo-
Alacaluf tSalaks
estornudar (sneeze) Zacapoaxtla (Aztec) ik£o:a
Quechua aiilnina
Mapuche efiwin
Mosetene aiikki
Cavinena (Tacanan) haii-
Chama (Tacanan) ati / aii-
Tacana (Tacanan) tido
Cashibo (Panoan) ’ atiSanki-

There are examples when one cannot decide whether the word
is a loanword or the similarities are coincidental. The form
reconstructed by Voegelin, Voegelin, and Hale, for ‘redondo’
(round, circle/circular) in U A is very similar to Spanish ‘circulo’ :
U A *cikuri.
Note the similar forms in South American languages:

redondo (round, circle) Uto-Aztecan *cikuri

Mapuche iiq kill
Mosetene fcihiriyes
Proto-Quechua *kururu (ball of yarn)
Cavinena (Tacanan) kwarero-da-ke
Tacana (Tacanan) perorota

Even more baffling is the morpheme ko meaning ‘water’

which appears in the Aztec of Mexico, down through the Tucanoan
languages of Columbia (Wheeler), through the Quechua languages,
and down through Chile, and possibly to the Fuegian languages.
Is this a coincidence, a loanword, or a cognate?
agua (water) Tetelcingo (Aztec) dtlah-ko (river)
Proto-Quechua *yaku
Mapuche ko, ko-we (pozo ‘ well’ )
Amahuaca (Panoan) wako’ -m i’
Proto-Panoan *paSko (possibly tributary)
Chama (Tacanan) 6ixoko (pantano ‘swamp’ )
Cavinefia (Tacanan) epokotana- (quebrada
‘ravine’ )
Alacaluf aqcolaj
Bolivian Spanish kurifci (pantano ‘swamp’)
Proto-Tucanoan *-ko It occurs in seven o f these
languages in various forms, in the glosses for
water, liquid, rain, milk, saliva, medicine, bev­
erage (Wheeler).

Another morpheme that appears to have widespread occur­

rence is the morpheme for female. Vivo Escoto documents the
use of Aztec siwat in place names throughout Central America.
mujer (woman) Zacapoaxtla (Aztec) siwa:-t
Tetelcingo (Aztec) sowa-tl
Quechua Sipas (muchacha ‘girl’)
Mosetene -si ~ -s (feminine gender)
Yaninahua (Panoan) 5iwi-ya

Also related?

abuela (grandmother) Tetelcingo (Aztec) isihtsi (su abuela)

Cashibo (Panoan) 6i£i
Mapuche £e£e (abuelo ‘grandfather’)
 Still another morpheme that appears to have widespread
occurrence:
casa (house) Uto-Aztecan *kali, *ki
Zacapoaxtla (Aztec) iia:n
Aymara (Jaqi) uta
Proto-Tacanan *e-tafe
Chama (Tacanan) e’kP
Amahuaca (Panoan) hatf*, tapa:s
Mosetene aka
Mapuche ruka, nikal
Selknam (Chon) ka-wj
Alacaluf at, ata

6. Genetic Relationships

In order to understand the stages of progress that comparative

studies of South American languages are in today, one can
imagine the various stages o f the development of Indo-European
work. I f South American Indian linguists were to land in Europe
before the time of Chaucer and before the time of extensive
communication and transportation, they would observe some
similarities between languages, but on the whole they would
feel that the continent contained a huge hodge-podge o f completely
different languages. If, by historical accident, one of them who
had lived in Spain for a year, went to Denmark, he or she would
note similarities between the languages. With great enthusiasm
the linguist could assemble a respectable file of resemblances.
Knowing what we know today about the Germanic and Romance
families, it is easy to see that there would be great confusion if
the hypothetical linguist would try to reconstruct a proto-sound-
system from Danish and Spanish. There would be additional
conflicts and confusion if another linguist from Poland had seen
similarities and tried to reconstruct a former language from
Danish and Polish. The wisdom of the centuries would indicate
that the linguist should have dealt with the families separately
before attempting reconstruction at a higher level.
In studying the distant relationships o f South American
languages, I have chosen not to attempt to reconstruct proto­
phonemes. A t this early stage of observation it is enough to
identify areas of possible phoneme correspondences that fall
within a certain phonological space. Regularity is seen in the
recurrence of sounds in words across family lines. A t this early
stage the understanding of sound change is not enhanced by right
or wrong guesses in setting up hypothetical proto forms. Looking
only at the reconstructed proto form may obscure other relation­
ships that occur on the periphery of cognate sets. Language change
is not as neat and tidy as recently contrived proto forms imply.
Though I believe it is premature to posit proto-sounds, the
sheer quantity and quality of resemblances between Uto-Aztecan
and the other languages listed above speak for a genetic relation­
ship. Th e vocabulary that I have amassed is basic vocabulary,
not the kind that lends itself to borrowing. The entries show
phonological resemblances between the languages. If, in time,
the genetic relationships are proven without a doubt, then the
phonological resemblances, or reflexes of a former state, will be
seen to be the correspondences o f the daughter languages. A sub­
stantial corpus of resemblances at the phonological, grammatical,
and even the semantic level could be assumed to be evidence of
distant relationship, even without working out the proto phonemes.

7. Structural Linguistic Features

Phonological and grammatical similarities occur between

Uto-Aztecan and South American languages. Further examples
and a word list of about 250 entries are given in a companion
article4. Here I will present a few examples of sounds andpossible
correspondenceswith languages which I believe to be related.
The presentation which follows is given from the point of view
of Proto-Uto-Aztecan and Proto-Aztecan. The following chart
is taken from Voegelin, Voegelin, and Hale; I have added two
5
sibilant phonemes from Aztecan, c and , because this accom­
modates the South American languages which are much more
complicated in the area of the sibilants. Proto-Quechua and
Proto-Tacanan, for example, each has six sibilant sounds.

Proto-Uto-Aztecan (plus 6 and i)

p t k kw a
c 6
s 5 h
m n i)
1
r
w y

4. Uto-Aztecan and South American Linguistic Affinities, manuscript, in preparation.

 Examples of resemblances are presented here. These selected
examples are taken from my larger files which contain material
from the following: Uto-Aztecan (Voegelin, Voegelin and Hale,
1962; Davis, 1966; Klein, 1959); Zacapoaxtla Aztec (Key and
Key, 1953); Tetelcingo Aztec (Brewer and Brewer, 1962);
Quechumaran (Orr and Longacre, 1968; Martin, 1969; Orr and
Wrisley, 1965; Clair-Vasiliadis, 1976); Tacanan (Key, 1968);
Panoan (d’Ans and den Eynde, 1972; Shell, 1965); Mosetene
(Armentia, 1901-1902; Anon, New Tribes Mission, n.d.); Mapuche
(Araucanian) (Erize, i960; Key, field notes, Key, 1978); Chon
(Selknam) (Najlis, 1975; Sudrez, 1973); Alacaluf (Key and
Clairis, 1978; Clairis, field notes, and Clairis, 1977); and other
material, all of which is documented in the Mapuche article
(Key, 1978).
The examples are presented in the following order:

Uto-Aztecan UA
Comanche
Hopi
Huichol
Papago
Northern Paiute NP
Shoshone
Aztec Az
Aztec—Zacapoaxtla AzZac
Aztec—Tetelcingo AzTet
Qjiechumaran Q -A
Quechua Q,
Aymara
Tacanan T ac
Cavinena Cav
Chama Chm
Huarayo Huar
Tacana Ta
Panoan Pan
Amahuaca Ama
Capanahua Cap
Cashibo Cshb
Chacobo Chac
Chaninahua Chan
Marinahua M ar
Mayoruna M ay
Shipibo-Conibo SC
Yaminahua Yam
 Mosetene ( = Tsimane = Chimane)
Mapuche (Araucanian) Map
Chon
Ona
Selknam Selk
Tehuelche Te
Alacaluf ( = Qawasqar) Ala
UA *p corresponds to p and b in South American languages. A t times it
corresponds to zero. Examples o f p:p are: agua (water) U A *pa;
PanMar p iikd; *Pan *paro (river). Camino/caminar (walk, path)
U A *po; *Q, *puri-; M ap ripi. Examples o f p:b are: cabello (hair)
U A *po; *Pan *boo; PanM ay -bo-. Ojo (eye) U A *pusi; TacChm -bosi
(face). Pesado (heavy) U A *piti; *T ac *bike-. Examples o f p:o are:
entrar (enter) U A *paki; *Pan *i’ ki-. Pesado (heavy) U a *piti; ‘ Pan
*’ iwi.
UA *t corresponds to t, and at times to a sibilant. Note that Chama (Tacanan)
t derives from *T ac *s. U A *t corresponds to M ap tr. There is also
involvement with k throughout the languages. Note that *T ac *t has
reflex k in TacChm , and in this language there is fluctuation between t
and k. Examples are: barro (clay) A zT et tloltsakti:-; Q, turu; T acT a
rutu; PanCshb fiua; M ap fotra. Ciego-(blind) AzZae m o tafa; PanAma
5
wst&?) M op lr«wm». Diente (tooth) U A *tama; ‘ Pan * ita. Escupir
(spit) UA-Comanche tusi-; *Q -A **thuCa-; Aymara thusa; PanCshb
tuiu-ka-. Fuego (fire) U A *tahi; A zZac tit; *T ac *-ti-; *Pan * ii’ i;
Mosetene tsi; M ap kitral.
UA *k corresponds to k and kw. It corresponds to *T ac *k which has both k
and kw as reflexes. Note also the involvement with t as mentioned above.
A t times there is involvement with fricative x and with h. Note that *T ac
*x has k, h, and x as reflexes. Examples are: boca (mouth) U A *kama;
*T ac *-kaca; PanAma han i’ . Buho (owl) U A *tukur(i); Q puku, kurkuku;
5
M apkoo. Carbon (charcoal) A zZac tekol; * Q k ’ il ,imia; TacChm koha’ia;
Mosetene kii; M ap kuyul.
UA *kw corresponds to k, kw, back velar q, w, 0, and globalized velars.
Examples are: cola (tail) U A *kwasi; Mosetene oSi; ChonOna ’ die;
ChonSelk k’ oi, ’ oi. Comer (eat) U A *kwa(’ )a; * Q * q a ra -; TacC av ara-;
ChonSelk q ir (hambre). Defecar (defecate, excrement) U A *kwita;
Q k i ia . Humo (smoke) U A *kwici; UA-Hopi kwi:ci(qwi ) ; *Q_ *q’ uini;
TacChm wiSafia-; PanCshb tsif kwi; PanChac ko’ i'ni. Bueno (good)
A zZac kwali; *Q, ‘ al^i; *Q,-A **waiyi; M ap kimey. O lor (odor) U A
•kwana; A zZac tahnekwi; *Q. *asna-; TacC av kwehi-.
UA *’ has not been identified yet with regular correspondences.
UA **c has reflexes c and t in the Aztecan languages.
**s has reflexes s and I in the Aztecan languages.
Proto Quechua has an affricate and a sibilant series which occur at three
points of articulation: alveolar, alveopalatal, and retroflexed. These
are also aspirated and globalized.
 Proto Tacanan also has a complex array o f reflexes. An analysis of the
Tacanan series shows the variety o f reflexes possible (Key, 1968,
P- 35):
*c: c c s *d: hd i *d: d c s
d d
*s: s d 4
* : h s h 4
* : S 5 5

Note that t is a possible reflex o f *s in TacChama. Throughout the

languages that are concerned here, there is a relationship between t and
the sibilants and fricatives.
This large range o f sounds in these areas has not been satisfactorily
analyzed in any o f the proto studies, but there are some patterns which
have been noted. Some o f the conditioning is in the environment o f front
vowels.

UA *c has sibilant correspondents. There may be regularity with c and s;

note that these are both reflexes of *T ac *c. Examples are: cabello (hair)
U A *co(ni) (head hair); T acT a -do-a (head). Costurar (sew) U A *coma;
TacC av toco-; TacChm soko-. Chico (small) A zZac cikci:n, -cin;
UA-Papago dim; * Q ,p iii; TacChm -di(’ )a; TacH uar -sisi; T acT a wai-
didi; *Pan *pi[s]t[i]a; M ap pidi. Delgado (thin) A zZac pica:wa-; Q,cinzu;
•T a c *oSori; TacChm eSfwiyo; PanAma dilni3; PanChac o|o; M ap
troqli. Espesar (thicken) A zZac -tecawa; *Q, *sankhu; *M ap tronge.
Perro (dog) A zZac ickwi:n-ti; Ala a’tsiyqans.

5
Proto-Aztecan *d has correspondents d, s, and , among other sibilants. Examples
are: dormir (sleep) A zZac kodi; *Pan *’ osa-; Mosetene kodi. Escama
(fish scale) A zZac itapad; M ap trawa-dallwa; ChonSelk -sa-. Frasada
(blanket) “ Spanish” poncho; Q, puncu; *T ac *m ala; PanYam acicA;
M ap pontro; ChonOna J6:n (cloak); ChonSelk sdonwn. Hacer (make, do)
A zZac -di:-; Q,-di- (fabricar); PanAma ti:’ -.

UA 4
*s often corresponds to *T ac *s and * ; it also corresponds to other
4 5
sibilants such as and . Examples are: bianco (white) U A *tosa; *Tac
4 4
*pa a-; *Pan *o o. Corazdn (heart) U A *sula; *Q, * unqu; T acT a 5
4
moesomo. Cortar (cut) U A *siki~ *sika; *T ac * iki. Hoja (leaf) U A
*sawa; *T ac *sawa- (green). Intestino (intestine) U A *si; T acC av -sere;
TacChm see; M ap killde. Llegar (arrive) U A *, asi ~ *’ asi; Mosetene
atsi. Maduro (ripe) U A *kw asi~ *kwasi (also cooked); TacC av esiri;
T acT a epohaha; PanAma wantonii’ ; Mosetene puisi. Piema (leg, thigh)
U A *kasi; TacC av -ca-ka; *Pan *ki i. 4
Proto-Aztecan *5
corresponds to several sibilants, but especially d. Examples
4
are: espiritu (spirit) A zT et mo ikooni:; PanAma yoSi; PanChac yoSini.
4
Estdmago (stomach) A zZac ipo ; Q, pusu. Funinculo (boil) AzTet
5
Ji: i:hkwepuni:; *Q, ‘ C’ upu; T acT a nacibo. Imperativo A zZac ii-;
Map -di. Partir (split, divide) A zZac -ielo:a; *Q,*C?iqta-; Mosetene det;
 5
Map trelolun; ChonSelk Jaje; Ala tsas. Pelar (peel) A zZac -piSka, - i:ma;
*T ac *poSi-; Mosetene ianak; M ap fiafln.

UA *h corresponds to h in a few examples: flecha (arrow) U A *-hu; Pan Chan

t&uhua. Levantar (raise) A zZac ki-ahakwi; *Q, *hatari-.

UA *m usually corresponds to m, as in examples: cabeza (head) U A *mo’ o;

*Q ,*um a; *Pan *mapo; Mosetene oflo. Comer (eat) U A *kumi ~ *kuma
(nibble); *Q,*mikhu-. D ar (give) U A *maka; ChonSelk m ah-ki (ofrecer).
5
Furtinculo (boil) A zT et ii: i:htomuni:; Mosetene onoy; M ap moy.
Golpear (hit) A zZac -maga; *Q,m aqa-. Mano (hand) U A *ma; U A-N P
mai; *Q ,*m aki; *T ac -me; T ac embai (brazo); *Pan *miklnl; ChonSelk
marr. Miedo (afraid) A zZac -mowti:a; *Q, *m aniaii-; TacC av moiya-;
Mosetene noyeye; ChonSelk mjo.

UA *n usually corresponds to n; it also corresponds to h. Note that *Tac

*n includes both n and n in its reflexes. Examples are: came (meat)
A zZac naka-; TacChm -yami; *Pan *nami. Encender (to light a fire)
U A *naya; TacChm inapokwa-; M ap iyimln. Necesitar (need) AzZac
moneki; Q m una-na; Mosetene emo&e; M ap euam yen. Nombre (name)
UA-Comanche (nk)hnia; PanCshb ani. Parecer (resemble) A zZac ne:si;
Mosetene efie; M ap femqey. Pedir (ask, beg) U A *tani; UA-Papago
taany ; *Q,*m ana-.

UA *q may correspond to n in some languages. It also appears to have a

relationship with a nasal which is articulated toward the front o f the
mouth, for example m or Mapuche fronted n. Note that M ap q very
often corresponds to T ac m and Pan m. Examples are: doblar (fold,
bend) U A *qola; T acC av no’ka-. Gente (people) U A *kuqa (spouse-
male); PanCap noki; M ap kona (servant, soldier), kure (spouse-female);
ChonTe kena. Lengua (tongue) U A *liqi; UA-Huichol nenf; *Tac
*-yana; *Pan *ana; Mosetene num; M ap kewin.

UA *1 usually corresponds to ly in Q,; to * f in *T ac; to r and n in Pan; to 1

11 1
and in M ap; and to in the Fuegian languages. Examples are: bramar
(roar) A zZac nana:lka; PanAma ward’ -; M ap waulutuy. Bueno (good)
A zZac kwali; *Q, *alyi. Carbdn (charcoal) A zZac tekol; *Q, kailyiinla;
Map kuyul. Casa (house) U A *kali; *T ac *-tare; M ap nikal. Chicharra
(cicada) A zZac £ikili:6i; T acT a reiri; M ap dille. Escribir (write, paper)
A zZac kihkwi:lo:a; *Q, *qilyqa-; T acC av kirika ~ kilika; Mosetene
kirika; M ap 6illka. Estrella (star) A zZac si:tali:n; ChonSelk telu; Ala
tsele-’sewe. Hombro (shoulder) A zZac iahko:l; ChonSelk q’ iol-te-nn.
Jalar (pull) A zZac -tila:na; T acT a rire-; PanChac nini-. Junco (reed)
A zZac to:li:n; Q,tutura. Mariposa (butterfly) A zZac pa:pa:lo:-t; *Q ,*pily-
pintu; PanSC pinpln.

UA *r corresponds to r and to zero. Note that zero is a reflex of *T ac *r

and *f. Examples are: buho (owl) U A *tukur(i); Q,kurkuku; M ap koo.
Frio (cold) U A *sl(pi); UA-Huichol S6(ri); *Q, *Ciri. Grande (large)
U A *w i-; A zZac we:(yi); UA-Huichol we(ri-); *T ac *afi-da (note
1
that *t is also a correspondence o f U A * ).
UA *w corresponds to w. It may be involved with r in some way. Note that
semivowels w and y are reflexes o f *T ac *r and *f. Examples are: cielo
(sky) A zZ ac elwiak; M ap wenu. Cuerno (horn) U A *’ awa; *Q,waqhra;
4
PanMar yiwaSinda. Decir (say) U A * , a\t ~ *’ awa; *Q_wiiya-; *T ac *a-;
PanMar yowiwi. Gordo (fat) U A *wi-; UA-Huichol w i:(y i); *Q, *wira.
UA *y corresponds to y in a few examples: hambre (hunger) A zZac maya:n;
A zTet maydna; *Q, *yarqa(6i)-; M ap qlfiln; ChonSelk karayne. Morir
(die) UA-Comanche (ti)yaai-; UA-Shoshone (ti)ye-; *Q, *aya (corpse);
•T a c *iye-.

University o f California at Irvine.

REFEREN CES

A non, n.d. [grammar and dictionary notes on Tsimane], New Tribes Mission,
Cochabamba, Bolivia.
D ’A n s , Andre-Marcel, and V a n d e n E y n d e , Els, 1972, Lixico Amahuaca (Pano),
Universidad Nacional M ayor de San Marcos, Peru, Centro de Investigacidn
de Lingiiistica Aplicada, Documento de Trabajo No. 6.
A r m e n t i a , F. Nicolas, igoi-1902, Los indios Mosetenes y su lengua, Anales de la
Sociedad Cientifica Argentina, Buenos Aires : vol. 52 (1901), pp. 145-160,
288-306; vol. 53 (1902), pp. 49-65, 150-157, 234-241, 292-297; vol. 54
(1902), pp. 49-60, 144-150, 181-201, 272-282.
B r e w e r , Forrest, and B r e w e r , Jean G ., 1962, Vocabulario Mexicans de Tetelcingo,
Morelos, Mexico, D.F. : Institute Lingiiistico de Verano, 374 p.
C l a i r - V a s i u a d i s , Christos [ = Clairis], 1976, Esquisse phonologique de l’aymara
parle au Chili, La Linguistique, 12, 2, pp. 143-152.
C l a i r i s , Christos [ = Clair-Vasiliadis], 1977, Premifcre approche du qawasqar :
identification et phonologic, La Linguistique, S3, 1, pp. 145-152.
D a v i s , Irvine, 1966, Numic consonantal correspondences, International Journal
o f American Linguistics, 32, 2 (April), pp. 124-140.
E r i z e , Esteban, i9 6 0 , Diccionario comentado: Mapuche-Espaftol: Araucano, Pehuenche,
Pampa, Picunche, Ranculche, Huilliche, Buenos Aires, Editorial Yepun, Bahia
Blanca, 5 5 0 p.
K e y , Harold and K e y M ary Ritchie, 1953, Vocabulario Mejicano: de la Sierra de
Zacapoaxtla, Puebla, Mexico, D.F.: Institute Lingiiistico de Verano, 232 p.
— 1967, Bolivian Indian tribes: classification, bibliography, and map o f present language
distribution, Summer Institute o f Linguistics, 128 p.
K e y , M ary Ritchie, 1966, Vocabulario Castellano regional, Vocabularios Boli­
vianos, No. 5, Cochabamba, Bolivia, Institute Lingiiistico de Verano, 62 p.
— 1968, Comparative Tacanan phonology: with Cavinena phonology and notes on Pano-
Tacanan relationship, The Hague, Mouton, 107 p.
— 1968, Phonemic pattern and phoneme fluctuation in Bolivian Chama
(Tacanan), La Linguistique, 2, pp. 35-48.
— 1976, La fluctuacidn de fonemas en la teoria fonoldgica, Signos, g, 1, pp. 137-
61
143, Universidad Cat ica de Valparaiso, Chile.
— 1978, Araucanian genetic relationships, International Journal o f American
Linguistics, 44, 4 (October 1978), pp. 280-293.
K ey, M ary Ritchie, 1979, The grouping o f South American Indian languages,
Tubingen, Gunter Narr, 170 p.
— 1979, Phoneme fluctuation and minimal pairs in language change, Linguis-
tique fonctiomelle : M ats et perspectives : pour Andrt Martinet, ed. by M ortlza
M a h m o u d i a n , Presses Universitaires de France, pp. 305-310.
— and C l a i r i s , Christos, 1978, Fuegian and Central South American language
relationships, Actes du X L II International Congress o f Americanists, Paris
(September 1976), pp. 635-645.
K l e i n , Sheldon, 1959, Comparative Mono-Kawaiisu, International Journal o f
American Linguistics, 25, 4 (October), pp. 233-238.
M a r t I n , Eusebia Herminia, 1969, Bosquejo de estrvetura de la lengua Aymara:
fonologla, morfologia, Buenos Aires, Universidad de Buenos Aires, Facultad
de Filosofia y Letras, 75 p.
M a t t e s o n , Esther, W h e e l e r , A lva, J a c k s o n , Frances L ., W a l t z , Nathan E .,
C h r i s t i a n , Diana R ., 1972, Comparative studies in Amerindian languages, The
Hague, Mouton, 251 p.
N a j l i s , Elena L ., 1975, Diccionario Selknam, Buenos Aires, Universidad del
Salvador, Facultad de Historia y Letras, Instituto de Filologfa y LingUfstica,
>59 P-
O l s o n , Ronald D ., 1964-1965, M ayan afHnities with Chipaya ofBolivia, I and II,
International Journal o f American Linguistics, 30, 4, pp. 313-324; 31, 1, pp. 29-38.
O r r , Carolyn, and L o n o a c r e , Robert E., 1968, Proto-Quechumaran, Language,
44, 3, PP- 528-555-
— and W r i s l e y , Betsy, 1965, Vocabulario Quichua: del Oriente del Ecuador, Serie
de Vocabulario Indigenas, No. 11, Instituto Linguistico de Verano, and
Ministerio de Educacidn, Qjuito, Ecuador, 152 p.
S h e l l , Olive Alexandra, 1 9 6 5 , Fano reconstruction, University o f Pennsylvania
Dissertation, 2 6 7 p .
S t a r k , Louisa R ., 1970, M ayan affinities with Araucanian, Papers from the
6th Regional Meeting, Chicago Linguistic Society, pp. 57-69.
— 197a, Maya-Yunga-Chipayan: a new linguistic alignment, International
Journal o f American Linguistics, 38, 2 (April), pp. 119-135.
S u Ar e z , Jorge A ., 1969, Moseten and Pano-Tacanan, Anthropological Linguistics,
11, 9 (December), pp. 255-266.
— 1973, Macro-Pano-Tacanan, International Journal o f American Linguistics, 33,
3 (Juiy)> PP- -
137 154 -
V rvo E s c o t o , Jorge A ., 1 9 7 2 , E l Pohlamiento Ndhuat en E l Salvadory otros Paises
de Centroamirica, Ministerio de Educacidn, San Salvador, El Salvador,
Central America.
V o e g e l i n , C . F ., V o e g e l i n , F. M ., and H a l e , Kenneth, 1962, Typological
and comparative grammar o f Uto-Aztecan: I (Phonology), International
Journal o f American Linguistics, 28, 1, Supplement (January), 144 p.
W h e e l e r , A lva, 1973, A comparison o f concepts and their formalizations in South
American Indian languages, American Anthropological Association, New
Orleans.