Uto-Aztecan comparative linguistics and etymological databases Текст научной статьи по специальности «Языкознание и литературоведение»

Ilia Peiros Santa Fe Institute

Uto-Aztecan comparative linguistics and etymological databases

The paper presents a general survey of Uto-Aztecan historical studies. It discusses the genetic classification of the family, phonological reconstructions of individual daughter languages and of Proto-Uto-Aztecan itself. Lexical material used in the paper is taken from electronic databases created for the "Evolution of Human Languages" Project (Santa Fe Institute).

Keywords: Uto-Aztecan languages, lexicostatistics, comparative linguistics, etymological databases.

The project "Evolution of Human Languages" was launched in 2002 by Murray Gell-Mann and Sergei Starostin. One of its main aims is to build a network of etymological databases for all language families of the world, in which, for each known family, we plan to have:

— a system of phonological correspondences connecting many, if not all, languages within this family;

— an extensive etymological corpus supporting these correspondences;

— a lexicostatistical classification based on reliable etymologies.

After this type of work was performed for some better known families (results are available at starling.rinet.ru), it became clear that in many cases, instead of a full-scale study of every language family, we can only conduct a preliminary survey based on existing material.

The current paper briefly presents the results of such a survey on Uto-Aztecan (UA), one the better known language families of Northern America. Due to the nature of the paper, no detailed discussion of literature and alternative views is included.

The family consists of about 60 languages. Its generally assumed classification is given in [Campbell 1997: 134]:

Uto-Aztecan

Northern Uto-Aztecan Numic

Western

Northern Paiute (Oregon, Idaho, Nevada) Monache (California) [obsolescent] Central

Shoshone, Panamint [obsolescent], Comanche [obsolescent] (Nevada, Utah, Wyoming, Oklahoma) Southern

Southern Paiute (Utah, Nevada, California, Arizona) Ute, Chemehuevi [obsolescent] (Utah, Colorado, California, Arizona) Kawaiisu [obsolescent] (California) Tubatulabal [moribund] (California)

Journal of Language Relationship • Вопросы языкового родства • 4 (2010) • Pp. 90-116 • © Peiros I., 2010

Takic

Serran: Serrano [moribund], Kitanemuk [extinct] (California) Cahuilla [moribund?], Cupeno [moribund] (California) Luiseno-Juaneno [obsolescent] (California) Gabrielino [extinct], Fernandeno [extinct] (California) Hopi (Arizona)

Southern Uto-Aztecan Pimic (Tepiman)

Pima-Papago (Upper Piman) (Arizona, Sonora) Pima Bajo (Lower Piman, Nevome) (Sonora)

Northern Tepehuan, Southern Tepehuan (Sonora, Durango, Jalisco) Tepecano [extinct] (Jalisco) Taracahitic

Tarahumaran

Tarahumara (Chihuahua) Guarijio (Chihuahua, Sonora) Tubar [extinct?] (Chihuahua) Cahitan (Yaqui-Mayo-Cahita) (Arizona, Sonora) Opatan

Opata [extinct] (Sonora) Eudeve [extinct] (Sonora) Corachol-Aztecan Cora-Huichol Cora (Nayarit) Huichol (Nayarit, Jalisco) Nahuan (Aztecan)

Pochutec [extinct] (Oaxaca) Core Nahua

Pipil (El Salvador, extinct in Guatemala and Nicaragua) Nahuatl: many dialects (Mexico).

"[Languages] known to have fewer than 10 speakers are specified as 'moribund'; languages with more than 10 but fewer than 100 speakers are labeled 'obsolescent'" (p. 133).

An additional list of extinct languages which may have belonged to UA is given on pp. 133-134 of Campbell's book as well as in other sources, like.

In a somewhat impressionistic way, typical of the rest of the book as well, Campbell does not justify his classification. The only modern UA classification that is based on complete, explicit presentation of the evidence is W. R. Miller's paper from 1984.1 Using lexicostatistics, Miller (p. 13) comes to the following conclusion: "thus there seem to be five groups that are more or less coordinate with each other: Numic, Tubatulabal, Takic, Hopi and SUA" (South Uto-Aztecan); this is supported by the matrix of 'cognate density'. Further investigation of this matrix (see the table on p. 114 below) suggests, however, that lexicostatistics allows to generate a different classification.

Let us reduce Miller's matrix to a compressed version, where each transparently obvious group is represented by the lowest percentage that one of its members shares with other groups. This version looks as follows:

1 Hale's earlier lexicostatistical classification [Hale 1958-59] also presents the data.

*Numic *Takic Hopi *Sonoran *Aztecan

*Numic X 22 22 23 14

*Takic 22 X 26 24 18

Hopi 22 26 X 26 24

*Sonoran 23 24 26 X 29

*Aztecan 14 18 24 29 X

It can be seen that *Numic and *Aztecan form two ends of the spectrum, while the other groups are somewhere in between. Normally, *Sonoran should be grouped with *Aztecan, but it shows much higher figures with *Numic. Such discrepancies are a clear indication that we are dealing with the effects of undetected borrowings within related language groups.

Restricting ourselves to reconstructed intermediate 50-item proto-lists2, we can overcome this obstacle and obtain the following Uto-Aztecan classification (utza.dbf3):

Uto-Aztecan

1. *Numic

2. Main group

a. Central:

(i) Tubatulabal, Hopi Kitanemuk

(ii) *Cupan

b. Southern:

(i) *Sonoran

(ii) *Aztec

The weakest point of this classification is that it places Hopi close to Tubatulabal; this may also reflect later contacts between the two, not detected so far.

According to glottochronological calculations4 based on the complete 100-item lists for attested languages, the Proto-UA language began to disintegrate approximately 4,500 — 4,200 years ago. It should be mentioned, however, that calculations based on Sergei Yakhon-tov/Sergei Starostin's 35-item list reveal an entirely different age — only 3,500 years. This is surprising, since usually calculations based on the shorter list generate slightly older dates than those based on the longer list. The reasons for such a discrepancy remain unknown.

Numic languages

The Numic group is usually divided into three branches, with two languages each — Southern: Kawaiisu and Ute (Chemehuevi, Southern Paiute and Ute); Central: Comanche and Shoshoni (Panamint and Shoshoni dialects); Western: Mono and Northern Paiute.

This classification is accepted by Miller (1984), but the matrix of similarities given is his paper suggests a different tree structure:

2 See [G. Starostin: 2009].

3 All of the databases mentioned in this paper will be available on EHL's website (http://starling.rinet.ru) in the near future and can also be obtained upon request from the author.

4 If not indicated otherwise, all the lexicostatistical and glottochronological results have been generated with the aid of STARLING, a computerized linguistic environment developed by S. Starostin.

Mon NPa Pan Sho Com Kaw Che Spa Ute

Mono X 77 59 58 57 52 50 53 52

N.Paiute 77 X 58 58 58 56 55 58 57

Panamint 59 58 X 87 79 54 61 62 59

Shoshoni 58 58 87 X 88 55 58 62 61

Comanche 57 58 79 88 X 49 54 59 59

Kawaiisu 52 56 54 55 49 X 75 79 76

Chemehuevi 50 55 61 58 54 75 X 86 78

S.Paiute 53 58 62 62 59 79 86 X 87

Ute 52 57 59 61 59 76 78 87 X

Numic (more than 49%)

A. Central-Northern Numic (more than 57%)

1. Western: Mono and Northern Paiute (77%)

2. Central: Panamint, Shoshoni, and Comanche (more than 79%)

B. Southern (more than 75%)

3. Ute: Chemehuevi, Sothern Paiute, and Ute (more than 78%)

4. Kawaiisu

The same classification is obtained using S. Starostin's 35-item list (numi.dbf).

Southern Numic languages

Phonologically, the Ute languages are very similar to each other. Southern Paiute, studied in [Sapir 1930], has the following system:

Initial consonants:

P- m- w-

t- n-

c- y- s-

k-

kw-

h-

Word-medial position (single consonants and clusters):

-v- -PP- -mp- -mm- -w-

-r- -tt- -nt-, -nc- -nn- -n-

-c- -cc- -nc- -y- -ss-

-Y- -kk- -gk- (-g-)

-yw- -kkw- -qkw- -gw-

-?- -h-

Final consonants: none

Vowels (short and long):

i, i: i, i: u, u:

e, e: o, o:

a, a:

The language is well known for its "consonant gradation". Sapir (1930, 62) talks about spi-rantization (S), gemination (G), and nasalization (N):

S: uywi-vi 'grass'+Abs

G: kucca-ppi 'ashes'+Abs

N: miyi-mpici 'gopher'+Abs

Morphophonemic [Sapir 1931], accentual [Miller 1980], and other explanations were suggested for these phenomena, but the real reason for "gradations" seems to be purely historical. A simple reconstruction may be based on the following considerations:

— the stem in Southern Paiute is usually formed by two syllables and thus may have the form of (a) CVPPV, where PP is any geminated stop, affricate or 1; (b) CVPV, where P is any other consonant or a cluster of nasal + stop;

— in most cases Gemination is associated with the first type of stems, whereas Spiranti-zation is mainly found within the second: (a) CVPPV + G, (b) CVPV + S;

— therefore, Spirantization and Gemination, predetermined by the consonants in the second syllable, must have developed later in the history of the languages;

There are, however, also other cases, each with a significantly smaller number of examples: (c) CVPPV + S: kicci-vi 'saliva' Abs, (d) CVPV + G: movi-ppi 'nose' Abs. Historically these can be regarded as showing traces of final consonants: *h (or *s) and *?. Nasalization can be explained as a trace of a lost final nasal: (e) *CVPVN, *CVPPVN.

As a result, we can identify five types of stems in Southern Paiute:

(a) *CVPPV#: *kucca# > kucca-ppi 'ashes'

(b) *CVPV#: *uywi# > uywi-vi 'grass'

(c) *CVPPVh: *kiccih > kicci-vi 'saliva'

(d) *CVPV7: *movi1 > movi-ppi 'nose'

(e) *CVPVN, *CVPPVN: miyiN > miyiN-mpici 'gopher', *wikkuN > wikku-mpicc 'buzzard'

The same three types of "consonant gradation" are found in the other Ute languages as well as Kawaiisu.

The Kawaiisu phonological system, defined on the original stems of the language, is found in [Zigmond et al. 1990, pp. 5-6]:

P v b m w

t r d n

c 3 s

c 3 s

k Y

kw yw

? h

i, i: i, i: u, u:

e, e: o, o:

a, a:

The Proto-Southern Numic system of consonants is reconstructed as follows: Finals:

Initial position:

*# *-? *-h *-N

*P *m *w

*t *n *s

*c *y

*k

*kw

*? *h

Medial position:

*-b- *-PP- *-m- *-mm *-w-

*-d- *-tt- *-n- *-s-

*-3- *-cc- *-y-

*-g- *-kk-

*-gW- *-kkW-

*-h-

The supporting correspondences are given in the following table:5

Kawaiisu Proto-S. Numic S. Paiute Chemihuevi Ute Note

P *P- P P P

v *-b- v v v

P *-PP- PP P P

b *-mb- mP mP P < *N-Pi

m *m- m m m

m *-m- mm m m

w *w- w w w

w *-w- # # #

w qW w w -VN, iwV

t *t- t t t

r *-d- r r r

5 The sources are as follows: Kawaiisu — [Zigmond, et al. 1990]; Southern Paiute — [Sapir 1931]; Chemihue-vi — [Press 1979]; Ute — [Givon 1979].

Kawaiisu Proto-S. Numic S. Paiute Chemihuevi Ute Note

t *-tt- tt t t

n *n- n n n

n *-n- n n n

c *c- c c c

3' 3 *-3- c c c

c *-cc- cc c c

s, s *s- s s s

s, s *-s- ss s s

y *y- , y y y

y *-y- y y y

k *k- k k k

g *-g- g g g

k *-kk- kk k k

Y *-nk- qk qk k

n *-q- q q, (u+m+u) #

kw *kw- kw kw kw

Yw *-gW- w gw Yw-, Y(+i)-

kw *-kkW- kkw Kw kw

? *?- #, ? #, ? #

? #, ? ? ?

h *h- ?, # h ?, #

h *-h- # h #

The etymological data supporting these correspondences can be found in the database numiet.dbf. The vocalic system of Proto-Southern Numic requires further investigation.

Central Numic languages

All Central Numic languages have six vowels (i, e, i, a, u, and o) that can be short or long. The systems of initial consonants of modern Central Numic languages are identical to those of Southern Numic, while in the medial position there occur many more phonemes (single consonants or clusters).

Medial position consonants of Shoshoni [SOD]:

b pp mb ?b m (mm) hm w hw

d tt nd ?d n nn hn ?n s ss

3 cc n3 y hy ?y

g kk qg q (hq)

gw kkw qgw qw

? h

It is possible that all clusters containing ?, h and nasal consonants are not actually found within roots, but are rather the results of various morphophonemic processes.

Another noticeable Central Numic feature is the existence of an additional type of "consonantal gradation". Thus, in Panamit one finds the following types of articulation:

— spirantized: [nayava?a] 'on the bighorn' (naka 'bighorn', pa?a 'on')

— geminated: [tuappa?a] 'on the son' (*tua? 'son', pa?a 'on')

— nasalized: [piyimba?a] 'on the duck' (piyin 'duck', pa?a 'on')

— aspirated: [hainci^a?a] on the friend' (haincih 'friend', pa?a 'on')

(Miller et al. 2005: 415)

Apparently the "aspirated gradation" corresponds to the final *-h that is reconstructed for Southern Numic.

The following reconstruction of Proto-Central Numic is based only on my analysis of four languages / dialects: Panamint [Dayley 1989], Shoshoni and Western Shoshoni [SOD], Comanche [Robinson and Armagost 1990].

The sets of initial and final consonants are identical to those of Proto-Southern Numic, with the exception of not found in PCN.

The set of PCN medial position consonants is as follows:

*-b- *-?b- *-PP *-mb- *-mm- *-hm- *-?m- *-w- *-hw-

*-d- *-?d- *-tt- *-nd- *-n- *-nn- *-hn- *-?n- *-ss

*-3- *-?3- *-cc *-n3- *-y- *-hy- *-?y-

*-g- *-?g- *-kk (*-hq-)

*_gw_ *-kkw- *-5gW- (*-gw-)

*? *-h-

It should be noted that some of the reconstructed medial consonants may in fact be clusters found at morpheme boundaries.

The following correspondences support this reconstruction:

Proto-C. Numic Panamint Shoshoni W. Shoshoni Comanche Notes

*P P P P P

*-b- P b P b / hP, p

*-?b- P :?b, (:b) P bV, :bV

*-PP- PP P PP hp

*-mb- mP mb mP P

*m- m m m m

*-?m- m :m, m m m

*-m- mm mm m m

*-hm- hm hw hm Hm

m ?w, :?w :w ?m, :mV 2 examples

*w- w w W w

*-w- w # # w

*-hw- hw hw, h h

Proto-C. Numic Panamint Shoshoni W. Shoshoni Comanche Notes

*t- t t t t

*-d- t t t r

-?d- t :?d, ?d t :r, (e:t)

*-tt- tt t t ht

*-nd- nt nd nt t

*n- n n n n

*-nn- nn nn n n

*-n- n n n n

*-hn- hn hn hn hn

*-?n- nn :?n ?n, :n

*c c c t c

*-3- c 3' 3 c c

*?3 c :?3' :3 c c

*-cc- cc c, c cc hc, :c

*-n3- nc n3' n3 c

*s- s s s s

*-ss- s s, s s s

*y- y y y y

*-y- y y y y

*-hy- hy hy hy hy

*-?y- y ?y y y

*k k k k k

*-g- k g k k

*-kk- kk k kk hk, (:k)

*-qg- qk ng nk k , :k

-?g- kV? :?g k :k

*-q- q n n n

*-?q- q nn n

(*-hq-) hm hn hm Hm 1 example

*kw kw / k(+u) kw / k(+u) kw / k(+u) kw / k(+u)

*_gw- kw gw / (u+)g ku / (u+)k kw / (u+)k

*-kkW- kkw / (u+)k kw / kw / (u+)hk

*-qgW- qkw ngw nkw kw

(*-qW- ) qw w 1 example

*?- # # # #

? ? ? ?

*h- h h h h

*-h- #, h h h h

The following reconstruction of Proto-Numic is based on a comparison of Proto-Central and Proto-Southern systems. A significant gap, however, is the absence of data on Western Numic languages.

A preliminary reconstruction of the vowel system seems to be quite simple (it is not quite clear if we need to reconstruct long vowels):

*i *i *u

*o

*a

The consonantal system of Proto-Numic is as follows: Initial consonants:

*P *m *w

*t *n *s

*c *y

*k

*kw

*? *h

Final consonants:

*# *-? *-h *-N

Medial consonants:

*-b- *-PP (*-mb-) *-m- (*-hm-) *-w-

*-d- *-tt- *-n- *-hn- *-s-

*-cc *-y- *-hy-

*-g- *-kk *-9g- *-hq-

*-gW- *-kkW-

*? *-h-

Some of the medial combinations seem to have been clusters, found on morpheme boundaries. It is possible that further investigation will fill in a few gaps in the reconstruction, such as an unexpected absence of *~3~.

Correspondences supporting the reconstruction are:

*Proto-Numic PCN PSN Notes

*P- *P- *P-

*-b- *-b- *-b-

*-PP- *-PP- *-PP- few examples

(*-mb-) *-mb- *-mb- alternatively: *-N-pa

*t- *t- *t-

*-d- *-d- *-d-/-c- c — next to *i

*Proto-Numic PCN PSN Notes

*-tt- *-tt- *-tt- few examples

*c- *c- *c

*-cc- *-cc- *-cc-

*k- *k- *k-

*-g- *-g- *-g-

*-kk- *-kk- *-kk-

*-qg- *-qg-

*kw- *kw- *kW

*_gw- *-gW- *-gW-

*-kkW- *-kkW- *-kkW-

*m- *m- *m-

*-m- *-m-,*-mm- *-m- / *-w-

(*-hm-) *-hm- *-m- 1 example

*n- *n- *n

*-n- *-n-, *-nn- *-n-

*-hn- *-hn- *-n-

*-hq- *-hq-

*-qW- *-gW-, *-m- *-w- / # # after *o

*w- *w- *w-

*-w- *-w- *-w- few examples

*y- *y- *y-

*-y- *-y- / # *-y- / # # after *i

*-hy- *-hy- *-y-

*s- *s- *s-

*-s- *-s-, *-ss- *-s-

*h- *h- *h-

*-h- *-h-, # *-h-, #

*?- *?-

Examples supporting these correspondences are collected in numiet.dbf. Further investigation is needed to finalize the set of PN medial consonants and firmly distinguish them from other possible morpheme-boundary clusters.

The reconstructed system allows for further 'fine tuning'. E. g., one can suggest that the two types of stops and affricates actually represent an opposition between plain and glottal-ized consonants: *p ~ *p', *t ~ *t', etc., which is neutralized in the initial position.

The EHL Numic etymological database (numiet.dbf) consists of more than 600 entries. Most of these represent modified versions of etymologies found on such lists as [Iannucci 1972] and [Miller 2003].

Glottochronological calculations suggest that the disintegration of Proto-Numic may have begun about 2,500 years ago.

Takic languages

Miller (1984) distinguishes two branches of Takic languages:

a. Serrano-Gabrielino

(1) Serranan: Serrano, Kitanemuk

(2) Gabrielino (Gabrielino, Fernandino)

b. Cupan

(1) Cupeno, Cahuilla

(2) Luiseno

His lexicostatistical matrix, however, does not support any reasonable classification of Serrano and Gabrielino, probably due to the quality of the wordlists:

Gabrielino Serrano Cahuilla Cupeno Luiseno

Gabrielino X 45 42 34 38

Serrano 45 X 50 38 35

Cahuilla 42 50 X 65 50

Cupeno 34 38 65 X 48

Luiseno 38 35 50 48 X

It is possible to improve the results for Serrano by including forms only from Kroeber (1907) and by adding Kitanemuk words collected by Harrington [Anderton 1988]. A new matrix (based on taki.dbf) is:

Kitanemuk Serrano Cahuilla Cupeno Luiseno

Kitanemuk X 75 50 50 50

Serrano 75 X 61 58 52

Cahuilla 50 61 X 80 61

Cupeno 50 58 80 X 61

Luiseno 50 52 61 61 X

Some differences in the percentages of lexical similarities between the languages are mainly due to differences in the wordlists. Nevertheless, the overall pattern remains the same, suggesting that Miller's classification (without Gabrielino) may be accepted:

Takic (more than 52%)

a. Kitanemuk + Serrano (75%)

b. Cupan (more than 61%)

Cahuilla + Cupeno (80%) Luiseno

The three Cupan languages are quite well described. The system of Luiseno can be presented as follows:

Vowels (spelled as i, u, e, a, o):

i u

E D

a

According to Bright (1965), all vowels may be short / long and stressed / unstressed, but certain structures predetermine the choice of length and / or strength. For example, 'in final accented position, only long vowels occur' (p. 344); 'when an initial syllable has a long vowel, it is always accented' (p. 345).

Consonants:6

P v m w

t n l s

r s

C y

k 5 x

kw kw

q

qw

? h

Elliott (1999: 14-15) writes that (all?) consonants between two short vowels are either long or accompanied by ? if the first vowel is stressed.

My analysis of Takic comparisons suggests that the vocalic length distinction in Luiseno is of later origin7. The rules to eliminate it are as follows:

— the root may consist of one or two syllables: CV, CVC(C), CVC(C)V, CVC(C)VC(C);

— in nouns, monosyllabic roots always have long vowels unless the final consonant is ?: ku:y 'husband', wi? 'fat, grease';

— in disyllabic words stress can be placed on the first or the second syllable. If a stressed syllable ends in a consonant8, the vowel is long, otherwise the vowel is short: ?aana 'ant', ?amuu 'mescal plant, Agave deserti', ?ecvu-s 'left hand', qawi? 'forehead';

— verbs are predominantly monosyllabic with short vowels. Disyllabic verbs follow the same rules as nouns: ?ula?q=i- 'to sew (vt.)'.

When modern Luiseno forms do not follow the suggested pattern, we assume that they may have recently lost a consonant: caay-i- 'to sift, winnow' < **ca?ay-i- (cf. Cahuilla ca?a-l 'sieve, trap'), qiin-i 'to plow' < **qiyin-i (cf. Cahuilla qiyne- id.).

6 I was not able to complete a list of consonantal clusters acceptable in Luiseno roots.

7 For other views on stress in Cupan languages see Hill and Hill 1968, Munro 1990, etc.

8 One should mention, however, that in some cases it is not clear how to divide a word into syllables, cf. kuumal 'poker' < ** kuu=mal (ku- 'fire').

The Proto-Cupan system is reconstructed as followed: Consonants:

*p *v *m *w

*t *n *l *s

*c *y

*k *x

*kw

*q

*qw

*? *h

Vowels:

*i *i *u

*e ( =*e) *o (=*D)

*a

The correspondences behind these proto-phonemes are mostly trivial.

The EHL etymological database (takiet.dbf) currently contains 562 entries that reproduce, with some modifications, Miller's list of Proto-UA etymologies [Miller 2003], with additional comparisons from [Munro 1990] and some other sources. No direct search for new etymologies has been conducted.

Glottochronology suggests that the Proto-Takic language may have begun to disintegrate about 3,200 years ago.

Sonoran languages

A list of Sonoran languages can be found in [Miller 1984]:

1. Tapiam: Upper Piman (Papago, Pima, Nevome), Lower Piman, Northern Tepehuan,

Southern Tepehuan (Southern Tepehuan, Tepecano)

2. Taracahitian

(a) Tarahumara (several dialects) Guarijio (Highland, Lowland)

(b) Opatian: Opata, Eudeve, Jova?

(c) Cahita (Mayo, Yaqui)

3. Tubar

4. Corachol: Cora, Huichol.

The Opatian languages are extinct and I was not able to obtain wordlists for any of them.

Tapiam. A brief description of Papago phonology can be found in [Mathiot 1973: 113-114]:

Bilabial Dental Alveolar Palatal Velar Glottal

Tense stop P t c k

Lax stop b d d 3 g ?

Fricative s s h

Nasal m n

Lateral l

Glide w

Vowels (short and long):

i i u

o

a

Stress is predictable.

The systems of other Tepiman languages are listed in [Bascom 1965].

A reliable reconstruction of Proto-Tepiman was conducted by Bascom in 1965 and is the main source of the Tepiam database (tepiet.dbf, with more than 300 etymologies). Additional lexical data are available only for Papago (Saxton & Saxton 1983; Mathiot 1973) and Nevome (Pennington 1979). Papago is now sufficiently well represented in the database. Nevome, known only from a dictionary composed in the XVIIth or XVIIIth centuries, has not been studied.

The following system is proposed for Proto-Tepiman:

Root structure: CVCV with no consonantal clusters.

Consonants (initial and medial):

*P *b *m *v

*t *d *n *r *s

*k *g

*? *h

Vowels (may be short or long):

*i *i *u

*o

*a

This reconstruction is based on the following set of phonological correspondences:

Tepiman N. Tepehuan S. Tepehuan U. Pima L. Pima

*P P P P P

*b b b b b

Tepiman N. Tepehuan S. Tepehuan U. Pima L. Pima

*m m m m m

*v v, # v v v

*t t / ty (i+, +i) t / ty (i+, +i) C / t (+i,i,u) t / C (i+, +i)

*d d / dy (i+, +i) d / dy (i+, +i) 3 / d (+i,i,u) d / 3 (i+, +i)

*n n / n(i+, +i) n / n(i+, +i) n/ n (+i,i,u) n, n (i+, +i)

*r r / l (i+, +i) r / ly (i+, +i) d / l (+i) r / l (+i)

*s s / s (i+, +i) s / s (i+, +i) s / s (+i) s / s (i+, +i)

*k k, h k k, h k

*g g, # g, ? g g

*? # ?

*h h, # h, # h, # h, #

*i i i i i

*i i i i i

*u u u u u

*o o o o o

*a a a a a

These correspondences are supported by reconstructions, collected in tepiet.dbf.

Tarahumaran. The two main sources for Tarahumaran languages are a dictionary of Central Tarahumaran [Hilton et al. 1993], and a grammar and dictionary of Highland Guarijio [Miller 1996].

The Guarijio system of phonemes is as follows:

Consonants (initial and medial):

P m w

t n l s

C y

k g

? h

Two types of medial clusters are found in the roots: -? + consonant and -h + consonant. Vowels:

i u

e o

a

Stress is usually placed on the second (last) syllable of the root, but there are a significant number of exceptions.

The system of Tarahumara is quite similar, as is the reconstructed Proto-Tarahumaran. Minor issues of the reconstruction are not discussed here.

The Tarahumaran database (taraet.dbf, more than 350 etymologies) is mainly based on Miller's materials. It should be mentioned, however, that the database includes only a small portion of possible etymologies for this group.

Mayo (Cahita). Lexical corpora for several closely related Mayo languages / dialects are found in three dictionaries: Mayo [Collard and Collard 1962], Yaqui [Estrada et al 2004] and Yoeme [Molina et al. 1999]. Only the Yoeme dictionary has been used for the time being. The phonological system of this variety is as follows:

Consonants (initial and medial):

bw p m v w

t n R l s

y

k g

? h

No consonant clusters are found within root morphemes. Vowels:

i u

e o

a

All vowels can be short and long; vowel length can sometimes fluctuate under certain conditions, probably caused by morphological factors that are not yet fully understood.

A small Mayo database was created with the sole aim of checking the validity of Yoeme forms used in comparisons with other Sonoran languages.

Corachol. For my study, two sources of Huichol data were available: a small dictionary of the language published as [McIntosh & Grimes 1954] and a manuscript of a much bigger dictionary prepared in 1982 [Grimes et al. 1982], which had never been published. It is this manuscript that is used as my primary source of information. The phonology of the language may be presented as follows:

Consonants (initial and medial):

P m W

t n r

c Y s

k

kw

? h

No consonantal clusters are found. Vowels (all vowels can be short or long):

i i u

e

a

Tones: H(igh) (V), L(ow) (unmarked). The origin of tones is not yet fully understood.

The Cora data is taken from [McMahon & McMahon 1959], but has not yet been properly analyzed.

A few additional extinct Sonoran languages are known with various degree of detaliza-tion. None of them have been used in the reconstruction, but some of their words mentioned in various etymological sources are given in the database.

The Proto-Sonoran system and a list of etymologies (about 350 entries) were produced by [Lionnet 1985]. The phonological system of the proto-language is as follows:

Consonants (initial and medial):

*P *m *w

*t *n *l *s

*c *y

*k

*kw

*? *h

Proto-Sonoran *l is not found in the initial position.

The suggested reconstruction is based on the following set of correspondences:

Proto-Sonoran *Tepiman Yoeme Huichol *Tarahumara

*P- *p, *v v' P h #

*kw *b bw kw *w, *kw(+u)

*w v v, w w *w

*t *t t t, c *t

*c *s c c *c

*s *h s s *s

*h h #

*? # #/ V?- #

*k *k k k *k

*m *m m m *m

Proto-Sonoran *Tepiman Yoeme Huichol *Tarahumara

*n *n n n *n

*l *r, *d l l *l

*y *d y y *y

Vowels:

*i *i *u

*o

*a

The vocalic correspondences are straightforward.

Although, as has already been mentioned, some Sonoran languages distinguish between long and short vowels, this distinction is not reconstructed for the proto-language.

Proto-Tepiman roots with long vowels normally correspond to Proto-Tarahumara roots with consonantal clusters. This observation permits to reconstruct Proto-Sonoran clusters as *-?C- and *-hC-. The root structure in the proto-language, best preserved in Guarijio, is *CV(H)CV. Further study is needed to clarify the origin of long vowels in Yoeme and Huichol.

The Proto-Sonoran etymological database (sonoet.dfb, about 400 entries) consists of all the acceptable comparisons from Lionnet's and Miller's sources. Additional research has also been conducted to find missing reflexes in various languages of the family.

Miller (1984) has identified four equal branches (groups) of the Sonoran languages. An alternative interpretation of his matrix, however, gives us the following tree:

Sonoran (34% and more):

A. Papago, Nevome, N. Tepehuan (79% and more)

B. 39% and more:

1. 51% and more:

(a) Guarijio and Tarahumara (83%);

(b) Opta and Eudeve (73%);

(c) Mayo and Yaqui (93%);

2. Tubar

3. Corachol (58%)

This classification is confirmed by my own lexicostatistical study based on 35 and 50 item-lists (sono.dbf). The 100-list gives, however, a different picture:

Nevo Papa Guar Tara Yoem Cora Huic

Nevome X 93 56 52 59 46 50

Papago 93 X 53 49 57 46 48

Guarijio 56 53 X 81 66 51 54

Tarahumara 52 49 81 X 61 47 51

Yeome 59 57 66 61 X 53 59

Cora 46 46 51 47 53 X 74

Huichol 50 46 54 51 59 74 X

The differences in classifications are presumably caused by etymological misinterpretations of the (so far) poorly understood data from Cora. Without it, the classifications are identical:

A. *Tepiam

B 1a. Tarahumaran 1b. Yeome 2. Huichol

Glottochronology suggests that Proto-Sonoran began to disintegrate about 3,200 years ago.

Aztecan languages

Miller classifies the Aztecan group into two branches: a. Pochutec, b. General Aztecan: Pipil, Aztec (Classical Aztec. Tetelcingo, Zacapoaxtla, and others). The extinct language Po-chutec, known from a short article by Boas (1917), was not included in Miller's matrix. Nevertheless, lexicostatistical calculations support his view:

35-item matrix:

N.Pu Tet Pip Zac Mec Poc

N.Puebla X 94 94 85 81 75

Tetelcingo 94 X 97 88 84 78

Pipil 94 97 X 91 88 79

Zacapoaxtla 85 88 91 X 81 76

Mecayapan 81 84 88 81 X 78

Pochutec 75 78 79 76 78 X

A detailed discussion of modern Aztec dialects can be found in [Canger 1988]. The comparative phonology of the family is well-known [Dakin 1982]. In 1978 Campbell and Langacker published the following set of Aztecan vocalic correspondences (p. 94):

C&L Pochutec Class.Azt. Tetelcino Zacapoaxtla Pipil

1 *a e a a a a

2 *a: a a: D a: a:

3 *o o o o o o

4 *o: u o: u o: u:

5 *i i i I i i

6 *i: i i: i i: i:

7 *i o i I i i

8 o,e,# a e e ee

9 *e: e e: ie e: e:

Dakin (1983) has shown that correspondence No. 7 is only found before consonants and thus does not reflect an additional Proto-Aztec vowel. A restricted distribution, how-

ever, is not a sufficient argument against a reconstruction, so the proto-system can be interpreted as:

*i (5) *i (7) *i: (6)

*e (8) *o (3) *e: (9) *o: (4)

*a (1) *a: (2)

The Proto-Aztecan consonantal system, suggested by Dakin [1979: 50], is formed by 15 consonants:

*P *m *w

*t *n

*A *l

*c *s

*c *y *s

*k

*kw

*?

This reconstruction is based on the following set of correspondences:

Proto Pochutec Class.Azt. Tetelcino Zacapoaxtla Pipil

*p P, (b) P P P P

*m m m m m m

*w w w w w w

*t t, (d) t t t t

*n n, n n n n n

*A t, (d) A A t t

*l l l l l l

*c c c c c c

*s s s s s s

*c c c c c c

*s s s s s s

*y y y y y y

*k g, (k) k k k k

*kw g, (k) kw kw kw kw

h ? h h h

The phonemes *c and *s are often found in proximity with *i and *i:; this indicates their secondary origin.

The Aztec database (azttet.dbf) contains about 350 entries, which obviously represents only a small portion of the proto-language lexicon. Glottochronological calculations suggest that the disintegration of Proto-Aztecan began about 2,000 years ago; Kaufman [2001] talks about 500 CE.

Hopi

The dialects of Hopi form a separate branch of UA. Whorf [1946] identifies four of them (a claim that I am not able to verify through lexicostatistics): First Mesa (or Whorf's Polacca), Mishongnovi (or Whorf's Toreva), Shipaulovi (or Whorf's Sipaulovi), Third Mesa (or Whorf's Oraibi). The best known is the dialect of the Third Mesa [Hopi 1998]. Its phonological system is as follows:

Consonants:

p v m w

t n l s

c * / r

ky n y

k

kw qw

q qw

? h

The opposition of velars and uvulars is lost before front vowels. Vowels (long and short) (the vowels are spelled as i, e, u, o, a, o):

i i

0 o

£ a

Stress is predictable.

Tones: the Third Mesa dialect has developed tonal distinctions on long vowels, diphthongs, and vowel + sonorant sequences (l, m, n, q, qw). The tones are either falling or level. The falling tone corresponds to one of several sequences in the Mishongnovi dialect recorded by Whorf [Manaster-Ramer 1986]: (a) vowel + preaspirated consonant; (b) vowel + voiceless sonorant; (c) vowel + h.

Proto-Uto-Aztecan

The resulting Proto-UA consonantal system can be reconstructed as follows: Initial consonants:

*p *m *w

*t *n *s

*c *y

*k

*kw *yw

*q

*qw

*? *h

Medial consonants:

*p *b (*mb) *m (*w)

*t *d *n *l *s

*c *3 *y

*k *g (*q)

*kw *gw V (V)

*q *G (*ÇG)

*q (*Nw)

*?

Proto-consonants in brackets are not supported by sufficient number of examples.

The system of final consonants is probably the same as that of Proto-Numic.

The consonantal correspondences are supported by etymologies in utatet.dbf. This database consists primarily of comparisons from Miller's list of 2003. The list, published on the Internet, is Hill's revised and expanded version of Miller's computerized database for Uto-Aztecan cognate sets of 1988. The list was rekeyed and made available for the purposes of the EHL project by Nikolayev; I have converted the list to the current database format, double-checked the actual language forms and often expanded the etymologies, using recently published dictionaries and other data sources.

The set of Proto-UA consonantal correspondences is as follows:9

*Proto UA *Sonor *Aztec Tubatulaba *Takic Hopi *Numic Notes

*P- *P *p, #V: P *P P *P

*-b- *P *P P, b *v v, P *b

(*-p-) *P *P *PP no good exx.

(*-mb-) *hp mb 1 example

*m-, *-m- *m *m *m *m m *m

(*w-) *w # w *w # *w 1 example

(*-w-) *w *w w *w w *w, #

*t- *t *À, *t t *t t *t

*-d- *t, ht *t, *1 d, l *l t *d

*-t- *ht t l l t *tt

*n-, *-n- *n *n n *n n *n

*-l- *l *l n *l n *hn

*s-, *-s- *s *s, *s s *s s *s

*c- *c *c, *c *c c *c

*c, *3 *c y, dz *y *y *У, *3

*-c- *c, *hc *c c *c *y

9 The set of correspondences is significantly different from the one given in [Voeglin, et al. 1962, Miller 1967].

*Proto UA *Sonor *Aztec Tubatulaba *Takic Hopi *Numic Notes

*y-, *-y- *y *y y *y y *y

*k- *k *k k *k k, ky *k

*-g- *k *k g *k *k *g

*-k- *k *k *k *k k, ky *kk

*-qg- *hk *qg *k k *qg

(*q-) *n *n n *q q *n, *q(w) no good exx.

( *-q-) *n n *n, q n, q *q

*-qw- *m n m m m *qw

(*-Nw-) *hw qw, kw *w, *qw qw no good exx.

*kw- *kw *kw w *kw kw *kw

(*-gW-) *kw *w kw *gw- 1 example

(*-kw-) *kw kw kkw 1 example

CV- ) *w # w *w l *w 1 example

№ ) *w *kw w w l #

*q *k *k h *q *q *k

(*-G-) *hk *k h *q q *g 1 example

*-q- *k *k h *x q *g

(*-qG-) hk k q *x q *qg 1 example

*qW- *kw *kw w *qw *kw *kw

(* -qw-) (*P) *kw *qw 1 example

*? *? #V?, # ?-, # *? # #

?-, # # #

*h *h #V? h *h h *h

This reconstruction is, naturally, far from being complete: some proto-phonemes are not supported by a sufficient number of examples, while others demonstrate unexplained double reflexes in daughter-languages. Much more work is needed to obtain a detailed proto-UA reconstruction and a better etymological dictionary. Some etymologies in utatet.dbf may not belong to the Proto-UA level, but rather represent other chronological levels, areal roots, or even borrowings. These problems require further investigation.

It is suggested that the homeland of the Proto-UA family was located somewhere in Arizona and neighboring areas of New Mexico and Mexico (data in [Flower 1983]), based on an analysis of plant and animal names). Jane Hill [2001] made an attempt to demonstrate that the speakers of Proto-UA were agriculturalists, but her etymological conclusions are not well supported, since none of the proposed "agricultural" etymologies can be attributed to the level of the proto-language.

Mn NP Pn Sh Cm Ka Ch SP Ut Tb Gb Sr Ca Cu Lu Ho Pp Nv NT Gu Ta Op Eu Ma Yq Tu Hu Co CA Te Za Pi

Mono X 77 59 58 57 52 50 53 52 39 26 26 29 28 26 33 23 24 25 26 23 26 28 27 29 28 25 25 18 19 17 16

N.Paiute 77 X 58 58 58 56 55 58 52 42 26 24 27 27 27 32 26 26 28 29 27 29 27 28 30 27 24 22 18 18 17 15

Panamint 59 58 X 87 79 54 61 62 59 37 26 24 27 24 25 27 25 24 26 23 21 21 23 25 26 27 23 22 16 16 15 14

Shoshoni 58 58 87 X 88 55 58 62 61 38 26 24 27 24 24 23 25 24 26 23 21 20 22 24 26 28 23 23 16 16 15 14

Comanchi 57 58 79 88 X 49 54 59 59 35 23 21 24 23 22 22 23 23 23 24 21 20 26 27 24 27 21 21 14 14 13 12

Kawaiisu 52 56 54 55 49 X 75 79 76 39 24 26 27 26 24 31 26 24 27 24 22 20 24 25 29 28 23 22 16 17 16 16

Chemehuevi 50 55 61 58 54 75 X 86 78 42 27 28 31 30 27 33 28 26 28 24 22 26 26 27 26 27 23 21 15 15 16 15

S.Paiute 53 58 62 62 59 79 86 X 87 39 26 27 31 31 27 31 28 26 30 25 23 24 24 25 29 30 24 22 16 16 17 16

Ute 52 52 59 61 59 76 78 87 X 40 27 27 29 28 26 32 30 27 29 28 26 23 26 27 30 31 25 23 16 17 18 17

Tübatulabal 39 42 37 38 35 39 42 39 40 X 40 35 38 37 34 38 35 35 37 36 32 33 27 28 35 33 32 30 24 25 26 24

Gabrielino 26 26 26 26 23 24 27 26 27 40 X 45 42 34 38 29 35 24 26 29 28 26 35 35 28 24 24 19 20 20 21 21

Serrano 26 24 24 24 21 26 28 27 27 35 45 X 50 38 35 29 27 27 30 34 34 31 26 29 32 28 28 21 22 22 24 23

Cahuilla 29 27 27 27 24 27 31 31 29 38 42 50 X 65 50 31 31 30 32 34 33 33 30 33 35 29 34 24 23 24 20 20

Cupenô 28 27 24 24 23 26 30 31 28 37 34 38 65 X 48 31 28 27 29 29 26 29 29 26 26 26 26 23 19 20 19 18

Luisenô 26 27 25 24 22 24 27 27 26 34 38 35 50 48 X 26 25 24 26 28 28 24 25 28 28 23 27 22 19 19 24 24

Hopi 33 32 27 23 22 31 33 31 32 38 29 29 31 31 26 X 32 33 33 32 28 33 35 34 36 30 28 26 24 24 31 30

l 'a p a go 23 26 25 25 23 26 28 28 30 35 25 27 31 28 25 32 X 85 79 44 41 40 45 43 45 40 41 34 29 30 31 30

Nelvome 24 26 24 24 23 24 26 26 27 35 24 27 30 27 24 33 85 X 79 47 42 44 47 45 47 41 43 34 29 30 32 29

N.Tepehuan 25 28 26 26 23 27 28 30 29 35 26 30 32 29 26 33 79 79 X 47 42 40 45 49 49 46 42 35 30 30 29 33

Guarijio 26 29 23 23 24 24 24 25 28 36 29 34 34 29 28 32 44 47 47 X 83 55 59 58 58 48 51 42 32 32 33 34

Tarahumara 23 27 21 21 21 22 22 23 26 32 28 34 33 26 28 28 41 42 42 83 X 54 52 51 51 44 48 38 33 34 35 38

Opata 26 29 21 20 20 20 26 24 23 35 26 31 33 29 24 33 40 44 40 55 54 X 73 53 55 42 48 35 39 38 39 40

Eudeve 28 27 23 23 22 26 24 26 27 35 26 30 34 29 26 35 45 47 45 59 52 73 X 61 62 51 49 42 40 40 37 39

Mayo 27 28 25 26 24 27 25 27 28 35 29 33 36 26 28 34 43 45 49 58 51 53 61 X 93 51 48 45 38 38 38 39

Yaqui 29 30 26 26 24 29 26 29 30 35 28 32 35 26 28 36 45 47 49 58 51 55 62 93 X 53 51 46 39 39 35 37

Tubar 28 27 27 28 27 28 27 30 31 33 24 28 29 26 23 30 40 41 46 48 44 42 51 51 53 x 41 39 36 35 35 37

Huichol 25 24 23 23 21 23 23 24 25 32 24 28 34 26 27 28 41 43 42 51 48 48 49 48 51 41 x 58 39 37 33 35

Cora 25 22 22 23 21 22 21 22 23 30 19 21 24 23 22 26 34 34 35 42 36 35 42 45 46 39 58 X 37 35 80 85

CLAztec 18 18 16 16 14 16 15 16 16 24 20 22 23 19 19 24 29 29 30 32 33 39 40 38 39 36 39 37 X 85 81 77

Telelcingo 19 18 16 16 14 17 15 16 17 25 20 22 24 20 19 24 30 30 30 32 34 38 40 38 39 35 37 35 85 x 85 81

Zacapoaxtla 17 17 15 15 13 16 16 17 18 26 21 20 24 20 19 24 31 31 32 29 33 35 39 37 38 35 35 33 80 85 X 77

Pipil 16 15 14 14 12 16 15 16 17 24 21 19 23 20 18 24 30 30 29 33 34 38 40 39 39 37 37 35 79 81 77 x

Matrix from Miller 1984

Literature

Anderton, Alice J. (1988). The Language of the Kitanemuks of California. PhD. Dissertation. University of California, Los Angeles.

Bascom, Burton William. 1965. Proto-Tepiam (Tepehuan-Piman). Ph.D. dissertation, University of Washington. Boas, F. 1917. El dialecto mexicano de Pochutle, Oaxaca. IJAL 1:9-44.

Bright, William, and Jane H. Hill. 1967. The linguistic history of the Cupeno. Studies in Southwestern Ethnolinguis-

tics, ed. By Dell Hymes, pp.351-371 Bright, William. 1965. Luiseño Phonemics. IJAL 31:342-345. Bright, William. 1965. Luiseno Phonemics. IJAL 31:342-345.

Campbell, Lyle, and Ronald W. Langacker. 1978. Proto-Aztecan vowels. IJAL 44:85-102, 197-210, 262-279.

Campbell, Lyle. 1997. American Indian Languages. Oxford — New York: Oxford University Press.

Canger, Una. 1988. Nahuatl dialectology: A survey and some suggestions. IJAL 54:28-72.

Collard, H. y E. Collard. 1962. Castellano-mayo, mayo-castellano. Istituto Linguistico de Verano.

Dakin, Karen. 1979. Phonological Changes in Nahuatl: the Tense / Aspect / Number Systems. IJAL 45:48-71.

Dakin, Karen. 1982. La Evolution Fonológica del Protonáhuatl. Colección Lingüística indígena. Mexico: Universidad

Nacional Autónoma de Mexico. Dakin, Karen. 1983. Proto-Aztecan vowels and Pochutec: An alternative analysis. IJAL 49:196-203. Dayley, Jon P. 1989. Tumpisa (Pamamint) Shoshone Dictionary. UCPL 115. Berkeley and Los Angeles: University of

California Press. Electronic edition 1998. Elliott, Eric B. 1999. Dictionary of Rincon Luiseno. Ph. D. dissertation.

Estrada Fernández, Zarina et al. 2004. Diccionario yaqui-español y textos: obra de preservación lingüística. Hermosillo

and México, D.F.: Universidad de Sonora and Plaza y Valdés Editores. Felipe S. Molina, Herminia Valenzuela, David Leedom Shaul. 1999. Yoeme-English English-Yoeme : with a comprehensive grammar of Yoeme language. New York: Hippocrene Books. Fowler, Catherine S. 1983. Some lexical clues to Uto-Aztecan prehistory. IJAL 49:224-57. Givon, T., ed. 1979. Ute Dictionary. Preliminary ed. Ignacio, Col.: Ute Press. Grimes, Jose. 1981. El Huichol. MS

Hale, Kenneth. 1958-59. Internal diversity in Uto-Aztecan: I. IJAL 24:101-7, II. IJAL 25:114-21. Hill, Jane H. 2001. Proto-Uto-Aztecan: a community of cultivators in Central Mexico? American Anthropologist 103:913-934.

Hill, Jane H, and Kenneth C. Hill. 1968. Stress in Cupan (Uto-Aztecan languages). IJAL 34:233-241). Hill, Jane H., and Rocinda Nolasquez. 1973. Mulu'wetam: The Fist People. Banning, Calif.: Malki Museum Press. Hilton, Simón K. Diccionario Tarahumara. Istituto Linguistico de Verano. Iannucci, David. 1972. Numic historical phonology. Ph.D. dissertation, Cornell University. Hopi-English Dictionary of the Third Mesa Dialect. 1998. Tucson: The University of Arisona Press. Kaufman, Terrence. 2001. The History of the Nawa language group from the earliest times to the sixteenth century: some initial results. MS

Kroeber, A.L. 1907. Shoshonean dialects of California. UCPAAE 4:65-165.

Lamb, Sydney M. 1964. The classification of the Uto-Aztecan languages: a historical survey. UCPL 34:106-25. Lionnet, André. 1985 Relaciones internas de la rama sonorense. Amerindia, 10.

Manaster-Ramer, Alexis. 1993. On Lenition in some Northern Uto-Aztecan Languages. IJAL 59:334-341. Manaster-Ramer, Alexis. 1986. Genesis of Hopi Tones. IJAL 52:154-160.

Mathiot, Madeleine. 1973. A Dictionary of Papago Usage. CV.1. Bloomington: Indiana University. McIntosh, Juan B. y Jose Grimes. 1954. Vocabulario Huichol-castellano. Castellano-huichol. Istituto Linguistico de Verano.

McLaughlin, John E. 1992. Counterintuitive Solution in Central Numic Phonology. IJAM 58:158-181. McMahon, Ambrosio y Miaria Aiton de McMahon. 1959. Cora y Espanol. Istituto Linguistico de Verano. Miller, Wick R. 1967. Uto-Aztecan Cognate Sets. UCPL 48. Berkeley and Los Angeles: University of California. Miller, Wick R. 1982. Southern Paiute and Numic Final Feature. IJAL 48:444-449.

Miller, Wick R. 1984. The classification of the Uto-Aztecan languages based on lexical evidence. IJAL 50:1-24. Miller, Wick R. 1996. Guarijío: gramática, textos y vocabulario. Mexico: Universitad nacional autónoma de México. Miller, Wick R. 2003. Uto-Aztecan Cognate Set. Revised and expanded by Kenneth C. Hill (ms.).

Miller, Wick R., Dirk Elzinga, and John E. McLaughlin. 2005. Preaspiration and Gemination in Central Numic. IJAL 71:413-44.

Miller, Wick R., James L. Tanner, and Lawrence P. Foley. 1971. A Lexicostatistic Study of Shoshoni Dialects. Anthropological Linguistics 13/4:142-164.

Munro, Pamela. 1983. Selected studies in Uto-Aztecan phonology. IJAL 49:277-99.

Munro, Pamela. 1990. Stress and Vowel Length in Cupan Absolute Nouns. IJAL 56:217-250.

Pennington, Campbell W., ed. 1979 Vocabulario en La Lengua Nevome, The Pima Bajo of Central Sonora, Mexico. Vol. 2. Salt Lake Sity: University of Utah Press.

Press, Margaret. 1979. Chemehuevi: Grammar and Lexicon. UCPL. 92. Berkeley and Los Angeles: University of California Press.

Robinson, Lila W., and James Armagost. 1990. Comanche Dictionary and Grammar.

Sapir, Edward. 1931. Southern Paiute Dictionary. Publications of the American Academy of Arts and Sciences 65, no. 3.

Sapir, Edward. 1931. The Southern Paiute language, Parts I, II, and III. AAASP 65.1-730.

Saxton, Dean, and Lucille Saxton. 1983. Dictionary: Papago and Pima to English, English to Papago and Pima. Tucson: University of Arizona Press.

Seiler, Hansjakob, and Kojiro Hioki. 1979. Cahuilla Dictionary. Banning, Calif.: Malki Museum Press.

SOD: Shoshoni Online Dctionary, at: www.shoshonidictionary.com.

Starostin, George. 2010. Preliminary lexicostatistics as a basis for language classification: a new approach. Journal of Language Relationship 3:79-116.

Voegelin, C. F. 1958. Working dictionary of Tubatulabal. IJAL 24:221-28.

Voegelin, C. F., F.M. Voeglin, and Kenneth Hale. 1962. Typological and Comparative Grammar of Uto-Aztecan: I (Phonology). IJAL Memoir 17.

Whorf, Benjamin. 1946. The Hopi language, Toreva dialect. Linguistic Structures of Native America, ed. by H. Hoijer, 158-183. New York : The Viking Fund, Inc.

Zigmond, Maurice L., Curtis G. Booth, and Pamela Munro. 1990. Kawaiisu. A Grammar and Dictionary with Texts. UCPL 119. . Berkeley and Los Angeles: University of California Press.

Статья представляет собой краткий обзор наработок по исторической фонологии юто-ацтекской языковой семьи. Основными предметами обсуждения являются генетическая классификация языков этой семьи, реконструкции фонологических систем праязыков ее отдельных ветвей и собственно праюто-ацтекская реконструкция. Лексический материал, на основании которого разработаны соответствующие реконструкции, взят из электронных баз данных, созданных в рамках проекта «Эволюция языка» (Институт Санта-Фе).