Text (preliminary draft only -- absolutely no quotation -- and footnotes
not yet supplied)
Copyright © 2003 David Shaul. All rights reserved
Used as ethnic markers, languages in the same geographic region may
to show little vocabulary (lexical) borrowing. They may show, however,
similarities in language structure (sound system, grammar) due to borrowing.
This paper presents nonlexical data to support the idea of contact between
some Yuman languages and Pima, and between Pima and Zuni in the Hohokam
regional system.
Beyond language data, there is convincing archaeological, ethnological,
ethnohistoric, and physical anthropological data in support of the multi-ethnic
model of Hohokam culture proposed here.
Language Data
Of the languages descended from Proto-Yuman, there is a phonemic contrast
between dental /t/ and retroflexed /_/ in Cocopa and Diegueno (Langdon
1970:540; Langdon and Munro 1980:Table 3). A /t/ versus /_/ contrast
is an areal feature of California languages (Langdon and Silver 1977),
but this marked contrast is phonemic in Yuman only in the Delta-Colorado
subfamily.
Of the Tepiman subfamily of Uto-Aztecan, only the northernmost (Tohono
O’odham, Akimel O’odham; both are called "Piman"
in this paper) located in southern Arizona have a retroflexed stop /d./
which is an anomalous development from Proto-Tepiman */r/ (Bascom 1965).
Geographic proximity suggests that Piman /d./ developed during a period
of intimate contact with Delta-California Yuman.
The most likely candidate for Yuman-Piman contact is Maricopa, in residence
in the Phoenix area with the Akimel O’odham in the late 1700s
and during the 1800s. However, Maricopa, a member of the River subfamily
of Yuman, has no /t/ versus /_/ contrast (Gordon 1981:6; Spier 1946:8-13;
Sunn and Harwell 1973:70-71). Shaul and Andresen (1989) suggested the
Hohokam archaeological culture as a likely location for the development
of Piman /d./ modeled on Delta-California /_/. It is notable that both
/d./ and /_/ have the same distribution (post-stress in roots).
Shaul and Hill (1998:380-381) document another, older, phonological
diffusion between the River subfamily of Yuman (Yuma, Maricopa, Mojave).
In these languages, Proto-Yuman */y/ developed as / _/, with an intermediate
stage of /d/ (Langdon (1975, 1978). (The /_/is the voiced th in the
English word then.)
A sound shift of /y/ to /d/ occurs in all Tepiman languages as part
of a larger sound shift.
Proto-Uto-Aztecan kw w y
Proto-Tepiman b g d
The general pattern is that the glide consonants changed into voiced
stops. This sound change is very rare in the world’s languages
(Pulleyblank 1989:388). The development of /y/ to /d/ in all Tepiman
languages and a parallel, isolated change in River Yuman suggests the
following chronology.
II /d./ develops in Piman after Delta-California Yuman /_/
I / _/ develops in River Yuman from */y/ after the
Tepiman shift of */y/ to /d./
This argues for prolonged contact by northern-most Tepiman (Piman) with
various Yuman languages over a long period of time, with southern Arizona,
the Hohokam heartland, being the most likely location.
Delta-California Yuman and Tepiman also share grammatical traits. Shaul
and Hill (1998), following Shaul (2000), show that Cocopa and Diegeuno
share a means of joining clauses with Tepiman. The affixes are as follows.
Tepiman Delta-California Yuman
-tV ~ -c -c
ma- -m
-k ~ -ku -k
In both systems, the shapes and position (suffixes) are strikingly similar.
In both Yuman and Tepiman, the word order is: Subject + Object + Verb
(SOV). If two clauses have the same subject, the second mention of the
subject is left out: for example, Suzi went shopping and bought a coat,
instead of Suzi went shopping and Suzi bought a coat. Imagine that Yuman
and Piman bilinguals notice the following clause-joined pattern where
the second subject is left out: SOV_OV (the blank is where the "deleted"
subject would have appeared). In both the Tepiman systems, -tv ~ -c
and –c are used to link sentences in which the subject (identical)
subject has been left out; such markers are know as "same-subject"
markers.
At the same time, bilinguals would have noted two other similar pairs
of affixes (Tepiman ma- and Yuman –m; Tepiman –k ~ ku- and
Yuman –k). Tepiman ma- is only used as a prefix, whereas Tepiman
–k ~ ku- can be used as a prefix or a suffix. In looking for a
marker for "different-subject"s, bilinguals would have preferred
–k to fit into the SOV_OV pattern to mark different subjects.
In Tepiman, -k marks different subjects (example: He went to the store,
and he [a different person] bought a new car). The different-subject
marker in Delta-California and River Yuman is –m. Thus, speakers
of Tepiman must have resolved the ambiguity by chosing –k over
ma-, since the latter only appears as a prefix. Hale (1983) argued that
the clause-initial use of the clause joining markers in Tepiman was
an innovation in that subfamily, as the cognates in other Uto-Aztecan
languages (Takic subfamily, Hopi, Numic subfamily) are all suffixes.
This innovation had to have happened before Tepiman contact with Yuman.
No other subfamily of Uto-Aztecan uses this system to mark same and
different subjects. This suggests that Yuman was a model for the Tepiman
system.
Northern-most Tepiman (Piman) shares another unusual grammatical feature
with some Yuman languages. A shortened form of subject pronouns ("clitics")
tend to appear as the second word in sentences through-out Uto-Aztecan
languages (Langacker 1977:37). In Tepiman languages, there are second-position
clitics for "I", "we" and "you." Only
in Piman are there second position clitics to mark "s/he",
"it", and "they." (Shaul 1986:chapter 4).
In northern Tepiman (Piman), se4cond position clitics are the basic,
obligatory way of keeping track of subjects. This is not true of any
other Tepiman languages (Shaul 1986:chapter 4). Steele (1979) reconstructs
a full set of second position subject clitics as a fundamental characteristic
of Proto-Uto-Aztecan. Shaul (1986:chapter 4) points out that a full
clitic set appears in only two subfamilies (Takic, Tepiman); however,
only northern-most Tepiman has this feature. Another subfamily (of one
language, Tubatalabal) has the system only in the singular ("I",
"you" and "s/he" ~ "it"). Only the Takic
subfamily has a complete set of subject clitics in all languages. It
is no accident that Takic, Piman, and Tubatalabal have subject marking
by clitics, and are next to the Yuman languages, which mark subject
clitic by subject prefixes that fit into each S of a SOVSV pattern when
only one verb is used. (Usually, the second V is either "be",
"do", or "say"; all verbs in Yuman fit into this
basic three-part classification.).
Moreover, the second SV is monosyllabic, and will always be the second
word in sentences, since subjects and objects expressed as nouns in
Yuman language use are usually dropped once they have been established
in discourse. This means there is a high frequency of the SV (subject
clitic plus classifying verb) as the second element in sentences in
actual usage. This is exactly the pattern of Takic, northern-most Tepiman
(Piman), and Tubatalabl; all three of these surround Yuman languages
(which as historically contiguous). Only these three groups in Uto-Aztecan
have this feature as an obligatory part of their grammars, making Steele’s
(1979) reconstruction of such a system for Proto-Uto-Aztecan impossible
(and very unlikely for Proto-Tepiman).
Ethnological, archaeological, ethnohistorical information, and data
from physical anthropology strongly supports the model of Piman and
Yuman prehistory presented here. This additional information also suggests
the Hohokam regional system as a possible location of language contact.
More resolution can be provided by the following relative linguistic
chronology.
II (4) /d/ vs. /d./ develops in Piman and Delta-California Yuman
(3) second position subject clitic system develops in Piman
I (2) /y/ develops as /d/ in Tepiman and River Yuman
(1) same s. differ rent subject marking develops
In stage I, /d/ became /_/ in River Yuman languages, which bordered
Tepiman at this time, the model for the change being available in any
Tepiman variety. This change must date to Proto-Tepiman times because
the change appears in all Tepiman varieties. The development of same
vs. different subject marking in all Tepiman must date to this time,
since the feature is also present in all Tepiman varieties.
In stage II, Piman developed subject marking by using second position
clitics, following a Yuman model. Piman also developed the distinctive
/d./, after a Yuman (Delta-California subfamily) model.
Thus, there are two distinct stages of Yuman-Piman contact. In stage
I, trait (1) originates in Yuman, and trait 2 originate in Tepiman.
This suggests egalitarian relations, both with language as an ethnic
marker (paucity of loanwords). In stage II, the changes are all from
Yuman into Piman, suggesting that Yuman was a prestige linguistic code.
Language continued as an ethnic marker, although the Delta-California
languages would be a promising place to look for loan words, since contact
was more recent, and because a Delta-California variety or varieties
may have served as a prestige code.
Gumperz and Wilson (1971) show that where language is an ethnic marker,
vocabulary tends to not be borrowed, although grammatical systems may
become nearly identical. Gumperz and Wilson worked in a village in India
with High and Low castes of Marathi (Indo-European family, Indic subfamily)
and Kannada (Dravidian language family). They found that the order of
prefixes and suffixes on words, as well as word order patterns had become
nearly identical in all four varieties (High and Low Marathi, High and
Low Kannada), with vocabulary serving the maintain the ethnic boundaries.
This suggests that in the Hohokam context, languages were used as ethnic
markers.
There is a third language that is part of linguistic interaction in
the context of Hohokam, Zuni. Evidence of contact is limited to loanwords
of ceremonial reference (next section). Borrowing appears to go both
ways, suggesting egalitarian relations, but the vocabulary is limited
(referring to ceremonial).
Zuni Linguistic Data
There is a small, but tantalizing, group of loanwords shared between
Piman and Zuni. (So far, no attempt has been made to do comparative
Tepiman work on the words below.)
Piman Zuni
siwañi, "rain shaman" (Underhill 1939) _iwani, "rain
priest" (Newman 1958:39); "rain cloud" (Bunzel 1935:373);
probably a loan into Zuni from Keresan _i:wana, "rain deity, kachina"
(Lange 1959; White 1942)
kihe/kiha, "brother or sister in-law" kihe, "ceremonial
brother"
kok’oi, "ghosts, spirits of the dead" kok:o, "kachina"
ho-wi, "banana yucca" (Yucca baccata) ho:-ts’ana, Y.
baccata
ho:-k’apa, Y. whipplei ("Spanish bayonet")
sema-, "bold/mean" _uma-‘kwe, "Shumakwe Curing
Society"
There is a loan from a Puebloan language into Piman ("rain priest/shaman,
rain cloud"). Other shared forms appear to have been borrowed into
Zuni from Piman. Only two Zuni words in Newman’s dictionary (1958)
begin with ki-, suggesting that the sequence is rare in Zuni (and thus
found only in loan words), kihe being one of these two. (The other word
is for "kiva" and probably a loan into Zuni from Hopi). The
Zuni word kok:o "kachina" has a doubled syllable (reduplication).
This word is clearly a borrowing from the Piman form (the reduplicated
form of koi "dead", a root which has a well-known Uto-Aztecan
ancestry). The suffix –wi (in the Piman word for "banana
yucca") is common in Piman plant names. The root is found in the
Zuni words for Yucca species, suggesting that the plant had some special
ceremonial meaning; better information is needed for comparative Tepiman
ethnobotany in order to determine direction of borrowing in this case.
The suffix –‘kwe is one of two ways of making nouns plural
in Zuni; again, better comparative Tepiman would resolve direction of
borrowing.
Stevenson (1904:545) states that the songs of the Shuma’kwe Society
are in Piman. Moreover, the curing action of the society (for convulsions)
involves untwisting a plaque made of yucca fiber (Parsons 1939:364).
The meanings of the other loanwords (whatever the direction of borrowing)
in ceremonial in nature. Ceremonial vocabulary is often shared, even
where language is otherwise an ethnic marker. For example, the Pueblos
are a classic case of loanword avoidance (Dozier), yet there is an entire
ceremonial vocabulary among the Pueblos (White 1942).
Shaul and Hill (1998:381-382) propose one loanword from Yuman into Tepiman
(and thus at stage I). A paucity of shared vocabulary between Yuman
and Tepiman/Piman could indicate use of language as an ethnic marker
and/or reflect the limited dictionaries of the languages involved. However,
a general lack of loanwords would mean that the Hohokam regional system,
made up of several ethnic groups, used language as an ethnic marker.
It would seems that this dates from stage I times (of the scenario developed
in the previous section).
Archaeological Data
Ceramic distribution of Hohokam pottery and Patayan II (Yuman) pottery
geographical overlap is focused in an area historically occupied by
Yuman cultures, between Gila Bend and Organ Pipe National Monument (Ezell
1954:figure 2). Gila Bend was a location where Hohokam and Patayan II
cultures coexisted during the Hohokam Sedentary phase (ca. A.D. 900-1100;
Teague 1981:15; Wilcox and Sternberg 1983:203-208, 238-239). Along the
Gila River, Patayan II pottery is a common intrusive ware in Hohokam
sites (Vivian 1965:139; Greenleaf 1975; Johnson 1965:53-69). In the
core Hohokam area around Phoenix, a large quantity of Patayan pottery
was found at the transitional site of Las Colinas (ca. A.D. 1050-1200).
The ceramic distribution data suggests that around A.D. 1000, Yuman
and Piman speaking people interacted within the Hohokam regional system.
Sea shell trade and production in Hohokam culture is another indicator
of Yuman-Piman interaction (summary in Shaul and Andresen 1989:118-119).
Shell trade networks developed in areas where Hohokam and Patayan cultures
overlapped.
[obsidian data]
Cremation with an Ethnological Correlation
Fewkes (1912:102) was the first to suggest that cremation, as a distinctive
feature of Hohokam culture, was linked to southern California where
cremation was prevalent prehistorically and historically. Sayles (1937:100)
argued that cremation was a "common bond" between the Hohokam
and historically known groups in southern Arizona and southern California.
Shaul and Anredesen (1989:116-117) point out that of all the Yuman cultures,
which all practice cremation, that only the Dieguenos (Heye 1919; Davis
1921) had a Hohokam-like (Johsnson 1964; Haury 1976:164) cremation complex:
reservation of ashes and bones from primary cremation in an urn that
was buried in a communal area. The correspondence of cremation details
between earlier Hohokam culture and historic Diegueno culture, along
with the linguistic data, suggests that the ancestors of Delta-California
Yumans range in prehistoric times much further east than their historic
locations. This is confirmed by ethnographic and ethnohistoric data.
Ethnographic and Ethnohistoric Data
Ethnographic data also suggested prolonged and intimate contact between
speakers of Yuman languages and Piman.
The Pimans have a ceremony called vi:kita which unites hunting and gathering
symbolism (deer hunt) with agriculture in a recreation of the world.
The Delta (Cocopa) and River Yumans also have a pre-Christian ceremony
the keruk, which is structurally parallel to the vi:kita. Bahr (1982)
has argued that this similarity indicates that Pimans and some Yumans
participated in a single religious and ideological system. Ellis (1968:69)
and Hedges (1973:32) suggest that some ethnographically documented Yuman
figurines are developments from Hohokam culture. Underhill (1954:654-655)
gives evidence for grouping Piman and Yuman cultures together as a culture
area; many of the factors given are not subsistence related. Nettl (1954)
shows that Piman musical style is in an intermediate position between
the Pueblo and Yuman-California styles, with the latter predominating.
Shimkin (1941:233) stated, in his study of Uto-Aztecan kinship terminology,
that "except for some Spanish changes among the Tarahumara, the
preponderant outside influence, especially among the Pima-Papago, was
Yuman."
Piman mythology (Saxton and Saxton, eds. 1973); Bahr et al. 1994; Bahr,
ed. 19xx) posits two native populations in southern Arizona, one conquering
and succeeding the other. The initial population, the Hohokam, was conquered
by the ancestors of the Pimans. In some versions, the surviving vanquished
become one of the two Piman moieties; in other versions, they became
the Yuman-speaking tribes along the Colorado River (the River subfamily
of Yuman). Thus, Piman mythology argues for intimate interaction of
Yuman speakers and Pimans in the past; the present paper argues that
the Yuman language(s) involved in Hohokam culture was Delta-California
Yuman, not River Yuman.
Physical Anthropological Data
The physical anthropological evidence for a multi-ethnic Hohokam cultures
comes from mitochondral DNA (mtDNA), blood groupings, and mutation on
the albumin gene. Malhi (et al 2003) found that the "distribution
of haplogroups [of mtDNA] in the [American] Southwest is structured
more by archaeological tradition than by language [group]" (Malhi
et al 2003:108). This data envisions the Hohokam and Anasazi as "geographically
widespread archaeological cultures produced by multiethnic groups"
(2003:108).
Following Lorenz and Smith (1996), Malhi et al (2003) used a larger
and more representative sample in terms of both groups sampled and mtDNA
sequence data. They found nearly identical distributions of mtDNA haplogroups
in speakers of all Yuman languages sampled, and in speaker of Piman
varieties. Moreover, a shared haplotype of the haplogroup B is shared
by Yuman (Pai) and Piman (Tohono O’odham) samples (2003:12). This
is consistent with blood group data (Brown et al 1958). It is also worth
noting that mtDNA haplogrouping of Pimans differs greatly from that
of other speakers of Uto-Aztecan languages (Hopi, the Numic subfamily,
Nahua/Aztec).
A mutation in the albumin gene (AL*Mexico) is also relevant to the model
of Hohokam as a multi-ethnic culture. AL*Mexico "is limited to
the southwestern United States, where it is reported in some (Pima,
Papago, Ute) but not all (e.g. Hopi) Uto-Aztecan speaking groups"
(Smith et al 2000:559). This distribution of AL*Mexico matches the mtDNA
grouping of Yumans, Pimans, and Zunis (high frequency of haplogroup
B, near absence of A). Distribution of mtDNA and AL*Mexico within Uto-Aztecan
suggests that Uto-Aztecan speakers are not closely genetically related.
Data from physical anthropology shows that Pimans (but not other Uto-Aztecans),
Yumans, and Zunis had prolonged prehistoric contact. It does not say
when, but the other data considered here suggests that the arena of
this contact was, at least in part, the Hohokam archaeological culture.
Conclusion
Pervasive ethnographic data links Yuman and Piman cultures, a view supported
by Piman ethnohistory. Two chronological stages of such interaction
are apparent from lingusitic data, with the second involving Pimans
and Delta-California Yumans. Cremation data corroborates Delta-California
contact with Piman culture, and archaeology suggests that they interacted
around A.D. 1000 in the context of Hohokam culture. Physical anthropological
data suggests, along with linguistic data, that speakers of Zuni may
have been part of the multi-ethnic Hohokam.
|