Ioway Cultural Institute : The Ioway Virtual Library

Convergent and Divergent Accommodation in Language Death

[A scholarly paper on the disappearance of Chiwere, the language of the Iowa-Otoe-Missouria]

An Excerpt: "...the development of "family" dialects, which are disparaged by those outside the family. The breakdown of speakers' accommodation to one another in a dying language can be seen to lead to dialect isolation and speaker intolerance of variability. Further, there is a politeness requirement of the society that mandates younger conversational partners to invite the discourse of older persons and to reply in brief, respectful responses to their elders, but not to engage them in vigorous turn-taking. That convergent accommodation has helped a generation acquire only limited active competency in Chiwere since its members learned the language from elders in a home setting and did not use it outside the home with peers. The practice sped the decline of the indigenous language."

Type of Accommodation:

Divergent
Convergent
Upward
Speaker: Acrolect
Speaker: Acrolect
Hearer: Basilect
Hearer: Acrolect
Downward
Speaker: Basilect
Speaker: Basilect
Hearer: Acrolect
Hearer: Basilect


Table 1. Convergence and Divergence in Accommodation
Theory
(Adapted from Giles 1973:92)

Actually, of course, one need not adopt an entire dialect to engage in this activity; mere adoption of some features of a dialect is sufficient, as the pop group example shows; or as Bell (1984) has demonstrated, the shift in a style of speaking is adequate, for example a shift from formal to informal style. In fact, a famous instance of downward divergence was signaled not with speech, but with dress when, at close of World War II, General McArthur took Japan's surrender wearing casual military attire rather than formal uniform to signal his contempt for the enemy.

Trudgill reported on the strong suggestion in the communications literature1 that positive behavioral accommodation is a universal trend (1986:2) and speculated that accommodation might also be a universal process by which linguistic change is diffused. For us, it is an easy step from that position to one that suggests that accommodation is also a source of change. In this paper, we examine the history and present-day setting of Chiwere Siouan, which are typical of many other American Indian languages. We propose that lack of accommodation -- or divergent accommodation, along with convergent accommodation, is one engine driving language death in Chiwere communities, and perhaps in similar Plains Indian settings.

THE PROCESS OF CHIWERE OBSOLESCENCE

In obsolescent American Indian languages, there can appear family dialects; the speech of non-family members is often disparaged. Such a situation prevails among Chiwere speakers. As Chiwere has declined, it has been used less as an every-day and more as a sacred medium; it lives today almost exclusively in its high register, what has ben termed its Latinate form (Hill and Hill 1986). Thus the breakdown of speakers' accommodation to one another in a dying language can be seen to lead to dialect isolation and speaker intolerance of variability.

In the face of the expressed desire to maintain the language by members of the two Chiwere-speaking tribes -- the Otoe-Missouria and Ioway -- the language continues to decline. Chiwere counts fewer than 10 fluent speakers, and up to about 30 other persons who have some speaking knowledge of the language. Some persons have a passive understanding of Chiwere, although they themselves either do not speak at all or only name vocabulary items. All fluent speakers are elderly, and all show some degree of language deficit. As documented elsewhere (Stanley and Furbee 1901), Chiwere lives primarily in its high register, what has been termed the Latinate form (Hill 1973; Hill and Hill 1986), or at least in a context appropriate to the high register. The language of every-day communication is English. The genres in which Chiwere is used include songs (including the creation of new songs), portions of naming ceremonies, prayers, stock phrases used in public oratory that otherwise is conducted in English, and so on. We know of one pair of friends who routinely converse informally in Chiwere; this pair, ages 94 and 86, are hinharo, or formal friends pledged by their parents to this life-long relationship in their boyhoods.

The language is now a precious medium. Possession of only a few phrases or a memorized story elevates the speaker in the community view. The language has become "objectualized," to use Silverstein's (1984) term, as it has reached heirloom status. Indeed, it appears that the Chiwere communities accept a strong version of the Whorf hypothesis, because its few fluent speakers are themselves regarded as heirlooms, as cherished reservoirs of knowledge of ancient ways just because they are fluent in Chiwere, and therefore are thought to have access to important ways of thought associated with life before contact with Europeans. In fact, of course, the fluent speakers' knowledge of the past life on the Plains is quite fragmentary. With two possible exceptions, none of them has had much exposure to the high, sacred language -- the primary venue through which the language lives today. As often as not, that Latinate form emerges in a memorized text from the lips of persons who are essentially non-speakers of Chiwere. For example, as a memorized old song sung by a singer who is not a fluent speaker of the language, or perhaps not even a "speaker" or "semi-speaker" in the linguistic sense at all. Thus, singers in the Tribe may memorize songs in Chiwere, just as they would songs in Ponca, Caddo, and other languages. So we have the curious situation of a living -- barely living -- high form of a language, but a living fluent population of speakers who know primarily the every-day, secular form of it. Further, new songs are constructed in consultation with these few fluent speakers, none of whom is fluent in the high register, which means that new sacred texts are now constructed in every-day Chiwere, and committed to memory by non-speakers of the language for use in sacred settings such as the powwow. The remainder of this paper treats how such a circumstance could have come about, and what lessons might be learned from it for the study of language death.

A SCENARIO FOR CHIWERE LANGUAGE OBSOLESCENCE

Following is a possible scenario for the process of language death in Chiwere and similar languages, following accommodation theory. These stages may be compared with those presented for Mexicano (Hill and Hill 1978).

Stage 1. Intra-tribal factionalism increases as cultural controls and practices are disturbed by increased white contact and government activity (+/-1800 - 1880).
This first stage that we can identify is that of growing factionalism. During this period, the Otoe-Missouria and Ioway tribes were being settled on reservations in Nebraska and Kansas, then later moved to lands in Oklahoma. As evidence of the seriousness of dissent that arose during this period, both tribes had splinter groups that refused for a time to move to the designated Oklahoma lands. On the other hand, the Ioway had some members who had left the Kansas and Nebraska lands early to join the Sac and Fox in Oklahoma. Certainly, however, the earlier pressure of the westward movement of white settlers also encouraged factionalization as traditional cultural controls and practices were disturbed by contact and governmental
activity.

Stage 2. Family dialects arise (1880-1920).

Through resettlement, education, andmissionizing, this first generation of residents to grow up in the Oklahoma lands were strongly discouraged in their attempts to maintain a traditional way of life. These people were educated in schools that stigmatized use of Indian languages. Following the now familiar pattern, many were sent to boarding schools where they were kept out of contact with their home culture, thus creating in the generation a subgroup poised for leadership, a group that was fluent in English, disinclined toward maintenance of the traditional culture, and eager to broker. The best speakers of Chiwere today are the survivors of this generation.

For this generation, success in the community no longer was achieved along traditional lines, a situation that sometimes pitted traditionally prominent families against newly important families. This was a period of "paper" chiefs, chiefs chosen by the government because they were cooperative rather than by traditional means, further rupturing intra-tribal alliances.

As speakers began to devalue the speech of those outside their own families, they would feel that they would have to "speak down" to non-family members if they were to accommodate in Chiwere, since for them their own dialect would be the acrolect and all others would be some kind of basilect. With the increase in factionalism, there was less and less reason to accommodate. At this stage of the process, many of the adults would have been monolingual. The political and social disagreements in the tribe promoted lack of accommodation, and since to any individual his or her dialect was the correct language, divergent accommodation became an excellent way to put-down non-family in political settings. To do so, one would need only use an elevated form of their language, in effect, "talk up" (Giles' upward divergence) to the non-family member, speaking in a manner that perhaps diverged more than usual in order to exaggerate differences. The scene is then set for two developments: First, the younger bilingual generation began to broker for the tribe as their elders ceased cooperating with one another, linguistically and otherwise. Second, everyday use of Chiwere began to be limited to the home, making it possible for "my Grandmother's language" to become the "best language," thus promoting the birth of so-called "family dialects" (Allan Taylor, personal communication, 1987).

Stage 3. The rise of English as the language of every-day accommodation.

(1920-1950).
It would have fallen to the next generation of tribe members to try to pull things together politically. This generation was fluent in English and marginally bilingual; much of it was educated, often in boarding schools that taught the white point of view, so for them there may have seemed more hope of rapprochement between the cultures than for their parents. Like all third generation immigrants, they had the luxury of cultivating an interest in their roots without endangering their place in the new society. They had some impetus for trying both to maintain their tribes and to find a place for them in a white-dominated America.

At this point, however, English moved toward an every-day language of accommodation outside the home, the neutral language. Aspects of this situation resemble Third World settings where a colonial language is used in public discourse just exactly because it neutralizes all the signals of region, class, and caste that the native languages may carry (e.g., see Ferguson and Gumperz 1960). So, at this stage in the death of Chiwere, if one is being polite to a member of the community who comes from outside one's family, one uses English, the neutral language, in order to accommodate convergently to that speaker. If one is not being polite, one uses the most divergent form of his family's dialect of Chiwere, and does not accommodate to the speaker.

The home would have remained Chiwere-speaking, but children did not become active Chiwere speakers at least in part because a politeness requirement of Chiwere society impeded their learning. That requirement dictates that the younger of a conversational partner invite the discourse of the elder partner but ought not engage in vigorous conversational turn-taking with someone older. Such a custom represents convergent accommodation. Children following this courtesy would gain little active speaking experience at home among elders, the very persons from whom they should learn the language and other traditional knowledge. Ordinarily, active speaking skills would be gained conversing with age mates outside the home, but since English by this time was the every-day language of accommodation outside the family, these youngsters grew up with good passive knowledge of Chiwere but little or no active skill in the language.

Stage 4. Bilingualism yields to spoken every-day English and specialist-controlled ritual Chiwere (1950-present).

The result was that English became the language of every-day discourse, and Chiwere remained the language of ritual. Since many of the public rituals had been destroyed in the suppression of late-19th century reservation life, the rituals that remained tended to be family ones: naming ceremonies, family songs, rituals that are "owned" by a family member and passed down to another family member. Dance groups, medicine societies, and other centers of public ritual activity fell into inactivity. Furthermore, as the number of competent speakers declined, there were fewer persons who could be said to control both the every-day and the Latinate registers of Chiwere. Now, in these final days of life for Chiwere, there is really no fluent speaker with control of both registers. Instead, there is some memorized Chiwere that is Latinate in use in high context (e.g., old songs); there is every-day Chiwere that is used very occasionally in every-day contexts (e.g., conversation between relatives or close friends), and most important, there is newly cast "formal" Chiwere for use in high contexts that has been composed in every-day Chiwere (e.g., newly constructed songs).

Let us look at Chiwere data from the standpoint of accommodation theory, and consider predictions for language change in Chiwere from the theory.

INTERDIALECTS AND CONSONANT COLLAPSE

Accommodation theory predicts that the process of accommodation will yield an interlanguage or interdialect that contains intermediate forms, forms that did not necessarily exist in the original languages or dialects. A large number of variations are produced which primarily involve salient features, features that speakers of both dialects or languages know to be characteristic of the other. For example, the flap t in American English is adopted by accommodating British English speakers because it is known to be an American feature; the same is true for numbers of lexical items, and some syntactic constructions (although in this instance for few morphological ones). Thus, we begin to see a very large repertoire of forms. This situation obtains for positive convergent accommodation, as in the English case, but from examination of Chiwere, seems also to hold true for divergent accommodation. In Chiwere, there also are a large number of well-known forms that differ between the Otoe-Missouria and the Iowa dialects. These may date to a time when positive accommodation existed between the two dialects. Examples of these differences include: (a) O-M / ñ / ~ I /o/, (b) O-M /s/ ~ I / çs /, (c) O-M second-syllable stress on some lexical items vs. Iowa first-syllable stress on the same items. In addition, it would appear that the lack of accommodation has also produced variants within and across the two dialects.

This state may have developed because family dialects changed in isolation from one another. For that reason, one can predict that the more factionalized a group is the more forms of markers there will be in its dialect; furthermore, change should proceed more rapidly in the more factionalized group where family dialects are stronger and less likely to be used with other Chiwere speakers. In the specific case of consonant collapsing, however, we can already see a more advanced state of change as indicated by fewer variants as a reduction in consonant inventory draws near to completion. Both dialects of Chiwere show a collapse of the contrast between the glottalized consonants and the lenis, unaspirated consonants, but the Iowa dialect shows greater loss than the Otoe-Missouria dialect. It may be hypothesized that reduction of variants signals imminent language death.

Chiwere has three series of consonants: (1) a tense aspirated series, (2) a lenis unaspirated series with both voiced and unvoiced alternates, and (3) a glottalized series. The lenis series of consonants has two alternates, voiced and unvoiced. The change that has been occurring is that the glottalized consonant series is merging with the voiced alternate of the lenis series.

The result is a reduction from three series of consonants to two series. These are the voiceless tense aspirate consonants and the lenis consonants, the latter with voiced and unvoiced alternates.

This change can be seen also in the related Dhegiha Siouan languages (Rankin 1978, Rankin and Koontz 1986). Dhegiha has four series of consonants: (1) tense voiceless gemminates, (2) tense voiceless aspirates, (3) lenis consonants with voiced and unvoiced alternates, and (4) glottalized. Over time, the related developments in Dhegiha have been those in 3:

3. (1) glottalized fricatives merged with plain voiceless counterparts
(2) C' > C' ~ C
(5) Ch retained
(4) C > [+voice]
[+lax]
(5) CC retained.

We can compare the Chiwere and Dhegiha collapses:

These changes would leave Dhegiha, eventually, with a tense/aspirate/lax set of contrasts in the consonants, which would have derived from a tense/aspirate/lax/glottalized series. In Chiwere, on the other hand, an original set of contrasts based on aspiration (also tense), laxness (with voiced and voiceless variants), and glottalization is nearly completely reduced, especially in the Iowa dialect, to a contrast between tense aspirates and lax unaspirated consonants; the latter have two variants, a voiceless one and a voiced one that has absorbed the former glottalized consonants. Although we think it unlikely that the Iowa dialect will persist into the next generation as a spoken language, we do think it possible for it to endure sufficiently long to develop this voice/non-voice contrast in the lax consonants. As Rankin (1978) and Rankin and Koontz (1986) note, the collapse of glottalized consonants and lax consonants into a single phoneme follows a pattern predicted by a universal hierarchy of phonological features, one that is paralleled by the contemporary situation in the Dhegiha languages where a four-way contrast is collapsing into three. If Chiwere were to persist in the Iowa-speaking community, we suggest that the language would develop the voicing contrast shown in 4; if so, it would then have a three-way contrast once again, but one based on tense consonants, lax voiced consonants, and lax unvoiced consonants. We suggest this path of development because it is our impression that there remains a difference in the voicing behavior between lexical items with the now lax consonants that derive from the original lax set and those that derive from the glottalized set. The latter, we perceive, rarely have devoiced versions, whereas the former have voiced and voiceless alternates.

NEED EXAMPLES HERE of lexical items (a) with lax consonants deriving from original lax set and (b) with lax consonants deriving from glottalized set.

We also suggest that there is a kind of communicative necessity that would drive such an innovation since homophony becomes very high with only two contrasting series of consonants; indeed, Dhegiha innovated its fourth series of consonants from what was apparently a proto-Siouan series of three.

EVIDENTIALITY AND THE SECOND PERSON PARADIGM

Concern for evidence permeates Chiwere grammar. It can been seen, for example, in the series of particles associated with the sentence. Often referred to as the "sentence determiners," these have gender-specific variants and include khe (m.)/khi (f.) declarative (e.g., warigrókhiwi khe 'I (masc.) pray you listen to me'); re (m.)/ræ ~ r´ (f.) (often now, only re, the masculine form, is used) polite request/polite command; ho (m.)/h£a (f.) hortative; hna (m.)/ hna (f.) yes/no question; ne (m/f) strong demand/command; ah (m.)/ a÷e (f.) irritation, ask£u (m.)/ asko (f.) quotative. The function of these particles appears to be more complex than simply labeling gender of speaker and type of speech act. They also signal source of information in the expression (quotative) or attitudes toward the information (quotative) or the addressee (irritation). These latter distinctions suggest the possibility of a previous independent system of evidentials, now conflated with the sentence determiners, although today concern for character of evidence distinctions can be identified in other areas of the grammar as well.

The deictic system, for instance, holds a distinction for going away from and back to the speakers 'home' or 'home base.' The 'home' category pervades Siouan languages and is cross-cut by the category of motion and arrival. These forms are conventionally identified as "vertitive verbs" (Taylor 1976). So, for example, the vertitive Chiwere verbs grí means 'to arrive here/home (motion at arrival)' and gú 'to arrive here toward home (motion prior to arrival)' may be compared with the nonvertitive *íi 'to arrive here (not home) (motion at arrival)' and hú 'to arrive here (not home) (motion prior to arrival)' (Taylor 1976:292).

We can couple these findings with possible changes in the second person paradigm. One fluent speaker freely uses second person forms for questions or commands, but generally resists its employ for declaratives except with the future tense (which he tends to translate as a polite command). He claims that the second person is never used with past or present in declaratives because one cannot tell someone what he or she did or is doing, only perhaps what the addressee will do. In fact, other speakers do use the second person with past and present, and it is reported in early studies, such as the Marsh and Dorsey manuscripts, and is included in Whitman's 1947 paper, the primary informant for which was the father of the speaker in question (Hopkins and Furbee 1991). Forms of the second person with present and past tense verbs were discovered also in the casual, non-elicited speech of the informant who denied their appropriateness, and he accepted several documentary and present-day contextualized examples of the second person in present and past tense from these texts, although he remained reluctant to extend general acceptance to all such cases.

Those observations suggested a loss of the second person category in some grammatical contexts and possibly some social ones. Explanations for such changes in language death offered in the past have ranged from their being patterned according to a reverse of the either the acquisitional sequence or the creolization process to, more generally, the elements unraveling from most marked to least marked. The last suggestion is the most direct statement of an implicit universal formulation and was the form tested in a study of change in the Chiwere second person paradigm reported by Lonsdorf and Furbee (1993) using a version of the noun phrase hierarchy proposed by Silverstein (1977, 1985). That work gave special consideration to accounting for the regularities seen at the interface of two planes of analysis, the phonological and morphosyntactic. Among the specific sets Silverstein treated were noun phrases and pronominalizations, leading to his developing a markedness-based typology of case marking that made particular reference to a possible dynamic between ergative and accusative systems. Lonsdorf and Furbee found some evidence that more marked Chiwere pronominal Forms were being lost in advance of less marked ones.

Table 2 shows relevant Chiwere pronominal forms arranged according to the pertinent categories and features in the noun phrase hierarchy. Only those features that pertain for Chiwere are included. The prediction is that the left-most forms will be the most stable and the most resistant to erosion, the right-most forms the most vulnerable.

Table 2 omitted due to lack of space]

From this hierarchy of noun phrase types in referential space, one can make several predictions: There is the universal typological prediction that a language having a category at the right of the figure is likely to possess all categories to the left of it. In addition, as in other instances of implicational hierarchies, the model predicts a growing complexity such that one might expect children learning a language to first acquire the leftmost categories, which are also the least marked. Conversely, one might test the idea that a person (e.g., an aphasic) is likely to lose linguistic categories from the most marked end of the scale before those that are less marked. Similar predictions can be drawn for the birth and death of languages themselves that in the growth and development of creole languages, categories should be added from left (least marked) to right (most marked) (see Bickerton 1981 &;1983, Sankoff and Laberge 1973), and conversely, that in the situation of language death, categories should be lost from right (most marked) to left (least marked).

The pronominal study compared forms from various documentary sources with those obtained from recent field work. The documentary sources included the Dorsey manuscript (n.d. [1890]), the Marsh manscript (n.d. [1935]), Voegelin's 1941 publication, Whitman's 1947 publication, and Good Tracks' 1991 dictionary, which is based largely on earlier manuscripts and publications. The documentary materials touched back into the 19th century and even the more recent recorded the language when it was widely spoken and healthy. The recent materials derived from field notes dating from 1988-1992. Both literature and field notes were surveyed for examples of pronominal use with verbs in both the agentive and stative verb conjugations. Some examples (adapted from Lonsdorf and Furbee 1993) follow:

Documentary Examples 4: Abbreviations used in these examples are 1 p (first person), 2
p (second person), ag (agentive), pl (plural), def (definite), sing (singular), masc (masculine),
dec (declarative), fem (feminine), pt (patient).



AGENTIVE CASE

a. ra...wi 2 person plural (definite)

hºi
yºinºa
ra
grèi
wi
'You have come back'
my
brothers
2 p ag
arrive back
pl def
(Marsh, n.d.)


b. ra - 2nd p singular

ra
nºèayºi
ra
nºèayºi
'You are standing'

nèayºi nèayºi
2 p sing
stand/be standing
(Marsh, n.d.)


c. ha - 1st p. singular

ha
skèaÇje
'I play'
1 p sing ag
to play
(Marsh, n.d.)


d. hºi . . . wi 1st p. plural inclusive (def)

hºi
*i
wi
kSe
'We have
come'
1p pl ag
come
[here]
pl def
masc dec
(Marsh, n.d.)


e. hºi - 1st p dual

hºi
gèi lo
kSe
'I'm happy'
1 p dual ag
happy,
glad
masc dec
(Marsh, n.d.)


STATIVE CASE

a. ri-. . .wi 2 person plural

ri
x'ºèa
~nºioe
w`asgºu
'you (pl) seem
to be tired.'
2 p pl pt
live; have
vitality
have
nothing
wi + asgºu
pl (def.)
perhaps
(Marsh n.d.,
Voegelin
1941)


b. ri- 2 person singular

ri
xèºa~ne
kSe
'you're big'
2 p sing pt
to be big
masc dec
(Marsh, n.d.)


c. hºi- 1st person singular

hºi
xwa~ne
'I am lost'
1 p sing pt
to be lost
(Marsh
n.d.,Voegelin 1941)


d. wa - wa . . . wi 1 person plural

wiwa
hge
hda
wi
~nºioe
kSe
'we shall
never
wa + i +
wa
like (as)
shall
pl
have none,
nothing
masc dec
be like/thus'
1p pl pt
(Marsh n.d.)


e. wa - wa 1 person dual

wawèa
~nºioe
'we have nothing'
1p dual pt
have none, nothing
(Marsh n.d.)


Recent Fieldwork Examples Abbreviations used in these examples are 1 p (first person),
2 p (second person), ag (agentive), pl (plural), def (definite), sing (singular), masc
(masculine), dec (declarative), fem (feminine), pt (patient).

AGENTIVE CASE

a. ra . . . wi 2 person plural (def.)

heda
warèu*
ar`e
÷sºu
ra
h``ºa
wi
'and you (pl)
even
and
food
it is
indeed,
even
2 p ag
cook
pl def
cook the food
'


b. ra - 2nd p singular

rèa
n`ayºi
kSe
'you stood up'
2 p sing ag
to stand
masc dec


c. ha - 1st p. singular

ha
çjèi
kS`i
'I came'
1 p sing ag
came,
arrive
fem dec


d. hºi . . . wi 1st p. plural inclusive (def)

hºi
mèa~ni
wi
kS`e
'we (2) are
walking'
1 p pl ag
to walk
(move)
pl
masc dec


e. hºi - 1st p dual

hºi
ma:~nìi
kSe
'2 or 3 that's
walking(moving)
1 p dual ag
walking(moving)
masc dec
(we're walking)'


STATIVE CASE

a. ri . . . wi 2 person plural

iyºa
hºa
rèi
gra*`e
wi
'we come to you
one
hºi + a
2 p pl pt
single out,
seek out
help
pl
(Bear Clan)'
1 p sing ag
on


b. ri - 2 person singular

ri
ruxèa we
}ne
kS`e
' they have
scratched you
'
2p sing
pt
'skin; peel
something'
3 p pl
masc. dec.
(you have a
mark)
rux'èi 'to
scratch'


c. hºi - 1 person singular

hºi
s'èage
kSi
'I am old'
1 p sing pt
to be old
fem dec


d. wa - wa . . . wi- 1 person plural

wawa
"èabed`a
wi
kSe
'we're
intelligent'
1 p pl pt
intelligent
pl
masc dec


e. wa - wa 1 person dual

hèine
urage
are
wèawa
~ni
asgºu
'We're the last ones
still
we 2
last, end
that is
1p dual pt
a~ni
perhaps/ it
seems
friends, we are the
two
to have
remaining ones.'


The study found some support of the idea that pronominal category loss proceeds from most marked to least marked, although the evidence was not conclusive due largely to the small number of examples of the first person dual and the second person plural categories in either data set. The first person dual category was the one predicted to be the most
resistant to loss, but it actually showed even fewer examples than the second person plural. The decline of the latter showed no time differential. As mentioned, however, examples are few for both these categories, and there are many fewer examples of the second person plural in both document and field note records of this category than of the second person singular, first person singular, or first person plural.

Nonetheless, what seems to be happening here, from an accommodation perspective, is that as Chiwere dies, selected aspects of its structure coalesce due, at least in part, to changes in the circumstances of its use, specifically the restricting of the every-day language register first to the home (where information sources and evidence are well known and widely shared) and then to memory alone. It is possible that a prior system of evidentials has deteriorated and been distributed, has specialized to exceptional cases, or possibly has collapsed in on the sentence determiners primarily. The newly reduced paradigm has become reinterpreted and rationalized. The reinterpretation now extends even into the pronominal categories, where folk etymological reasoning can be offered for a decline in distinctions made there. Accommodation theory provides a rationale by which one can retrodict shifts as hypotheses. We are pointed by the theory in this instance to look for evidence of a previously strong system of evidentials, possibly associated with the tense or aspect morphology.

DISCUSSION AND CONCLUSIONS

In this paper, we have argued that both positive, convergent accommodation and negative, divergent accommodation can impel language change. In the case of Chiwere, the two processes would seem to have accelerated the demise of the language. Indeed, the Chiwere data lend support to the social cause for language maintenance and language loss, and may point to a different direction for studies of language death.

The thrust of research on languages in the process of death has been toward identifying the properties of various languages in the state of decline or toward finding the universal characteristics that govern their decline and associating those with models that have known universal currency. For example, the types of phonological reductions seen in Chiwere and Dhegiha language death correlate well with the relative hierarchy of sound types found by Greenberg (1960a, pp. 63-66) in that the sounds are lost in a series from most marked to least marked (Rankin 1978 and Koontz 1986, pp. 20-21; see also Dressler 1973 for a discussion of phonological change). Nevin (n.d., p.18) has claimed that the deacquisition explanation requires an assumption that successively younger semi-speakers of a dying language cease learning at successively earlier stages of the acquisition process. Whether interrupted learning is the only means of obtaining deacquisition is perhaps questionable, but it is clear that what is known of Stage 3 of Chiwere death is congruent with that hypothesis. The findings from Chiwere pronominal loss also support a universalist claims.

Nonetheless, it is generally agreed that language internal factors never predominate as causes of language change (Swadesh 1948, Hill 1983), even though their pattern may reflect an internal set of predispositions, as in Sapir's concept of drift. Still, structural study of language change has tended to center on language-internal examination of properties of change in a dynamic system glimpsed in mid-motion. Functional studies, on the other hand, do identify aspects of the process of change that might be thought of as impelling primary factors, but such first causes are nearly always macrolevel ones external to the speech situation: For example, the pressure of a colonial language on an indigenous one (Hill and Hill 1986), the pressure of a salient dialect feature for identification with a lifeway, as in Labov's Martha's Vineyard study (1965), or the integrative importance of a dying language in a community (Dorian 1978). Of course, functional studies also identify causes of change within dynamic systems, as in studies of syncretism in language (Hill and Hill 1986). In addition, interesting general hypotheses have been derived from the functional perspective that may be said to make universal claims; for example, that language death is a case of decreolization (Dressler and Wodak-Leodalter 1977, Trudgill 1978, Jones-Jackson 1984), that it is deacquisition (Voegelin and Voegelin 1977, Mithun and Henry 1979), or that it is
code complexity reduction (Hill 1973).

Under accommodation theory we can hypothesize about possible microlevel first causes for language change. Furthermore, these products of accommodation are themselves amenable to analysis as following universal tendencies, and so are identifiable also with what might be thought of as an advantage associated with study of language death from the point of view of loss of competence. For example, Trudgill demonstrates that British English speakers accommodate to American English by altering their pronunciation according to a hierarchy of features such that those with the fewest or weakest inhibiting factors are accommodated to first. The sequence he discusses is (1) -/t/- > -/d/- in Peter etc., (2) /a:/ > /æ/ in dance etc., (3) [C] > [ a ] in top etc., and (4) Ø > /r/ /__ { C .

Looking back at the Chiwere and Dhegiha data, we can see the same processes at work. Furthermore, we can use such information to frame hypotheses about the direction and character of language change. It is our position that we should reorient study of language death toward more extensive examination of individual cases and their unique features prior to any wholesale attempt to relate them to universal tendencies. For example, reduction of syntactic complexity and massive borrowing from the dominate language are thought to be two traits that universally herald language obsolescence. Chiwere and some other dying Plains languages show the first, but not the second. They have almost no borrowing from English. Clearly social and political factors are involved in Chiwere's resistance to loan words. Perhaps linguistic factors are involved as well, but further investigation is needed to determine what they might be.

We would therefore urge a longer period of concern with the unique aspects of individual cases of language death because to do so offers two kinds of information: First, it may give a more accurate appraisal of the general universal processes involved in language death. And second, it may hold particularly useful clues to what a moribund language was like in the full vibrancy of life. Perhaps, before a language dies it may be revealing of a widespread importance attached to information validity and source. The situation with the second person forms points one to an increased weighting of the importance of the other expressions in the language that pertain to information source: the gender-specific sentence determinitives and the vertitive verbs. The latter may also be associable with semantic features of evidence, experience, or volition, and there may be other verb sets in Chiwere that may be similarly interpreted, such as those identified for Lakhota by Rood and James (1991) that are constrained in their possible use with first person agents. Not only may it be the case that an obsolescent language such as Chiwere retains unmarked elements in expression of the universal grammar, perhaps just before it dies such languages also reveal what was most important about them individually in life.

It is probably an important Chiwere fact functionally that it is the second person forms in present and past that one consultant has lost, since he claims one cannot tell someone what that person knows or did, but can only ask or command the addressee. The fact is that Dorsey in the 1880s, Marsh in the 1930s, and Whitman in the 1930s and 40s, all give these second person forms, and some other present-day speakers of both dialects use them too. Their loss is a new development, but it may tell us also something functionally important about Chiwere < a way of accommodating to the hearer, if you like, that may be relatable structurally to the universal feature hierarchy (Silverstein 1976) since the decline in the second person appears to be predicted by the person and animacy feature hierarchy.

We would suggest that, in general, it is better to measure all processes against a model based on universal characteristics, for example , the implicational hierarchies, rather than to jump to explanation of language death as representative of a process such as decreolization, or deacquisition, which in itself is a middle-level theoretical hypothesis drawn from ideas about universal grammar but not fully demonstrated. Such an approach resembles one in descriptive statistical analyses called Connonical Correlation. In the statistical instance, in order to measure the character and degree of similarity between two or more models, one easures each against some other model, which may be an independently motivated model, or may be a construct that lies somewhere between the two or more models being compared. That is, it may be theoretically motivated, or it may be derived. What we would urge is that we develop a model of language death grounded in universal principles against which we measure actual instances of languages in decline. In so doing, we may work back and forth between the model being constructed and the one developed from the data. We may also measure each against models derived fairly directly from analogic cases (decreolization, deacquisition, etc.). Further, in a case such as Chiwere, we would anticipate that the Connonical Universal Model would inform attempts to construct a full descriptive model of the language from the partial ones obtainable from individuals.

FOOTNOTES

1 For support of the Missouri Chiwere Language Project, from which this paper derives, we thank the Alumni Development Fund, the Faculty Research Council, and the Department of Anthropology of the University of Missouri-Columbia, the Penrose Fund of the American Philosophical Society, and the National Science Foundation. We are especially grateful to the Otoe-Missouria and Ioway communities of Oklahoma and several persons within those groups who have lent their time and counsel. 2 This literature is reviewed in Gatewood and Rosenwein (1981).

REFERENCES

Bell, A. (1984). Language style as audience design. Language in Society 13:145-204.

Bickerton, D. (1981a). Roots of language. Ann Arbor : Karoma.

_____________ (1981b). Pidgin and creole languages. Scientific American 249(1):116-22.

Comrie, B. (1989). Language universals and linguistic typology. 2nd edn., Chicago : University of Chicago Press.

DORSEY, J. OWEN. n.d. (1890). Various MSS on Chiwere. Smithsonian Institution.

Dressler, W. (1972). On the phonology of language death. Papers from the 8th Regional Meeting, Chicago Linguistic Society. Chicago: Chicago Linguistic Society. 448-457.

Dressler, W., &;Wodak-Leodolter, R. (1977). Language preservation and language death in Brittany. International Journal of the Sociology of Language 12:33-44.

Ferguson, C. A., &;J. Gumperz, J. J., eds. (1960). Linguistic diversity in South Asia: Studies in regional, social, and functional variation. Research Center in Anthropology, Folklore, and Linguistics Publication 13. International Journal of American Linguistics 26, #3, part III.

Furbee, L., &;J.D. Hopkins. 1991. The Chiwere verb word. Paper presented at the 90th Annual Meeting of the American Anthropological Association, Chicago.

Gatewood, J., &;Rosenwein, R (1981). Interactional synchrony: Genuine or spurious? A critique of recent research. Journal of Nonverbal Behaviour 6:12-29.

Giles, H. (1973). Accent mobility: A model and some data. Anthropological Linguistics 15:87-105.

________ (1979). Ethnicity markers in speech. In K. R. Scher &;H. Giles (eds.), Social markers in speech. Cambridge: Cambridge University Press. 151-189.

Giles, H., &;Johnson, P. (1987). Ethnolinguistic identity theory: A social psychological approach to language maintenance. International Journal of the Sociology of Language 68:69-99.

Giles, H., &;Smith, P. (1979). Accommodation theory: Optimal levels of convergence. In H. Giles &;R. St Clair (eds.), Language and social psychology. Oxford: Blackwell.

Giles, H., Taylor, D.M., &;Bourhis, R. (1973). Towards a theory of interpersonal accommodation through language: Some Canadian data. Language in Society 2:177-192.

Greenberg, J. H. (1966a). Language universals: With special reference to feature
hierarchies. Janua Linguarum, Series Minor, 59. The Hague: Mouton.


_______________ (1966b). Some universals of grammar with particular reference to the order of meaningful elements. Universals of language, 2nd edn., ed. by Joseph H. Greenburg, 73-113.Cambridge, MA : MIT Press.

Dorian, N. C. (1973). Grammatical change in a dying dialect. Language 49:413-438.

_______ (1978). The fate of morphological complexity in language death. Language 54:590-609.

Good Tracks, J. G. (1992). Baxoje - jiwere - nyut'aj < Ma'unke Iowa - Otoe - Missouria language to English. Boulder, Colorado: Center for the Study of the Languages of the Plains and the Southwest.

Grice, H. P. (1975). Logic and conversation. In P. Cole &;J. Morgan (eds), Syntax and
semantics, vol. 3. Speech Acts. New York: Academic Press. 41-58.

Hill, J. H. 1973. Subordinate clauses and language function. In C. Corum et al. (eds.), You Take the High Node and I'll Take the Low Node, Chicago: Chicago Linguistic Society.
33-52.

Hopkins, J.D., &;L. Furbee. 1991. Indirectness in the interview. Journal of Linguistic
Anthropology 1(1). 63-77.

Keenan, E. L. 1976. Towards a universal definition of subject. Subject and topic, ed. by Charles N. Li, 303-33. New York : Academic Press.

___________ &;B. Comrie. 1977. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8, 63-99.

____________________ 1979. Data on the noun phrase accessibility hierarchy. Language 55, 333-51.

Koontz, J. E. 1991. 'Suddenly' in Mississippi Valley Siouan. Paper presented at the 90th
Annual Meeting of the American Anthropological Association, Chicago.

__________ &;R.L. Rankin. (1986). ?????????

Labov, W. (1965). On the Social Motivation of a Sound Change. Word ?????
MARTHA'S VINEYARD STUDY ????????

________ (1966). The social stratification of English in New York City. Washington, DC: Center for Applied Linguistics.

________ (1969). The logic of nonstandard English. In W. Labov. Language in the inner city. Philadelphia: University of Pennsylvania Press. 201-240.

Le Page, R.B. (1968). Problems in description in multilingual communities. TPS 1968, 189-212.

_______ (1975). Polarizing factors: Political, social, economic, operating on the individual's choice of identity through language use in British Honduras. In J.G. Savard &;R. Vigneault (eds.), Les États multilingues. Quebec: Laval University Press.

_________ (1978). Projection, focussing, diffusion. Society for Caribbean Linguistics Occasional Paper 9.

Le Page, R.B., Christie, P., Jurdant, B., Weekes, A.. &; Tabouret-Keller, A. (1974). Further report on the sociolinguistic survey of multilingual communities. Language in Society 3:1-32.

Marsh, G. H. n.d. (1935) Materials for a study on the Iowa language, MSS 1-6, 9-10, 13-15. Philadelphia: American Philosophical Society.

Mithun, M., &; Henry, R. (1979). Notes on moribundity: Incipient obsolescence of Oklahoma Iroquois. Paper presented at the International Congress of Americanists, Vancouver.

Miller, W. (1971). The death of language or serendipity among the Soshoni.
Anthropological Linguistics 13:114-120.

Nevin, B. n.d. Obsolescence in Achumawi: Was Udall working with semi-speakers? Ms.

Nordenstam, K. (1979). Svenskan i Norge. Gothenberg: University Press, cited in P. Trudgill, Dialects in contact. 24-28.

Rankin, R. L. (1978). The unmarking of Quapaw phonology: A study of language death. Kansas Working Papers in Linguistics 3:45-52.

Rood, D.S., &;E. James. (In press). Lessons from the Lakhota dictionary. Proceedings of the Mid America Linguistics Conference, 1991. Stillwater, OK: Oklahoma State University

Sankoff, G., &;S. LaBerge. (1973). On the acquisition of native speakers by a language. Kivung 6(1). 32-47.

Silverstein, M. (1976). Hierarchy of features and ergativity. Grammatical categories in Australian languages, ed. by R. M. W. Dixon (Australian Institute of Aboriginal Studies (Canberra) Linguistic Series, no. 22), 130-71. Atlantic Highlands, NJ : Humanities Press, Inc.

_____________ (1987). Cognitive implications of a referential hierarchy. Social and functional approaches to language and thought, ed. by Maya Hickman, 125-164. Orlando: Academic Press, Inc.

____________ (1984). The 'value' of objectual language. Paper presented at the 83rd Annual Meeting of the American Anthropological Association, Denver, November 14-18, 1984.

Stanley, L. A., &;Furbee, L. (1991). The "relevant" language in Native American language planning. Proceedings of the Mid America Linguistics Conference, 1989. Cedar Falls, IA: Northern Iowa University.

Trudgill, P. (1983). On dialect. Oxford: Basil Blackwell.

____________ . (1986) Dialects in contact. Oxford: Basil Blackwell.


Voegelin, C. (1941). Internal relationships of Siouan languages. American Anthropologist 43. 246-49.

Whitman, W. 1947. Descriptive grammar of Ioway-Otoe. International Journal of American Linguistics 12.233-48.

Return to top
Return to Ioway Virtual Library main page