Focality and Topicality Marking in Biloxi

The purpose of this paper is to explore the use of two clitics that appear as discourse markers in Biloxi (ISO 639-3 bll) narratives, specifically –di and –yą. While a few Siouanists and other linguists have scratched the surface in identifying and analyzing these clitics, this will be the first attempt at an in-depth analysis of these clitics in Biloxi narratives as focality and topicality markers. Einaudi first examined these particles in her dissertation, A Grammar of Biloxi (1974). Unfortunately, her brief examination concludes with the rather discouraging remark that these clitics “remain the thorniest problem of Biloxi syntax” (1974: 149). With this paper, I attempt to make this problem of Biloxi syntax a little less thorny. Unfortunately, the nature of analyzing an extinct language in which we can no longer elicit help from native speakers for their perceptions about the use of particular clitics means that this philological analysis is, to some degree, necessarily subjective and theoretical. The only data left for us to analyze are the narratives from Biloxi oral tradition and some elicited sentences. The data in this analysis are far from black and white and are indeed rather messy, leaving many unanswered questions and seeming to imply a greater degree of complication and complexity in the language than we may now, without the benefit of native speaker intuition, be able to know for certain. However, these clitics will be analyzed within the theoretical framework called the Givenness Hierarchy (Gundel, 1993), which will hopefully aid in our understanding of their use. Biloxi is an extinct Siouan language. Specifically, it is a member of the Ohio Valley, or Southeastern, branch of the Siouan language family. Its closest known linguistic cousins are Ofo and Tutelo, also extinct. Biloxis were first encountered by Europeans in southern Mississippi in 1699. Later, Biloxis came to inhabit Louisiana and eastern Texas. The remaining Biloxis currently share a small reservation with Tunicas, a linguistically unrelated people, in Marksville, Louisiana. The last known native semi-speaker of Biloxi, Emma Jackson, died in 1934. Biloxi is the best-documented member of the southeastern branch of Siouan, but this is the least studied and documented branch of the Siouan language family. Thus, the analysis and study of Biloxi is of crucial importance, not only for its own sake, but for the knowledge yet to be gained from this little-studied branch of Siouan and the contribution it can make to Siouan studies and to linguistics in general. Biloxi is an agglutinative, head-marking, subject-object-verb (SOV) language. Verbs are the most highly inflected category and are subject to noun incorporation. Biloxi lost the activestative split common in other Siouan languages. In Biloxi, there is no part of speech designated


Introduction
The purpose of this paper is to explore the use of two clitics that appear as discourse markers in Biloxi (ISO 639-3 bll) narratives, specifically -di and -yą.While a few Siouanists and other linguists have scratched the surface in identifying and analyzing these clitics, this will be the first attempt at an in-depth analysis of these clitics in Biloxi narratives as focality and topicality markers.
Einaudi first examined these particles in her dissertation, A Grammar of Biloxi (1974).Unfortunately, her brief examination concludes with the rather discouraging remark that these clitics "remain the thorniest problem of Biloxi syntax" (1974: 149).With this paper, I attempt to make this problem of Biloxi syntax a little less thorny.
Unfortunately, the nature of analyzing an extinct language in which we can no longer elicit help from native speakers for their perceptions about the use of particular clitics means that this philological analysis is, to some degree, necessarily subjective and theoretical.The only data left for us to analyze are the narratives from Biloxi oral tradition and some elicited sentences.The data in this analysis are far from black and white and are indeed rather messy, leaving many unanswered questions and seeming to imply a greater degree of complication and complexity in the language than we may now, without the benefit of native speaker intuition, be able to know for certain.However, these clitics will be analyzed within the theoretical framework called the Givenness Hierarchy (Gundel, 1993), which will hopefully aid in our understanding of their use.
Biloxi is an extinct Siouan language.Specifically, it is a member of the Ohio Valley, or Southeastern, branch of the Siouan language family.Its closest known linguistic cousins are Ofo and Tutelo, also extinct.Biloxis were first encountered by Europeans in southern Mississippi in 1699.Later, Biloxis came to inhabit Louisiana and eastern Texas.The remaining Biloxis currently share a small reservation with Tunicas, a linguistically unrelated people, in Marksville, Louisiana.The last known native semi-speaker of Biloxi, Emma Jackson, died in 1934.Biloxi is the best-documented member of the southeastern branch of Siouan, but this is the least studied and documented branch of the Siouan language family.Thus, the analysis and study of Biloxi is of crucial importance, not only for its own sake, but for the knowledge yet to be gained from this little-studied branch of Siouan and the contribution it can make to Siouan studies and to linguistics in general.
Biloxi is an agglutinative, head-marking, subject-object-verb (SOV) language.Verbs are the most highly inflected category and are subject to noun incorporation.Biloxi lost the activestative split common in other Siouan languages.In Biloxi, there is no part of speech designated "adjective."Biloxi lexemes commonly translated by English adjectives are actually verbal in nature.
Currently published materials on Biloxi include the Dorsey-Swanton Dictionary of the Biloxi and Ofo Languages (1912), two articles by Haas: "The Last Words of Biloxi" (1968) and "Swanton and the Biloxi and Ofo Dictionaries" (1969), A Grammar of the Biloxi Language (1974) by Einaudi, and my own article: "A Reanalysis of the U-circumflex and U-brève in Biloxi" published in the Kansas Working Papers in Linguistics (KWPL), volume 28 (2006).

Narratives Analyzed and Methodology
The primary narratives analyzed in the preparation of this paper are the following: 1) A Fox Story 2) A Ghost Story 3) Ant, Katydid, and Locust 4) How Kuti Mąkde Made People2 5) The Crow and The Hawk 6) The Earth Rolled3 7) The Opossum and the Raccoon 8) The Rabbit and the Bear 9) The Red-Winged Blackbird 10) The Wolf That Became A Man Of the texts listed above, it should be noted that at least a couple (How Kuti Mąkde Made People and The Earth Rolled) are not originally Biloxi but rather translations of narratives widely circulated among various southeastern indigenous cultures, some believed to be of African origin diffused via African-American slaves (Dundes, 1969).
What is apparent from the narratives is that Biloxis, at least in formal discourse and in oral tradition, coded certain nouns with clitics in an effort to transmit pragmatic information to the hearer.These clitics relate to a speaker"s "assumptions about the addressee"s knowledge and attention state" and serve as processing signals to the hearer (Gundel, 1993: 275).In other words, these clitics are firstly "signposts" to comprehension, serving to clarify or disambiguate referents in cases where the speaker may fear possible misunderstanding or confusion on the part of the hearer.Secondly, these clitics serve the pragmatic function of emphasizing, highlighting, or otherwise bringing some element of personal feeling or emotion to the speaker"s relating of the narrative.
My method of analysis was to find the occurrences of these clitics primarily from the above listed narratives.Next, I analyzed their distribution in these narratives, and, equally as important, where they did not occur when expected.Further, I examined their distribution in accordance with the theoretical framework called the Givenness Hierarchy.

3.
The Givenness Hierarchy (Gundel, 1993): The Givenness Hierarchy (GH) is a theory in which different determiners and pronominal forms correlate with different cognitive statuses reflecting information about location in memory and attention state, which then enables the hearer to restrict the set of possible referents (Gundel, 1993).Gundel proposes that there are six cognitive statuses relevant to the form of referring expressions in natural language discourse and that these are related in the GH shown below (ibid, 1993).
The statuses are ordered from most restrictive (in focus) to least restrictive (type identifiable).However, it must be mentioned that this can be rather confusing, since the GH reflects cognitive focus rather than pragmatic focus.Thus, what is considered pragmatic focus (such as in first mention, where in English we would use the indefinite article) would actually be under "type identifiable" in the GH scale.But, while the GH may not be the most optimal framework for sorting out the Biloxi clitic data, I believe it does at least give us a working theoretical framework through which the distribution and usage of these clitics becomes a bit more comprehensible and a bit less mysterious.
There are six cognitive statuses relevant to the form of referring expressions in natural language discourse according to the GH.However, Gundel informs us that, "Not all six statuses are required for all … languages" (1993: 284).In Biloxi, it appears "type identifiable" and "referential" can be conflated into one status.This also appears true for "familiar" and "uniquely identifiable," meaning that only four statuses appear to be necessary and sufficient in Biloxi.As for the definition of givenness, Brown and Yule state: Given information is specified as being treated by the speaker as "recoverable either anaphorically or situationally" … and "new" information is said to be focal "not in the sense that it cannot have been previously mentioned, although it is often the case that it has not been, but in the sense that the speaker presents it as not being recoverable from the preceding discourse (1983: 179, italics mine).
I find the italicized words in the quote above crucial to understanding this use of -di andyą in Biloxi.These clitics are types of pragmatic markers used when the speaker wishes or feels the need to emphasize or give more information to the hearer.Thus, these clitics are analyzed from a discourse rather than a grammatical perspective.

Focus and topic
My statistical analysis of the data reveals that, out of the ten narratives examined, the clitic -di occurs as first mention focus marker in six of them, and -yą occurs after first mention as a topicalizer-definite article in seven of them, but, as stated before, not necessarily on a consistent basis.These Biloxi clitics have been previously identified as case markers (Drechsel, 1997), but it must be stressed that, given the statistical occurrences in the data, these clitics are most definitely not case markers, since, unlike in true case-marking languages like Latin or Russian, these clitics do not consistently occur.Given this analysis, I find it best to consider -di a type of focus marker and -yą a type of topic marker or definite article, that is, forms of discourse marking.
Gatschet (1886) makes the revealing statement in his handwritten field notes that "the suffixed -ya is often indicative of distance, -di, -ti of nearness being close to the one speaking" (his underlining).Thus, it appears these two clitics had a literal meaning expressing actual physical distance from the speaker, but they grammaticized into discourse devices to express a speaker"s abstract feelings of distance, nearness or remoteness from a subject in discourse or narrative.
Interestingly, this use of -di and -yą seems to correlate with the southern American English dialectal usage of "this (here)" (first mention) and "that (there)" (the one already introduced, the one that the hearer already knows), as in the sentence: This (here) coyote was howling outside my house last night, and that (there) (same) coyote caught a rabbit.

a. Focus -di
The clitic -di apparently descends directly from Proto-Siouan (PS) -ri, a focus marker also found in Hidatsa.This clitic is used in Biloxi as both a focus marker and as a type of intensifier.In the first case, it is sometimes (about six times out of ten in the data analyzed) used for first mention when objects or characters are first introduced into a story, thus signaling FOCUS or new information.Example 1 below is the first sentence from the story, The Opossum and the Raccoon.However,  marking also occurs in the case of first mention.In example 2, neither ąyaaxohi nor axka occurs with focus marker -di: (2) An old woman put a large quantity of persimmons out to dry (D-S 1912:67:1) In the second case, the clitic -di seems to be an alternative to the more common -xti or -sti intensifying suffix: ( Circle was lying on top of the rock in her form as a circle Note that, in the above Mandan example, kamixe, "circle", appears both with and without the focus marker -re.

b. Topic -yą
Einaudi states that "yą is sometimes glossed as "the" and other times has a quasidemonstrative force to it meaning "yonder"" (1974: 133).She also states that "… yą often has the value of "that", although it usually has a neutral meaning "the"" (ibid.: 151).
Indeed, per Gundel, "A number of researchers … have observed that in various languages expressions referring to topics are necessarily definite " (1988: 213).Therefore, we can say that -yą is a form of definite article (as suggested by Einaudi) that statistically occurs most frequently when the noun to which it is suffixed has already been introduced into a story, thus marking TOPIC or old information.In both examples above, the marked nouns, ątatka and ąyaaxohi, were mentioned earlier in the discourse.However, -yą also sometimes occurs in the first sentence of a story, particularly with natural objects, such as a pond.This seems to indicate that -yą"s use as a topicalizer or definite article extends to objects even in the position of first mention when the object is culturally well known or salient (i.e, known to both speaker and hearer).Example 1 above provides a case in point, in which Dorsey revealingly translates yohi-yą as "a certain pond."The pond"s marking with -yą in the first sentence of the story seems to indicate that this isn"t just any pond, translatable with the English indefinite article, but instead it seems to imply that this is a specific pond known to both speaker and hearer in a geocultural context. 5 corollary use of topic marking occurs in Haida, an Athapaskan (Na-Dene) language of Alaska: (8)

Ham Cove guu l xaogaanggank
Ham Cove TOP he fish-PAST Ham Cove is where he fished. 6(Edwards 1979: 155) Edwards states, "… topicalized elements in Haida sentences are … items known to the speaker and hearer from context…."(1979).Thus, in the above Haida sentence, the topic marker guu is used "for the recall of general knowledge" (in this case recalling to mind Ham Cove, a location known to both speaker and hearer in a geocultural context).It appears that the same use of topic-marking occurs in Biloxi in which -yą is often enclitic to place names, e.g, Amǫyix-yą, Baton Rouge; Rapid-yą, Rapides Parish; Ayixi Xuhe-yą, Roaring Creek; Tą Nithąąyą, New Orleans.This is an example of grammaticization and extension of -yą"s original topical meaning, recalling to the hearer"s memory a place of which they both have general knowledge, whether a particular town or geographical feature.
As for why the clitic -yą alternates with  marking, Gundel states, "The use of topic markers appears to be optional, a fact which distinguishes them from case marking particles" (1988: 217).In a 1985 study of Hausa, a Chadic (Hamitic) language, Jaggar (in Lyons, 1999) offers a correlate.He writes: while the appearance or omission of the [Hausa definite] article cannot be fully predicted, it is largely determined by the accessibility of the referent.If the previous mention of a referent is considerably far back in the discourse it is less easily activated by the hearer, and if other, similar, referents have occurred in the intervening discourse, these can interfere with the hearer"s identification of the intended referent.In such circumstances the speaker tends to use the "heavier coding" of an article-marked noun phrase as a way of alerting the hearer to the need to find the referent and thus helping him in the task.In other words, a bare noun phrase is used when the referent is judged to be easy to access, and a definite-marked noun phrase when more effort seems to be required.
The last sentence in the quote informs us that, in Hausa, nouns may be left unmarked if the speaker feels the referent is easily accessed by the hearer and there is no need for "heavier coding" of a cliticized noun.Similarly, in Biloxi, neutral non-marked nouns are sometimes sufficient for the speaker"s purposes, but the markers can be optionally employed if the speaker believes that more effort is required for recognition or identification on the part of the hearer.

Conclusion
We have seen that, in the Biloxi data, the clitic -di appears to correlate fairly well with "type identifiable" or "referential" positions, while -yą appears to correlate fairly well with "uniquely identifiable" and "familiar."Further,  marking seems to occur with all statuses when the speaker deems the heavier "coding" supplied by -di or -yą as unnecessary or superfluous.It must be remembered that discourse marking is speaker-centered, largely based on the speaker"s attitude about the hearer"s knowledge and attention state.
This paper was a first attempt at an in-depth analysis of two Biloxi clitics, the focality and topicality markers -di and -yą.For the most part, these clitics have remained a mystery to Siouanists and other linguists over the past century.Unfortunately, discourse analysis of an extinct language is difficult and the conclusions are certainly not concrete.As we have seen, the data in this analysis are far from black and white and are indeed rather messy, leaving many unanswered questions and seeming to imply a greater degree of complication and complexity in the language than we may now, without the benefit of native speaker intuition, be able to know for certain.
It is evident that, as regards discourse and narrative analysis, language and cannot be easily separated.Discourse and narrative analysis requires a deep knowledge and understanding of the culture and the relationship between speaker and hearer underlying the discourse strategies a speaker employs and the clitics s/he uses.Unfortunately, Biloxi culture is now largely lost to us.
That being said, I have attempted in this paper to synthesize the data available in the best way possible according to our current understanding of it.I have also employed the GH (Gundel, 1993) as a theoretical background upon which to frame these important but enigmatic focus and topic clitics for a better understanding of their use in Biloxi discourse and narrative.Despite the inherent difficulties in analyzing complex data from an extinct language, I hope that I have achieved my intended purpose in giving us a deeper understanding of the language in discourse and narrative and to hopefully pave the way for further research and analysis.

Fig. 1
Fig. 1 Six cognitive statuses of the GH