Emmon Bach
UMass (Amherst) / SOAS
2004-02-08

0. Introductory:

Smalgyaxian (Tsimshianic): located along Nass River, the lower reaches of the Skeena and on coastal and island communities around and below mouth of Skeena and in Metlakatla, Alaska. Four languages: Coast Tsimshian (CT), Southern Tsimshian (ST); Nisg̱aʼa (Ni), Gitksan (Gi). The punctuation indicates subgrouping: Nisg̱aʼa and Gitksan are very close, the distinction being more political than linguistic. Southern Tsimshian (Sgu̎u̎x) is spoken by only a few people (in one family, as far as I know). By Upriver Tsimshian I mean the varieties of Coast Tsimshian spoken in Kitselas and (especially) Kitsumkalum, both near Terrace, BC. There has previously been no linguistic work on Upriver or Riverine Tsimshian as far as I know. All groups refer to their own language as Smʼalgyax (= `real language'). The family is not known to be related to any other in the world, except possibly Penutian (recent work by Marie-Lucie Tarpent).

Burden of this presentation: Outline of some facts about bound pronominals and `connectives' in Coast Tsimshian, hitherto (we believe) misunderstood or misanalyzed.

Moral: Morphological and phonological analysis is a necessary requisite for syntactic analysis (at least as much as vice versa).

Caveat: I believe that the findings reported here apply to all varieties of CT, but I may be wrong. There are many questions still to be answered. This presentation deals primarily with forms used in the current everyday language. A fuller set of connectives and other grammatical forms are found in the older or more formal style (see Boas, 1911; Mulder, 1995, and texts).

General characteristics of Smalgyaxian languages: Verb-initial but with a lot of grammatical and lexical stuff coming before verb. Compared to some other languages of the NW linguistic area that I know (principally Wakashan), Smalgyaxian languages make extensive use of `particles' (as opposed to affixes or other word-internal operations) both in phrase grammar and word grammar: examples below. Argument marking is Ergative-Absolutive (sometimes): discussion below. All of the languages make use of socalled `connectives' that glue together words in various constructions, suffixed to one word but controlled by characteristics of the following word or phrase.

Vocabulary for today's lesson, with orthographic / phonetic notes:

baa	run (singular)
hanaʼa(x)	woman	ʼ glottal stop x back (`uvular') velar spirant (x̱ / x̄)
ḵʼoł	run (plural)	ł voiceless l ḵ back k ([q])
łmoom	help	Cʼ / ʼC glottalized C (stop)
noo	mother
ʼnu̎u̎m	we, us	u̎ high back unrounded vowel
ol	bear (black Ursus americanus)
smʼooygit	chief, man of high rank
waap	house
ʼyuuta	man (male)	ʼR glottalized R (sonorant)

Three initial examples:

1. Yagwa̱n łmoomsm. I am helping you guys.
   
yagwa̱ -n- łmoom- -sm
PROG 1s-ERG help 2p-ABS

2. Ndm al łmoomsm. I am going to (will) help you guys.
   
-n- dm al łmoom -sm
1s-ERG FUT emph help 2p-ABS

3. Yagwa̱ baayu. I am running.
   
yagwa̱ baa -(y)u
PROG run 1s-ABS

A central distinction: Determinate vs. Nondeterminate nominals:

Determinate are personal names, independent pronouns, some close-kin terms: Meli, Dzon, ʼnu̎u̎m `we, us', nooyu `my mother'

Nondeterminate are all others (thus in particular all non-kin common noun nominals): hanaʼa(x) `woman,' ʼyuuta `man,' ol `bear,' waap `house,' smʼooygit `chief'

1.1 Bound Pronominals: Forms

(Plain) Absolutives:

-(y)u (1s) `me, I' (-yu after vowels)
-n (2s) `you'
-m (1p) `us, we'
-sm (2p) `you guys'
-t (3) `him, her, it, them; he, she, they'


Ergatives:

-n- (1s) `I'
-m- (2s) `you'
dip (1p) `we'
-m- -sm (2p) `you guys'
-t- (3) `he, she, it, they'


Marked Absolutives (term from Dunn, 1979):

-ʼnu (1s) `me'
-ʼnm (1p) `us'
-ʼnsm (2p) `you guys'


(the remainder are not distinguished, i.e.:

-n (2s) `you'
-t (3) `him, her, it, them, he, she, they')

Independent or free pronoun forms:

ʼnu̎u̎yu (1s) `I, me' (u̎ is a high back unrounded vowel)
ʼnu̎u̎n (2s) `you'
ʼnu̎u̎m (1p) `we, us'
ʼnu̎u̎sm (2p) `you guys'
ʼniit (3s) `him, her, it, he, she'


Note: these free pronoun forms are all treated as determinate nominals with 3rd person linking.

Notation: S: Subject of intransitive; A : Agent (subject) of transitive; O : Object of transitive (Dixon, 1979); Vi : intransitive verb; Vt : transitive verb; DN : Determinate Nominal; NN : Nondeterminate Nominal; C : Connective

1.2 Bound Pronominals: Functions:

Plain Absolutives:

Possessives: (na) Possessee-ABS(pronominal):

nawaabu `my house'
nawaabn `your house' etc.


If the possessor is expressed by a nominal, then Connective + Nominal in place of the ABSolutive suffix (see below).

Subject (sometimes): Yagwa̱ baayu. `I am running' [ baa : `run' singular subject]
...baan you...
...baat she/he/it...
...ḵʼołm we... [ ḵʼoł : `run' plural subject]
...ḵʼołsm you guys...
...ḵʼołt they...


Object (sometimes):

4. Yagwa̱n łmoomn. I am helping you.
5. Yagwa̱ dip łmoomsm. We are helping you guys.
6. Ndm łmoomsm. I'll help you guys.

Agent (sometimes):

7. Dm łmoomu. I'll help her/him/it.
Note: no overt form for 3 Object here.


Ergatives:

Agent (sometimes): Examples (12 -- 14)

Marked Absolutives: Subjects (sometimes), Objects (sometimes).

8. Dm baaʼnu. I'll run.
9. Dmt łmoomʼnsm. She/he/it will help you guys.

SO: one main question: what means `sometimes'?

1.3 Connectives: Forms: (partial story!)

Set 1: Before Absolutive Nominals:

Nondeterminate Connective (NC): -a (-ZERO after vowels and sonorants)
Determinate Connective (DC): -s ( -is / -as)

Functions: EXACTLY AS ABOVE for plain absolutives: so these are ABSOLUTIVE markers. Examples:

10. nawaaba ʼyuuta the man's house (Possessive)
11. Yagwa̱ baas Dzon. John is running.(Subject)
12. Yagwa̱n łmooms Meli. I'm helping Mary.(Object)
13. Dm łmooms Meli. Mary'll help him/her/it.
(compare Example (15): No overt 3 affix.)

So what is the -t on this example?

14. Dm łmoomt. She'll help him. (etc.)
Answer: it represents the Agent!

Set 2: Elsewhere (sometimes): Nondeterminate: same as above: -a

Determinate: -t ( -it / -at):

Functions: At least: where Nominals are used in sentences and are not absolutives. Examples:

15. Dm łmoomdit Meli. She's gonna help Mary.
16. Dm łmooms Melit Dolati. Mary'll help Dorothy.
17. Dm łmoomut Meli. I'll help Mary.
18. Dm łmoomu smʼooygit. I'll help the chief.
19. Dm łmooms Matt-a ʼyuuta. Matt will help the man.

(to understand orthography: voiceless stops are voiced before vowels, and final -x is replaced by the corresponding stop: -g̱ [note: x in CT means back or uvular x̱ ([x̩])]

2. Additions, Elsewheres, Sometimeses:

2.1 Additions: There are more forms for bound pronominals:

20. Yagwa̱m łmoomdu. You are helping me.
21. Yagwa̱m łmoomdm. You are helping us.
22. Mdm łmoomdu. You'll help me.
23. Mdm łmoomdm. You'll help us.
24. Dm łmoomdn. You'll help him/her/it.

Hang on to these forms, as they are important to the argument below.

2.2 Elsewheres:

In some sentences with conditionals, negatives, and questions, we find a different Nondeterminate Connective -ł. This connective occurs also in other places (e.g. with conjunctions like di- and (older) g̱an-.

2.3 Sometimeses:

SO what principles govern the choice among these various options: a large part of the not very extensive literature on CT is devoted to this question: Boas, 1911; Dunn, 1979a/b/c; Mulder, 1994 (and 1988). There are two domains of relevance:

Mode: choice of tense, subordinating conjunctions, etc. Boas uses the terms `indicative' and `subjunctive' (followed by Mulder). Dunn tries to establish a `continuum' from `marked' to `less marked' temporal items and conjunctions. We are fairly sure about two modes, which I call noncommittally yagwa̱-pattern and dm-pattern. (The latter comes roughly under Boas's indicative and Dunn's marked.)

Person to person: a kind of hierarchy (Mulder especially pursues this idea) reminiscent of rankings of persons etc. seen in other languages, but there does not seem to be a consistent single ranking in Tsimshian. The following principles seem to work:

I. In the dm-pattern with an intransitive verb, the Subject is expressed with a marked absolutive form:

25. Yagwa̱ baayu. I am running.
26. Dm baaʼnu. I'm gonna run.

II. In the dm-pattern, if the Object is a third person, then the Agent is expressed as a Plain absolutive:

27. Ndm łmoomn. I'm gonna help you.
28. Dm łmoomu. I'm gonna help her.
29. Dm łmoomnt Meli. You're gonna help Mary.

III. In a transitive sentence, if the Object is a first person form, then a marked absolutive or one of the forms -du -dm is used. Exact distribution of choices and optionality still open to investigation.

30. Yagwa̱m łmoomdu. You are helping me.
31. Dmt łmoomʼnm. He/she/it'll help us.

IV. Full nominal expressions, except for Absolutives, are best understood as linked to pronominal affixes on the verb. (Cf. Tarpent on Nisgha, but the `except' clause does not hold there). So it is best to start with bound pronoun only examples and then see how to expand them into sentences with independent nominals. Examples (subscripts show linking):

32. Yagwa̱t łmoomt. She is helping him.
33. Yagwa̱t łmoomd_i ʼyuuta_i. She is helping the man.
34. Yagwa̱t_i łmoomd_jit Meli_i. Mary is helping him.
35. Yagwa̱t_i łmoomd_jit Meli_it Dzon_j. Mary is helping John.
36. Dm łmoomt. She will help him.
37. Dm łmooms Meli. Mary will help him. [ABS Argument / ZERO Object!]
38. Dm łmoomda ʼyuuta. She will help the man. [idem.!]
39. Dm łmooms Melit Dzon. Mary will help John.

3.0 Alternatives:

Dunn and Mulder treat the sequences -dit , -da as unanalysed alternative `connectives.' In all these cases, it seems as if these items are combinations of the 3 affix plus a connective (with epenthesis).

Dunn and Mulder analyse the special forms for first person affixes -du and -dm as sequences of 3 Object and 1s and 1p A. Examples above seem to show this is wrong.

Boas, 1911, seems to come closest to stating the correct generalization about use of absolutives for Agent: when the Object is third person. Boas's statements about the connectives are confusing because of a failure to separate clearly the formal notions of ERGATIVE and ABSOLUTIVE affixation and direct linking of nominal arguments from the functional notions expressed in Dixon's terminology of Agent, Subject, Object.

4.0 Nonfinal words:

There is a a lot to do still: there are too many -t suffixes, and conditions for deletions and epenthesis have not been worked out in enough detail (cf. Tarpent's work on Nisg̱aʼa).

Acknowledgements.

This report represents ongoing and joint work with Mildred Roberts of Kitsumkalum. We are grateful to members of our classes, UNBC FNST 137 and 138, for feedback, and to Morris Mason (Kitselas) in addition for contributing his knowledge of his language.

References.

Boas, Franz. 1912 Tsimshian Texts. New Series. Publications of the American Ethnological Society 3.

Boas, Franz. 1916. Tsimshian Mythology. Thirty-first Annual Report of the Bureau of American Ethnology.

Boas, Franz. [n.d.] Tsimshian Texts. Smithsonian Institution: Bureau of American Ethnology, Bulletin 27. Reprinted by Native American Book Publishers, 5884 Winans Lake Road, Brighton, Michigan, 48116 USA [actually Nisgha texts, collected at "Kinkolith"]

Dixon, R. M. W. 1979. Ergativity. Language 55: 59--138.

Dunn, John Asher. 1978. A Practical Dictionary of the Coast Tsimshian Language. Ottawa: National Museum of Canada. Canadian Ethnology Service. Paper No. 42. (CTs)

Dunn, John Asher. 1979a. A Reference Grammar for the Coast Tsimshian Language. Ottawa: National Museum of Canada. Canadian Ethnology Service. Paper No. 55.

Dunn, John A. 1979b. Tsimshian connectives. IJAL 45:131--40.

Dunn, John A. 1979c. Pronominal concord in Coast Tsimshian. IJAL 45:224--31.

Dunn, John Asher. 1995. Smʼalgax / A Reference Dictionary and Grammar of the Coast Tsimshian Language. Seattle and London: University of Washington Press / Juneau: Sealaska Heritage Foundation. [Reprint of Dunn, 1978 and 1979.]

Mulder, Jean Gail. 1994. Ergativity in Coast Tsimshian (Sm'alga̱x). Berkeley / Los Angeles / London: University of California Press. (= University of California Publications: LINGUISTICS Volume 124) [revision of 1988 UCLA dissertation with same title)

Sasama, Fumiko. 1995. Coast Tsimshian Plural Formation with Phonological Introduction. [MA Thesis ?? UNBC]

Tarpent, Marie-Lucie. 1987. A grammar of the Nisgha language. Ph.D. dissertation: University of Victoria.