Eung-Do COOK

The University of Calgary

1. It has been assumed, implicitly or explicitly, by many linguists and philosophers that every language has a verb (or two) which is equivalent to the English verb be in its meaning and grammatical function. The lexico-semantic as well as grammatical features of this verb have been considered most interesting not only to the linguist but also to the scholars of many related disciplines such as philosophy and logic. Verhaar (1967), editing a series of studies entitled The verb "be" and its synonyms states the objectives of the series as follows:

"These studies will provide some of the necessary foundational material for research in logic, the theory of knowledge and ontology; and possibly in other philosophical disciplines".

In Sarcee, as in other Athapaskan languages, there are two verb roots which are comparable to English be, namely t'a (imperfective) 'to be' and lh (imperfective) 'to become, to exist'[1]. In this paper, I shall deal primarily with the former, since it is this verb which is comparable in its grammatical function to the so-called 'copula' of many well-known languages.

2. The verb stem t'a belongs to what I call 'the minor paradigm', as it inflects only for the imperfective and the perfective, whereas verbs of 'the major paradigm' inflect for the continuative and/or the iterative as well as the imperfective and the perfective (Cook 1972). The inflection of the verb stem t'a is marked phonologically by tonal alternation: the mid tone for the imperfective (t'a) and the low tone for the perfective (t')[2].

A verb root of an Athapaskan language is preceded by a number of prefixes which include the so-called 'classifiers', namely s (< ), l[3], d, and , pronominals (subject or object), postpositions, aspects, modes, themes, and adverbials. The function of the classifier is opaque in many cases, and it is perhaps reasonable to consider it grammatically nonfunctional as it has become 'lexicalized' (see Section 3). The distinction between a thematic prefix and an adverbial prefix is not always clear. However, it seems reasonable to consider that a verb must contain at least the following prefixes, although some may be phonologically zero, in that order:

(1) mode - asp(ect) - sub(ject) - c(lassifier) root[4].

The simplest sentences with the root t'a include the following examples

(2) а. nst'a a - ni - s - - t'a 'I am so'

adv asp subj c root

b. nt'a a - ni - ni - - t'a 'you are so'

adv asp subj c root

c. t'i a - - - - t'a 'he is so'

adv asp subj c root she


The analyses of the above forms are straightforward except the second (2b) where the subject morpheme ni 'you' is deleted causing the tonal sandhi on the preceding syllable; namely the aspect morpheme ni which happens to be homophonous with the subject morpheme. The constraints which govern the deletion of the subject morpheme ni and other prefixes are quite complicated, and shall not be the concern of this paper[5].

The forms presented in (2) may occur alone as an utterance in a proper context. Assuming that each of the above Forms is a sentence, it is reasonable to consider that an underlying constituent which is a complement has been deleted. In fact, such an assumption is substantiated by such sentences as;

(3) а. sn nst'а 'it's mе'

I аm

b. nn nt'а 'it's уоu'

you you are

c. t t'a 'it's water'

water it is

The noun phrase constituent of the above sentences, namely sn, nn, and t, are no doubt complements which might originate from an underlying sentence in a more abstract analysis. Incidentally, the sentences in (3) are not the so-called 'equational' type.

Consider the following sentences:

(4) а. tst'n stih 'I am a Sarcee'

b. tsgsn stih 'I am a soldier'

c. xkdj stih 'I am a chief'

The verb stih 'I am' has the following constituent structure:

i - s - - liN

peg subj c root

The element i, which Sapir called 'peg', is essentially epenthetic (Cook 1971c), the morpheme s 'I' is the subject, is a classifier, аnd the verb root is underlyingly liN. The underlying sequence of s and l is realized by the surface [t], and the underlying nasal becomes [n8] by a rule of inflection, and is finally realized by [h] in the imperfective and by [n] in the perfective (see Cook 1972).

Although the sentences of (3) and those of (4) are structurally parallel on the surface, it is quite obvious that only the sentences of (4) are comparable to the equational type of English and many better known languages. Furthermore, it is very interesting to note that it is only in a sentence of (3), whose verb root is -t'a, where a sentence followed. by the vocalic suffix -a may substitute the complement noun phrase as exemplified by the following sentences:

(5) a. dich 'I will go'

b. dich t'a 'I will really go'

(6) a. ycdjn 'he sang'

b. ycdjna t'a 'he really sang'

(7) a. tc'dinst'a 'I am holy'

b. tc'dinst'a t'a 'I am holy indeed'

This vocalic suffix which follows an embedded sentence in the (b) sentence is a nominalizer (one might consider this a complementizer) whose underlying form i is assimilated to the following vowel or gets deleted after a vowel. Therefore, the underlying constituent structure of (5b) can be represented as follows:

[[[dich] S2 [i] N] NP [t'a] VP] S1

What makes the above analysis particularly interesting is that the complex construction illustrated by the (b) sentences of (5)-(7) is strikingly similar to those of English, Japanese, Korean, and Chinese, among others, where the verb of the matrix sentence is the copula in each case:

(8) English

а. John walks

b. [It is John [whо walks]S2]S1

(9) Japanese

а. Watakusi wа аrukimasu 'I walk'

b. [[Watakusi wа aruku]S2 nо desu]S1 'It is the case (true) that I walk'

(10) Korean

а. Nа nun ketnunta 'I walk'

b. [Nа nun ketnun] S2 kes ita]S1 'it is the case that I walk'

(11) Chinese

а. Wŏ b q 'I am not going'

b. [Wŏ sh [b q] S2 ]S1 'it is not me who is going'

Details aside, one can observe common syntactic and semantic properties in the above complex sentence constructions. In each pair of sentences above, the (a) sentence is embedded as S2 wich is nominalized in the (b) sentence, and the verb of each matrix sentence (S1) is a copula. Semantically, in each pair, (a) is a 'colorless' statement, whereas (b) is a 'colored' one with a certain part of the sentence being 'emphasized' or 'focussed'.

3. The syntactic properties of the verb t'a seem to reflect some sort of universal or near universal feature of the copula. Then, what about the semantic or lexical properties of the verb? On the basis of what has been presented so far, one could say that the root t'a is semantically 'empty' like the copula of many other languages including English. However, consider the following examples:

(12) a. k'ts'nt'a 'everyone will be there'

b. k'nst'a 'we all will be there'

c k'anst'a 'we were all there'

(13) a. kdit' 'he's seriously ill'

b.       kdnt'a 'I'm seriously ill'

c.       kdint' 'you are seriously ill'

(14) a. tc'dinst'a 'I am holy'

b. tc'diyist' 'I was holy'

c. tc'dist'a 'I will be holy'

The sentences given in (12)-(14) are structurally the same in that every form contains a thematic or adverbial prefix, namely k' in (12), k in (13), and tc' in (14). All the other prefixes are pronominals and aspects which are irrelevant to the 'lexical' meaning of the sentence.

Two questions might arise immediately in the minds of those who are not familiar with the morphology of the Athapaskan languages. First of all, how does one know that the root in (12) - (14) is the copula t'a? In other words, are they not different verb roots which happen to be homophonous with the copula?

It is a well-known fact that in the Athapaskan languages the meaning of a verb root is modified by an adverbial prefix, an incorporated noun, or what is often called a 'thematic' prefix, to such an extent that meanings so modified do not reflect any transparent semantic relationships. For example, in Chipewyan the root θer has the basic meaning 'to do, to act' which forms a theme 'to stay' together with prefix n, and the root -/a 'to handle a round solid object' forms a theme 'to own' together with the same thematic prefix n (Li 1944). In Sarcee the verb root ti 'to handle a long object' forms a theme 'to carry' with prefix na, a second theme 'to put into' with prefix ti. In view of this lexical process, it is not difficult to see that the root t'a forms different themes with different prefixes as exemplified by the exemples in (12)-(14). Having provided this explanation for the lexical process and the morphological structure of the sentences in (12)-(14), one might be satisfied that the job of a linguist has been finished. However, this explanation would not satisfy a more inquisitive mind. An obvious question would be what does this lexical process imply? That is, why are the particular themes, namely 'to be there' (12), 'to be seriously ill' (13), 'to be holy' (14), and possibly three others (which will be discussed shortly as they exhibit somewhat different morphological structures) built upon the verb t'a? Perhaps, this is more of a philosophical question rather than a linguistic one.

The exemples presented in (15)-(17) have a different morphological structure than those examined above in that each contains a pronominal prefix:

(15) a. gk'ts'nist'a 'I am pregnant'

b. gk'ats't'a 'she is pregnant'

(16) a. saγnst'a 'have pity on me!'

b. nγst'a 'I pity you'

(17) a. snst'a 'I am like you'

b. mnst'a 'you are like me'

Stences (16) and (17) are structurally identical in that each contains what might be called a 'pоstpositional phrase' consisting of an object pronominal prefix followed by a postposition (pp). The structural analyses of (15) and (16) are given below as (15a) and (16a) respectively.

(15а) gu - k'а - ts'а - а - ni - s - - t'а 'I аm pregnant'

obj рр ? adv asp subj с root

(16a) si - γa - a - ni - - t'a 'Have pity on me!'

obj pp adv subj c root

In (15a), the indefinite pronominal gu plus the postposition k'a constitutes pоstpositional phrase, while in (16a) the first person pronominal (obj) si (the vowel of which assimilating to a of γa) plus the postposition γa constitutes a pоstpositional phrase. The meanings of the postpositions are difficult to define; k'a may be translated as 'on top of' and γa as 'towards'. Furthermore, the element ts'a in (15a) is yet to be identified. It is not clear whether one can postulate a semantic rule that will account for the respective meaning which is, as it were, built upon the meaning of the verb root t'a or the forms with different meanings are fossilized idioms. In any case, an extremely interesting question is: How are these concepts related, if the two root forms (15) and (16) as well as those in (12)-(14) are not the case of true homonyms?

A second question that might be asked with respect to the forms presented in (12)-(14) is : if the root t'a is a copula, is the thematic prefix or other elements) in the prefix complex not adjectival which bears the basic lexical meaning of each sentence? The answer is 'no'. In Sarcee, like in other Athapaskan languages, there is no 'true' adjective distinct from verbs. In other words, there are verb roots which are semantically comparable to English adjectives, but their morphosyntactic behavior is not different from that of the other verbs. The following exemples illustrate the point:

(18) a. tyinsl 'you are hot'

b. dk'z 'it is red'

c. ssdj 'I am old'

The roots sl (a), k'z (b), and dj (c) mean 'to be hot', 'to be red', and 'to be old' respectively. Clearly, the root t'a has nothing to do with the 'adjectival' expression in sentences (12)-(14).

The forms in (17) are structurally simpler than those in (16) in that each has a pronominal prefix preceding nst'a which is the simplest form of the sentence as presented in (2) except that the classifier preceding the root is s rather than . For example, sentence (17a) has the following underlying structure:

si - a - ni - ni - s - t'a

obj adv asp subj c root

The subject ni 'you' is deleted raising the tone on the preceding syllabe, namely the aspect morpheme ni. As mentioned at the beginning, it is very difficult to pin down the function of the classifiers in Sarcee as well as in other Athapaskan languages, but it is generally recognized that the classifier s (<*) adds the force of transitivity to the verb. If this is the case, the meaning of the stem t'a plus the classifier s in (17a) can be translated literally as 'to resemble'. In other words, the meaning 'to resemble' is derivable from the meaning of t'a 'to be' and that of the classifier s.

4. In what precedes I have provided an account of purely 'linguistic' analysis, in its narrow sense, of the forme whose verb root is t'a. The semantic differences among the forms analyzed in this paper are explained in terms of different prefixes associated with the root, namely thematic and adverbial prefixes, classifiers, and postpositional phrases. These accounts of formal structures, however, do not provide for any explanation of how the different concepts are derivable from, or related to, the meaning 'to be'. In Section 2, I examined grammatical characteristics of the verb t'a in complex sentence structures. Perhaps, one could conclude that the Sarcee copula t'a compares well with copulas of other languages in its grammatical properties, while the lexico-semantic properties reflect something unique in the Sarcee language and culture. This conclusion in itself is not unexpected, but the details of the grammatical and lexico-semantic features which have been analyzed in this paper will provide bases for further study. The remaining question which I cannot answer (and perhaps many ordinary grammarians are not even interested in answering) is why are only a half a dozen concepts, such as 'to be holy', 'to be seriously ill', 'to resemble', etc. expressed by a verb phrase whose root is the copula in Sarcee?


COOK, E.-D. 1971a. Vowels and tones in Sarcee. Language 47.164-179.

COOK, E.-D. 1971b. Morphophonemics of two Sarcee classifiers. International journal of. American Linguistica 38.231-233.

Cook, E.-D. 1971c. Phonological constraint and syntactic rule. Linguistic Inquiry 2.465-478.

COOK, E.-D. 1972. Sarcee Verb Paradigma. Mercury Series N 2. Ottawa: National Museum of Man.

HOIJER, H. 1946. Introduction. In: Linguistic Structures of Native America. Viking Fund Publications in Anthropology N 6.

HOIJER, H. 1971. Athapaskan morphology. In: Studies in American Indian Languages, ed. by Jesse Sawyer. University of California Publications in Linguistics 65. Berkeley and Los Angeles University of California Press.

HYMES, D. 1956. Na-Dene and positional analysis of categories. American Anthropologist 58.624-38.

LANDAR, H. 1967. Two Athapaskan verbs of 'being'. In: The verb 'Be' and its Synonyms, Part I, ed. by John W.M. Verhaar. Foundations of Language Supplementary Series Volume 1. Dordrecht, Holland : D. Reidel Publishing Co.

LI, F.-K. 1930. A study of Sarcee verb stems. International Journal of American Linguistics 6.3-27.

SAPIR, 1925. Pitch accent in Sarcee, an Athapaskan Language. Journal de la Socit des Amricanistes de Paris, n.s., l7.185-205.

VERHAAR, J.W.M. 1967. Editorial preface. In: The Verb 'Be' and its Synonyms. Foundation of Language Supplementary Series Volume 1. Dordrecht, Holland: D.Reidel Publishing Co.

* This paper was originally schedulded to appear in the Journal de la Socit des Amricanistes de Paris (Vol. 62), but due to difficulties in the publication schedule of the Journal, the editors and the author have agreed to publish this paper in this review.

Sarcee, which is probably most closely related to Sekani and Beaver (see Holier 1946), is one of the northern Athapaskan languages still spoken by a handful of elderly people living on a reserve bordering the South-West boundary of Calgary, Alberta, Canada.

[1] Landar (1967) compares these two verb roots of several Athapaskan languages under the title 'Two Athapaskan verbs of "being"'. Unfortunately, however, he does not present any significant syntactic or semantic properties of these verbs. What one finds in his article is a general morphosyntactic description of the Athapaskan languages, which one does not anticipate from the title.

[2] The transcription used in this paper is self-explanatory except the following: b, d, g, dz, etc. are invoiced and unaspirated whereas p, t, k, ts, etc. are unvoiced and aspirated; c and j in digraphs and elsewhere are equal to and respectively; o is phonetically [ə̣] which is "velarized and dark-timbered" (Sapir 1925); geminate vowels with different tones each (e.g. , ) are realized phonetically as a long vowel with an inflected tone (e.g. , ǐ), whereas geminate vowels with a same tone on each (e.g. , ) are realized as an over long vowel (e.g. .a, ..i); high and low tones are marked by and ` respectively, and mid tone is unmarked. Forms with no fixed tone are cited in the text without any tone mark (see Cook 1971a for further details).

[3] This classifier is always realized on the phonetic surface as a or i; see Cook (1971b) for a phonological account of this classifier.

[4] See Li (1930) and Cook (1971e) for some details of the prefix morphology of Sarcee, Hymen (1956) for a comparative study of Athapaskan verb prefixes, and Hoijer (1971) for a broader description of Athapaskan morphology.

[5] See Cook (1971c) for some discussion which has significant bearing upon the notion of 'global rule' in generative grammar.