<<

Chris Donlay LING 236: Advanced Language Change Final Paper November 30, 2009

Demonstratives and Their Reflexes: A Study in from Eyak

The has long been cited as prime material for grammaticalization. Its very nature makes prone to further evolution. Because it is deictic, its meaning is general and use frequent – two common preconditions for grammaticalization (Hopper and Traugott 2003). This flexibility allows it to appear in a variety of constructions, which then leads to the development of polysemy and perhaps morphological and phonological change. And indeed often follow this path. In a cross-linguistic study of 85 languages, Diessel (1999) shows may grammaticalize into a number of other functions, from personal and to , copulas and focus markers. No language will develop all of these uses, but most will develop some. And given the importance of the independent demonstrative, its original use remains even as it evolves other functions. This phenomenon, in which older and newer uses coexist within a language, is called layering in grammaticalization (Hopper and Traugott 2003:124-126). While Diessel instantiates these many functions with examples from a variety of languages, does not focus on the resultant layering that exists in each. To that end, in this paper I analyze the grammaticalization of demonstratives in a single Native American language: Eyak, a relative of the Athabaskan family. Despite the fact that there are no historical records for the language, Eyak provides an ideal laboratory for this kind of study. Its demonstratives have clearly evolved in function but not to the point where morphological and phonological change have rendered them unrecognizable. As a result, it is possible to identify the demonstrative origins of certain pronouns, articles, copulas, focus markers and discourse markers in the language. This transparency also highlights two findings not discussed by Diessel: a dual demonstrative that simultaneously refers to two referents, and discourse markers that do not originate from adverbial demonstratives. Donlay, Demonstratives in Eyak, p. 2

Eyak

Eyak is an extinct language formerly spoken by communities lived on the Gulf of Alaska, roughly between the towns of Yakutat and Cordova (Krauss 1982). The language descends from the Athabaskan-Eyak family, which is believed to have existed approximately three millennia ago (Krauss 1982). Eyak split from Proto-Athabaskan at some point before the latter came into being about two thousand years ago. It is therefore a distant relative of Slave, the lone Athabaskan language included in Diessel’s study. Syntactically, Eyak is similar to its Athabaskan relatives. It is a -final language with complex verbal morphology. The verb structure consists of a stem and a series of templatic affix positions, both prefixes, suffixes and enclitics. All other elements in a clause occur before the verb and order among them is flexible. Third person subjects are canonically not marked on the verb (Krauss 1965). The demonstrative system in Eyak is a simple one, containing four :

al nonhuman proximal aw nonhuman distal/unmarked anh human singular ahnu∙ human

They are further described in the next section. The previously-mentioned dual demonstrative, unhaw, only exists in the focus construction and is discussed in that section. Eyak has not yet been fully documented. There is no comprehensive grammar, but a dictionary (Krauss 1970a) and annotated folktales (Krauss 1970b, Krauss 1982) have been published. The data used in this paper are gathered from the later work. And although the structure of narratives favor third-person constructions, there is enough reported speech in the tales to see how personal deictics function in the language. The spelling system used in the narratives is shown in Table 1 alongside their IPA equivalents, though only the former is used in this paper. The vowels in Eyak are [i, e, ǝ, a, u]; the shwa is transcribed as a. Long vowels are indicated with a raised dot (e.g. i∙ [i:]). Vowels may also be nasalized, which is shown in the transcription by n following the vowel (Krauss Donlay, Demonstratives in Eyak, p. 3

1970b, Krauss 1982). Interlinear glosses and literal translations are provided by me unless otherwise noted1.

Table 1: Consonant inventory of Eyak

Plain d [t] dl [tɬ] dz [ʦ] j [ʧ] g [g] gw [gʷ] g [q] Aspirated t [tʰ] tl [tɬʰ] ts [ʦʰ] ch [ʧʰ] k [kʰ] q [qʰ]

Glottalized t’ [t’] tl’ [tɬ’] ts’ [ʦ’] ch’ [ʧ’] k’ [k’] q’ [q’]

Fricative ɫ [ɬ] s [s] sh [ʃ] x [x] xw [xʷ] x [χ] Sonorant w [w] m [m] l [l] n [n] y [j]

Glottal h [h] ’ [Ɂ]

(adapted from Krauss 1970b, Krauss 1982)

Defining a Demonstrative

The origin of demonstratives is a controversial . The process of grammaticalization rests on the idea that all grammatical items evolve from lexical expressions. Yet demonstratives, which clearly count as grammatical items, were until recently believed to be linguistic primitives. Diessel (1999) takes this view, suggesting that demonstratives may exist in language prior to the beginning of grammaticalization and, therefore, prior to any other grammatical form. More recently, Heine and Kuteva have shown that demonstratives may be derived from locative and even such as ‘go’ or ‘see’ (2007:84). However, this does not completely disprove the earlier view. The fact that a language evolves new demonstratives does not mean that there were none previously, nor that all demonstratives originate in this way. Indeed, the literature is full of examples of new constructions supplanting older synonymous ones (e.g. Hopper and Traugott 2003). Their fundamental usefulness to communication suggests that ‘demonstratives might form a class of deictic expressions that belong to the basic vocabulary of every language’ (Diessel 1999:8). The data from Eyak does not resolve this issue; the origin of its demonstratives remains unclear.

11 Abbreviations: CONT contrastive, COP , CUST customary, DEM demonstrative, DIST distal, DU dual, FILL filler, FOC focus, HUM human, NEG negative, OBJ object, PFV perfective, PL plural, POSS , PROG progressive, PROX proximal, REP repetitive, SG singular, SUBJ subject.

Donlay, Demonstratives in Eyak, p. 4

Despite these questions, Diessel’s study of demonstratives is instructive. For example, he takes a broad view in defining the demonstrative, highlighting both its syntactic and pragmatic uses:

They are primarily used to focus the hearer’s attention on objects or locations in the speech situation (often in combination with a pointing gesture), but they may also function to organize the information flow in the ongoing discourse. More specifically, demonstratives are often used to keep track of prior discourse participants and to activate specific shared knowledge. The most basic function of demonstratives is, however, to orient the hearer outside of discourse in the surrounding situation. (Diessel 1999:2)

In the study, Diessel examines demonstratives in terms of four parameters: morphology, semantics, and pragmatics. Each is discussed in turn, covering both Diessel’s generalizations and how they relate to the specifics of Eyak.

Morphology

Cross-linguistically, the morphology of demonstratives varies considerably (Diessel 1999:22). In some, demonstratives are invariable, in others they are highly inflected. Typically, they remain unbound, though cliticization is common. In Eyak, the four demonstratives are invariable and are largely unbound. They do take postpositions as suffixes, however. Unbound and bound demonstratives are shown in (1) and (2) respectively.

(1) Aw qit dasa’iɫinu∙, DEM.DIST downhill dump-PFV-3PL They [literally ‘those’, i.e. ravens] dumped them down, (Krauss 1982:51)

(2) Awq’ ya∙naht ate∙k’. DEM.DIST-on down.towards lie.down-CUST It would lie down on top of them [literally ‘on those’, i.e. ducks]. (Krauss 1982:49)

Semantics

Semantically, demonstratives typically contain two types of information: distance and reference (Diessel 1999:35-36). The former registers the relationship between the target item Donlay, Demonstratives in Eyak, p. 5 and the speaker’s deictic center, differentiating between, for example, proximal and distal. The latter encodes characteristics about the referent, which may include , gender and humanness. In Eyak, the demonstratives are semantically asymmetric. The two nonhuman demonstratives encode both distance and reference, but number is not differentiated. Al is proximal, meaning ‘this one’/‘these’ and aw is distal meaning ‘that one’/‘those’. These forms also encode basic parameters about the referent, which may be animate but nonhuman, inanimate, or propositional. Furthermore, the distal has also developed an unmarked meaning (Krauss, personal communication), a common cross-linguistic occurrence (Diessel 1999). These three uses are instantiated in the following examples.

(3) animate but nonhuman

aw daɫaxwe∙k DEM.DIST groundhog those groundhogs (Krauss 1982:112)

(4) inanimate

aw t’ik’ɫ sixa’ DEM.DIST bow-and-arrow 1SG-POSS my bow-and-arrow [literally, ‘that bow-and-arrow of me’] (Krauss 1982:85)

(5) propositional

Awleht q'uhnu∙ aw shaxwa∙shɫ. DEM.DIST-because.of FOC-3PL DEM.DIST butcher-PFV That's why [literally, ‘because of that’, i.e. previous events] they butchered it. (Krauss 1982:52)

The human demonstratives, on the other hand, only encode humanness and number, but not distance or gender. Anh refers to a singular human and ahnu∙ plural humans. They are also invariable and typically unbound, although they too may take postpositions as suffixes. Examples are shown below.

Donlay, Demonstratives in Eyak, p. 6

(6) Anh utl’ yi∙nhinh awa∙ uxaht q’e’ sdiyahɫ. DEM.HUM.SG 3P-with be-3SG however 3P-without back go-PFV His companion [literally, ‘that one (who) is with him’], however, had gone back without him. (Krauss 1982:110:7)

(7) Anhch’ k’ushiyah ula’x di∙’yahɫ. DEM.HUM.SG-with angry 3P-over become.angry-PFV is infuriated with him. (Krauss 1982:91)

Syntax

In terms of syntax, Diessel (1999:57-58) distinguishes between the syntactic context and categorial status of demonstratives. This means that if a demonstrative is used in different contexts it then belongs to different grammatical categories. The four possible combinations are shown in Table 2. The terminology is fairly self-explanatory: independent demonstratives that appear in argument positions are pronouns; those that co-occur with are determiners; those that modify verbs are adverbial; and those that appear in copular and nonverbal phrases are identifiers.

Table 2: Distribution and category of demonstratives Distribution Category Pronominal demonstrative Demonstrative pronoun Adnominal demonstrative Demonstrative Adverbial demonstrative Demonstrative Identificational demonstrative Demonstrative identifier (Diessel 1999:4)

As in most languages, the demonstratives in Eyak can be used as independent pronouns, as shown in (8) and (9), where they mean ‘this one’ or ‘those’. Indeed, this is likely their original function. Their use as demonstrative pronouns, determiners and identifiers is described in the next section along with other grammaticalized functions. Eyak does possess demonstrative adverbs, such as aant ‘here’ and uut ‘there’, but since they are unrelated to the four basic demonstratives they are excluded from the present analysis.

Donlay, Demonstratives in Eyak, p. 7

(8) Al, ux k’uqu’xɫshehyu∙, DEM.PROX hunting.things These, things I hunt with,

ich’ qu’xɫah, 2P-to give I’ll give them to , (Krauss 1982:42)

(9) Awya∙’t ɫtsinhtk, daax q’uhnu∙, ahnu∙, DEM.DIST-out toss-REP and FOC-3PL DEM.HUM.PL She was tossing them out, and then, they [literally ‘those ones’],

a∙ngala’a∙n’ ahnu∙ saɫxe’t∙ ɫ, river-to DEM.HUM.PL carry.on.shoulder-PFV with it on their shoulders they came to a river (Krauss 125:14)

Pragmatics

The primary purpose of demonstratives is to ‘orient the hearer in the speech situation, focusing his or her attention on objects, locations, or persons’ (Diessel 1999:93). However, they may also refer to conceptual ideas, propositions and even portions of discourse. Figure 1 shows the relationship between these uses.

Figure 1: Pragmatic uses of demonstratives

pragmatic uses

exophoric endophoric

anaphoric discourse deictic recognitional (Diessel 1999:6)

Exophoric demonstratives refer to concrete entities in a speech situation. Endophoric demonstratives, on the other hand, refer to elements within the speech itself. For example, anaphoric demonstratives point to phrases, discourse deictics refer to propositions, and Donlay, Demonstratives in Eyak, p. 8 recognitional demonstratives point to entities that belong to the shared knowledge of the interlocutors (Diessel 1999:93-112). In Eyak, all but the recognitional use is attested in the texts, although it is possible that the latter may have existed in everyday speech. An example of each is presented below.

(10) exophoric use

Al, ux k’uqu’xɫshehyu∙, DEM.PROX hunting.things These, things I hunt with,

ich’ qu’xɫah, 2P-to give I’ll give them to you, (Krauss 1982:42)

(11) anaphoric use

Ya∙kuts’k ɫe’t’ya’ch’, little box-into Into a little box,

aw iɫt’a’ ɫya∙k’. DEM.DIST gather-CUST they used to gather them. (Krauss 1982:54)

(12) discourse deictic use

Awyaxa∙q' q'al a∙nda' q'e' xsdiyahɫ. DEM.DIST-by.means.of FOC-PROX here-to back come-PFV That's [i.e. referring to previous anecdote] how I managed to get back here. (Krauss 1982:43)

In summary, then, the demonstrative paradigm in Eyak is relatively simple. There are four forms – al (proximal nonhuman), aw (distal nonhuman/unmarked), anh (human singular), and ahnu∙ (human plural) – which are invariable and unbound, except when suffixed by postpositions. Between them, they semantically differentiate distance, humanness and number, though not symmetrically. Syntactically, they may be used independently or in constructions; these other uses are described in the next section. They may refer exophorically and Donlay, Demonstratives in Eyak, p. 9 endophorically, and with regard to the latter, they may point to either noun phrases or propositions. In short, they exhibit the same properties as their counterparts in many other languages.

Grammaticalization of Demonstratives

Now that the basic function of Eyak demonstratives has been described, we can examine the additional uses that have grammaticalized in the language. In general, grammaticalization explains the evolution through which a lexical item takes on a grammatical use or a grammatical marker takes on additional grammatical functionality. This process is accomplished through common mechanisms of language change. Semantically, a word is extended to new contexts, which makes the meaning more general and eventually more abstract. Morphologically, an independent word may become a clitic and then a bound morpheme. Syntactically, the item becomes more rigidly constrained but also more frequent through more generalized use. And phonologically, reduction may occur, further differentiating the new use from the old (Hopper and Traugott 2003, Heine and Kuteva 2007). Diessel’s study describes the wide variety of functions that may evolve from demonstratives. Many of them occur in Eyak, such as pronouns, articles, copulas, focus markers and discourse markers. Diessel’s list, and the corresponding uses that exist in Eyak, are presented in Table 3. Each of the functions in Eyak are discussed below. Based on these functions and the general tendencies within grammaticalization, Diessel provides eight criteria that specifically address the evolution of demonstratives. These criteria, reproduced in Table 4 below, are used to evaluate the functions that evolved from Eyak’s demonstrative paradigm. Evaluating the independent demonstratives in Eyak against these criteria shows that they, too, have changed functionally over time. As the examples in the previous section attest, they have become semantically more general, though unchanged in syntax, morphology and phonology. Originally referring to concrete entities in the real world, they are now freely used to point to text-based arguments, such as subjects and objects. And, the nonhuman forms can also be used to refer to discourse passages, such as propositions or situations. These changes allowed

Donlay, Demonstratives in Eyak, p. 10

Table 3: Grammaticalized demonstratives, cross-linguistically and in Eyak Cross-Linguistic Uses Uses in Eyak

Third person pronouns Third person pronoun Relative pronouns Complementizers Verbal enclitic/ Sentence connectives Definite articles Definite Indefinite articles Noun class markers Linkers Determiners Number markers Copulas Copula Focus markers Focus marker Expletives Discourse markers

(adapted from Diessel 1999:115-150)

Table 4: Criteria for the grammaticalization of demonstratives Change Type Description 1) Grammatical items that developed from demonstratives are no longer used to Functional focus the hearer’s attention on entities in the outside world. 2) They are deictically non-contrastive.

3) Their occurrence is often restricted to a particular syntactic context. Syntactic 4) They are often obligatory to form a certain grammatical construction.

5) They are usually restricted to the distal or, less frequently, the proximal form. Morphological 6) They may have lost their ability to inflect.

7) They may have undergone a process of phonological reduction. Phonological 8) They may have coalesced with other free forms. (Diessel 1999:118)

the demonstratives to be used more widely and thus likely set the stage for grammaticalization within specific constructions. Donlay, Demonstratives in Eyak, p. 11

Definite Articles

As shown above, demonstratives may be used in combination with a noun or noun phrase in Eyak. Diessel describes how this adnominal use may lead to the development of definite articles. Typically, demonstrative pronouns are used to refer to arguments in an emphatic or contrastive way. When they become obligatory and used for any kind of argument, they lose their marked meaning, becoming definite articles (Diessel 1999:128-129). This process is not complete in Eyak, but the trajectory is apparent. There is no article per se in the language. are often used where English would combine a noun with a definite or indefinite article, as (13) and (14) show. New arguments are often introduced in this way.

(13) Wax q’aw aw ɫma∙tk’, thus FOC-DIST DEM.DIST cook-CUST Thus he cooks them,

tse∙le∙ quh . octopus the octopus. (Krauss 1982:101)

(14) Łundiyahs, lah, mouse lo A mouse, lo,

k’ulax q’e’ isɫi’anhɫinu∙, gu∙dik. one-at more see-PFV-3PL again they saw another one, again. (Krauss 1982:40)

But, as already described, demonstratives can be freely used with noun phrases. Often the use is emphatic or, as in (15), constrastive. But at other times the noun phrase is neither new nor unexpected and so no emphatic purpose is evident. For example, (16) is from the end of a short passage that contains only two characters, a man and his child. Since they are given information, additional mentions can be accomplished with a bare noun. Yet the human singular demonstrative is used with ‘child’, even though the possessive adds further clarity. The emphasis here is pragmatically unnecessary.

Donlay, Demonstratives in Eyak, p. 12

(15) Łinh ih awa∙, adax, dik’ udagaleh k’a’ɫe∙q… one 3P.of however NEG sense have-NEG One of them, however, had no sense…

Anh ɫinh ih awa∙ adax k’udzu∙. DEM.HUM.SG one 3P.of however good The other one however was good. (Krauss 1982:67)

(16) Sasinhɫ, anh saqe∙ts’akihdik uxa’. die-PFV DEM.HUM.SG child-too 3P-POSS He died, and the child too with him. (Krauss 1982:117)

The frequency of these unemphatic demonstratives suggests that they are in the process of losing their marked meaning and merely serve to point to an active argument a definite article. Its optionality, however, suggests that it has not yet rigidified into an obligatory marker. As a result, the demonstrative’s meaning is likely ambiguous; speakers may or may not interpret it in an emphatic way. This kind of ambiguity is a necessary ingredient for reanalysis (Hopper and Traugott 2003:52), and it indicates that the grammaticalization process was not complete at the time the language died out.

Third person pronouns

Given their deictic nature, demonstratives are essentially a type of pronoun. Within a clause, they may be used to emphatically or contrastively highlight verbal arguments that are already given information. Over time, this use becomes less emphatic and the demonstrative may be reanalyzed as an obligatory third person pronoun, and even eventually becoming a verb agreement marker. As part of this process, the pronominal demonstrative changes from independent morpheme to clitic to affix (Diessel 1999:120). This process is underway in Eyak. First and second person subjects, as well as direct objects in any person, may be marked on the verb (Krauss 1965). But third person subjects are canonically zero, which may have led speakers to routinely use demonstrative pronouns to emphasize the subject. While not obligatory, this is a frequent pattern in the data. For example, (17) shows the human singular anh being used as a subject marker, and (16) shows the unmarked aw, here referring to a giant rat, doing the same. Donlay, Demonstratives in Eyak, p. 13

(17) Anh uch’axalya xda x aw saɫt’ak’ɫ. DEM.HUM.SG 3P-underarm DEM.DIST shot-PFV He shot it through its underarm. (Krauss 1982:85)

(18) Aw q’aw adix q’aw aw utl’ sha’a’ch’ɫ. DEM.DIST FOC-DIST inside FOC-DIST DEM.DIST 3P-with go-PFV Then it went inside with them. (Krauss 1982:48)

Direct objects may also be optionally instantiated by demonstrative pronouns. In (19), the unmarked aw refers to a whale that is killed, and in (20) anh refers to a woman who is kidnapped.

(19) uga∙kgayu∙’a∙ aw shashehɫ. 3P-uncles-for DEM.DIST kill-PFV they killed it for their uncles. (Krauss 1982:105)

(20) Aw tse∙le∙xq uh anh saɫlahɫ. DEM.DIST octopus DEM.HUM.SG abduct-PFV The octopus abducted her. (Krauss 1982:100)

The difference between the two uses, independent demonstrative versus demonstrative pronoun, is mainly pragmatic. In its clearest use, the former is used alone in a separate intonation unit and points to an entity outside the speech event. The latter takes an argument position within a clause and represents a noun phrase already mentioned. Semantically, the line is blurry since they both convey the meaning of ‘that one’ in English. And, further, there are no morphological or phonological differences. It is unclear how separate these functions were for native speakers. Interestingly, independent personal pronouns do exist in Eyak, but they are solely used in a contrastive identificational way. For example, in (21) aa, which means ‘she herself’, is used to distinguish the subject from the subject of the previous sentence.

(21) A∙ awa∙ xatl’ tsu∙tk’inh. 3P.CONT however night sleep-CUST-3SG She herself however would sleep at night. (Krauss 1982:126)

Donlay, Demonstratives in Eyak, p. 14

The replacement of pronouns with demonstratives is even more common with postpositional constructions. There is a third person object prefix u- that may be used with postpositions. It is semantically general and can refer to humans and nonhumans, both singular and plural; an example is shown in (22). However, it is much less common than demonstratives in these constructions. Examples (23) through (26) show how each of the four demonstratives combines with a postposition in Eyak.

(22) uleht dik’ k’u∙lishe∙k’q. 3P-because.of NEG anything-kill-CUST-NEG why you don’t kill anything. (Krauss 1982:110)

(23) Alt’a’x q’aw, DEM.PROX-behind FOC-DIST Behind this it is, (Krauss 1982:128)

(24) awxa’ adawiɫ idasaɫiɫ. DEM.DIST trouble have-PFV they made war over it. (Krauss 1982:53)

(25) anhxaht a∙k’inh. DEM.HUM.SG-without go-CUST-3SG She kept going off without him. (Krauss 1982:86)

(26) ahnu∙tl’ wax i∙t’inhinh. DEM.HUM.PL-with thus live-3SG He lived with them. (Krauss 1982:104)

The replacement of nouns or pronouns with demonstratives in Eyak follows the common grammaticalization channel noted by Diessel above. Their use has generalized to a broad pronominal use, filling any argument slot outside the verb complex. The sheer frequency of their use indicates that they have lost a great deal of their emphatic or contrastive meaning, although their optional nature suggests they may be slightly more emphatic than the zero use. Aside from the semantics, however, there are few other signs of grammaticalization – they remain morphologically unbound and phonologically unchanged.

Donlay, Demonstratives in Eyak, p. 15

Verbal enclitic / complementizer

Closely related to the function, but showing a higher degree of grammaticalization, is the verbal enclitic. In this use, the human demonstratives have become argument markers that attach to the verb. Most often they refer to subjects, as in (27) and (28), though they can also point to objects, as in (29), or even oblique arguments (Leer 1991; Krauss, p.c.). They are not obligatory, however; they are used whenever the speaker wishes to highlight the argument. There is no nonhuman counterpart.

(27) Ahnu∙ sasuhɫinh. DEM.HUM.PL kill-PFV-3SG She killed them. (Krauss 1982:71)

(28) La’tnu∙ ɫila∙’gayu∙ q’uhnu∙, iɫaduxinu∙. two men FOC-3PL trap-PROG-3PL.SUBJ Two men it was, they were trapping. (Krauss 1982:109)

(29) Aw sa’ehɫinh. DEM.DIST marry-PFV-3SG.OBJ It married her. (Krauss 1982 101:31)

Diessel does not expand on the relationship between third person pronouns and verb agreement markers, but he does cite Givon’s diachronic cline through which a demonstrative pronoun becames a third person pronoun, cliticizes and eventually affixes to the verb as an agreement marker (Givon in Diessel 1999:120). In Eyak, the verbal enclitics have not yet reached the final, obligatory form, but they are certainly more advanced in grammaticalization than the demonstrative pronouns. For example, morphological, syntactic and phonological changes are apparent. They can no longer stand on their own, but must be cliticized to the verb. Cliticization is also an indication of syntactic change, since they are no longer free to occur anywhere in the clause. Their vowel quality has changed, further differentiating them from the original form. And like the demonstrative pronouns, their semantic use is endophoric only. These changes demonstrate that they have grammaticalized further than the pronominal use. Krauss (p.c.) states that these verbal enclitics were originally complementizers. In other words, the form sasuhlinh originally meant ‘that one who killed’ rather than ‘she killed’. In other languages, where similar constructions arose, it was a result of two independent clauses Donlay, Demonstratives in Eyak, p. 16 merging, and in the process a demonstrative is interpreted as a complementizer (Diessel 1999:123-125). It is easy to see how this may have developed in Eyak as well. Clause chaining is quite common in the narratives as is the use of anti-topics following the verbal clause, as shown in (30). If an independent demonstrative were used instead of a noun phrase, as in (31), a change in intonation boundaries would clear the way for cliticization to the verb and reanalysis as an argument marker rather than an anti-topic, as shown in (32).

(30) Aw q’unhaw, q’e’ sdiyahɫ, anh ɫila∙’. then back go-PFV DEM.HUM.SG man Then, he went back, that man. (Krauss 1982:88)

(31) q’e’ sdiyahɫ, anh back go-PFV DEM.HUM.SG went back, that one

(32) sdiyahɫ, anh > sdiyahɫ anh > sdiyahɫinh go-PFV DEM.HUM.SG go-PFV-3SG went, that one > went that one > went he

Judging by the translations, the complementizer use appears to be rare in the narratives. The verbal enclitics are instead treated as personal pronouns. However, the structure of English seems to force a distinction that is not there for Eyak speakers. In several footnotes Krauss supplies literal translations that clearly show the complementizer role of the enclitic, as in (33) and (34).

(33) ma∙gadatl’alah saquhɫinu∙ lake-upper.end-around stay-PFV-3PL lake dwarves literally ‘they (who) stay/sit around (the upper end of) a lake’ (Krauss 1982:39)

(34) Di’wax q’aw ts’iyuh’e∙x yax iɫa’a∙nxinh q’a’anh, just FOC-DIST black.bear-for around look-PROG-3SG COP.3HUM.SG Just looking around he was for black bears, literally ‘he being one who was just/still going/walking around looking for black bear’ (Krauss 1982:45)

Donlay, Demonstratives in Eyak, p. 17

It is unclear whether native speakers viewed these as distinct constructions or as a single function with potentially ambiguous meaning. Either way, they show that the human demonstratives grammaticalized into a more specific use through rigidification of syntax and the attendant morphological and phonological change.

Copula

In a number of languages, nonverbal copulas are grammaticalized from demonstratives. The process involves two adjacent noun phrases that refer to the same referent, the second including a demonstrative. Over time the two phrases are seen as a single equational clause, and the demonstrative is reanalyzed as a copula. Diessel (1999:147) provides the following schematic to illustrate the process:

[NP] [DEMi NPi] > [NP COPi NPi]

In Eyak, the demonstrative origin of the copula construction is still transparent. The four copular forms consist of a prefix q’a’ and a corresponding demonstrative, as the examples in (35) through (38) attest. The origin of the prefix is unclear, though it is likely related to the focus marker q’ discussed in the next section (Krauss, p.c.).

(35) Xu∙ q'a'al. 1SG COP.PROX This is me. (Krauss 1982:130)

(36) K'uguda'luw ma∙ q'a'aw. big lake COP.DIST It was a big lake. (Krauss 1982:87)

(37) qe'ɫ q'a'anh, woman COP.3HUM.SG a woman it was, (Krauss 1982:124)

(38) Gayaqa∙ ga' daxunhyu∙ q'a'ahnu∙. 3PL-like people COP.3HUM.PL They're people like us. (Krauss 1982:137)

Donlay, Demonstratives in Eyak, p. 18

There is little doubt that the process Diessel describes is the grammaticalization channel that led to the development of copulas in Eyak. Two common discourse patterns likely aided the change. First, it is quite common to refer to the same noun phrase repeatedly in adjacent clauses, as shown in (39). Second, verbless sentences are also not uncommon, as in (40). This syntactic environment is amenable to the adjacent positioning of two noun phrases that are eventually reinterpreted as a copular construction.

(39) Tse∙le∙xquhyaquhyu∙ da’saɫxa’ ɫinh, octopus-young have-PFV-3SG She had octopus-young,

k'uyaquhyu∙, some-young some young,

la’dih uyaquh, two young two young,

tse∙le∙xquhyaquhyu∙ . octopus-young octopus-young. (Krauss 1982:104)

(40) Daɫaxwe∙k, groundhog groundhog

datli∙ daɫaxwe∙k. already groundhog already a groundhog. (Krauss 1982:114)

Focus markers

Demonstratives may also grammaticalize into focus markers. Diessel suggests this is an extension of the development of copulas. In his view, two independent clauses – a nonverbal copular clause and a separate verbal clause – are merged into one complex clause. The first is then reanalyzed as a focal argument of the following verb phrase (Diessel 1999:148-149). Donlay, Demonstratives in Eyak, p. 19

One can easily see how this happened in Eyak by looking at (41). We have already noted that clause chaining is frequent in the language. Here the subject of the sentence is introduced in the first clause, and clarified in the second with the copula. The third clause describes what the subject is doing; note that the verb is not marked for subject.

(41) Daxunh, person A person,

qe’ɫ q’a’anh, woman COP.3HUM.SG it’s a woman,

xi∙ t xalaqt aw giyahga ladaqt sadahɫ. yonder upland DEM.DIST water-above sit-PFV is sitting above the water yonder upland. (Krauss 1982:130)

Now consider (42), which describes the moment when an octopus turns into a human. The subject noun is the same as above, but it is followed by a focal marker rather than the copula. Just as Diessel describes, the new construction appears to combine two clauses into one sentence. Recall that because third person subjects and given objects may be zero marked, the verb alone can serve as an independent clause. The focus construction, too, can exist as a separate clause.

(42) Daxunh q'aw saɫe'ɫ. person FOC-DIST become-PFV He [literally ‘it’] had become a person. (Krauss 1982:104)

The morphology of the focus marker, however, complicates the matter. Like the copular construction, it consists of a prefix q’ attached to one of the four demonstratives: q’al, q’aw (~ q’uw), q’unh, quhnuu. The prefix here is slightly different from that of the copula q’a’. Its origin is unknown although, as mentioned, the two are undoubtedly related (Krauss, p.c.). Unlike the copula, however, the demonstrative is not necessarily coreferential with the focused element. In (42), for example the nonhuman aw refers to the subject ‘octopus’ rather than the human object ‘person’. Donlay, Demonstratives in Eyak, p. 20

Indeed, it seems that the q’ alone marks the focal element, which may be an argument of the verb, a locative or temporal phrase, or even a demonstrative pronoun. This q’ cannot stand alone however, it must always be followed by a demonstrative. Again, the latter may refer to any argument in the phrase whether core or oblique. It is believed that the human enclitics in this construction, unh and uhnuu, are allomorphs of the verbal enclitics inh and inuu (Leer 1991; Krauss, p.c.). In other words, the latter typically attach to the verb but when the focal construction is used they obligatorily attach to the q’ instead. Examples (43) through (45) demonstrate four versions of the focus construction.

(43) U∙dax q'aw ɫi'q' ya∙yu∙ uxa' awa∙ lagax'ah da∙x, then FOC-DIST all things 3P.POSS 3PL-for make.off when So then when I make off with all their things for them, (Krauss 1982:93)

(44) u∙dax q'unh idehdah awlah u'satahɫ. then FOC-3SG really DEM.DIST-about find-PFV then he really found out about it. (Krauss 1982:115)

(45) U∙dax q'uhnu∙ u∙ch' saqehɫ. then FOC-3PL there-to boat-PFV Then they went there by boat. (Krauss 1982:51)

There is a fifth enclitic that may attach to the focal q’, and that is the unusual dual unhaw (~ unhuw). This enclitic is a concatenation of the human and nonhuman demonstratives (Krauss, p.c.), and is used to refer to two separate referents. In actual use, these referents may either be the expected combination of human and nonhuman referents, or two human referents. Examples (46) and (47) demonstrate these uses. This enclitic exists only in this construction; there is no counterpart anywhere else in the language.

(46) human + nonhuman

U∙dax q’unhaw, then FOC-DU Then,

awxa’ wax i∙t’inhinh. DEM.DIST-with thus live-PFV.3SG she lived with him [literally ‘it’]. (Krauss 1982:101)

Donlay, Demonstratives in Eyak, p. 21

(47) human + human

anh q'unhuw anhlax isaɫ'anhɫ. DEM.HUM.SG FOC-DU DEM.HUM.SG-to see-PFV it was he she saw. (Krauss 1982:130)

Given this complexity, the grammaticalization channel is unclear. Two possible interpretations must be considered. In the first, the focus construction developed from the copula, a pattern that is cross-linguistically attested (Diessel 1999). In the beginning, the focal element and the following demonstrative were tightly linked as they are still in the copular construction. Over time, as more and more elements were used in the focal position, the demonstrative’s reference became decoupled from the focal element allowing it to point to any verbal argument. The development of the dual enclitic could only happen after this semantic decoupling took place. In the alternative view, the focus construction consists only of the q’ marker. For some reason – perhaps to differentiate it from the homonymous postposition q’ ‘on’ – this marker was not allowed to stand alone. The verbal enclitics, which were always free to refer to any argument, were fronted to morphologically and phonologically support the focus marker. The nonhuman demonstratives did the same, likely through analogy. Cross-linguistically, the second constituent position, commonly known as the Wackernagel position, is favored for the placement of clitics (Anderson 1993), and this may have played a role here. Or perhaps analogy was the cause, helping to create a construction that mirrors the copula. And again, the dual enclitic arose after the cliticized focus construction became the norm. At this juncture, there is not enough information to tilt the scale towards either explanation. Regardless of the channel followed, however, the focus construction clearly shows another grammaticalized use of the demonstratives in Eyak. Semantically, they refer without emphasis or contrast. They are both obligatory and restricted, since they must attach to the q’ marker if it is present in the clause. Morphological and phonological changes are also evident. The unmarked demonstrative has an allomorph in this construction that does not exist elsewhere, and the dual enclitic shows further development of this particular paradigm. Furthermore, the human enclitics unh and uhnuu exhibit a vowel change further removed from either the original demonstratives (anh, ahnuu) and the verbal enclitics (inh, inuu).

Donlay, Demonstratives in Eyak, p. 22

Discourse markers

As previously mentioned, in Eyak the focus construction may take a demonstrative as its focal element. Often, the unmarked demonstrative aw occurs in this way, and it typically refers to a previous proposition. However in two instances idioms have emerged that have taken on discourse functions: aw q’aw and aw q’unhaw. Diessel does not address discourse markers directly. But he does note that temporal adverbs, such as ‘then’ and ‘now’ which also have discourse uses in English, are often grammaticalized from adverbial demonstratives (Diessel 1999:139-140). In this regard, the Eyak discourse markers highlight a grammaticalization channel not found in Diessel’s cross-linguistic study. (A similar channel is cited by Heine and Kuteva (2007:212), however.) The idiom aw q’aw literally means ‘that’s that’ (Krauss 1970b:13). In this sense, the focal element refers to given information, just as its English counterpart does. Krauss (1970b:13) notes that it may be used as a discourse connective meaning ‘then’. This use, temporally separating two consecutive events, is the only function it serves in the narratives, where it is frequent. It can be argued that as a temporal marker the focal aw is no longer referential, a dramatic development for a deictic marker. And true to form, the enclitic after the q’ marker is referential, pointing to any argument in the clause. The examples (48) through (51) show four of the possible choices; aw q’al is not attested in the narratives.

(48) Aw q'aw anh qe'ɫ idah aw li∙k'. then-DIST DEM.HUM.SG woman good DEM.DIST prepare-CUST So then the woman would prepare it nicely. (Krauss 1982:74)

(49) Aw q'unh al ta∙sga ɫaht'a'x anh yax saɫtehɫ. then-3SG DEM.PROX belt-inside DEM.HUM.SG down put-PFV Then he stuck him down in under his belt. (Krauss 1982:41)

(50) Aw q’uhnu∙, q’e’ shdi’a’ch’ɫinu∙, then-3PL back went-PFV-3PL Then, they went back, (Krauss 1982:51)

(51) Aw q'unhuw q'e∙ya∙naht i∙nsditahɫ. then-DU back-down-towards move.head-PFV Then he put his head back down. (Krauss 1982:87)

Donlay, Demonstratives in Eyak, p. 23

In addition to serving as a temporal , aw q’unhaw has another use. When it appears in its own intonation unit, it serves as a discourse filler (Krauss, p.c.), an empty phrase used as a speaker plans the next meaningful utterance. Examples are presented in (52) and (53). Although translated as ‘then’, in this use the phrase is not referential, just like the fillers ‘well’ and ‘so’ in English.

(52) Aw q’unhaw, FILL Then,

sahɫinh. go-PFV-3SG she went off. (Krauss 1982:86)

(53) Aw q'unhuw, FILL Then,

q’e’ qu’daqeh. back boat they were going to boat back. (Krauss 1982:72)

These two constructions provide the most grammaticalized uses of demonstratives in the language. Syntactically, they are restricted to very specific constructions that are only used in certain ways, such as between verb phrases or in a separate intonation unit. Morphologically, only the unmarked (distal) form can sit in the focal position. And semantically, they are arguably non referential – surely the endpoint in the life of a deictic expression.

Conclusion

The varied functions of the demonstratives in Eyak provide a case study for how grammar emerges in a language. Their use in specific constructions gave rise to new functions, which in turn led to further evolution. The grammaticalization channels I outline for Eyak are schematized in Figure 2. (The two possible origins of the focus marker are indicated by dotted lines.) It is not surprising that the demonstrative’s frequency and flexibility in use create an environment ripe for grammaticalization. Indeed, the process in Eyak mirrors that of many other Donlay, Demonstratives in Eyak, p. 24

Figure 2: Grammaticalization of demonstratives in Eyak

demonstratives

definite article third person pronoun copula

verbal enclitic/complementizer

focus marker

discourse markers

languages, as Diessel’s (1999) cross-linguistic study demonstrates. The Eyak case shows how the resulting functions interrelate as well as how they productively coexist in the language. In addition, Eyak has provided evidence for two developments not predicted by Diessel. The first is the development of a dual demonstrative that simultaneously refers to two individual referents. The second is the creation of two discourse markers, one temporal and one filler, from nonadverbial demonstratives. Sadly, because of the language’s recent extinction, we will never know how these constructions might have evolved in the future.

Donlay, Demonstratives in Eyak, p. 25

References

Anderson, Stephen. 1993. Wackernagel’s revenge: clitics, morphology, and the syntax of second position. Language 69:68-98. Diessel, Holger. 1991. Demonstratives: Form, function and grammaticalization. Amsterdam: John Benjamins Publishing Company. Heine, Bernd and Tania Kuteva. 2007. The genesis of grammar: A reconstruction. Oxford: Oxford University Press. Hopper, Paul J. and Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press. Krauss, Michael E. 1965. Eyak: A preliminary report. Canadian Journal of Linguistics 10.167- 187. Krauss, Michael E. (ed.) 1970a. Eyak dictionary. University of Alaska and Massachusetts Institute of Technology. Krauss, Michael E. (ed.) 1970b. Eyak texts. University of Alaska and Massachusetts Institute of Technology. Krauss, Michael E. (ed.) 1982. In honor of Eyak: The art of Anna Nelson Harry. Fairbanks: Alaska Native Language Center. Leer, Jeff. 1991. Evidence for a northern Northwest Coast language area: promiscuous number marking and periphrastic possessive constructions in Haida, Eyak, and Aleut. International Journal of American Linguistics 57.158-93.