User:James Crippen/Tlingit verb
From Wikipedia, the free encyclopedia
The morphology and syntax of the Tlingit verb is far more complex than for the noun. The verb is comprised of a root and classifier, thematic affixes, valence or transitivity affixes, aspectual affixes, derivational affixes, and inflectional affixes. The affixes are regularly modified to fit phonological restrictions which can result in certain affixes not being realized, or being transformed into non-syllabic morphemes. These phonological restrictions create the sense that the verb morphology is fusional, but analysis shows that the verb morphemes are always decomposable into a agglutinative unrealized substrate.
[edit] Overview of verb morphology
The Tlingit verb can be analyzed according to a sort of slot-and-filler (or “position class”) template morphology, a template of affixes which are attached to the verb root in specific positions to produce individual verb conjugations. Story and Naish (1966) developed the initial descriptions of affixes in the Tlingit verb, then revised them in their verb dictionary (1972). Subsequently, Leer (1991) developed a far more extensive description which he presented in his dissertation. This article builds upon the analysis of verb affixes given by Cable in his “A metrical analysis of syncope in Tlingit” (2004).
[edit] Affix slot numbering
The table below gives rough correspondences between the different template schemata, but note that some of the slots were originally given in different orders than shown. This accounts for the seemingly peculiar numbering of the Naish and Story versions.
-
+15 +14 +13 +12 +11 +10 +9 +8 +7 +6 +5 +4 +3 +2 +1 0 -1 -2 -3 -4 -5 Leer 1991 +8 +7 +6 +5 +4 +3 +2 +1 0 -1 -2 -3 -4 -5 N/S 1966 3 2 1 50 60 40 30 20 10 STEM 110 120 N/S 1972 1 2 3 5 4 6 pf 7 8 9 10 11 12
Not all of the above affix slots are orthographically represented as part of the verb; in particular the first five are usually written as separate words. However their appearance and form are always obligate with certain verbs and in many cases there is no clear phonological word boundary between them, thus they must be analyzed as integral parts of the verb complex.
[edit] Slot names
Within each of these verb affix schemata there are accompanying names or descriptions. The list below gives the descriptions for the slot numbers as used in this article.
-
+15 adverbial proclitics +14 reciprocal woosh +13 3rd person subject number agreement +12 incorporated object pronouns +11 incorporated alienable nouns +10 incorporated inalienable nouns +9 theme prefix ya +8 theme prefix ka +7 outer aspect conjugations +6 irrealis +5 inner aspect conjugations +4 aspect +3 distributive +2 incorporated subject pronouns +1 classifier 0 root -1 derivation -2 duration -3 inner modes -4 outer modes -5 epimodes
Naish and Story originally attempted to name individual affixes rather than slots as is done in many other agglutinative languages. Leer largely abandoned this practice in favor of generalized slot names. The reasoning for this is due to the complex interdependencies between various affixes. For example, although the +5 na- is identifiable as a progressive aspect prefix, it has a number of different functions depending on the use of other affixes in the verb. As such it is not always an indicator of progressive aspect, so the name is somewhat misleading. Applying names to only the slots avoids this confusion, however it makes it difficult to discuss affixes individually and complicates glossing of them. In this article we attempt a compromise, providing names for the slots as well as names for individual verb affixes with the caveat that they are not always indicative of the actual function of the given affix.
[edit] Leer’s names
The following naming system is that used by Leer in his dissertation The Schetic Categories of the Tlingit Verb (1991). It should be noted that Leer’s terminology varies in his other works on Tlingit, however none of them are as comprehensive (or as difficult) as his dissertation. Consequently his terminological variations which appear elsewhere have been ignored here.
-
+8 proclitic adjunct phrases +7 number prefixes +6 incorporated object pronominals and incorporated alienable nouns +5 incorporated inalienable nouns +4 schetic prefixes +3 distributive prefix +2 subject pronominals +1 classifiers 0 root -1 root suffix -2 durative suffixes -3 inner modal suffixes -4 outer modal suffixes -5 epimodal suffixes
[edit] Naish and Story’s names
As noted previously, Naish and Story in their 1966 dissertations preferred names for each individual affix and simply described the slots in terms of generalized groups. However in their Tlingit Verb Dictionary (1972) they gave names and orders to each slot as well as revised names for particular affixes. The table below lists only those used in the verb dictionary.
-
1–3 theme prefixes 4 irrealis 5 continuative 6 aspectival 7 distributive 8 subject 9 extensors (i.e. classifiers) 10 stem (i.e. verb root) 11 consonantal 12 functional
[edit] Valence
All verbs belong exclusively to one of four valence classes. The class can be determined by the incorporated pronouns in the verb, i.e. whether the verb contains subject pronouns, object pronouns, neither, or both. Valence does not have a distinct morphological representation, instead it is an implicit quality of a verb based on the combination of its root and other prefixes.
The four classes are transitive, intransitive, objective (stative per Story and Naish), and impersonal. The organization of these classes is visualized in the following table.
-
-Object +Object -Subject impersonal objective +Subject intransitive transitive
In the examples given below a null morpheme (e.g. 0.SUBJ “zero subject”) is used to indicate the lack of an object or subject pronoun, however this should not be taken as arguing for the actual existence of these morphemes. Instead they are merely a notational convenience to clearly show where a pronoun would otherwise be expected.
[edit] Transitive Valence
A transitive verb has both subject and object.
-
ixhsixhán i-xha-si-xhaan 2SG.OBJ-1SG.SUBJ-(s, -D, +I)-love “I love you”
[edit] Intransitive Valence
An intransitive verb has a subject but no object.
-
yaa nxhaagút yaa=0-na-xha-ÿa-goot along=0.OBJ-PROG-1SG.SUBJ-(0, -D, +I)-go “I am going along”
[edit] Objective Valence
An objective verb has an object but no subject. In Tlingit these verbs may correspond to English verbs in which the subject is a dummy pronoun it, or to verbs where the Tlingit object would be represented as an English subject. An example of the latter is the verb ÿa-k'éi “to be good”.
-
xhat yak'éi xhat-0-ÿa-k'ei 1SG.OBJ-0.SUBJ-(0, -D, +I)-good “I'm good”, “I'm okay”
We could describe this verb formally with a template, thus OBJ-(0, -D, +I)-k'ei. This signifies that an object is required in the conjugation of this verb.
Leer notes a subtype of objective valence verbs in which the indefinite human subject pronoun du- has been lexicalized. In these verbs the subject pronoun has no syntactic role. This also occurs in the impersonal valence with a corresponding subtype. These are called the du-objective and du-impersonal, respectively.
-
wuduwaníkw ÿu-du-ÿa-neekw PERF-INDH.SUBJ-(0, -D, +I)-blow “The wind blew”, “The wind blows”
[edit] Impersonal Valence
An impersonal verb has neither subject nor object. These are often comparable to English weather verbs which take a dummy subject pronoun it, e.g.:
-
awdigaan a-wu-0-di-gaan 3NEU.OBJ-PERF-0.SUBJ-(0, +D, +I)-shine “The sun is/has been shining”, “It's sunny”
In this example the third person neutral prefix a- fills the slot of an incorporated object pronoun, however it does not represent an object per se. Instead a- is a sort of thematic (i.e. syntactically meaningless) prefix, which along with khu- is indicative of weather phenomena. Compare this with another impersonal verb khu-ÿa-k'éi “to be good weather”:
-
khuwak'éi khu-0-ÿa-k'ei IDFH.OBJ-0.SUBJ-(0, -D, +I)-good “The weather is good”
-
seighán kei khukghwak'éi seighán kei=khu-[ga-u-gha]-0-ÿa-k'ei tomorrow up=IDFH.OBJ-[ASCN-IRR-DESCN]-0.SUBJ-(0,-D,+I)-good “It will be good weather tomorrow”
[edit] Verb prefix slots
The verb prefix slots are those to the left of the verb root, numbered positively. Together they make up the prefix complex of the verb and account for most of its morphology. The slots are presented below in increasing order, that is from nearest the verb root to the leftward. This roughly accords with their semantic importance, that is the tendency for prefixes nearer the verb root to have more significant meanings in the conjugation than the prefixes further from the root.
[edit] Classifier (+1)
The classifier is a characteristic verb prefix of all Na-Dene languages and is well preserved in Tlingit. It shows the tripartite structure also found in Eyak and reconstructed in Proto-Athabaskan.
Story and Naish (1966; 1973) used the term “extensor” to describe this morpheme. Such usage is inconsistent with both prior and concurrent work on other Athabascan languages, hence linguists have returned to Sapir’s and Goddard’s original term.
The Tlingit classifier is not exclusively a classifier affix. Instead it is a verb prefix which indicates a variety of verb qualities such as valency, argument structure, aspect, and agreement with other verb affixes in addition to its true classificatory uses such as indicating the physical shape or internal complexity of the verb’s arguments. The Tlingit classifier could be described as the zenith of the portmanteau morpheme because it encodes a very complicated and somewhat variable set of qualities in a single morpheme.
Although it is more complex than a mere classificatory prefix, the original name given by Sapir and Goddard, inspired specifically by its unproductive use in classifying physical qualities of verb arguments, has stuck. Indeed, it was extended by them from their work on Tlingit to describe the similar and at the time hypothetically related prefixes they found in the Athabaskan languages. However in those languages this prefix has mostly lost its classificatory function, leading to Athabaskanists to disparage the term despite its continuing use. The preponderance of work on Athabaskan languages in comparison to work on Tlingit has led to the use of a sort of rhetorical trope among the Athabaskan linguistic community where the term is introduced and then immediately derided for its seeming inappropriateness. This is apparently perpetuated out of ignorance with the Tlingit classifier prefix and its historical relationship with the Athabaskan languages.
[edit] Analysis and notation
Although the classifier is not further decomposable, there is some syntactic and semantic symmetry in its use as well as some phonological regularity among the various instances. The classifier can thus be analyzed as a mathematical function of three variables, notationally CL : (S × D × I) → C.
An instance of a classifier is notationally represented as an ordered triple (S, D, I). If the value of a particular dimension is variable then it is shown in the triple as an unadorned capital letter, as given previously. Otherwise the dimension is given with its value, thus the classifier ya- is notated (0, -D, +I).
The three dimensional classifier space C is given in the following table.
-
-D +D -I +I -I +I S 0 0- ya- da- di- l la- li- l- dli- s sa- si- s- dzi- sh sha- shi- sh- ji-
The S-dimension – what Leer (1991) calls the “series component” – has four values which roughly represent the nucleus consonants of the classifier, represented by the set S = {0, l, s, sh}. A particular row of the classifier space is called a “series” following Leer’s usage, thus the “0 series” or the “sh series” for example.
The D-dimension approximates the existence of the consonant <d> in the classifier except for certain irregularities. The D-dimension is notationally represented as D = {+D, -D}. The two classifiers l- (l, +D, -I), s- (s, +D, -I) and sh- (sh, +D, -I) are irregular where we might have predicted *dl-, *dz-, and *j-. This phenomenon may be explainable by appealing to the phonological complexity of a consonant cluster with no following vowel, thus the actual classifiers are simply easier to say than the predicted ones. The ji- (sh, +D, +I) classifier is not irregular, it can be considered to arise from <d> + <sh> + <i>. The da- classifier is somewhat abnormal since it has the vowel <a> rather than no vowel, again this may be due to practical difficulty.
The I-dimension determines the existence of the vowel <i> in the classifier. If its value is positive then the classifier has <i> as its vowel, but if it is negative then the classifier either has <a> or no vowel. The ya- classifier appears to be irregular, but it can be considered as a reduction of the vowel <i> to the related consonant <y> with a neutral vowel inserted to maintain syllabic rhythm, i.e. initially [ɪ] > [j] then [j] + [ə] > [jə]. This particular classifier ya-, perhaps the most frequently used, varies in pronunciation between a carefully spoken [jʌ] to a very reduced [j] heard in rapid speech.
In verb theme templates the classifier is usually underspecified and thus has variable values. In such cases a variable dimension is notated by the uppercase letter ‘S’, ‘D’, or ‘I’. For example a classifier with the S-dimension of 0 and the D-dimension of +D but an undetermined I-dimension might be represented as (0, +D, I). This template could resolve to either of the da- or di- classifiers, and similarly the classifier template (S, D, -I) could resolve to either sa- or s-.
For those who have some difficulty interpreting the three variables of the classifier space as given notationally, an exhaustive list of rules from the triples to the classifiers is given below.
- 0 series
- (0, -D, -I) → 0-
- (0, -D, +I) → ya-
- (0, +D, -I) → da-
- (0, +D, +I) → di-
- l series
- (l, -D, -I) → la-
- (l, -D, +I) → li-
- (l, +D, -I) → l-
- (l, +D, +I) → dli-
- s series
- (s, -D, -I) → sa-
- (s, -D, +I) → si-
- (s, +D, -I) → s-
- (s, +D, +I) → dzi-
- sh series
- (sh, -D, -I) → sha-
- (sh, -D, +I) → shi-
- (sh, +D, -I) → sh-
- (sh, +D, +I) → ji-
[edit] Usage
[edit] I-dimension
The I-dimension is realized as +I in all perfectives, the realizational, stative imperfective, conative imperfective, the "yoo#[I]g" imperfective, and both the irrealis and realis assertive submodes of the potential mode.
[edit] D-dimension
The D-dimension is realized as +D in a number of different forms.
- thematic element in intransitive
- thematic element in transitive OBJ-SUBJ-(0, +D, I)-naa “SUBJ drink OBJ”.
- required with indefinite human subject pronouns “du” except with 0 series classifiers
- with reflexive object pronominal
- with reciprocal object pronominal when coreferential with the subject (Leer 1991, p. 97)
- antipassive (S, +D, I)-ROOT
- self-benefactive ga-(S, +D, I)-ROOT
- plural verb suffix (S, +D, I)-ROOT-x'
- locomotive ka-(S, +D, I)-ROOT “moving by/while doing”
- dissimulative sh-kh'é-(l, +D, I)-ROOT “pretending to do”
- certain nonproductive instances
- OBJ-(S, +D, I)-ROOT-x' “OBJ has plural noun”
- OBJ-SUBJ-(S, +D, I)-ROOT-(y)ákhw “undoing, removing”
- ash-ka-PERF-SUBJ-(l, +D, I)-ROOT-aa “SUBJ plays at N/V-ing”
[edit] Allomorphy
Some classifiers have allomorphs depending on certain surrounding phonological features. In the case of ya- this is caused by the spread of rounding from round vowels and consonants in the preceding syllable, producing wa-. This is a result of the historical change of the ÿ phoneme to y and w.
Characteristic of Southern Tlingit speech is the loss of the <i> vowel in most phonological contexts of the classifier. Thus si-tee is realized in Southern Tlingit as stee rather than the Northern sitee. This vowel loss is sometimes cited by Northern speakers as an explanation for why Southern speech seems more rapid.
[edit] Story & Naish
In this section we will present a summary of the notation and analysis of the classifier which appears in Story’s and Naish’s works on Tlingit under the term “extensor”.
[edit] Leer
This section summarizes Leer’s notation of the classifier.
[edit] Incorporated subject pronouns (+2)
-
Prefix Gloss xha- ~ xh- 1SG.SUBJ too- ~ tu- 1PL.SUBJ ee- ~ i- 2SG.SUBJ yee- ~ yi- ~ y- 2PL.SUBJ 0- 3NEU.SUBJ du- IDFH.SUBJ
[edit] Distributive (+3)
-
dagha- ~ daxh- DIST
The distributive prefix slot has only one value, dagha-. There is some variation with the 3SUBJ number agreement prefix daxh- at +13. Naish and Story called these two prefixes together the “distributive prefix”.
The +3 distributive prefix refers to the referent of the absolutive argument, i.e. the object of a transitive or objective verb, and the subject of a intransitive verb. It indicates that the referent is composed of multiple members or parts which exist or act separately. Thus it can apply to a group of people considered as a unit, a collection of objects, a series of events, etc.
This prefix is usually used with affirmative stative imperfectives, the +13 daxh- prefix being used elsewhere. There is a preference for dagha- in less complex verbal prefix strings, and daxh- is used with more complex prefix strings. This is probably associated with their positions, daxh- being further from the root and hence less likely to collide or interfere with the phonological processes applied to the other inner prefixes.
It should be noted that daxh- is never used with the 1SUBJ and 2SUBJ prefixes.
There is a regular contraction of dagha- to daxh- before all classifiers with CV shape other than (0, -D, +I) ÿa-.
[edit] Perfective (+4)
-
ÿu- PERF gha- APERF
[edit] Inner aspect conjugations (+5)
-
Prefix Gloss 0- TEL na- PROG gha- DESC
[edit] Irrealis (+6)
-
u- ~ w- ~ 0- IRR
The irrealis in Tlingit is morphologically complex, consisting of the verb prefix as well as constraints on the classifier and various stem changes for open or closed verb roots. There is no corresponding realis affix per se, instead the realis is merely a lack of irrealis in the verb. It should be noted that the irrealis is not in itself a mood in Tlingit, it is rather an integral component of other aspects and modes, however it is clearly associated with the idea of counterfactual “unreality” as in other languages.
The irrealis prefix is normally represented as u- although it has various phonological realizations associated with labialization or the close back vowel <u>.
Appearance of the irrealis prefix is associated with -I in the classifier except in the potential and attributive verb forms. Together these two features may be considered the “grammatical” irrealis form.
The irrealis prefix cannot cooccur with any of the subject agreement prefixes that have close vowels. These prefixes are 2SG i-, 2PL yee-, 1PL tu-, and INDH du-. Possibly related to this close vowel constraint imposed by the subject agreement prefixes are the close vowel constraints from the perfective prefix ÿu- and the future prefix cluster ga-u-gha-. In the case of these two latter constraints the irrealis is said to be “zeroed” or “phonologically overridden” rather than missing because the -I classifier constraint and the root changes imposed by the irrealis are still present.
The irrealis often requires the negative particle l. This is seen as a proclitic l= in dependent clauses. It also appears in the word tléil or tlél which is proposed as a contraction of tléik' “no” and l “NEG”. This tléil is occasaionally pronounced as héil or hél, but not usually written as such.
In forming simple negative statements the irrealis is required. Compare the following two utterances:
-
yéi yatee yéi=ya-tee thus=(0, -D, +I)-be “It is”, “It is so”
-
tléil yéi utee tléi-l yéi=u-0-tee no-NEG thus=IRR-(0, -D, -I)-be “It is not”, “It is not so”
Note the appearance of both the irrealis prefix and the change of classifier to -I in the negative utterance. In comparison the perfective negative statement below has a zero irrealis, but the -I conditioned by it remains apparent.
-
awsikóo a-ÿu-si-koo 3NEU.OBJ-PERF-(s, -D, +I)-know “(he) knew it”
-
tléil awuskú tléi-l a-0-ÿu-sa-koo no-NEG 3NEU.OBJ-IRR-PERF-(s, -D, -I)-know “(he) did not know it”
[edit] Outer aspect conjugations (+7)
-
ga- ASC
[edit] Theme prefix ka- (+8)
-
ka- KTHM
[edit] Theme prefix ÿa- (+9)
-
ÿa- YTHM
[edit] Incorporated inalienable nouns (+10)
-
Prefix Gloss Derivation ji- hand < jín “hand” xh'a- ~ kh'a mouth < xh'é “mouth” tu- inside, mind < tú “inside, mind” sha- head < shá “head” lu- nose < lú “nose” se- voice < sé “voice” sa- ? xha- fur < xhaaw “fur” shu- end < shú “end, tip” gu- base < gú “base, butt” ta- bottom < tá “bottom” daa- surrounding < daa “surrounding, periphery” xhoo- midst < xhoo “midst” t'aa- landward < t'áa(k) “landward side” gin- tail < geen “tail flipper” s'an- limb < *s'aan “limb” (obs.) lidíxh'- neck < lidíxh' “neck” wakh-ka- eye-KTHM < waakh “eye” s'akh-ka- bone-KTHM < s'aakh “bone” xh'us-ka- foot-KTHM < xh'oos “foot, leg” s'ee-ka- eyebrow-KTHM < s'ee “eyebrow” s'ukh-ka- rib-KTHM < s'óokh “rib” duk-ka- skin-KTHM < dook “skin” laka- mouth < laká “inside of mouth” tl'ikh-sha- finger-tip < tl'eikh “finger” kiÿ-sha- knee-tip < keeÿ “knee” xh'a-tu- mouth-inside < xh'é “mouth” tukh-xh'e- ~ tukh'e- anus < tookh “anus” daa.it- joint < daa.it “joint”
[edit] Incorporated alienable nouns (+11)
-
Prefix Gloss Derivation ÿaan- hunger < ÿaan “hunger” shakux- thirst < shakoox “thirst” ÿa-ta- sleep < tá “sleep” xh'asakw- breath < xh'aséikw “breath, life” ghaxh- crying < ghaaxh “crying, weeping” xei- ~ xee- shadow < *xei “dusk, shadow” (obs.) khei- ~ khee- dawn < *khei “dawn” (obs.) ÿei- ~ ÿee- time < *ÿee “time” (obs.) kanik- report < kaneek “report, news” ÿakha- reproach < ÿakhá “reproach, curse” saa- name < saa “name” aan- land < aan “land, village” naa- clan < naa “clan, nation” sha.axhw- bundle < sha.aaxhw “bundle” yakw- boat < yaakw “canoe, boat, vessel” hin- water < héen “water, river” tux'- knee < *tóox' “knee” (obs.)
[edit] Incorporated object pronouns (+12)
-
Prefix Gloss xhat- ~ axh- 1SG.OBJ haa- 1PL.OBJ i- 2SG.OBJ yee- 2PL.OBJ a- 3NEU.OBJ ash- 3SAL.OBJ sh- RFLX.OBJ khu- ~ khaa- IDFH.OBJ at- IDFN.OBJ aa- PART.OBJ
[edit] 3rd person subject number agreement (+13)
-
has- ~ s- 3PL.SUBJ
[edit] Reciprocal (+14)
-
Prefix Gloss woosh- ~ wooch- ~ chush- RECIP daxh- DIST
The distributive prefix daxh- is normally positioned at +3, but is occasionally found at +14.
[edit] Adverbial proclitics (+15)
The adverbial proclitics are numerous with fairly context-dependent meanings, and as such cannot be easily abbreviated in glosses. The best option is usually to offer a rough English translation in the form of an appropriate adverbial phrase or clause. Although all the adverbial proclitics are placed in a single slot, they can be organized thematically into different categories.
-
Manner yéi~yei thus, so yóo thus Thematic ÿaa mental state or activity Directionals yoo back and forth, to and fro, up and down ÿaa along, down, obliquely kéi up yéi down ÿeekh~ÿeikh~eekh down to beach or shore daakh up from beach or shore, away from open, off of fire or heat daak out to sea, out into open, onto fire or heat, (falling) downwards Invariable locationals khut astray, getting lost ÿaxh to completion yuxh outside yaaxh into a canoe, boat, or other vehicle héenxh into water uxh + kei out of control, blindly khwáaghxh + daakh by mistake, wrongly yetxh~yedaxh starting off, taking off ÿanaxh~ÿaanaxh underground ghunaÿéi~ghunéi beginning aa there áa + ÿaxh turning over shóo + ÿaxh turning over endwise héeni into water gági from hiding into the open éeghi from woods onto beach or shore dáaghi from water onto beach or shore Variable locationals ÿan~ÿaxh~ÿándei ashore, onto ground, to rest neil~neilxh~neildéi inside, into a house, home haat~haaxh~haadéi hither, toward speaker yóot~yóoxh~yóodei thither, away, to an indefinite location khuxh~0~khúxhdei (returning) back kux~kuxxh~kúxdei aground, into shallow water
The manner proclitic yéi “thus, so” is the least marked (i.e. most common) of the adverbial proclitics, and in some verb themes has become semantically empty, functioning merely as an additional thematic element . This is best seen in the verb theme N-x' yéi=OBJ-(0,d,i)-tee “OBJ is at N”, as in the following phrase.
-
Sheet'kax' yéi xhat yatee Sheet'ka-x' yéi=xhat-ÿa-tee Sitka-LOC thus=1SG.OBJ-(0,-D,+I)-be “I am in Sitka”.
[edit] Verb suffix slots
The right verb slots are those which occur to the right of the verb root and are negatively numbered. Taken together they compose the suffix complex of the verb, and are less frequently used.
[edit] Derivational Suffixes (-1)
-
-: -ÿ -ee -aa ... ...
