User:James Crippen/Tlingit grammar/Verb

The morphology and syntax of the Lingít veb is far more complex than for the noun. The verb is comprised of a root and classifier (theme per Leer), 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 synthetic unrealized substrate.

Verb morphology
The Lingít verb can be analyzed according to a sort of slot-and-filler morphology, a template of affixes which are attached to the verb root in various ways to produce individual verb conjugations. Story and Naish (1966) developed the initial descriptions of affixes in the Lingít verb, then revised them in their verb dictionary (1972). Subsequently, Leer (1991) developed a more extensive description which he presented in his dissertation. This article uses the analysis of verb affixes following Cable (2004) which is the most descriptive available.

Affix slot numbering
The table below gives rough correspondences between the different analytical 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 schemata.


 * {|class="wikitable"

!  !!+15!!+14!!+13!!+12!!+11!!+10!!+9!!+8!!+7!!+6!!+5!!+4!!+3!!+2!!+1!! 0 !!-1!!-2!!-3!!-4!!-5
 * - align=center
 * Leer 1991||+8 ||colspan=2|+7 ||colspan=2|+6 ||colspan=3|+5 ||colspan=4|+4 ||+3||+2||+1|| 0 ||-1||-2||-3||-4||-5
 * - align=center
 * N/S 1966||colspan=5| || 3 || 2 || 1 || 50 || 60 ||colspan=2|40 || 30 || 20 || 10 ||colspan=3|STEM||colspan=2|110||colspan=2|120
 * - align=center
 * N/S 1972||colspan=5| ||1 || 2 || 3 || 5 || 4 || 6 || pf || 7 || 8 ||  9 ||colspan=3|10  ||colspan=2|11 ||colspan=2|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.

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.




 * align=right|+15|| adverbial proclitics
 * align=right|+14|| reciprocal woosh
 * align=right|+13|| 3rd person subject number agreement
 * align=right|+12|| object agreement
 * align=right|+11|| unreduced incorporated nouns
 * align=right|+10|| reduced incorporated nouns
 * align=right|+9 || theme prefix ya
 * align=right|+8 || theme prefix ka
 * align=right|+7 || outer aspect conjugations
 * align=right|+6 || irrealis
 * align=right|+5 || inner aspect conjugations
 * align=right|+4 || aspect
 * align=right|+3 || distributive
 * align=right|+2 || subject agreement
 * align=right|+1 || classifier
 * align=right| 0 || root
 * align=right|-1 || derivational
 * align=right|-2 || durative
 * align=right|-3 || inner modes
 * align=right|-4 || outer modes
 * align=right|-5 || epimodes
 * }
 * align=right|+4 || aspect
 * align=right|+3 || distributive
 * align=right|+2 || subject agreement
 * align=right|+1 || classifier
 * align=right| 0 || root
 * align=right|-1 || derivational
 * align=right|-2 || durative
 * align=right|-3 || inner modes
 * align=right|-4 || outer modes
 * align=right|-5 || epimodes
 * }
 * align=right|-2 || durative
 * align=right|-3 || inner modes
 * align=right|-4 || outer modes
 * align=right|-5 || epimodes
 * }
 * align=right|-4 || outer modes
 * align=right|-5 || epimodes
 * }
 * }

Leer's names
The following naming system is that used in Leer (1991).




 * align=right|+8 || proclitic adjunct phrases
 * align=right|+7 || number prefixes
 * align=right|+6 || incorporated object pronominals and incorporated alienable nouns
 * align=right|+5 || incorporated inalienable nouns
 * align=right|+4 || schetic prefixes
 * align=right|+3 || distributive prefix
 * align=right|+2 || subject pronominals
 * align=right|+1 || classifiers
 * align=right|0 || root
 * align=right|-1 || root suffix
 * align=right|-2 || durative suffixes
 * align=right|-3 || inner modal suffixes
 * align=right|-4 || outer modal suffixes
 * align=right|-5 || epimodal suffixes
 * }
 * align=right|0 || root
 * align=right|-1 || root suffix
 * align=right|-2 || durative suffixes
 * align=right|-3 || inner modal suffixes
 * align=right|-4 || outer modal suffixes
 * align=right|-5 || epimodal suffixes
 * }
 * align=right|-3 || inner modal suffixes
 * align=right|-4 || outer modal suffixes
 * align=right|-5 || epimodal suffixes
 * }
 * align=right|-5 || epimodal suffixes
 * }

Naish and Story's names
Naish and Story in their 1966 dissertations preferred names for each individual affix and simply described the slots in terms of generalized groups. The names given below are as used in their verb dictionary (1972).




 * align=right| 1–3 || theme prefixes
 * align=right| 4 || irrealis
 * align=right| 5 || continuative
 * align=right| 6 || aspectival
 * align=right| 7 || distributive
 * align=right| 8 || subject
 * align=right| 9 || extensors (i.e. classifiers)
 * align=right| 10 || stem (i.e. verb root)
 * align=right| 11 || consonantal
 * align=right| 12 || functional
 * }
 * align=right| 9 || extensors (i.e. classifiers)
 * align=right| 10 || stem (i.e. verb root)
 * align=right| 11 || consonantal
 * align=right| 12 || functional
 * }
 * align=right| 11 || consonantal
 * align=right| 12 || functional
 * }
 * }

Valence
All verbs belong exclusively to one of four valence classes. The class can be determined by the incorporated pronominals in the verb, i.e. whether the verb contains slots for subject pronominals, object pronominals, 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.


 * {|class="wikitable"


 * ||-Object ||+Object
 * -Subject|| impersonal  || objective
 * +Subject|| intransitive || transitive
 * }
 * +Subject|| intransitive || transitive
 * }

Objective Valence
An objective verb has an object but no subject. In Lingít these verbs may correspond to English verbs in which the subject is a dummy it, or to verbs where the Tlingit object would be represented as an English subject. An example of the latter is the verb ya-k'éi “to be good”.




 * colspan=3| xhat yak'éi
 * xhat || ya- || -k'éi
 * 1 SG.OBJ || (0, -D, +I) || good
 * colspan=3|“I'm good”, “I'm okay”
 * }
 * colspan=3|“I'm good”, “I'm okay”
 * }
 * }

We could describe this verb formally with a template, thus  OBJ -(0, -D, +I)-k'éi. This signifies that an object is required in the conjugation of this verb.

Leer notes a subtype of objective valence verbs in which an indefinite human subject pronominal du- has been lexicalized. In these verbs the subject pronominal 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.




 * colspan=4|wuduwaníkw
 * wu- || -du- || -ya- || níkw
 * ? || INDEFH.SUBJ || (0, -D, +I) || blow
 * colspan=4|“The wind blows”, “The wind is blowing”
 * }
 * colspan=4|“The wind blows”, “The wind is blowing”
 * }
 * }

Impersonal Valence
An impersonal verb has neither subject nor object. These are often comparable to English weather verbs which take the dummy subject it, e.g.:




 * colspan=4|awdigaan
 * a- || -wu- || -di- || -gan
 * 3 NEU || PERF || (0, +D, +I) || shine
 * colspan=4|“The sun is/has been shining”, “It's sunny”
 * }
 * colspan=4|“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 pronominal, however it does not represent an object per se. Instead a- is a sort of thematic prefix, which along with khu- is indicative of weather phenomena. Compare this with another impersonal verb khu-ya-k'éi “to be good weather”:




 * colspan=3|khuwak'éi
 * khu- || -ya- || -k'éi
 * INDEFH . SUBJ || (0, -D, +I) || good
 * colspan=3|“The weather is good”
 * }
 * colspan=3|“The weather is good”
 * }
 * }




 * colspan=6|Seighán kei khukghwak'éi
 * seighán || kei || khu- || ga-u-gha-0- || -ya- || -k'éi
 * tomorrow || ADV || INDEFH . SUBJ || CONJ - IRR - ASP -0. SUBJ || (0, -D, +I) || good
 * colspan=6|“It will be good weather tomorrow”
 * }
 * colspan=6|“It will be good weather tomorrow”
 * }
 * }

Intransitive Valence
An intransitive verb has a subject but no object.




 * colspan=6| yaa nxhagút
 * yaa || 0- || -na- || -xha- || -0- || -gút
 * along- ADV || 0. OBJ || ASP || 1 SG.SUBJ || (0, -D, -I) || walk
 * colspan=6|“I am walking along”
 * }
 * colspan=6|“I am walking along”
 * }
 * }

Transitive Valence
A transitive verb has both subject and object.




 * colspan=4| ixhsixhán
 * i- || -xha- || -si- || -xhán
 * 2 SG.OBJ || 1 SG.SUBJ || (s, -D, +I) || love
 * colspan=4|“I love you”
 * }
 * colspan=4|“I love you”
 * }
 * }

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.

Classifiers (+1)
The classifier is a characteristic of all Na-Dene languages and is well preserved in Lingít. It shows the tripartite structure also found in Eyak and reconstructed in Proto-Athabaskan. The classifier’s structure being tripartite, it can be best conceived of as a cube (actually a sort of rectangular solid) where each portion is a dimension and taken together the coordinates in each dimension fix the realization of a classifier into a single point in the cube and thus a single form.

The three dimensions of the classifier are the I-dimension, the D-dimension, and the S-dimension. The S-dimension (Leer’s “series component”) has values in the set {0, l, s, sh}, the D-dimension is the set {+D, -D}, and the I-dimension is the set {+I, -I}. Coordinates in the three dimensions are given in the order (S, D, I). Below is the mapping from coordinates to classifiers.


 * (0, -D, -I) → 0-
 * (0, -D, +I) → ya-
 * (0, +D, -I) → da-
 * (0, +D, +I) → di-
 * (l, -D, -I) → la-
 * (l, -D, +I) → li-
 * (l, +D, -I) → l-
 * (l, +D, +I) → dli-
 * (s, -D, -I) → sa-
 * (s, -D, +I) → si-
 * (s, +D, -I) → s-
 * (s, +D, +I) → dzi- (< d- + s- → dz-)
 * (sh, -D, -I) → sha-
 * (sh, -D, +I) → shi-
 * (sh, +D, -I) → sh-
 * (sh, +D, +I) → ji- (< d- + sh- → j-)

Although it is difficult to represent the cubic classifier space in two dimensions with a three dimensional graphic, it can be done with a two dimensional table since two of the dimensions have only two discrete values.


 * {|class="wikitable"

!rowspan=2|  !!colspan=2| -D  !!colspan=2| +D ! -I !! +I !! -I !! +I
 * 0 || 0-   || ya-  || da- || di-
 * l || la-  || li-  || dl- || dli-
 * s || sa-  || si-  || s-  || dzi-
 * sh || sha- || shi- || sh- || ji-
 * }
 * s || sa-  || si-  || s-  || dzi-
 * sh || sha- || shi- || sh- || ji-
 * }
 * }

Underspecified Templates
In template forms of classifiers where the dimensions take variable values they are notated by a lone lowercase 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 be resolved to either of the da- or di- classifiers depending on other constraints. The classifier template (s, d, -I) could resolve to either sa- or s-.

The 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.

The 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 classifiers
 * with reflexive object pronominal
 * with reciprocal object pronominal when coreferential with the subject (Leer 1991, p. 97)
 * antipassive
 * self-benefactive ga-
 * plural verb suffix -x' 
 * locomotive ka-(s, +D, i)- “moving by/while doing”
 * dissimulative sh-kh'é-(l, +D, i)- “pretending to do”
 * certain nonproductive instances
 *  OBJ -(?, +D, i)- ROOT -x' 
 *  OBJ - SUBJ -(?, +D, i)- ROOT -(y)ákhw “undoing, removing”
 * ash-ka- PF - SUBJ -(l, +D, i)- ROOT -aa “ SUBJ plays at N/V-ing”

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 cited by Northern speakers as an explanation for why Southern speech seems more rapid.

Subject agreement (+2)
There are six possible values for the subject agreement slot in the verb. They are listed in the Lingít Syntax section but are given here as well.


 * {|class="wikitable"


 * 1 SG   ||align=center|xha-
 * 1 PL   ||align=center|too-
 * 2 SG   ||align=center|ee-
 * 2 PL   ||align=center|yi-
 * 3 NEU  ||align=center|0-
 * INDEFH ||align=center|du-
 * }
 * 3 NEU  ||align=center|0-
 * INDEFH ||align=center|du-
 * }
 * INDEFH ||align=center|du-
 * }

Distributive (+3)
The distributive prefix slot has only one value, dagha-.

Irrealis (+6)
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 prefix per se, instead the realis is indicated by lack of indicated 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 represented as u- although it has various phonological realizations associated with rounding or the close back vowel.

Appearance of the irrealis prefix is associated with -I in the classifier except in the potential and attributive verb forms.

The irrealis prefix cannot cooccur with any of the subject agreement prefixes that have close vowels. These prefixes are 2 S i-, 2 P yee-, 1 P 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 stem 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:




 * colspan=2| yéi yatee
 * yéi || ya-tee
 * thus || (0, -D, +I)-be
 * colspan=2|“It is”, “It is so”
 * }
 * colspan=2|“It is”, “It is so”
 * }
 * }




 * colspan=3| tléil yéi utee
 * tléi-l || yéi || u-0-tee
 * no- NEG || thus || IRR -(0, -D, -I)-be
 * colspan=3| “It is not”, “It is not so”
 * }
 * colspan=3| “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.

Object agreement (+12)
There are eleven object agreement prefixes given below.


 * {| class="wikitable"


 * 1 SG   ||align=center| xhat, axh
 * 1 PL   ||align=center| haa
 * 2 SG   ||align=center| i-
 * 2 PL   ||align=center| yee-
 * 3 REC  ||align=center| a-, 0-
 * 3 NEU  ||align=center| a-, 0-
 * 3 SAL  ||align=center| ash
 * RFLX  ||align=center| sh-, 0-
 * INDEFH ||align=center| khu-
 * INDEFN ||align=center| at-
 * PART  ||align=center| aa-
 * }
 * 3 SAL  ||align=center| ash
 * RFLX  ||align=center| sh-, 0-
 * INDEFH ||align=center| khu-
 * INDEFN ||align=center| at-
 * PART  ||align=center| aa-
 * }
 * INDEFN ||align=center| at-
 * PART  ||align=center| aa-
 * }
 * }

Right Verb 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.