This paper outlines the complex system of noun classification in Dakelh. It is particularly concerned with the existence of multiple classificatory subsystems and with the fact that the categories that they make use of are not homomorphic.
The dialect discussed here is the Nak'albun/Dzinghubun (Stuart/Trembleur Lake) dialect. The details vary from dialect to dialect, but overall the classificatory systems are fairly similar.
Words or sentences preceded by a star (*) are ungrammatical.
The classificatory system is comprised of eight subsystems, viz.:
The first of the subsystems affects numbers and quantifiers, which take five different forms. These are exemplified for the numbers from one to ten and for two quantifiers in table (1).
(1) Numbers and Quantifiers
|9||'ilho hooloh||'ilho hoolohne||'ilho hooloh||'ilho hoolodun||'ilho hoolowh|
The generic series, illustrated by (2), is used for counting most physical objects. It is also this series that is used for counting in the abstract, e.g. when reciting the numbers, and for telephone numbers, addresses, and so forth.
|He lost five dollars.|
The human series is used for counting human beings (3). As is generally the case in Dakelh, the grammatical category of human beings includes dogs (4), but is rarely extended to other animals.
|He is married to two women.|
|He is going to give me three dogs.|
The multiplicative series often refers to numbers of times, as in (5). However, it is also used with some units of measurement, such as weeks (6).
|It was four nights that we camped where the Hudson's Bay factor used to live.|
|We were there for three weeks.|
The areal series is used for counting sets of discrete areas. It is not sufficient that it refer to something areal. For example, in (7), although land is areal, the appropriate form of the quantifier "all" is the abstract form. The areal form in (8) is ungrammatical. Appropriate uses of the areal forms are illustrated in (9) and (10), where what is counted are a number of discrete areas.
|This land is all Indian land.|
|This land is all indian land.|
|These houses are all indian houses.|
|On my father's trapline there are many houses. (nt john 14.1)|
The abstract series is used for counting things that have no physical form, such as kinds and ideas, as in (11).
|We have four clans here.|
A second system of noun-classification is found in the possessive prefixes attached to nouns, which also serve as pronominal objects of postpositions. Here the distinction is binary, between a generic category and an areal category. this distinction is made only in the third person singular. In (12) the third person singular possessor is marked by u-, but in (13) it is marked by wh- since "village" is areal.
|My father's trapline.|
|The chief of our village.|
The same distinction in third person singular objects of postpositions is exemplified by (14) and (15). In (14), the postposition -ando "above" takes the generic prefix b-, while in (15) it takes the areal prefix wh.
|The cupboard is above the stove.|
|(15)||Nuk'atun dot'en-i nut'o-i||ooyoh||whandoh||nut'o|
|The helicopter is flying around above his house.|
The main absolutive prefixes are d, which refers typically to stick-like objects, n, which refers typically to round objects, and wh, which refers typically to things having areal or spatial extent. In the case of intransitive verbs, these prefixes agree with the subject. Many verbs therefore have four forms, according as they have one of these three prefixes or none. An example is the verb "to float". it may have no classifier prefix, as in (16), where the subject is a human being, or any of the three, as in (17)-(19).
|A young woman is floating around.|
|A log is floating around.|
|A ball is floating around.|
|A house is floating around.|
When a verb is transitive, the classifier prefix agrees with its object. The verb "to eat", for example, may take no classifier prefix, as in (20), n for something round like an apple as in (21) or d for something stick-like like a stick of pepperoni, as in (22). It may also take the prefix wh as in (23). This form could refer to eating something like a house (e.g. in the story of hansel and gretel), but in this example is most naturally interpreted as a kind of indefinite.
|I am eating char roe.|
|I am eating an apple.|
|I am eating a sausage.|
|I am browsing (at a smorgasbord).|
Some verbs distinguish only three of the four categories. most of the colour terms, for example, listed in (24), have no distinct /d/-class form. Where other verbs would take on a /d/-class form, these take on the generic form.
The term for "white", however, has the full set of four forms: lhyul, dulyul, nulyul, whulyul.
Still other verbs have only two distinct forms. these are always the wh form and a generic form. An example is the verb nduda "to hurt". While most parts of the body, such as the hand, take the generic form as in (25), areal/spatial body parts such as the chest take the areal form, as in (26). Stick-like body parts (27) and round body parts (28) nonetheless take the generic form, even if they take a distinct form where one is available, as when the colour of the face is described (29).
The absence of distinct forms for some noun-classes can often be attributed to the fact that the generic form already contains one or more of the classificatory prefixes, though without its classificatory meaning. For example, the absence of distinct /d/-class forms for the colour terms other than "white" can be attributed to the fact that they already contain a d-qualifier prefix. Since "white" contains no /d/-qualifier, it can take the full range of classificatory forms. Similarly "to hurt" inherently contains both /d/- and n-qualifier prefixes and so has no distinct /d/- and n-classifier forms. the absence of classificatory forms cannot always be accounted for in this way. Moreover, although Athabaskanists freely refer to classificatory and homophonous non-classificatory (thematic) prefixes as if they were instances of the same morpheme (the so-called "qualifier" prefixes), It is clear from the most elementary theoretical considerations that they cannot be: if a morpheme pairs sound and meaning, it cannot be meaningless in one context and meaningful in another. A satisfactory account of these phenemona therefore remains to be developed.
|My hand hurts.|
|My chest hurts.|
|My leg hurts.|
|My face hurts.|
|My face is red.|
Dakelh has a total of nine demonstratives. They may function as demonstrative adjectives, modifying a noun (e.g. ndi kwun "this fire"), or as pronouns, standing by themselves (e.g. ndi "this"). One dimension of the system is whether the item referred to is human or non-human. as in other aspects of dakelh grammar, dogs fall into the human category, while other animals generally do not. there is some variation in how people treat cats and other animals with which people may be closely associated. If the item referred to is human, there are distinct singular and plural forms. there is no distinction of singular and plural for non-humans.
The other dimension of the system is the location of the item. While english has a two term system, in which this and these refer to things near or associated with the speaker, and that and those refer to items distant from or not associated with the speaker, Dakelh makes a further distinction between among things not near or associated with the speaker. These may be associated with the person spoken to, in which case nyoo, nyoon and nyoon-ne are used, or they may be distant from or not associated with either the speaker or the addressee, in which case nghun-i, nghun-un and nghun-ne are used.
|non-human||human singular||human plural|
|away from both||nghun-i||nghun-un||nghun-ne|
Agentive deverbal nouns exhibit the same pattern as demonstratives. There are three suffixes that may be added to a verb to form a noun meaning "the one who Vs" or the like. The suffix -un is human singular, so nudaih-un means "dancer". (Note that ere is a homophonous suffix -un meaning "place where", so nudaih-un also means "place where one dances", i.e. dance hall or ballroom.) The suffix -ne is human plural, so nudaih-ne means "dancers". The suffix -i is non-human, so that that from 'udulht'oos "it peels off bark", we derive 'udulht'oos-i "barker", that is, the machine that strips bark from logs.
In Dakelh a relative clause follows the head noun. there is no complementizer, but the verb optionally takes a relativizing suffix. Three of the suffixes reflect the same categories as with the demonstratives and agent nouns, namely human singular, as in (31), human plural, as in (32), and non-human, as in (33).
|he brought back his wife, who was sick.|
|The dog barks at everyone who goes by.|
|The berries you picked are big.|
The suffix -un has, however, an additional use as a relatiziver. When the relativized argument of the verb is areal it is -un that is used, not the expected non-human -i. Here are some examples:
|The red house is our house.|
|They saw large grizzly tracks.|
|He is reaching the top of the steep mountain.|
|The burning house is collapsing.|
|They straightened out the crooked road.|
|The bear crawls back into the big cave.|
|We live in a beautiful country.|
|The house that burned down is still smoking.|
Notice how in each case the relativized verb contains the areal absolutive classifier prefix wh- as well as the relativizing suffix -un. For example, in (34) the verb meaning "it is red" takes the areal form whudulk'un rather than the generic form dulk'un. In (35) and (39) hooncha is the areal equivalent of the generic ncha. In (35) the main verb, which governs the relative clause, is inflected for an areal object; if its object were non-areal it would take the form han'en. \enumfootnote
The interrogative "how many?" has five forms the use of which depends on the kind of thing whose quantity is questioned. These are:
(42) How Many?
Another classificatory system is the set of classificatory verbs, used to describe the handling and location of objects of different types. The bases of classificatory verbs convey no information other than the type of object; what action is performed is determined by the choice of derivational prefixes. (43) illustrates the range of verbs that may be derived from the stem for handling two dimensional flexible objects (e.g. shirts).\enumfootnote
(43) Derivatives of "Handle Two-Dimensional Flexible Object"
|behanaitilhchus||he is going to take it out|
|didutalhchus||he is going to hold it up|
|dughaitalhchus||he is going to hang it up|
|k'italhchus||he is going to put it on (the table)|
|k'unaitalhchus||he is going to put it back on (the table)|
|k'unaitilhchus||he is going to take it off (the table)|
|sanaitilhchus||he is going to bring it back|
|yughatilhchus||he is going to give it to her|
|yughutilhchuz||he is going to lend it to her|
|nutilhchuz||he is going to carry it around|
|'atilhchus||he is going to bury it|
|tatilhchus||he is going to submerge it|
|natilhchus||he is going to put it on the ground|
|yaiyutilhchus||he is going to bring it ashore|
The categories into which objects are divided are illustrated in (44), where forms of "he will give me" appropriate to a variety of objects are given. Notice that some but not all of these bases permit cross-classification by means of the absolutive prefixes.
(44) He Will Give Me
|long rigid object||generic||canoe||sghatitelh|
|contents of open container||generic||cup of tea||sghatikalh|
|2-dimensional flexible object||generic||shirt||sghatilhchus|
The choice between the plural and non-plural default verbs is not entirely straightforward. A single object calls for the non-plural verb, three or more for the plural verb. Two objects usually, but not always, take the non-plural verb.
The plural default verb is also used for certain single items namely ropes and fishnets, perhaps because these are considered to consist of multiple coils and meshes.
The category "contents of an open container" also includes beds, for mysterious reasons.
|bed||still||it is folded||while||they put it in|
|They placed the folded bed in the truck.|
The category described as consisting of uncountable objects calls for comment. The objects included in this category are usually small, e.g. grains of sugar or sand, berries, or toothpicks. Consequently, this category has sometimes been described as consisting of a "quantity of minute objects". However, larger objects may fall into this category. An example is potatoes, which take this verb if in sufficient quantity. Even smaller objects, such as coins or berries, only fall into this category if there are enough of them, typically three or four.
It appears that this category contains effectively uncountable objects. In the case of numerous minute objects, we do not expect to treat them as individuals and to count them. Larger objects often are individuated, but need not be. For example, when we are dealing with a heap or bushel of potatoes we do not individuate them and usually do not know how many there are.
The same notion is found elsewhere in the language. There is a verb theme l-dilh which usually occurs with the n absolutive classifier prefix, in which case it is usually translated "to eat berries off the bush". Although this describes one situation in which this verb may be used, it is not a sufficiently general translation. This verb is also appropriate if a person eats berries one at a time out of a bowl, but not if he eats many berries at once with a spoon. It is also used to describe taking a pill, in which case it is appropriate even if the subject takes only a single pill. Without the n absolutive classifier prefix this verb may be used to describe a bear tearing up an anthill and eating ants. I suggest that the true meaning of this verb is: "eat effectively uncountable objects one at a time". If this is correct, the meaning of this verb makes use of the same concept of effective uncountability as does the classificatory verb system.
With the exception of the verb for uncountable objects, which is intrinsically plural, the specialized verbs are used only for single objects. Plural objects that as individuals would fall into one of the specialized categories take the plural default verb. For example, although "I will give you a fish" is lho nghatistelh, "I will give you three fish" is ta lho nghatislilh. Younger speakers often violate this rule and use the specialized verb even with a plural object.
Nouns do not fall into fixed noun classes. Different choices of classifier are possible for the same referent depending on precisely what is said about it. For example, a rope may be treated as /d/-class, as seen in (46) and (48. In these cases the generic equivalents in (47) and (49) are ungrammatical.
|The rope is thick.|
|The rope is thick.|
|The rope is thin.|
|The rope is thin.|
However, when the rope is considered from another perspective generic forms become grammatical, as in (50) and (52), and the corresponding /d/-class forms become ungrammatical, as in (51) and (53).
|The rope is long.|
|The rope is long.|
|The rope is short.|
|The rope is short.|
That this is not attributable to the non-existence of /d/-class forms for these verbs is illustrated by (54) and (55).
|The song is long.|
|The board is short.|
Dakelh appears to be in the process of innovating an additional absolutive classifier ta-, derived from the compounding form of "water". This prefix is used with reference to non-directed bodies of water, that is, lakes and ponds as opposed to rivers and streams. This prefix may be used alone as in (56) and (57) or in combination with the /d/-prefix as in (58). Another apparent instance of this prefix is illustrated in (60). "to be shallow" does not occur without the ta-.
|The ocean is wide.|
|Williston Lake is full.|
|Stuart Lake is wide.|
|The waves are big because the lake is deep.|
It is possible for the absolutive classifier prefixes to co-occur. Co-occurence of the wh- prefix and the d- prefix has the expected semantics. In (61) the waist is regarded as /d/-class because the figure is stick-like and as wh-class because the waist is an area of the body, comparable to the back, which is also wh-class. The waist may also be treated as purely wh-class, as in (62). As (63) shows, the waist may also be treated as generic.
|White women are frantic for narrow waists.|
|She has a narrow waistline.|
|The woman is small and also slender.|
Co-occurence of the d- and n- prefixes, in contrast, appears to be idiosyncratic and non-compositional. Objects such as round logs, which we might expect to be treated as simultaneously round and stick-like, are purely /d/-class, as illustrated by (64) and (65). These two prefixes do co-occur, but only with reference to the size of approximately circular openings like the mesh of fishnet (66), the spaces between threads in textiles (67), and the size of snares (68).
|This log is long.|
|This log is long.|
|This fishnet has a fine mesh.|
|This cloth is coarse.|
|Grizzly bear snares are large.|
The categories employed by the several classificatory sub-systems in Dakelh are not homomorphic. This non-homomorphism manifests itself in several ways:
Although several of the classificatory verb bases permit cross-classification by absolutive classifier prefixes, this cross-classification is not entirely compositional. The verb base for non-plural generic objects permits the use of the d- classifier prefix in limited circumstances. Most of the typical uses of this prefix are not possible here because they require the special verb base for long rigid objects. One of the few remaining uses is for names, songs, speeches and the like, which are considered /d/-class but are not long, rigid objects and therefore do not permit the use of that verb base. Although the inclusion of these items in the /d/-class may be anomalous, this usage is nonetheless compositional. What is unexpected is that the generic verb bases may be used with the d- prefix in reference to rocks, which are not otherwise /d/-class. This usage is non-compositional.
Yinka Déné Language Institute © 2006