• No results found

CONCLUSION ON PARTITIVE CLASSIFIERS Some general comments are made here

In document The Kiriwinan classifiers (Page 180-183)

15 classifiers in this subcategory, in contrast to the topographical classifiers, which specify parts of

4.4.3.6 CONCLUSION ON PARTITIVE CLASSIFIERS Some general comments are made here

about the partitive classifiers.

First, we may not properly refer to them as item specifiers. While there is a sense in which anything identified as part of, or formed by division of, a larger whole, becomes and is identifiable as a new item, yet the use by a speaker of a classifier from this group of partitives specifies not the new entity as an item in isolation; but what is specified is generally the original whole item modified by

partition. This applies whether we are speaking of undivided parts of a larger whole or of pieces

separated from a larger item. The domain of Cl 3 kai- "rigid/long" may include reference to whole sticks of tobacco only; once a stick of tobacco is divided, the remnant may be indicated only by classifiers which specify division. Merely the fact of division may be specified, or else the fact of division plus the mode of its division; or alternatively the fact of division plus the size of the divided piece. It is clear therefore that the relation between the partitive specifiers and the noun is more complex than the

pronominal role of the item specifiers. The partitive classifiers may supplement the item specifiers by

being used in context with them, and thus they add to the noun phrase the meaning component of partition obligatorily, plus optionally the component of either the activity of division or the size of the piece. Figure II on page 60 pictures amongst other things the possible divisions which may be applied to a

stick of tobacco, as an example of the way partitive specification may be applied to whole items.

The multiple role of the activity

specifier subgroup has been noted above.87 We may note also that a similar sort of dual role is

evident in these partitive classifiers which specify quanta (see for example the repeater-like function of Cl 90 katupo- in example 72 above), and that

further some hierarchy of size may be specifiable by some of them. The hierarchical arrangement may be seen in garden division classifiers, the size of cuts of meat from a carcase, the size of serves of food, and the general proportions of anything divided, from a half portion down to the useless crumb or scrap.

It should be noted that although these partitive classifiers do not have a basic role of item specification, yet by some extensions of their par­ titive reference, particularly in reference to modern or trans-cultural items, they do in some cases function as item specifiers. The broad function of the

classifier Cl 91 pita- "part" illustrates this, as has been discussed above.88 The classifier pila-

has a very wide domain, and in its adaption to modern categories of specification it approaches in the

partitive area the breadth of the Basic Property Specifier Cl 5 kwai- "thing".

Finally, some reference must be made to the "nounfree" function which is broadly in evidence throughout the partitive classifiers. Thus the

classifiers which specify the activity of division may be used with the name of the activity, As

Cl 66 tabudo- "room" in ma-TABUDO-na taboda "that

87

88

See p 138 above. See p 164 above.

division, that wall"; but its usual form is simply the noun-free phrase ma-TABUDO-na "that (room of a house)", there being no separate word for "room". Or Cl 65 livisi- "shelf", when used in reference to a shelf or drawer is used as a noun-free expression LIVISI-tala "shelf-one", as here again there is no noun for "shelf"; or else, in reference to the

contents of a shelf it may be used with the name of the contents of the shelf, as ma-LIVISI-na taitu "that-shelf taitu" "That shelf of taitu yams". As multiple specification is a regular component of the partitive classifiers, such nounfree classifier use is the most usual way in which the partitive classi­ fiers occur. A similar pattern is seen in the other Group II classifiers, where each classifier will carry a potential of multiple specification.

In fact many of the classifiers included in the list of "partitive" classifiers will really be seen as partitive in one of their specifications, and arrangement-specific in the other, with the reverse being true of some in the "arrangement classifier" subgroup. Thus, the partitive classifier Cl 65

livisi- "shelf" is "partitive-specific" in reference to the shelf, but "arrangement-specific" in reference to the group of yams on the shelf. Likewise the

arrangement classifiers Cl 113 dodiga- "load" or Cl 114 kaiyuvai- "layer" are "arrangement-specific" in reference to the things which constitute the load or layer, but may be "partitive-specific" in reference to the layer or load itself. In many instances besides those already cited, it may be shown that the

difference between a partition or an arrangement of items is merely that of extreme points of a semantic

continuum, or different semantic polarities.89 Thus much of the ordering of the classifier data within these subgroups of classifiers must be seen

as an ordering of heuristic convenience, to enable a comprehensive study of the whole lexicon of

classifiers. Further comment summarising these subgroups is left until completion of discussion on the arrangement classifiers.90

In document The Kiriwinan classifiers (Page 180-183)