Asian Languages and Linguistics - Volume 3, Issue 2, 2022

Classifiers are morphemes which occur under specifiable conditions and which categorise nominal referents in terms of their animacy, shape, and other properties. The most widely represented type is numeral classifiers, which occur next to a number word or a quantifier. Further types include noun classifiers, verbal classifiers, classifiers in possessive constructions, and deictic classifiers. One language can have more than one type of classifier. In some, the same set of classifiers occurs in several classifier contexts, corroborating the unity of the phenomenon. Classifiers categorise nouns, and have to be distinguished from verbal action markers used to categorise and count actions. Classifiers have a variety of functions, and are never semantically redundant. Classifiers mirror social attitudes and hierarchies, physical environment and means of subsistence, and are susceptible to change in language contact situations. Contributions to this issue adress the systems and the functions of numeral classifiers and also classifiers in multiple contexts across Asia and beyond, including Austronesian languages of Taiwan, a selection of Tibeto-Burman (or Trans-Himalayan) languages, Zhuang, a Tai-Kadai language, and Kazakh, a Turkic language.

This paper is a typological exploration of the nominal classification systems found in the Austronesian languages of Taiwan, with special focus on their meanings, forms and functions within the Austronesian typological context. Based on Aikhenvald’s defining properties and classifications of classifiers (2000, 2017, 2019, 2021, forthcoming), the Austronesian languages of Taiwan investigated in this study display a set of relatively small classifier systems consisting of numeral classifiers and other classifier-like terms, such as measure words and verbal action classifiers. Numeral classifiers can be free or bound morphemes. The choice of a numeral classifier is based on the semantic features of a noun referent, mostly involving ‘humanness’. Though not every referent has to be assigned to a numeral classifier, numeral classifiers always accompany numerals. In these languages, none of them have numeral classifiers without presupposing the presence of sortal classifiers.

Dimasa (ISO 639-3 dis) is a Bodo-Garo language within the Tibeto-Burman family of languages. Like other Tibeto-Burman languages of Northeast India, it has a system of numeral classifiers which occur prefixed to their respective numerals. Across the dozen or so languages of Bodo-Garo, it has been noted that the classifier-numeral word can be placed either before or after the counted noun. The existence of numeral classifiers, as well as the clf-num order may have arisen in this area due to contact with Tai languages during the Ahom kingdom period from the 1200’s to 1800’s. In the Tai languages, classifiers precede the number “one”. By comparing occurrences of classifiers in Dimasa texts, it can be seen that the order N clf-num corresponds to definite nouns, while the order clf-num n marks indefinites. In addition to classifier placement, Dimasa definiteness is also marked by demonstratives, case marking, topic marking, and/or combinations of these strategies.

This paper introduces noun categorization devices in nDrapa. It argues that nDrapa usually requires both numeral classifiers and verbal action classifiers to express classificatory techniques. Morphologically, classifiers are fused with numerals; without the involvement of classifiers, numerals cannot be directly associated with nouns. Classifiers can derive from the pragmatic function, in which a considerable number of classifier-derived quantifiers and adverbs reveal the adverbial function with respect to the softening voice and politeness. Historically, the majority of classifiers can trace their etyma back to nouns. This paper’s five sections explore the grammatical properties of classifiers in nDrapa. In Section 2 and Section 3, I provide background information on numerals and numeral classifiers, and describe the characteristics of verbal action classifiers in nDrapa. Section 4 explores expressions relevant to classifier-related grammatical categories. In Section 5, I summarize the pragmatic and functional properties of classifiers in nDrapa. In Section 6, I introduce the origin and development of classifiers in this language. To conclude this paper, I propose that the limited number of classifiers in nDrapa may exhibit an infant development of classificatory devices in this language.

Classifiers in Tani languages are similar in scale and semantic contents to the systems of better-known classifier languages such as Thai and Mandarin. Yet they are unusual in co-existing with an ancient and well-grammaticalised referential management system including both definite and indefinite articles, in lacking a generic classifier, in tending not to use a classifier with human referents, and in occurring exclusively to the right of head nouns, in the order [N CLF NUM]. They are also relatively more lexeme-like, occurring less frequently and with more semantic control than do the classifiers of many other East/Southeast Asian languages. This article will present a basic description of Tani classifier systems, and argue for their relatively recent development through the mechanism of a repeater construction functioning within a pre-existing [A-B B-C] template for taxonomic compound formation. Although this development is similar to pathways attested for other Asian languages, Tani classifiers do not share the same set of functional and distributional outcomes.

The Zhuang language employs a system of nominal classification, including numeral classifiers, modifier classifiers, class terms, kin prefixes, gender markers, prefixable morphemes for topographic names, among others. Quite often, the same set of morphemes is used as numeral classifiers and with demonstratives. The numeral classifier system is particularly salient, comprising numeral terms and noun classifiers, with dozens of classifiers featuring fine semantic distinctions. The modifier classifier system involves the use of classifiers in nominal modification of all kinds. The kinship system consists of kinship terms which take prefixes and other human terms as prefixes to personal names. Gender markers are found for human nouns, some kinship terms and animal nouns or plant names. Toponyms or place names are headed by morphemes describing the topographic or geographic features of the places or locations in question. Discourse contexts may also help to differentiate different types of Classifier Phrase.

Many languages distinguish numeral classifiers in their grammar systems categorizing nominals in terms of their inherent nature, such as animacy, shape, form, and arrangement. Kazakh, a Turkic language of Central Asia, distinguishes numeral classifiers, which are used in numeral classifier constructions consisting of a numeral, a numeral classifier and a specified noun. Numeral classifiers in such constructions can be optional. This paper aims at a preliminary description of the features, types, and usage of numeral classifiers in Kazakh.

In many languages, nouns need a numeral classifier to combine with number words or quantifiers. Even though languages may have an elaborate system of numeral classifiers, there usually are exceptional nouns that combine directly with number words. This paper deals with the following questions: Which nouns trigger a numeral classifier when counted? Which nouns are more likely to occur without an additional classifier in a numeral context? Data from 32 languages reveal a classifiability scale of nouns depending on their semantics. The more animate a noun, the more likely is the occurrence of a classifier. With inanimate nouns, classifiers often may or must be omitted. The investigation extends to units, with units of time being much more likely to be accompanied by a numeral classifier than other units. These findings are surprising in the light of the predominant view that numeral classifiers are individuating units in languages where nouns appear to be uncountable mass nouns. I discuss several approaches, of which animacy and the form-frequency correspondence principle are most promising.