Publishing Partner: Cambridge University Press CUP Extra Publisher Login

New from Cambridge University Press!


Voice Quality

By John H. Esling, Scott R. Moisik, Allison Benner, Lise Crevier-Buchman

Voice Quality "The first description of voice quality production in forty years, this book provides a new framework for its study: The Laryngeal Articulator Model. Informed by instrumental examinations of the laryngeal articulatory mechanism, it revises our understanding of articulatory postures to explain the actions, vibrations and resonances generated in the epilarynx and pharynx."

New from Oxford University Press!


Let's Talk

By David Crystal

Let's Talk "Explores the factors that motivate so many different kinds of talk and reveals the rules we use unconsciously, even in the most routine exchanges of everyday conversation."

We Have a New Site!

With the help of your donations we have been making good progress on designing and launching our new website! Check it out at!
***We are still in our beta stages for the new site--if you have any feedback, be sure to let us know at***

Review of  Genders and Classifiers

Reviewer: David Douglas Robertson
Book Title: Genders and Classifiers
Book Author: Alexandra Y. Aikhenvald Elena I. Mihas
Publisher: Oxford University Press
Linguistic Field(s): Typology
Anthropological Linguistics
Issue Number: 31.1646

Discuss this Review
Help on Posting

(xxi + 310 pp.) This ninth and latest in Oxford's very fine Explorations in Linguistic Typology series both summarizes and offers case studies in the ways in which grammars are known to develop and use overt categorization of nouns. Because classifiers in particular may be unfamiliar to many linguists, I will herein emphasize the factual contents of the book over my evaluation thereof.

Its Preface (viii-ix) broadly previews the volume's typological findings and describes the 2017 workshop from which its chapters emerged. Note is made that “[t]he analysis is uniformly cast in terms of basic linguistic theory” (BLT), which, far from a simplification for novices, is “the cumulative typological framework which provides the foundation for sound empirically-based descriptive and analytic works.” (See RMW Dixon's 3-volume opus (2010-2014) for a BLT summary of what linguists know.) Notes on the Contributors, with contact information, occupy pages x-xiii. Abbreviations (xiv-xxi) complete the introductory matter.

Chapter 1, “Noun Categorization Devices: A Cross-Linguistic Perspective” by Aikhenvald (1-29), is that editor's summary of research on the topic. Gender systems are always rooted in biological gender, and typically apply without exception to all of a language's nouns. Marking of gender can be overt (i.e. on the noun) or covert (on some agreeing constituent(s)). In many languages, it can interact with a classifier (CL) morpheme system. Six types of these are distinguished by their scope, the argument categorized, and its semantics: (1) Numeral CLs sort quantified nominals by salient inherent qualities of those entities. (2) Noun CL systems explicitly subdivide nominals per se into multiple semantic classes. (3) Possessive CLs similarly categorize Possessee arguments. (4) Verbal CLs are bound morphemes within the verb stem referring to a trait of the Object or Subject (hardly ever Agent) argument thereof. (5) Locative CLs, a rare type, index properties of objects that are referred to by a language's location markers. (6) Deictic CLs are even rarer, being mandatory references by a demonstrative to the positional orientation of an argument. CL systems typically do not apply to all nouns in a language, and more than one of the above kinds can coexist. Some languages allow several CLs on one noun, giving a more specific meaning to the head. “Multiple CL” languages deploy the same classifier in more than one syntactic context, a spectactular example on page 17 being a Tariana sentence of five words, each bearing the CL suffix for 'house'. Noun categorization devices bear a significant functional load, distinguishing meanings of a single stem, conveying speaker attitudes towards it, tracking its use in discourse, and so on. Noun classification devices can historically originate in bound nouns, verbs, or pronouns, and often are etymologically transparent, indicating their recent vintage; their presence is typically an areal phenomenon, implicating contact as a frequent means of their introduction.

Chapter 2, by editor Mihas, examines “Genders and Classifiers in Kampa (Arawak) Languages of Peru” (30-66). Gender marking is obligatory on various words agreeing with nouns (i.e. in the third person), as well as on some noun heads; animacy too is overtly distinguished in a subset of these languages. Cultural assumptions influence formal gender assignment, with myths explaining why e.g. stars are of masculine but landscape features of feminine gender. A multiple-CL categorizes nominals primarily by shape, but some reference shape or physical arrangement. CL use is however not necessarily frequent, occurring roughly once per 5 to 9 sentences, depending on the individual language, genre, etc. A primary function of CLs is for referent introduction and tracking, but they also allow for affective reference, both negative and positive, to fellow humans through unexpected CL choices.

In Chapter 3, Pilar M. Valenzuela investigates “Classifiers in Shiwilu (Kawapanan): Exploring Typologically Salient Properties” by (67-102). In this language, there is no grammatical gender, and its “multiple'' CLs are not obligatory. There is a significant degree of homophony, with three pairs of identical CLs and one CL identical to a freestanding noun. Statistics show that CL use is rather infrequent, calculated at once per roughly 28 spoken words. The most frequent CL function is to derive new vocabulary, contrary to previous typological generalizations that discourse tracking is central. Other functions include a “concord-like” and an adjective-like usage. More than once CL can occur per word, the final one functioning as head; a single CL can reduplicate, connoting descriptor root plurality. Animate-referring CLs on verbs tend to be interpreted as Agents, inanimates as Objects. Two CLs in a verb are interpreted respectively as Subject and Object.

Editor Aikhenvald presents Chapter 4 “A View from the North: Genders and Classifiers in Arawak Languages of North-West Amazonia” (103-143). Here she illustrates pertinent traits of these languages with respect to their tendency toward an unusual occurrence of gender and CL systems. Both are reconstructed to Proto-Arawak; yet a good deal of diversity occurs, with gender less or more independent of CL use in each language. Four sets of these languages are distinguished, depending on whether and to what extent they use CLs, which seem to be a recent historical development, grammaticalizing from largely distinct selections of nouns from language to language. Major functions of these CLs are (again) disambiguation of related forms, coining new forms, individuating nominals, and anaphoric reference. Appendices to the chapter explain the genetic subgroupings of these languages and tabulate the CL systems used in various of them.

Chapter 5 introduces another language family and another region with “Possessive Classifiers in Zamucoan” of the Northern Chaco, by Luca Ciucci and Pier Marco Bertinetto (144-175). CLs in these languages are unique in simultaneously expressing gender as well as a Zamucoan peculiarity, an obligatory morphosyntactic category called 'form' (predicative usage versus indefinite and definite argumental function; perhaps approximately a continuum of topicality). As seen of other languages in this volume, CLs here appear to be a young system. An important function they serve is to express possession of nouns that are ordinarily uninflectable for possession.

Shifting back to northwest Amazonia, with Chapter 6 Katarzyna I. Wojtylak describes “The Elusive Verbal Classifiers in 'Witoto'” by (176-196). This (noun-)genderless language uses a large (≥110-member) multiple-CL system taxonomizing nouns by physical properties, natural gender, and abstractness, as well as providing generic CLs and innovating new ones as needed. Most are monosyllabic; disyllabic ones tend to be more specific in meaning, a seeming sign of their origin as two CLs on one noun. The primary function of CLs is generating new noun stems. A notable feature is the non-productive occurrence of CLs in verbs, indexing an argument that then need not be explicitly stated.

Chapter 7 examines one of the rarer CL types, with Cristina Messineo and Paola Cúneo discussing “Multifunctionality of Deictic Classifiers in the Toba Language (Guaycuruan)” (197-221). This Argentinian language's CL system has three each of strictly deictic (proximal, distal, unseen) and of positional (standing, lying, sitting) members, all six identical in morphosyntactic behavior, thus justifying their consideration as a single “deictic CL” class. Agreeing in number and gender with their noun heads, and apparently obligatory with them, these CLs also participate in generating third-person and demonstrative pronouns, introduce relative clauses, and connect multiverb predicates (e.g. 3.want CL:proximate = 'wanted to eat'). The deictic CLs categorize nouns both in intuitive ways (for example, trees take the 'standing' CL, beds the 'lying', and pots the 'sitting' CL), and in conventional metaphorical extensions (plural for instance can be conveyed by the 'lying' CL, as though several entities were lined up in a row). As seen in other languages, CLs are amenable to further creative uses, such as distinguishing the mammal 'fox' from the fish called by the same noun stem, and even conveying an affective stance, as when 'dog' takes the 'standing' CL which is prototypically associated with humans (resulting in a joke or insult: 'You are a dog!'). Temporal reference is connoted by the strictly-deictic CLs, so that 'proximate' = 'present', 'distal' = 'recent past', and 'unseen' = 'remote past' or 'future'. Discourse structure too interacts with CL choice; as a rule 'distal' encodes Agents (a more topical function) but 'proximal' Objects/Subjects (which are more focused). The authors additionally show that CL choice varies by, and indicates, textual genre.

With Chapter 8, “Classifiers in Hmong” (222-248), Nathan M. White reviews the literature on the topic and presents four new empirical observations. First, there is a previously unrecognized CL subtype in the language, the locative classifier (viz. Aikhenvald's Chapter 1). Second, the Hmong CLs constitute an open class, one that readily recruits new members even from foreign languages. Third, certain syntactic phenomena are now more clearly analyzed; these include what has previously been seen as “verbal classifiers” but which are here argued to be nominals, and an “apportioning” construction more or less equivalent to English 'each'. Lastly, White shows that Hmong is another language in which CL choice helps structure discourse in terms of definiteness, specificity, genericity, and so forth.

Another East Asian language is the topic of Chapter 9, “Numeral Classifiers in Japanese” by Nerida Jarkey and Hiroko Komatsu (249-281), examining the standard variety. The authors find that “counters” accompanying Japanese numeral words pattern into numeral classifiers, conventionalized measures, and quantifiers, categories that are not fully discreet but instead occupy a “quality-quantity continuum”. Numeral CLs are obligatory except when counting with large numbers, which are foreign (Chinese) in origin. While the class is large (100 or more members by previous counts), a mere five numeral CLs total 82% of the tokens in one study sample cited, and average speakers' CL inventories appear to be shrinking through time. Living beings tend to be referred to either by 'human' or 'nonhuman animal' CLs, while inanimate nominals normally are classified by physical dimensions or cultural uses. Etymological sources of the majority of CLs not known to trace to ancient Japanese include native nouns and some verbs, as well as quite a few Chinese loans. Among the functions of the CLs is expression of specificity and indefiniteness, as well as signaling stylistic choice. As seen in other languages above, a given noun stem can take various CLs, economically connoting such descriptive information as size, shape, and even robot status, as well as speaker attitude (as when the CL -hiki 'small animal' is used to count 'people' in a disdainful way).

Chapter 10 concludes the case studies with another Asian language, in “Numeral Classifiers in Munya, a Tibeto-Burman Language” by Junwei Bai (282-298). As in Japanese, there is no nominal gender, so that Munya CLs are the only noun-classification device available. They are bound roots, of either sortal or mensural semantics. There are two interchangeable generic sortal CLs as well as ones for human/nonhuman living status, shape, unit of time, and so on. Mensural CLs subdivide into those for containers, groups, length, weight, and kind. Again we find this is a language using its CLs not only for quantification and classification, but also for discourse and pragmatic functions including specificity, definiteness, and anaphora. In addition, certain numeral + CL expressions can be used adverbially, sometimes with the sense of a verbal situation occurring 'N times'.

The book closes with Indexes of Authors (299-302), Languages, Language Families, and Linguistic Areas (303-305), and Subjects (306-310).


Like the other books in this series, this one is a valuable communication of what is now known in terms of the linguistic typology of a given phenomenon. Field workers as well as more theoretically oriented typologists will find its presentation of the possibilities of noun categorization in a single convenient volume to be enlightening. The combination of a sort of checklist of CL parameters to be probed by researchers (particularly Aikhenvald's Chapter 1) with illustrative cases from diverse regions amounts to an effective course in what, for linguists working in some parts of the world, will be a relatively exotic morphosyntactic device, thus remedying a gap in many colleagues' training. Certain chapters in this book are exceptionally concise, clear, and digestible. I want to single out Chapters 7 on Toba, 9 on Japanese, and 10 on Munya for their achievements in this regard; each would make a fine weekly reading in, say, a university seminar on classifiers.

Because it is an anthology of many authors presenting specialized expertise, the diversity of symbolic devices used in the chapters has to have been a challenge for the editors and their assistants to bring into a semblance of uniformity. Their solution, the thorough table compiling Abbreviations in the start of the book, is wise and useful. Almost inevitably, some entries in it are opaque (such as AT 'postposition “dine” in Murui'); some seem superfluous (EP 'epenthetic' in Kampa and Shiwilu, and LK 'linker' in 'Witoto', look purely phonological and might have been better left unsegmented); and certain identical abbreviations bear divergent glosses (like M 'masculine' and 'middle pronominal marker'), an issue that only grueling communication with the contributors and self-sacrificing editorial effort could forestall. A more substantial request for any future edition would be the inclusion of symbols in this same list, as nowhere is an explanation made of what is meant by “~” (seemingly, variable ordering of constituents, p. 224) and “.” (evidently, conjoinment of multiple syllables in a morph, ibid., but also denoting parsing within morphologically complex stems, p. 257).

Thanks to the fact that most textbooks and research are published by speakers of Western European languages, gender systems are probably fairly familiar to linguists worldwide; colleagues may be a good deal less aware that classifier systems, too, are of fairly wide occurrence. While one appreciates this volume's focus on two fascinatingly rich regions of CL use, it would be rewarding to see the editors discuss where else these devices occur. For example, they are not uncommon in North America, where the Na-Dene languages employ verbal classifiers in profusion (Krauss 1968), and Salish languages are well known for their ramified numeral classifier systems (Gerdts and Hinkson 2004). Beyond areal considerations, sign languages also are routinely discussed in terms of classifier use (viz. Engberg-Pedersen 2010), and a number of pidgins/creoles feature CLs as well (Maurer et al. 2013); both are typically omitted from typological surveys, and both likely would have contributed further to the editors' generalizations.

Naturally the approaches and terminology vary from chapter to chapter, mostly without much consequence, but sometimes an author's insight implies deeper issues that could profitably be pursued. One instance is the dichotomy claimed in Chapter 2 between “non-possessable” and possessed nouns. On its surface this seems like a novel and promising area for typological research. For Kampa Arawakan, an admirably precise list of the former class encompasses “entities from the domains of natural elements, supernatural...forces, and astral entities...also...proper names, vocative/address kinship terms, and some orientation and geomorphic terms” (p. 34). Upon reflection, however, this seems like an enumeration of the sorts of nominals that are pragmatically unlikely to be possessed in typical speech in any human language. The notion of “non-possessable”, then, may deserve further research in its own right as the counterpart to the great amount of typological work on kinds of possessa; it might provide a useful further point along the continuum usually thought of in terms of alienables vs. inalienables.

Various merely implicit or seemingly offhand observations in this volume would seem to merit more discussion and further crosslinguistic research. In this regard it is probably worth pointing out the uncommented-on phonological variation in CL shape that we see in some of the Kampa (e.g. p. 39) and Shiwilu data. One wonders whether certain functional uses of CLs might be more prone to phenomena such as phonological reduction and to influence from neighboring morphemes' segments. Another occurrence sporadically noted among the chapters is the use of “two consecutive number words accompanied by one numeral classifier, with an approximate reading” (p. 12); the chapter on Munya uniquely specifies that this pattern can extend to numerals as high as six (p. 290), and it would be interesting to learn its extent in a worldwide sample.

In summary, this volume is highly recommended as the most current summation of what has been learned by researchers into noun-classification systems in recent years. It is an excellent tool for any scholar seeking to solidify their grasp of the grammatical possibilities of languages and for those aiming to document languages in as full and intercomparable a fashion as possible.


Dixon, RMW. 2010-2014. Basic Linguistic Theory (Vols. 1-3). Oxford: Oxford University Press.

Engberg-Pedersen, Elisabeth. 2010. Factors that form classifier signs. Pp. 252-283 in Brentari, Diane (ed.), Sign Languages. Cambridge: Cambridge University Press.

Gerdts, Donna and Merdedes Q. Hinkson. 2004. Salish numeral classifiers: A lexical means to a grammatical end. STUF – Language Typology and Universals 57(2-3):247-279.

Krauss, Michael. 1968. Noun-classification systems in Athapaskan, Eyak, Tlingit and Haida verbs. International Journal of American Linguistics 34(3):194-203.

Maurer, Philippe and the APiCS Consortium. 2013. Sortal numeral classifiers. In: Michaelis, Susanne Maria & Maurer, Philippe & Haspelmath, Martin & Huber, Magnus (eds.), The atlas of pidgin and creole language structures. Oxford: Oxford University Press. Online at, accessed February 20, 2020.
David Douglas Robertson, PhD, works on Pacific Northwest languages and linguistic contact history, in particular the pidgin-creole Chinuk Wawa (Chinook Jargon) and Tsamosan (Soutwest Washington State) Salish languages. He specializes in the repatriation of indigenous intangible cultural property from archival language materials.

Format: Hardback
ISBN-13: 9780198842019
Pages: 336
Prices: U.S. $ 90.00