Studies in Language. International Journal sponsored by the Foundation “Foundations of Language” - Volume 41, Issue 2, 2017

The corpus-based analysis of Teop word classes demonstrates that lexical multifunctionality is not incompatible with the grammatical distinction between verbs, nouns, adjectives and adverbs, because this distinction does not manifest itself in their syntactic functions of heads of phrases, but in the kinds of modifier the words can combine with. Consequently, the Teop word class system falsifies all word class typologies that assume that a formal differentiation of event, object and property words presupposes a distinctive distribution across the head positions of determiner-marked referential expressions and TAM-marked predicative expressions. In addition to the multifunctionality of verbs, nouns and adjectives, the Teop lexicon shows regular patterns of conversion. The paper concludes with an assessment of the results and the limitations of the corpus-based approach and suggests four topics for further research: (1) the development of elicitation methods to supplement corpus-based analyses; (2) a typology of formally distinguished phrase types; (3) regular patterns of conversion across languages; and (4) possessive comparative constructions.

The issue of lexical flexibility is best tackled as the articulation of two separate mappings: one that assigns lexical items to word classes; another one that associates these word classes with the syntactic functions they can access. A language may endow its lexemes with more or less multicategoriality, and its word classes with more or less multifunctionality: these are two distinct facets of lexical flexibility, which should be assessed separately. Focusing on Hiw, an Oceanic language of northern Vanuatu, I show that lexical flexibility is there mostly due to the high multifunctionality of its word classes, each of which can regularly access a broad array of syntactic functions. Conversely, Hiw ranks relatively low on the scale of multicategoriality: most of its lexemes are assigned just one word class. This is how a language can be grammatically flexible, yet lexically rigid.

In constrast with Nêlêmwa (Oceanic, New Caledonia) whose lexemes are most generally subcategorised as nouns or verbs and undergo category-changing derivations, in Amis (Formosan), roots are pervasively categorially neutral, yet they contain semantic features and instructions that allow or disallow combination with primary derivational affixes which specify their class and category. Lexical categories are expressed after roots are derived into morphosyntactic words projected in a syntactic frame; they are then quite rigidly subcategorised as verbal, nominal or adjectival-modifying heads. Still, word forms display some functional flexibility; for instance, nouns and derived nouns, pronouns, numerals may be predicative in equative, ascriptive and focus constructions, simply by being in the syntactic position of the verb. Such functional flexibility is asymmetrical and does not apply to derived verb stems which must be nominalised to achieve argument function.

Hoava sa and ria have been analysed as pronouns in some contexts, and articles in others, an apparent case of flexibility in functional categories. However, this analysis depends on an assumption that pronouns are NP head. An alternative analysis employing the Determiner Phrase (DP) demonstrates that in all contexts sa/ria occupy the same syntactic position: DP head. They are always pronouns, alternating with articles in D, an analysis supported by evidence that 1st/2nd pronouns behave in an identical way. This unified analysis gives no grounds for positing membership of separate categories. In contrast, in Standard Fijian (SF) articles and pronouns occupy different syntactic positions: SF pronouns are not in D, but in N. The paper concludes that structures such as DP have considerable descriptive power; pronouns behave variably across Oceanic; and Hoava sa/ria are pronouns in all contexts. Their apparent flexibility was an artefact of earlier analyses, not a feature of the grammar.

Tongan is an Oceanic language belonging to the Polynesian subgroup. Based on previous work ( Churchward 1953 , Tchekhoff 1981 , Broschart 1997 ), Tongan has been classified as a 'flexible' language by various typological approaches on word classes ( Hengeveld 1992 , Rijkhoff 1998 , Croft 2001 ). This means that lexical items are per se not categorised in terms of major word classes, but they can function as noun, verb, adjective and manner adverb without morphosyntactic derivation. However, not all lexemes are entirely flexible occurring within all these constructions. So the crucial issue of how flexible Tongan really is remains. This question will be addressed by a survey based on a combination of syntactic and semantic word class criteria – basically following Croft's prototype approach ( 2000 , 2001 ) but also considering Hengeveld & Rijkhoff's work ( Hengeveld 1992 , Hengeveld, Rijkhoff & Siewierska 2004 , Hengeveld 2013 ) Evans & Osada's work (2005) . It reveals the scope of lexical flexibility for various lexemes and semantic groups.

Like many languages of the world, the Oceanic language Daakaka (Vanuatu) uses idiomatic combinations of body-part terms and verbs to express emotions, medical conditions and related concepts. However, languages differ in how they express the same concepts nominally. I will contrast the nominalization strategy found in Daakaka with other languages and discuss the differences. I will argue that the nominalization strategy in Daakaka is less transparent than its alternatives but that it allows for the formation of a paradigm that also includes meteorological expressions. This phenomenon highlights the need to look beyond individual lexemes when comparing lexical classes and derivational processes cross-linguistically. 1

Like other New Caledonian languages (see Ozanne-Rivierre 1998 : 33–34 for Nyelâyu; see Bril 2002 : 89–95, 2009 , this volume for Nêlêmwa; see also Moyse-Faurie 2004 : 15–61), Caac displays little categorial flexibility and, based on formal grounds, one can clearly identify two main syntactic categories: nouns and verbs, in addition to other small classes such as adverbs, adjectives or prepositions. Nouns, however, have the ability to be polyfunctional, and can function as the head of referential expressions as well as the head of predicative expressions in equative constructions, and in a certain type of presentative and spatial constructions, without undergoing any morphological change. By contrast, verbs require deverbal derivation in order to function as the head of referential expressions, a process mainly used for word creation purposes. There is in addition a small number of lexical bases which can function as the head of predicative and referential expressions indifferently. An analysis of the syntactic context in which they occur enables us to interpret them in a particular utterance. Similar lexemes in neighbouring languages have been analysed as flexible lexemes ( Bril 2009 : 2; in press ). In this paper, I would like to explore the extent to which those lexemes can be differentiated from nouns (notably indirectly possessed free nouns) and verbs in Caac, depending on whether one puts the emphasis on formal or semantic criteria.