Volume 9, Issue 3

Mundari, an Austroasiatic language of India (Munda family), has often been cited as an example of a language without word classes, where a single word can function as noun, verb, adjective, etc. according to the context. These claims, originating in a 1903 grammar by the missionary John Hoffmann, have recently been repeated uncritically by a number of typologists. In this article we review the evidence for word class fluidity, on the basis of a careful analysis of Hoffmann’s corpus as well as substantial new data, including a large lexical sample at two levels of detail. We argue that in fact Mundari does have clearly definable word classes, with distinct open classes of verb and noun, in addition to a closed adjective class, though there are productive possibilities for using all as predicates. Along the way, we elaborate a series of criteria that would need to be met before any language could seriously be claimed to lack a noun-verb distinction: most importantly strict compositionality, bidirectional flexibility, and exhaustiveness through the lexicon.

John Peterson There’s a grain of truth in every “myth”, or, Why the discussion of lexical classes in Mundari isn’t quite over yet   Kees Hengeveld and Jan Rijkhoff Mundari as a flexible language   William Croft Word classes, parts of speech, and syntactic argumentation

1. Introduction Our three commentators raise such a host of deep and interesting issues that we cannot hope to answer them all within the time and space at our disposal. To begin with, we would like to thank them for pushing us to articulate the reasons for our arguments more clearly, and for getting us to spell out a number of assumptions and intermediate argumentative steps that we did not make sufficiently clear or explicit in our original article. We will deal with the three commentators one by one, limiting ourselves to a few key points.

Delocutive verbs can be defined as verbs derived from a base X which mean ‘by saying or uttering “X” (to someone) to perform an act which is culturally associated with the meaning or force of X’, where X is a variable ranging over types of things that can be said or uttered – 2nd person pronouns and other terms of address, words for asking and answering questions, formulaic expressions for social acts like greetings, various kinds of expressives, characterizations of speech peculiarities. Although originally identified as such in, and illustrated exclusively from, Indo-European languages by Debrunner (1956) and Benveniste (1958), delocutives are not confined to this family, but show a wide genetic and areal spread. The aim of this paper is to delineate the systematic possibilities for crosslinguistic diversity and for historical change in delocutive formations, and in particular to relate derivational delocutives to equivalent syntactic constructions. In such a wider typological and diachronic view, delocutives are seen not to be cases of ordinary quotation, nor a rare peculiarity at the margins of ordinary word formation, but to be one variation on the theme of complex predicates, instructively bearing on the general question of where verbs can come from. Their closest affinities, synchronic and diachronic, are to predications of existential causation ( doing/making , often found to subsume saying ).