The ‘say, do’ verb in Nyulnyul, Warrwa, and other Nyulnyulan languages is monosemic

A well-known feature of many Australian Aboriginal languages is that the same verb form can mean either ‘say’ or ‘do’ (Rumsey 1990; Dixon 2002: 121). The received story is that this represents a polysemic verb with two distinct though not unrelated meanings. Few linguists, however, have presented arguments for this position. An exception is Knight (2004), who presents a set of arguments couched within the theoretical and methodological framework of Natural Semantic Metalanguage (NSM) that the inflecting verb -MA in Bunuba (Bunuban, Kimberley) is polysemous. Many of these arguments apply to other Australian languages, including languages of the nearby Nyulnyulan family. In this paper I focus on the two Nyulnyulan languages that I have studied most intensively, Nyulnyul (Western Nyulnyulan) and Warrwa (Eastern Nyulnyulan), and discuss the range of meanings and uses of the inflecting verb -J(I) ~ -D(I). These include the senses ‘say’, ‘do’, ‘think’, among a number of more specific senses. I present arguments for a radically monosemic account (e.g., Ruhl 1989; McGregor 1997) of this verb in its lexical uses, arguing in the process that the type of evidence adduced by Knight (2004) does not in fact argue for polysemy. I argue, that is, that a single core meaning is associated with the inflecting verb -J(I) ~ -D(I), and the senses ‘say’ and ‘do’ represent contextual senses or interpretations engendered by the context of use. A consequence of my approach is that the semantic meaning of lexical verbs is more abstract than often thought; indeed, it is so abstract that it is impossible to distinguish between the semantics of lexical and grammatical items in terms of the degree of abstraction.