Moving along the grammaticalisation path: Locative and Allative marking of non-finite clauses and secondary predications in Australian languages

Abstract

This paper examines three grammaticalised instances of local case-marking in Indigenous languages of northern Central Australia, coded as follows: (AN) Allative expressing object control on NP’s - the use of allative instead of locative in the meaning of ‘locative’ in a secondary predication where the subject of that predication has the same reference as an object or oblique in the main predication; (LS) Locative marked subordination the use of locative case-marking on the verb and other elements to mark types of non-finite subordinate clauses; (AS) Allative marked subordination expressing object control, combining features of A and LS, usually where LS is also present. The different distributions of these properties are plotted in a number of Central and northern Australian languages to provide a picture of current distribution and hypotheses about the origin of these constructions. The hypothesis proposed here is that the AN construction arose in a group of Pama-Nyungan languages in a restricted area of North-central Australia and partially overlapped with the presence of the LS construction in a wider grouping of Pama-Nyungan languages. This cooccurrence produced the AS construction, which subsequently diffused to a few neighbouring languages, including some Non-Pama-Nyungan languages. Both AN and AS can be called grammaticalisation since they depart from the semantic functions of the locational cases to mark control phenomena between predications. The marking of subordinate clauses by local cases, and particularly locative case, is relatively common cross-linguistically outside Australia and arguably maintains some cognitive metaphorical link with static location and motion, perhaps primarily through the near-universal ‘space=time’ metaphor. In AN, we see a much rarer development in which the metaphorical link between the concrete local meaning and the grammatical function is attenuated, although the possibility that AN involves ‘fictive motion’ as in Finnic languages is discussed.