Closing puzzle: Indefinites in Maori (Chung and Ladusaw 2004)

In chapter I, we advanced the idea that an indefinite noun phrase can be semantically combined as an argument to a predicate by two different modes of composition: Restrict and Specify. Restrict composes the property supplied by the indefinite with the predicate, limiting the semantic domain of the predicate but leaving it unsaturated. Specify shifts the property to an individual, which then saturates the predicate's argument, rendering the predicate (more) semantically complete. Although these operations often lead to results that are truth-conditionally equivalent, we believe that the difference in composition is linguistically significant. In particular, our theory of restriction and saturation makes an empirical prediction that we explore in this chapter. If indefinites and other semantically incomplete noun phrases can indeed be composed via different operations, then one might expect that the morphosyntax could explicitly reflect which composition operation is employed. A language could, for instance, use distinctive determiners or distinctive inflectional morphology to flag whether a given indefinite should be composed via Restrict or via Specify. We show here that this prediction is borne out in remarkable detail in Maori, a Polynesian languag of New Zealand. Maori has two articles, he and tetahi (plural etahi), that are widely recognized to be indefinite in some sense.1 These articles contrast, but not truth-conditionally, in sentences like the following (see appendix B for a list of abbreviations used in the Maori examples):