Where are the Papuan languages?

The easternmost province in Indonesia is Papua, but not all of the languages within its borders are Papuan, according to most linguists. While the languages of most of Indonesia belong to the Austronesian family, most (but not all) of the languages of Papua (and Irian Barat) are not members of this family. At the same time, not all of the languages located west of New Guinea are Austronesian: there are pockets of languages in north and south Maluku, Nusa Tenggara Timor and (historically) Nusa Tenggara Barat, that cannot be assigned to the Austronesian family. The term ‘Papuan ’ is often used to refer to these languages that are not Austronesian, but (almost) everyone who uses the term in a linguistic context acknowledges that it is a negative category, not a positive one: a language is said to be ‘Papuan ’ if it is spoken in the area near New Guinea, and is neither Austronesian nor Australian.1 On the other hand, there is common (albeit mistaken, to at least some degree) agreement about the existence of a ‘Papuan ’ linguistic type: a number of features have been proposed as ‘typical ’ of ‘Papuan ’ languages, such as SOV word order, a predominance of serial verb constructions and switch reference morphology; agglutinative morphology and case marking; and a reduced number of phonemic liquids, with complex syllable structures and phonemic tone or stress. Haiman (1980), Reesink (1987, 2000), de Vries (1993, summarising them), Donohue (1997, 2005a, 2006), and especially Foley (1998, 2000) all discuss some range of grammatical properties that are found in a number of languages of New Guinea, especially in the highlands (and middle Sepik). But many of these features are not (even in a south-west Pacific context) uniqu