Timothy Usher, Santa Fe Institute
The Trans New Guinea family was proposed by McElhanon and Voorhoeve (1970,) and has since been broadly accepted, though its evidentiary basis had never been well-defined. A number of disparate lines of reasoning have been used over the years to include or disinclude particular language families in Trans-New Guinea, resulting in opacity and confusion.
Awaiting the conclusion of lexical comparison between its reconstructed subfamilies, the most compelling evidence for Trans-New Guinea is the formal resemblances between the personal pronouns of its members. While many or most of McElhanon and Voorhoeve's lexical comparisons are specious, the significance of the pronominal comparisons is difficult to deny, particularly when it's appreciated that their bases, when prefixed, very often share the common functions of indicating both the inalienable possessor of a nominal and the undergoer of a verb.
Ross (2005, Pawley, Ross and Osmond 2005) has proposed a substantially reduced version of Trans New Guinea upon the basis of the presence or absence of these pronouns, an approach previously used to succesfully delineate its westernmost subfamilies from the unrelated families of the West Papuan Phylum (Capell 1975, Voorhoeve 1975.) In the eastern portions of New Guinea, Ross retained a number of subfamilies which do not clearly display these characteristic pronominal patterns. It is possible that some of these families do indeed descend from Proto-Trans New Guinea, and their pronouns have merely changed or have undergone regular phonological changes which have not been recognized, but this remains to be demonstrated by some other form of evidence (i.e. comparisons in the basic lexicon) which Ross does not attempt.
Another vein of comparison is found in the vocalic gender suffixes of South Bird's Head, Ok and Fly River, which mandate a common origin. Like the foundational pronoun chain, this is unimpeachable evidence, but here it's not clear if these reflect proto-Trans New Guinea or one of its higher-order subgroups.
Among the immediate purposes of this project is to establish a lexical foundation for Trans New Guinea within a framework of regular phonological correspondences. As this is currently underway, the delineation of Trans New Guinea adopted here is provisional, and subject (most probably) to expansion.