William Thurston (1982, 1987, 1989, 1992, 1994) analyzes the history of the languages of the northwest area of New Britain. This history has included much contact among the area's languages, all of which are Oceanic Austronesian with the exception of the Papuan language Anem. Thurston, however, took the position that all linguistic speciation is brought about by language contact, especially by language shift. In this paper, the comparative method is applied to Thurston's (and others') data to reconstruct a partial history of the languages of the area, exemplifying how the comparative method may be applied in contact situations. Reanalysis of his data shows that a number of his conclusions about the histories of the area's Austronesian languages are wrong, but validates his claim that language shift is manifested in copied specialist vocabulary, a conclusion that is important for historical contact linguistics, as such cases may provide few or no other clues that shift has occurred.