This paper addresses the questions, Do bilingually induced and shift-induced change have different outcomes? If they do, can these differences assist us in reconstructing the prehistoric past, specifically the linguistic prehistory of the (smallscale neolithic) societies of Melanesia. A key to better interpreting differences in the outputs of contact-induced change is to understand how such change in smallscale societies actually occurs. I argue that it is important to know the life-stage loci of change. I suggest that language shift has two life-stage loci, one in early childhood, where evidence of shift, if any, is restricted to specialist lexicon, and one in adulthood. Adult language shift appears to have been rare in Melanesia. I also suggest that bilingually induced change, which entails the syntactic restructuring of one's heritage language on the model of a second language, takes place among preadolescent children-a claim which is supported by various kinds of evidence. This understanding helps us in turn to interpret the outcomes of contact-induced change and to infer prehistoric events, since adult second-language learning typically leads to simplification, whilst childhood language learning may lead to an increase in complexity.