Dealing with variability in loanword integration
This chapter tests a strong loanword integration hypothesis: that donor-language material that has been borrowed will display variability in morphosyntactic integration paralleling that of the recipient language. This requires explicitly marshalling the recipient language as the benchmark for comparison, an innovation implemented here for the first time. Illustrating with the typologically different Tamil-English language pair, word order and case-marking of English-origin objects of Tamil verbs are analyzed. English indirect objects are overwhelmingly inflected with Tamil dative markers, but direct objects tend not to be marked for the accusative. Comparison reveals that this patterning reflects the case-marking variability inherent in the recipient-language benchmark, compelling us to recognize even these apparently bare forms as borrowed, and supporting the Nonce Borrowing Hypothesis. This demonstrates that the facts of variability must be taken into account to identify which forms have been borrowed and which have been code-switched.