Uncovering Mandarin Speaker Knowledge with Language Game Experiments

Abstract

Mandarin Chinese offers many intriguing puzzles for linguists because it has a shortage of morphophonological alternations. This has resulted in indeterminacy in various aspects of its phonological grammar, triggering much debate on syllable structure and allophonic mapping. The ambiguity of the data is also a problem for children acquiring Mandarin since alternative grammars can account for the surface forms equally well.

In order to find out what Mandarin speakers have learned about the phonology of their language, I conducted two language game experiments based on fanqie secret languages. It was found that markedness and faithfulness constraints are psychologically real for Mandarin speakers. Furthermore, the interactions between markedness and faithfulness constraints are shown to have an effect on glide movement in the language game. In addition, much speaker variation was observed in the experiment. I demonstrate that it is the result of constraint ranking variation. Nevertheless, general population-level trends on constraint ranking could still be identified. These trends lead to insights on phonological learning beyond Mandarin, showing evidence for naturalness bias and lexicon optimization.