What would probabilistic models predict: Experiments on locative inversion in Mandarin Chinese
Yuting Li
Department of Philosophy and Humanities, Freie Universitt Berlin, Germany
The question of whether human grammatical knowledge is better explained by a cat- egorical grammar or a probabilistic grammar has been a topic of debate among linguists, psy- chologists, and cognitive scientists for decades. Based on a corpus analysis of locative inversion in Mandarin Chinese using probabilistic models, this research conducted two experiments that provide straightforward evidence in favor of the idea that human grammatical knowledge is prob- abilistic in nature. Locative inversion, characterized by a preposed locative and a postposed NP that functions as the subject in the canonical sentence, has been documented in various languages (e.g.,English,Chichewˆa,Chinese),consider Mandarin Chinese:
The first experiment examined whether corpus probabilities of locative inversion can predict sentence ratings. The result showed that there is a positive correlation between corpus probabil- ities and acceptability ratings, suggesting participants are sensitive to the likelihood of certain linguistic patterns. The second experiment was to investigate the hierarchy of weights for the in- dividual factors used in corpus probability pattern. Following up Rosenbach (2003), we designed a 2 × 2 × 2 experiment with newness (new/given), definiteness (indefinite/definite) and heavi- ness (long/short) as independent factors, the resulting 8 conditions are placed in a way as to as- sume that newness is more important than definiteness and heaviness in a sequence. The results confirmed the existence of the hierarchy, which was consistent with the predictions made by the corpus model. These two experiments are encouraging as a response to what linguistic general- izations can be abstracted from a model, an issue addressed in Francis (2021) and many others.
References
- J. Bresnan. Is syntactic knowledge probabilistic? Experiments with the English dative alterna- tion, pages 75–96. De Gruyter Mouton, Berlin, New York, 2007.
- J. Bresnan, A. Cueni, T. Nikitina, and R. H. Baayen. Predicting the dative alternation. In Cogni- tive foundations of interpretation, pages 69–94. KNAW, 2007.
- E. J. Francis. Gradient acceptability and linguistic theory, volume 11. Oxford University Press, 2021.
- C.-T. J. Huang. Existential sentences in chinese and (in) definiteness. The representation of (in) definiteness, 226:253, 1987.
- A.Rosenbach. Aspectsoficonicityandeconomyinthechoicebetweenthes-genitiveandthe of-genitive in english. Topics in English Linguistics, 43:379–412, 2003.