概率与音系——兼述Goldsmith模式

Probability and Phonology

概率音系学通过研究语音的概率分布来研究音系学，是音系学的新流派。通过对单词部分的概率统计数据，我们可以推断出该单词的合格性。概率统计数据帮助语言学家预测和解释音系模式。Goldsmith提出的单构词成分模式和双构词成分模式，比较有效地解决了语音配列问题。概率的比较在科学推断和语言学习中起着十分重要的作用。

Probabilistic phonology is a new approach to phonology, which studies phonology with the help of probability distributions. The well- formedness of words can be inferred by the statistical data of probability on lexicon. The statistical data of probability can assist linguists to predict and explain phonological patterns. The unigram model and bigram model put forward by Goldsmith mange to solve the problems in phonotactics to some extent. Comparison of probabilities plays a crucial role in scientific inference and language learning.