The role of distributional factors in learning and generalising affixal plural inflection: An artificial language study

Inflectional morphology has been intensively studied as a model of language productivity. However, little is known about how properties of the input affect the emergence of productive affixation. We examined effects of three factors on the learning and generalisation of plural suffixation by adults in an artificial language: affix type frequency (the number of words receiving an affix), affix predictability (based on phonological cues in the stem), and diversity (the number of distinct phonological cues predicting an affix). Higher type frequency and predictability facilitated the acquisition of trained inflections. Type frequency contributed to participants’ inflections of untrained words early during learning, while reliance on diversity emerged gradually, alongside knowledge of phonological cues. Diversity as well as type frequency contributed to the emergence of default-like inflections, including minority defaults. The results elucidate the role of affix diversity and its interaction with other factors in the emergence of productive linguistic processes.

屈折形态学（Inflectional morphology）作为语言能产性的研究模型已得到深入探讨。然而，学界对于输入特征如何影响能产性词缀化的产生仍知之甚少。本研究考察了三类因素对成人学习者在人工语言中复数后缀化的学习与泛化表现的影响：词缀型频率（即接收某一词缀的词汇数量）、词缀可预测性（基于词干的语音线索）以及多样性（预测某一词缀的不同语音线索的数量）。较高的词缀型频率与可预测性可促进已训练屈折形式的习得。在学习初期，词缀型频率会推动学习者对未训练词项的屈折处理，而对多样性的依赖则会伴随语音线索知识的积累逐步显现。多样性与词缀型频率共同促成了类默认屈折形式的产生，其中包括少数派默认形式。本研究结果阐明了词缀多样性及其与其他因素的交互作用在能产性语言过程产生中的作用。