Algorithms for acquiring phonological features in Substance-Free Phonology and Emergent Feature Theory

大多数音系学理论认为，语言能力的语音成分是基于语音特征的固有单位，特征的声音和/或发音内容也是与生俱来的。这是基于规则的理论中的一个基本假设，例如SPE(Chomsky&Halle，1968)和后来的公式，其中特征被赋予了层次结构，例如特征几何(Clements，1985)，以及从最优化理论开始的基于约束的理论(Prince&Smolensky，1993)。然而，最近的研究对这些假设提出了挑战，例如，无物质音位学认为自然特征的语音内容与语音计算无关，以及涌现特征理论认为特征的内容是后天获得的，而不是与生俱来的。我的研究兴趣在于这一领域，即语音结构可以习得的程度，以及习得本身的认知机制如何影响这些结构的性质。

Most phonological theories have held the position that the phonological component of thelanguage faculty is based around the innate unit of the phonological feature, and that the acousticand/or articulatory content of features is also innate. This is a foundational assumption in rule-basedtheories such as SPE (Chomsky & Halle, 1968) and later formulations in which features are givenhierarchical structure, such as Feature Geometry (Clements, 1985), as well as constraint-basedtheories starting with Optimality Theory (Prince & Smolensky, 1993). However, recent approacheschallenge these assumptions, such as Substance-Free Phonology, in which the phonetic content ofinnate features is irrelevant to phonological computation, and Emergent Feature Theory, whichclaims that the content of features is acquired, rather than innate. My research interest is in thisarea, namely the extent to which phonological structures can be acquired, and how the cognitivemechanisms of acquisition itself might influence the nature of those structures.