This paper extends a notion of local grammars in formal language theory to autosegmental representations, in order to develop a sufficiently expressive yet computationally restrictive theory of well-formedness in natural language tone patterns. More specifically, it shows how to define a class ASL$$^g$$g of stringsets using local grammars over autosegmental representations and a mapping g from strings to autosegmental structures. It then defines a particular class ASL$$^{g_T}$$gT using autosegmental representations specific to tone and compares its expressivity to established formal language grammars that have been successfully applied to other areas of phonology.

中文翻译：

---

自分句语法的表现力

本文将形式语言理论中的局部语法概念扩展到自动分段表示，以便在自然语言音调模式中开发一种表达能力足够强但计算限制性好的理论。更具体地说，它展示了如何在自动分段表示上使用本地语法以及从字符串到自动分段结构的映射 g 来定义字符串集的类 ASL$$^g$$g。然后，它使用特定于声调的自动分段表示定义一个特定的类 ASL$$^{g_T}$$gT，并将其表达能力与已成功应用于其他音系领域的已建立的形式语言语法进行比较。