In Head-driven Phrase Structure Grammar (HPSG), the phonological module has been strikingly underinvestigated compared to other parts of the grammar. To remedy this situation, the aims of this study are both theoretical and empirical. We employ revised Typed Feature Structures (TFSs) to describe a number of phonological structures in Mandarin Chinese, mostly concerned with syllable structure and tone, based on the X-bar approach to the syllable structure of Mandarin (van de Weijer, 2012, van de Weijer and Zhang, 2008, Zhang, 2008). Specifically, we integrate tone features in the HPSG framework, and analyze tone sandhi (TS) by way of TFSs and constraints. We demonstrate that the enriched definitions and constraints using TFSs in phonology, like in morphology and syntax in HPSG, can successfully capture the relevant phonological generalizations and are therefore a step on the way toward a more comprehensive HPSG account of Mandarin grammar, illustrating its advantages such as simplicity, constructional consistency, modularity and facilitating interface description.

在 Head-driven Phrase Structure Grammar (HPSG) 中，与语法的其他部分相比，语音模块的研究明显不足。为了纠正这种情况，本研究的目的既是理论的，也是实证的。我们采用修改后的类型特征结构 (TFS) 来描述普通话中的许多语音结构，主要与音节结构和声调有关，基于对普通话音节结构的 X-bar 方法（van de Weijer，2012，van de魏杰和张，2008，张，2008）。具体来说，我们在 HPSG 框架中集成了音调特征，并通过 TFS 和约束来分析音调变奏（TS）。我们证明了在语音学中使用 TFS 的丰富定义和约束，如 HPSG 中的形态学和句法，