A general issue in pragmatics concerns the definitions of speech act (SA) types. Cognitive linguists agree that a directive SA involves a speaker exerting a force towards her addressee’s (A) performance of some action, and the subtypes of directives have been approached in terms of a metaphorical grounding based on force image-schemas. These idealized cognitive models include graded features, the values and the centrality of which differ across directive subtypes. I address the relationship between the form of utterances used as directives and the ontology of directives, and I discuss recent experiments supporting a view of SA s as graded categories. I show that these approaches enable adopting an empirically adequate distinction between the levels of pragmatic meaning and semantic meaning, which raises interesting possibilities for further experimental work on speech act recognition in cognitive linguistics.

语用学中的一个普遍问题涉及言语行为（SA）类型的定义。认知语言学家认为，一个指令SA涉及扬声器向她施加力收件人（一个有所行动的）性能和指令的亚型基于力图像模式提供一种隐喻接地方面进行过接触。这些理想化的认知模型包括分级特征，这些特征的值和中心性在指令亚型之间有所不同。我讨论了用作指令的言语形式与指令的本体之间的关系，并讨论了最近的支持SA观点的实验 s作为分级类别。我表明，这些方法能够在实证意义和语义意义的层次之间采用经验上适当的区分，这为在认知语言学中进行语音行为识别的进一步实验工作提出了有趣的可能性。