国内对指令言语行为（包括祈使句）研究已有不少成果，包括按照说话人的支配权或权势， 以及语句的语气等要素进行内部分类和界定。不过在实际交际中并不是所有的指令言语行为都根据 这几条因素决定其形式的，有大量违反以往研究结论的指令表达。因此，从单一标准对指令言语行 为进行分类和判断是不合理的。本文基于原型范畴理论，根据更为复杂的、有机组成的、多层次的 语境因素来考察指令言语行为范畴的选择机制，包括说话人受到的恒定因素和可变性因素，两种因 素的综合影响，决定了指令行为选择的范围大小，以及具体的形式的确定。

The studies on Speech Acts of Directives (and the traditional study of Imperative Sentences) of local scholars have achieved some valuable results, including the definition and the inner classification of Speech Acts of Directives based on the speaker’s authority and power, the tone of sentence. However, it is observed that in speech communication, a directive sentence is given by the speaker based on some other influences besides those factors, which is against the former studies. Therefore, there should be a more complicated mechanism in the formation of a speech act of directives, not just formed by an individual factor. This article examines the selection process of the directive speech acts based on the theory of prototypic categories, and take more complicated, organic organized contexts into consideration. The context factors can be divided into constant factors and changeable factors. The combination of the two factors decides the range of selection of directive speech acts and the final form of output of the speech act.

1. 引言
2. 指令言语行为与原型范畴
1) 原型范畴理论
2) 指令言语行为范畴成员的界定
3. 指令言语行为表达形式的选择机制
1) 指令言语行为交际的语境因素
2) 指令言语行为表达形式的生成
4. 结论

• 金宝兰(南京大学 文学院 语言学及应用语言学 博士研究生) | 김보라