本文基于互动视角将附加问句“X不X”的非疑问功能统摄起来，从与始发句的搭配互补分布的知域类“X₁不X₁”（是不是、对不对）和言行域类“X₂不X₂”（好不好、行不行）出发，通过观察口语会话语料，首先描写“X不X”交际中的非疑问功能类型及形式特征；其次对比分析各项功能在日常、访谈等非机构性会话以及在教学、谈判等机构性会话中的分布差异，获取各项功能在不同会话中的优势分布以及“X不X”的两类个案（是不是VS.对不对，好不好VS.行不行）的使用差异。之后，基于立场理论，归纳附加问句“X不X”的核心功能，并总结非疑问功能的总体框架；辅以对比附加问句“X吧”“X吗”，进一步明确“X不X”格式的功能特性。最后综合探讨影响“X不X”非疑问功能浮现的互动因素，概括出其倾向性的对话模式，以应用于对外汉语教学。

       从功能类别和形式特征来看，知域类“X₁不X₁”的非疑问功能主要表现为引发话题、寻求认同（含观点类和行事类）、邀请反馈和强化语气，言行域类“X₂不X₂”的非疑问功能主要为寻求同意（含请求、提议和命令等行为）、邀请反馈以及强化语气。综合来说，知域类和言行域类“X不X”功能特性上同中有异，相同点是二者的后两项功能相似，且都有行事类的“求同”功能。功能类型相似，话语分布也相近。相异点主要在功能的细微差异上：1）行事类的求同功能：前者有很强的语境依赖性，后者被识解的稳定性更强，没有明显的语境依赖；2）邀请反馈功能：前者是在陈述类、提议类、评价类始发句基础上，侧重于言听双方的言语互动，后者则是在祈使类表达后发出，更包含言后的行为反馈；3）强化语气功能：相比起前者，后者多一层制止、结束对立性会话的祈使效果。

       从不同会话类型的分布情况来看，首先，同一功能在不同话语类型中的分布存在差异。“X₁不X₁”在日常、访谈中以寻求认同、引发话题功能为优势分布；教学中则邀请反馈与师生互动特点更契合；谈判中功能总体的使用频率偏低。“X₂不X₂”的寻求同意功能在四类会话中均占优势；另外访谈中邀请反馈也较常见；教学中表命令的求同功能可出现；谈判中双方立场对立，强化语气功能也能浮现。总体上“X不X”的功能分布对互动场景比较敏感，较多用于非机构性会话。其次，“X不X”两类个案内部在用法上也存在差异。知域类中，比起“对不对”，“是不是”表引发话题和强化语气时更常见。言行域类中，“好不好”在各功能中均占多数。个案内部的分布差异可能与词汇义、语境特点等有关。

        总体来看，无论是知域类还是言行域类的“X不X”，附加问句均是以期待构建立场认同/一致为最基本的互动出发点，其核心功能是寻求（观点）认同或（行为）同意，而其他的非疑问功能则是在这两类功能基础上分别发展而来。此外，比起同样位于附加问句位置的“X吧”“X吗”，“X不X”有其功能特性，如在表达行事类的求同、强化语气功能上“X不X”更为典型，且更容易被听话者识解成带有言者的负面情态。

        影响附加问句“X不X”非疑问功能浮现的互动因素主要是社会关系中的权位关系、熟悉程度、言者认知/情感立场以及话语类型。“X不X”非疑问功能的分化属于语用变体而非语法化产物，可以提炼出不同互动情境下的倾向性话语模式，以辅助于对外汉语教学。

The paper studies the non interrogative function of the tag question “X bu X” from the perspective of interaction. Starting from the knowledge domain category “X₁ bu X₁” (shi bu shi or dui bu dui) and the speech and action domain category “X₂ bu X₂” ( (hao bu hao or xing bu xing), the paper first describes their non interrogative functional types and formal features by observing the oral corpus, and then compares and analyzes the distribution differences of the functions in non institutional conversations (sunch as daily and interview) and in institutional conversations (such as teaching and negotiation), so as to obtain the advantage distribution of the functions in different conversations, and the differences in the use of “X bu X” cases. Then, based on the Stance Triangle Theory, the paper summarizes the core functions of “X bu X”, and summarizes the overall framework of its non interrogative functions. By comparing the tag questions “X ba” and“X ma”, the functional characteristics of “X bu X” are further clarified. Finally, the paper discusses the interactive factors that affect the emergence of “X bu X” non interrogative function, and summarizes its tendentious dialogue mode for TCFL.

From the perspective of functional categories and formal features, the non interrogative functions of “X₁ bu X₁” are mainly manifested as initiating topics, seeking approval (including opinions and actions), inviting feedback and strengthening tone, while the non interrogative functions of “X₂ bu X₂” are mainly manifested as seeking approval (including requests, proposals and orders), inviting feedback and strengthening tone. On the whole, there are similarities and differences in the their functional characteristics. The similarities are that the latter two functions of the two are similar, and both of them have the function of “seeking common ground” of action class. The differences mainly lie in the subtle differences of functions: 1) In the similarity seeking function of action class: the former has strong context dependence, while the latter is more stable to be understood. 2) In invitation feedback function: the former focuses on the verbal interaction; the latter includes the behavioral feedback after the speech. 3) In strengthening the mood function: compared with the former, the latter has more imperative effect of stopping and ending the opposite conversation.

From the distribution in different conversations: first, there are differences in the distribu- -tion of the same function in different conversations. “X₁ bu X₁” has the advantage of seeking identity and initiating topic function in daily and interview; In teaching, the function of inviting feedback is more; In negotiation, the frequency of all functions are low. The consent seeking function of “X₂ bu X₂” is dominant in all four types of conversations. In addition, it is common to invite feedback in interviews; The function of seeking common ground in teaching also exists; The function of strengthening mood can also emerge in negotiation. On the whole, these functional distributions are more sensitive to interaction scenarios and are mostly used in non institutional conversations. There are also differences in the usage of “X bu X” in the two cases. Compared with “dui bu dui”, “shi bu shi” is more common in arousing topic and strengthening tone. And “hao bu hao” is in the majority of all functions. The differences in the distribution of individual cases may be related to lexical meaning and contextual characteristics.

Generally speaking, the most basic interaction point of tag question “X bu X” is to expect to construct a consistent stance. Its core function is to seek (opinion) agreement or (behavior) approval. Other non interrogative functions are developed on the basis of these two functions. In addition, compared with “X ba” and “X ma”, which are also located in the adjunct question position, “X bu X” has its own functional characteristics. For example, it is more typical in the function of seeking common ground and strengthening mood, and it is also easier for the hearer to recognize it as a negative modality with the speaker.

The interactive factors that influence the emergence of “X bu X” function are mainly related to the social / intellectual power relationship between the speaker and the hearer, the familiarity between the speaker and the hearer, and the speaker’s cognitive / emotional stance. The differentiation of its non interrogative functions is the product of pragmatics. To sum up, a kind of discourse model can be summed up as to be applied to TCFL.