关键词: 语用推理, 非外显意义, 预设, 理性言语行为模型, 得体性, 触发标记, 共同背景

Abstract:

Presupposition refers to the non-explicit assumption or belief held by both the listener and the speaker (the “common ground”). When encountering a presupposition message, the listener is required to infer what the speaker implies from the specific linguistic marker (or presupposition trigger) and its constrained object (or computational point). For instance, the sentence “Zhang Ming published a paper_{computational point} again_trigger” generates a presupposition “Zhang Ming published a paper before”. The listener relies on the trigger to access the common ground of both sides of the communication, and infers the presupposed content based on the computational point; subsequently, the listener relates the generated presupposition to the common ground and then updates their mental representation. These processes may be modulated by the word order between the presupposition trigger and the computational point, which would not change the critical role of the presupposition trigger in generating presupposition, but may affect the difficulty of presupposition processing.
Presupposition comprehension was often considered as involving the processing of pragmatic felicitousness. According to whether the presupposition generated by the listener is consistent with the common ground, the processing of felicitousness can be classified as “presupposition satisfaction” and “presupposition violation”. In the former, the presupposed content is consistent with common ground, while in the latter, the presupposed content is inconsistent with the common ground. According to whether the listener can rationalize the violated presupposition, the presupposition violation is further divided into “presupposition failure” and “presupposition accommodation”. Presupposition failure refers to the understanding person’s lack of ability to rationalize the inconsistent presupposed content, let alone integrate it into an existing mental representation. Presupposition accommodation means that the inconsistent presupposed content can be rationalized. The mental model can be updated and reconstructed. This categorization lays the foundation for investigating the cognitive processes of presupposition comprehension.
Numerous studies have demonstrated that the perceived common ground has an immediate impact on presupposition trigger and computational points. Accordingly, we propose a two-stage processing hypothesis regarding the contextual effect on presupposition comprehension. Specifically, in the first stage, the listener integrates the common ground with the presupposed meaning at the presupposition trigger; in the second stage, the concrete presupposed content is integrated with the common ground in the preceding context. Different processing stages may involve distinct cognitive processes. Furthermore, we reviewed several factors affecting the processing of presupposition at different cognitive stages, such as the linguistic types of the trigger, the semantic relatedness between the common ground and the presupposition, the forms of common ground presented, and the level of involvement in the experiment. Firstly, these triggers introducing similar and specific event structures may lead to similar cognitive efforts. Secondly, when integrating the common ground with the presupposed content, the listener does not simply judge whether the common ground and presupposed content are matched but carries out a more fine-grained processing according to the degree of matching between them. Therefore, the difficulty of integrating the two components may be affected by the degree of matching. Thirdly, the common ground of interlocutors can be established by linguistic co-presence (linguistic materials presented to the listener), visual co-presence (visual scene presented to the listener), or general world knowledge/community membership (common sense formed by the community). These manipulations differed in the modality of providing common contextual information and in the memory (such as short-term memory and the long-term memory) for retrieving these common grounds. Therefore, these specific cognitive processes are different when the listener integrates different types of common grounds and presupposed content. Finally, different experimental paradigms, such as those facilitating reading and those demanding interpersonal interaction, which demand different levels of motivation and engagement, may affect the extent of presupposition processing.
Future studies can explore the cognitive basis of presupposition processing from the following three perspectives: (1) using computational modeling to quantify the processes (such as perspective-taking) of the listener’s understanding of presupposition during language communication; (2) using brain imaging to reveal the neural basis of presupposition processing; (3) to validate and, when necessary, to modify the neurocognitive models of presupposition processing with data from special populations.

Key words: pragmatic inference, non-explicit meaning, presupposition, rational speech-act model, felicitousness, trigger, common ground