Agentivity, Animacy and Telicity: Pragmatic Inferences in Intransitive Clauses

和许多其他语言一样，德语不及物动词分为非宾格动词，如entkommen（逃避）和非否定动词，如reden（谈话）。在以前的研究中，非宾格被认为是指位置或状态的变化和选择病人角色的远程动词；非否定动词被分析为具有代理作用的动态动词。因此，在分裂不可及性中，能动性和telicity呈负相关。不及物动词的助动词选择和钝化是紧密联系在一起的：BE是可以接受的，而非宾格动词的HAVE和钝化是不可接受的。相反，BE是不可接受的，而HAVE和钝化是可以接受的。然而，能动性（通过animacy测试）和telicity之间假设的负相关关系以及辅助选择和非人格化之间假设的密切匹配是值得怀疑的。我们将系统地探讨这些问题，重点关注不及物动词，这些不及物动词在能动性、遥性和辅助选择方面是灵活的，比如Ger。schwewen是“漂浮，悬停”，schwimmen是“游泳，漂浮”。有了这些动词，能动性、动性和目的性可能会根据上下文变化，并可能出于实验目的而被操纵。本项目将结合快速可接受性判断和电生理数据（ERP）等先进的实验方法，为“xprg .de：基于实验证据的新语用理论”的主要目标做出贡献，以探索之前被忽视的语境信息对分割不可及性现象的影响问题。在更一般的层面上，我们讨论了语用学在语言架构和语言处理中的作用，正如Schumacher（2012, 2013）的句法-话语处理模型所设想的那样。根据该模型，语境依赖信息在语言加工过程中与语法信息在两种截然不同的操作中以增量方式交互使用。这个模型与最近的性能兼容的语法方法是一致的，其中不同类型的信息可以相互作用。

As in many other languages, German intransitive verbs are split into unaccusatives such as entkommen 'escape' and unergatives such as reden 'talk'. In previous research, unaccusatives are assumed to be telic verbs that denote a change of location or state and to select a patient role; unergative verbs are analysed as atelic verbs with an agent role. Thus, agentivity and telicity are inversely correlated in split intransitivity. Auxiliary selection and passivization of intransitive verbs are closely linked as follows: BE is acceptable, while HAVE and passivization are unacceptable with unaccusatives. In contrast, BE is unacceptable, while HAVE and passivization are acceptable with unergatives. However, the hypothesized inverse correlation beween agentivity (tested via animacy) and telicity and the assumed close match between auxiliary selection and impersonal passivization are questionable. We will explore these issues systematically by zooming in on intransitive verbs that are flexible with respect to agentivity, telicity and auxiliary choice, such as Ger. schweben 'float, hover' and schwimmen 'swim, float'. With such verbs, agentivity, animacy and telicity may change in a context-dependent way and may be manipulated for experimental purposes.This project will contribute to the main goals of the Priority Programme XPrag.de: New Pragmatic Theories based on Experimental Evidence by using a combination of advanced experimental methods such as speeded acceptability judgements and electro-physiological data (ERP) in order to explore the previously neglected issue of the impact of contextual information in split intransitivity phenomena. On a more general level, we address the role of pragmatics in the architecture of language and in language processing, as en¬visaged by Schumacher's (2012, 2013) Syntax-Discourse Processing Model. According to this model, context dependent information is used incrementally and interactively with grammatical information in two rather distinct operations during language processing. This model is in line with recent performance-compatible approaches to grammar in which different types of information may interact.