Gradient perception of grammaticality: On the Interaction of grammar, frequency of usage and processing complexity

在过去的几十年里，大多数形式句法理论都认为，句法知识是一个离散的符号系统，它在语言理解和语言产生过程中被使用，但在其他方面有别于执行机制。这一假设最近在几个方面受到了挑战。第一个挑战来自对这样一个事实的新认识，即语法性判断不是二元的，而是梯度的。第二个挑战来自关于语法和使用频率之间关系的新发现。这导致了句法理论的发展，用数字信息来扩充符号语法系统，这些数字信息通常被认为来自语言使用中句法结构的出现频率。本提案的目的是以德语的论元和动词线性化作为主要研究领域来评估这些挑战。主要的研究策略将是在基于权重的优选性理论的框架内构建所选语法现象的模型，并根据使用频率的数据和语法判断的实验调查数据来评估这些模型。如果无法从先前的研究中获得相关数据，将在拟议的研究项目中获得这些数据。

An assumption shared by most formal syntactic theories over the last decades has been that syntactic knowledge ist a discrete symbolic system which is put ot use during language comprehension and language production but is otherwise distinct from the performance mechanisms. This assumption has been challenged recently in several ways. A first challenge derives from a new appreciation of the fact that grammaticality judgments are not binary but gradient. A second challenge comes from new findings concerning the relationship between grammar and frequency of usage. This has led to the development of syntactic theories which augment the symbolic grammar system with numerical information which is often assumed to be derived from the frequencies with which syntactic structures occur in language use.The aim of the current proposal is to evaluate these challenges with argument and verb linearization in German as the principal areas of investigation. The main research strategy will be to construct models of the selected grammatical phenomena within the framework of weight-based Optimatility Theory and to evaluate these models with regard to data about frequency of usage and data from experimental investigations of grammaticality judgments. When the relevant data are not available from prior research, they will be obtained within the proposed research project.