Abstract not found

Many theories have been proposed to explain difficulty with center embedded constructions, most attributing the problem to some kind of limited capacity short-term memory. However, these theories have developed for the most part independently of more traditional memory research, which has focused on uncovering general principles such as chunking and interference. This article attempts to gain some unification with this research by suggesting that an interesting range of core sentence processing phenomena can be explained as interference effects in a sharply limited syntactic working memory. These include difficult and acceptable embeddings, as well as certain limitations on ambiguity resolution, length effects in garden path structures, and the requirement for locality in syntactic structure. The theory takes the form of an architecture for parsing which can index no more than two constituents under the same syntactic relation. A limitation of two or three items shows up in a variety o...

Abstract not found

A prevalent trend in modeling human sentence processing has been to account for both initial attachment preferences and reanalysis behaviors with minimal extensions to a presumed set of initial parsing operations. Here, an entirely different formulation of the initial attachment and revision processes is suggested. Rather than assuming that all parsing is (as much as possible) initial attachment, the opposite approach is advocated: that all parsing---even initial attachment---is restructuring. The realization of parsing as restructuring arises from a set of independently motivated computational assumptions within the competitive attachment architecture, a hybrid connectionist model of the human sentence processor. Central to the model is a unique parallel attachment operation that simultaneously attaches the current input phrase, while reattaching previously structured phrases. Within this model, reanalysis is not a separate process or module, but rather a side effect of the primary means of forming syntactic structures. The ease of performing possible reanalyses is therefore determined by the same conditions, such as recency and lexical preferences, that affect initial attachments. Furthermore, independently motivated constraints on the network structure determine the allowable syntactic configurations that may undergo restructuring within the competitive attachment operation. The model thus also provides a computational explanation of gardenpath sentences, in which automatic reanalysis is impossible.

This paper investigates the processing of long-distance filler-gap dependencies in Japanese, a strongly head-final language. Two self-paced reading experiments and one sentence completion study show that Japanese readers associate a fronted wh-phrase with the most deeply embedded clause of a multi-clause sentence. Experiment 1 demonstrates this using evidence that readers expect to encounter a scope-marking affix on the verb of an embedded clause in wh-fronting constructions. Experiment 2 shows that the wh-phrase is already associated with the embedded clause before the embedded verb is processed, based on a Japanese counterpart of the Filled Gap Effect (Stowe, 1986). Experiment 3 corroborates these findings in a sentence completion study. These findings clarify the factors responsible for Ôactive fillerÕ effects in processing long-distance dependencies (Crain & Fodor, 1985; Fodor, 1978; Frazier & Clifton, 1989; Stowe, 1986) in ways not possible in head-initial languages. The results provide evidence that the processing of filler-gap dependencies is driven by the need to satisfy thematic role requirements of the fronted phrase, rather than by the need to create a gap as soon as possible. The paper also discusses implications of these findings for theories of reanalysis.

xii Chapter 1 INCREMENTALITY AND PARSING........................................................................ 1 1.1

This paper presents a massively parallel parser that pre- dicts critical attachment behaviors of the human sentence processor, without the use of explicit preference heuristics or revision strategies. The processing of a syntactic ambiguity is modeled as an active, distributed competition among the potential attachments for a phrase. Computationally motivated constraints on the competitive mechanism provide a principled and uniform account of a range of human attachment preferences and garden path phe- nolnena.

We recorded event-related brain potentials (ERPs) during the processing of unambiguous German sentences containing different types of filler-gap dependency. Both topicalization constructions and wh-questions were found to elicit a left-anterior negativity (LAN) prior to the processing of the subcategorizing verb, relative to a gap-free control condition. At the subcategorizing verb, sentences containing a wh-dependency produced a parietal positivity (P600) relative to topicalization structures. These results support the claim that separable parsing processes are involved in the processing of syntactic dependencies, with working memory based processes being reflected in a LAN, and the relative difficulty of integrating the filler with its subcategorizer reflected in a P600. Integration cost but not memory cost was found to be influenced by the type of filler-gap dependency involved.

Ambiguity and the Computational Feasibility of Syntax Acquisition by William Gregory Sakas Advisor: Professor Virginia Teller The thesis presents a framework that can be used for empirical and formal analysis of parameter setting models of language acquisition. Such models attempt to mirror computationally the process by which children acquire the grammar of their native language. Research into formal language learning theory standardly focuses on issues of learnabili - Under what conditions is learning possible? The thesis contributes to the important, but under-investigated question offiasibili - Is acquisition possible within a reasonable amount of time and/or with a reasonable amount of work? The proposed framework formalizes existing notions such as the rate of parametric ambiguity and parametric expression within a generally defined parameter space, so that different types of learning algorithms and grammar spaces can be explored. Two influential learning algorithms are examined in detail: The Triggoing Learning Agodthm (Gibson and Wexler, 1994) and The Structural Tdggers Leame (Fodor, 1998). Empirical results indicate that the Triggering Learning Algorithm's simple hill-climbing search heuristics are sufficient to acquire the target grammar without the learner's consumption of an unreasonable number of input sentences when the learning space contains a strong correlation between the similarity of languages and the grammars that generate them. The results also indicate that the Structural Triggers Leamer's use of structural information lying beneath the surface word order of an input sentence allows for feasible learning when the rate of parametric expression varies across the input sentences encountered by the learner. Notably, however, both models are acu...

Syntactic parameter setting has proven extremely difficult to model. The original switch-setting' metaphor failed because parametrically relevant properties of a natural language sentence cannot be recognized withou t considerable structural analy sis.