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...

We introduce finite-state registered automata (FSRAs), a new computational device within the framework of finite-state technology, specifically tailored for implementing non-concatenative morphological processes. This model extends and augments existing finite-state techniques, which are presently not optimized for describing this kind of phenomena. We first define the model and discuss its mathematical and computational properties. Then, we provide an extended regular language whose expressions denote FSRAs. Finally, we exemplify the utility of the model by providing several examples of complex morphological and phonological phenomena which are elegantly implemented with FSRAs. 1.

What precisely is the universal nature of the human syntactic parser, such that it copes easily with some embedded structures, yet fails so dramatically on others (e.g., classic double center-embeddings)? A theory is proposed in the form of an architecture for parsing based on two simple ideas. The first is that human short-term memory is an indexing structure which can give rise to interference effects (storage limitations) when contents overlap with respect to the indices. For parsing, the contents are syntactic structures, and the indices are potential structural relations. The second idea is that the capacity of STM is the minimum capacity required to support the basic functions of parsing. The theory successfully accounts for the contrasts between over 50 difficult and acceptable constructions from English, French, German, Hebrew, Japanese, Mandarin, and Spanish. The theory has independent psychological and computational motivation, and is a functional part of a broader cognitive ...

In this paper, we present an efficient algorithm for parsing natural language using unification grammars. The algorithm is an extension of left-corner parsing, a bottom-up algorithm which utilizes top-down expectations. The extension exploits unification grammar's uniform representation of syntactic, semantic, and domain knowledge, by incorporating all types of grammatical knowledge into parser expectations. In particular, we extend the notion of the reachability table, which provides information as to whether or not a top-down expectation can be realized by a potential subconstituent, by including all types of grammatical information in table entries, rather than just phrase structure information. While our algorithm's worstcase computational complexity is no better than that of many other algorithms, we present empirical testing in which average-case linear time performance is achieved. Our testing indicates this to be much improved average-case performance over pre...

We introduce finite state registered automata, a new computational model within the framework of finite state (FS) technology that accounts for nonconcatenative morphological processes such as word formation in Semitic languages. It extends existing FS techniques, which are presently suboptimal for describing such phenomena. We define the new model, prove its mathematical properties and exemplify its usability by describing some non-concatenative phenomena. 1

Parsing with unification grammars is inefficient due to the expressive power of the grammars. Most unification-based parsing algorithms are extensions of context-free (CF) parsing algorithms, and few have been specially designed for unification-style grammars. We have developed an efficient parsing algorithm for unification grammars which takes full advantage of the expressiveness of the grammar. Our algorithm (called LC) is a variation of Left-corner parsing, and it exhibits significantly improved average-case performance as compared with previous unification-based parsers. Efficiency of our LC algorithm comes from two factors. First is the representation and architecture of LINK. LINK is a syntax-semantics integrated unification-based system which dynamically combines syntax (grammar) and semantics (domain knowledge). And LINK utilizes all available information at any given point during parsing. Second is the expectation-based Left-corner parsing strategy. By utilizing expectations, ...

The idea of a perfectly competent but resource limited language user is the basis of many models of sentence comprehension. It is widely assumed that linguistic competence is a) uniform; b) generative; c) autonomous; d) automatic and e) constant. It is also believed that the free expression of these properties is frustrated by limits in the availability of computational resources. However, no firm experimental evidence for the classical language user appears to exist. Negative evidence for each assumption is reviewed here and the notion of resource limitations is shown to be suspect. An experiment is reported which tested each of the five assumptions underlying the conventional notion of linguistic competence. It was found that native speakers of English a) differed in grammatical skill; b) often failed to display productivity; c) violated syntax in favour of plausibility; d) expended conscious effort to comprehend some sentences and e) appeared to adapt to novel structures as the experiment progressed. In line with previous studies, a relationship was found between comprehension skill and formal education. A new finding is that highly educated non-native speakers of English can outperform less educated native speakers of English in comprehending grammatically challenging English sentences. The results indicate that the classical language user is an inaccurate model of real language users, who appear to differ considerably in linguistic skill. A number of specific questions for further research are raised. †

We present a technique of finite state parsing based on vectorization and describe the application of this technique to a well-known problem of natural language processing, that of extracting relational information from English text. We define Vectorized Finite State Automata, the theoretical model behind the applied system, and discuss their significance.

Attributed Models of Executable Specifications Attributed automata (AA) were introduced as a formalism for executable specification of knowledge using regular syntax with attributes representing contextual relations as well as semantic properties of concepts. AA can be treated as a generalization of a state transition network with attributes and computational relations attached to states and transitions respectively. The formal definition and general properties of the AA are presented. The problems of composition/decomposition and minimization are considered in the paper. A short overview of applications is given. Contents 1 Introduction 2 2 Attributed Automata 4 2.1 Definition : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 4 2.2 Behavior of Attributed Automata : : : : : : : : : : : : : : : : 6 2.3 Attributed Automata as Recognizers : : : : : : : : : : : : : : 8 2.4 Expressive power of attributed automata : : : : : : : : : : : : 10 3 Compositions of Attributed Models 12...

Natural language processing has been in existence for more than fifty years. During this time, it has significantly contributed to the field of human-computer interaction in terms of theoretical results and practical applications. As computers continue to become more affordable and accessible, the importance of user interfaces that are effective, robust, unobtrusive, and user-friendly -- regardless of user expertise or impediments -- becomes more pronounced. Since natural language usually provides for effortless and effective communication in human-human interaction, its significance and potential in human-computer interaction should not be overlooked -- either spoken or typewritten, it may effectively complement other available modalities, such as windows, icons, and menus, and pointing; in some cases, such as in users with disabilities, natural language may even be the only applicable modality. This chapter examines the field of natural language processing as it relates to humanc...