Publications

Abstract

Certain categories of language learners need feedback on the grammatical structure of sentences they wish to produce. In contrast with the usual NLP approach to this problem—parsing student-generated texts—we propose a generation-based approach aiming at preventing errors (“scaffolding”). In our ICALL system, students construct sentences by composing syntactic trees out of lexically anchored “treelets” via a graphical drag&drop user interface. A natural-language generator computes all possible grammatically well-formed sentences entailed by the student-composed tree, and intervenes immediately when the latter tree does not belong to the set of well-formed alternatives. Feedback is based on comparisons between the student-composed tree and the well-formed set. Frequently occurring errors are handled in terms of “malrules.” The system (implemented in JAVA and C++) currently focuses constituent order in German as L2.

Abstract

We present an overview of several corpus studies we carried out into the frequencies of argument NP orderings in the midfield of subordinate and main clauses of German. Comparing the corpus frequencies with grammaticality ratings published by Keller’s (2000), we observe a “grammaticality–frequency gap”: Quite a few argument orderings with zero corpus frequency are nevertheless assigned medium–range grammaticality ratings. We propose an explanation in terms of a two-factor theory. First, we hypothesize that the grammatical induction component needs a sufficient number of exposures to a syntactic pattern to incorporate it into its repertoire of more or less stable rules of grammar. Moderately to highly frequent argument NP orderings are likely have attained this status, but not their zero-frequency counterparts. This is why the latter argument sequences cannot be produced by the grammatical encoder and are absent from the corpora. Secondly, we assumed that an extraneous (nonlinguistic) judgment process biases the ratings of moderately grammatical linear order patterns: Confronted with such structures, the informants produce their own “ideal delivery” variant of the to-be-rated target sentence and evaluate the similarity between the two versions. A high similarity score yielded by this judgment then exerts a positive bias on the grammaticality rating—a score that should not be mistaken for an authentic grammaticality rating. We conclude that, at least in the linearization domain studied here, the goal of gaining a clear view of the internal grammar of language users is best served by a combined strategy in which grammar rules are founded on structures that elicit moderate to high grammaticality ratings and attain at least moderate usage frequencies.

Abstract

Experimental sentence comprehension studies have shown that superficially similar German clauses with verb-final word order elicit very different garden-path and ERP effects. We show that a computer implementation of the Unification Space parser (Vosse & Kempen, 2000) in the form of a localist-connectionist network can model the observed differences, at least qualitatively. The model embodies a parallel dynamic parser that, in contrast with existing models, does not distinguish between consecutive first-pass and reanalysis stages, and does not use semantic or thematic roles. It does use structural frequency data and animacy information.

Abstract

This paper investigates the time and space complexity of word order computation in the psycholinguistically motivated grammar formalism of Performance Grammar (PG). In PG, the first stage of syntax assembly yields an unordered tree ('mobile') consisting of a hierarchy of lexical frames (lexically anchored elementary trees). Associated with each lexica l frame is a linearizer—a Finite-State Automaton that locally computes the left-to-right order of the branches of the frame. Linearization takes place after the promotion component may have raised certain constituents (e.g. Wh- or focused phrases) into the domain of lexical frames higher up in the syntactic mobile. We show that the worst-case time and space complexity of analyzing input strings of length n is O(n5) and O(n4), respectively. This result compares favorably with the time complexity of word-order computations in Tree Adjoining Grammar (TAG). A comparison with Head-Driven Phrase Structure Grammar (HPSG) reveals that PG yields a more declarative linearization method, provided that the FSA is rewritten as an equivalent regular expression.

Abstract

We present the design, implementation and simulation results of a psycholinguistic model of human syntactic processing that meets major empirical criteria. The parser operates in conjunction with a lexicalist grammar and is driven by syntactic information associated with heads of phrases. The dynamics of the model are based on competition by lateral inhibition ('competitive inhibition'). Input words activate lexical frames (i.e. elementary trees anchored to input words) in the mental lexicon, and a network of candidate 'unification links' is set up between frame nodes. These links represent tentative attachments that are graded rather than all-or-none. Candidate links that, due to grammatical or 'treehood' constraints, are incompatible, compete for inclusion in the final syntactic tree by sending each other inhibitory signals that reduce the competitor's attachment strength. The outcome of these local and simultaneous competitions is controlled by dynamic parameters, in particular by the Entry Activation and the Activation Decay rate of syntactic nodes, and by the Strength and Strength Build-up rate of Unification links. In case of a successful parse, a single syntactic tree is returned that covers the whole input string and consists of lexical frames connected by winning Unification links. Simulations are reported of a significant range of psycholinguistic parsing phenomena in both normal and aphasic speakers of English: (i) various effects of linguistic complexity (single versus double, center versus right-hand self-embeddings of relative clauses; the difference between relative clauses with subject and object extraction; the contrast between a complement clause embedded within a relative clause versus a relative clause embedded within a complement clause); (ii) effects of local and global ambiguity, and of word-class and syntactic ambiguity (including recency and length effects); (iii) certain difficulty-of-reanalysis effects (contrasts between local ambiguities that are easy to resolve versus ones that lead to serious garden-path effects); (iv) effects of agrammatism on parsing performance, in particular the performance of various groups of aphasic patients on several sentence types.