Edward Gibson

Linguistic complexity: locality of syntactic dependencies

This paper proposes a new theory of the relationship between the sentence processing mechanism and the available computational resources. This theory--the Syntactic Prediction Locality Theory (SPLT)--has two components: an integration cost component and a component for the memory cost associated with keeping track of obligatory syntactic requirements. Memory cost is hypothesized to be quantified in terms of the number of syntactic categories that are necessary to complete the current input string as a grammatical sentence. Furthermore, in accordance with results from the working memory literature both memory cost and integration cost are hypothesized to be heavily influenced by locality (1) the longer a predicted category must be kept in memory before the prediction is satisfied, the greater is the cost for maintaining that prediction; and (2) the greater the distance between an incoming word and the most local head or dependent to which it attaches, the greater the integration cost. The SPLT is shown to explain a wide range of processing complexity phenomena not previously accounted for under a single theory, including (1) the lower complexity of subject-extracted relative clauses compared to object-extracted relative clauses, (2) numerous processing overload effects across languages, including the unacceptability of multiply center-embedded structures, (3) the lower complexity of cross-serial dependencies relative to center-embedded dependencies, (4) heaviness effects, such that sentences are easier to understand when larger phrases are placed later and (5) numerous ambiguity effects, such as those which have been argued to be evidence for the Active Filler Hypothesis.

Processing Syntactic Relations in Language and Music: An Event-Related Potential Study

In order to test the language-specificity of a known neural correlate of syntactic processing the P600 event-related brain potential (ERP) component, this study directly compared ERPs elicited by syntactic incongruities in language and music. Using principles of phrase structure for language and principles of harmony and key-relatedness for music, sequences were constructed in which an element was either congruous, moderately incongruous, or highly incongruous with the preceding structural context. A within-subjects design using 15 musically educated adults revealed that linguistic and musical structural incongruities elicited positivities that were statistically indistinguishable in a specified latency range. In contrast, a music-specific ERP component was observed that showed antero-temporal right-hemisphere lateralization. The results argue against the language-specificity of the P600 and suggest that language and music can be studied in parallel to address questions of neural specificity in cognitive processing.

The influence of referential processing on sentence complexity

This paper reports the results of five experiments designed to investigate the effects of referential processing on sentence complexity. Gibson (Cognition, 68 (1998) 1) suggested that sentence complexity is related to the locality of integrations between dependent syntactic heads, and that an appropriate measure of locality is the number of new discourse referents intervening between the endpoints of those integrations. The experiments in this paper test, modify and extend Gibsons (1998) claims. Each experiment manipulated noun phrases (NPs) in the subject positions of object-extracted relative clauses in order to determine how different types of NPs affected sentence complexity. Experiments 1, 2 and 3 used questionnaires to gauge sentence complexity, whereas Experiments 4 and 5 used self-paced reading. The results from Experiments 1, 2, 4 and 5 suggest that the complexity of the experimental items was more closely related to the Givenness status of the embedded subject in the Givenness Hierarchy than to whether the embedded subject was old or new to the discourse. Experiment 3 compared materials in which a quantifier was rotated through subject positions of a nested relative clause structure. The results of this experiment support a discourse-processing-based distance metric for computing locality and provide evidence against a pure similarity-based account of structural complexity such as proposed by Bever (Bever, T. G. (1970). The cognitive basis of linguistic structures. In J. R. Hayes (Ed.), Cognition and the development of language (pp. 279-362). New York: Wiley).

Consequences of the Serial Nature of Linguistic Input for Sentenial Complexity

All other things being equal the parser favors attaching an ambiguous modifier to the most recent possible site. A plausible explanation is that locality preferences such as this arise in the service of minimizing memory costs-more distant sentential material is more difficult to reactivate than more recent material. Note that processing any sentence requires linking each new lexical item with material in the current parse. This often involves the construction of long-distance dependencies. Under a resource-limited view of language processing, lengthy integrations should induce difficulty even in unambiguous sentences. To date there has been little direct quantitative evidence in support of this perspective. This article presents 2 self-paced reading studies, which explore the hypothesis that dependency distance is a fundamental determinant of reading complexity in unambiguous constructions in English. The evidence suggests that the difficulty associated with integrating a new input item is heavily determined by the amount of lexical material intervening between the input item and the site of its target dependents. The patterns observed here are not straightforwardly accounted for within purely experience-based models of complexity. Instead, this work supports the role of a memory bottleneck in language comprehension. This constraint arises because hierarchical linguistic relations must be recovered from a linear input stream.

Word lengths are optimized for efficient communication

We demonstrate a substantial improvement on one of the most celebrated empirical laws in the study of language, Zipfs 75-y-old theory that word length is primarily determined by frequency of use. In accord with rational theories of communication, we show across 10 languages that average information content is a much better predictor of word length than frequency. This indicates that human lexicons are efficiently structured for communication by taking into account interword statistical dependencies. Lexical systems result from an optimization of communicative pressures, coding meanings efficiently given the complex statistics of natural language use.

Representing Discourse Coherence: A Corpus-Based Study

This article aims to present a set of discourse structure relations that are easy to code and to develop criteria for an appropriate data structure for representing these relations. Discourse structure here refers to informational relations that hold between sentences in a discourse. The set of discourse relations introduced here is based on Hobbs (1985). We present a method for annotating discourse coherence structures that we used to manually annotate a database of 135 texts from the Wall Street Journal and the AP Newswire. Alltexts were independently annotated by two annotators. Kappa values of greater than 0.8 indicated good interannotator agreement. We furthermore present evidence that trees are not a descriptively adequate data structure for representing discourse structure: In coherence structures of naturally occurring texts, we found many different kinds of crossed dependencies, as well as many nodes with multiple parents. The claims are supported by statistical results from our hand-annotated database of 135 texts.

Processing relative clauses in Chinese

This paper reports results from a self-paced reading study in Chinese that demonstrates that object-extracted relative clause structures are less complex than corresponding subject-extracted structures. These results contrast with results from processing other Subject-Verb-Object languages like English, in which object-extracted structures are more complex than subject-extracted structures. A key word-order difference between Chinese and other Subject-Verb-Object languages is that Chinese relative clauses precede their head nouns. Because of this word order difference, the results follow from a resource-based theory of sentence complexity, according to which there is a storage cost associated with predicting syntactic heads in order to form a grammatical sentence. The results are also consistent with a theory according to which people have less difficulty processing embedded clauses whose word order matches the word order in main clauses. Some corpus analyses of Chinese texts provide results that constrain the classes of possible frequency-based theories. Critically, these results demonstrate that there is nothing intrinsically easy about extracting from subject position: depending on the word order in the main clause and in a relative clause, extraction from object position can be easier to process in some circumstances.

Recency preference in the human sentence processing mechanism

Cuetos and Mitchell (1988) observed that in constructions in which a relative clause can attach to one of two possible sites, English speakers prefer the more recent attachment site, but Spanish speakers prefer the least recent attachment site, in violation of the proposed universal principle Late Closure (Recency Preference), which favors attachments to the most recent sites. Based on this evidence, Cuetos and Mitchell concluded that Late Closure is not a universal principle of the human sentence processing mechanism. In this paper, we provide new evidence from Spanish and English self-paced reading experiments on relative clause attachment ambiguities that involve three possible attachment sites. The results of our experiments suggest that a principle like Late Closure is in fact universally operative in the human parser, but that it is modulated by at least one other factor in the processing of relative clause attachment ambiguities. We propose that the second factor involved in the processing of these and related constructions is the principle of Predicate Proximity, according to which attachments are preferred to be as structurally close to the head of a predicate phrase as possible, and we further consider the origins and predictions of the theory combining these two factors.

The relationship between intonational phrasing and syntactic structure in language production

In this paper, we evaluate several theories of how syntactic/semantic structure influences the placement of intonational boundaries in language production (Cooper & Paccia-Cooper, 1980; Gee & Grosjean, 1983; Ferreira, 1988). Although the theories that we tested are shown to be quite successful, they are complex, and furthermore, they are incompatible with recent evidence for incrementality in sentence production. In light of these problems, we propose a simpler incremental model called the Left hand side/Right hand side Boundary hypothesis (LRB). According to this hypothesis, two factors that underlie the successful performance of the algorithms from the literature contribute to the likelihood of producing intonational boundaries at word boundaries: (1) the size of the recently completed syntactic constituent at a word boundary; and (2) the size of the upcoming syntactic constituent. These factors are further constrained by syntactic argument relationships. We demonstrate that the LRB performs as well as previous models with respect to the data from Experiment 1. In Experiment 2, we present evidence that the LRB outperforms previous models in certain instances. In Experiment 3, we demonstrate that the discourse status of relative clauses is an additional factor in intonational boundary placement.

Constraints on sentence comprehension

The process of comprehending sentences involves the integration of a variety of different information sources, constrained by the available computational resources. This paper surveys the evidence for four types of constraints on sentence comprehension: (1) lexical constraints, (2) contextual constraints, (3) locality-based computational resource constraints, and (4) phrase-level contingent frequency constraints. These four constraints, in combination with grammatical phrase-formation information and prosody, are sufficient to explain how interpretations are constructed for sentences as they are encountered, including complexity effects in unambiguous sentences and interpretation preferences in ambiguous sentences.