Julie A. Van Dyke

Computational principles of working memory in sentence comprehension

Understanding a sentence requires a working memory of the partial products of comprehension, so that linguistic relations between temporally distal parts of the sentence can be rapidly computed. We describe an emerging theoretical framework for this working memory system that incorporates several independently motivated principles of memory: a sharply limited attentional focus, rapid retrieval of item (but not order) information subject to interference from similar items, and activation decay (forgetting over time). A computational model embodying these principles provides an explanation of the functional capacities and severe limitations of human processing, as well as accounts of reading times. The broad implication is that the detailed nature of cross-linguistic sentence processing emerges from the interaction of general principles of human memory with the specialized task of language comprehension.

Distinguishing effects of structure and decay on attachment and repair: A cue-based parsing account of recovery from misanalyzed ambiguities

This paper presents the cue-based retrieval theory of parsing and reanalysis and illustrates how this account can accommodate a number of key results about parsing and reanalysis, including effects due to structure, distance, and type of structural change. Three offline experiments and one online experiment permit establishing the locus of these effects as due to properties of the initial parsing processes or to the repair mechanism. Specifically, the data reported here suggest that a structural factor specific to the operation of the parser, retrieval interference, affects attachment uniformly across ambiguous and unambiguous sentences and serves to create a limit on successful repair. In addition, these experiments suggest that distance of the head of an ambiguous phrase from its disambiguator affects repair processes—and not attachment processes—independently of the interference effect. These results are interpreted with respect to alternative models of reanalysis, which are contrasted with the cue-based retrieval account, which requires no distinct repair mechanism to account for the current results. A further contribution of this article is to suggest a statistical correction for individual variance in reading rates. Statistical analyses on individual subject data confirmed previous speculations regarding a possible increase in reading rates as subjects move through a sentence. While this individual variation limits fair comparisons of reading times in sentence regions that appear in non-identical serial positions, we demonstrate that such comparisons become meaningful when the appropriate regression analyses have been performed.

Interference Effects From Grammatically Unavailable Constituents During Sentence Processing

Evidence from 3 experiments reveals interference effects from structural relationships that are inconsistent with any grammatical parse of the perceived input. Processing disruption was observed when items occurring between a head and a dependent overlapped with either (or both) syntactic or semantic features of the dependent. Effects of syntactic interference occur in the earliest online measures in the region where the retrieval of a long-distance dependent occurs. Semantic interference effects occur in later online measures at the end of the sentence. Both effects endure in offline comprehension measures, suggesting that interfering items participate in incorrect interpretations that resist reanalysis. The data are discussed in terms of a cue-based retrieval account of parsing, which reconciles the fact that the parser must violate the grammar in order for these interference effects to occur. Broader implications of this research indicate a need for a precise specification of the interface between the parsing mechanism and the memory system that supports language comprehension.

Argumentation in Psychology: Background Comments

Argumentation constitutes 1 of the most common forms of human interaction. Yet despite its pervasiveness, relatively little psychological research has been conducted on the topic. This article serves as an introduction to this research and has 2 goals. One is to discuss a number of general issues relevant to the study of argumentation, including the definition, goals and functions, structure, evaluation of arguments and argumentation, and the relation of narrativity and argumentation. The 2nd goal is to describe some examples of the existing psychological research on argumentation, with emphasis on articles in this special issue. Topics include argumentation by children, argumentation skill, writing argumentative text, argumentation and case-based change, argumentation and critical thinking, and argumentation and narrativity in a legal context.

Low working memory capacity is only spuriously related to poor reading comprehension

Accounts of comprehension failure, whether in the case of readers with poor skill or when syntactic complexity is high, have overwhelmingly implicated working memory capacity as the key causal factor. However, extant research suggests that this position is not well supported by evidence on the span of active memory during online sentence processing, nor is it well motivated by models that make explicit claims about the memory mechanisms that support language processing. The current study suggests that sensitivity to interference from similar items in memory may provide a better explanation of comprehension failure. Through administration of a comprehensive skill battery, we found that the previously observed association of working memory with comprehension is likely due to the collinearity of working memory with many other reading-related skills, especially IQ. In analyses which removed variance shared with IQ, we found that receptive vocabulary knowledge was the only significant predictor of comprehension performance in our task out of a battery of 24 skill measures. In addition, receptive vocabulary and non-verbal memory for serial order-but not simple verbal memory or working memory-were the only predictors of reading times in the region where interference had its primary affect. We interpret these results in light of a model that emphasizes retrieval interference and the quality of lexical representations as key determinants of successful comprehension.

Reassessing Word Frequency as a Determinant of Word Recognition for Skilled and Unskilled Readers

The importance of vocabulary in reading comprehension emphasizes the need to accurately assess an individuals familiarity with words. The present article highlights problems with using occurrence counts in corpora as an index of word familiarity, especially when studying individuals varying in reading experience. We demonstrate via computational simulations and norming studies that corpus-based word frequencies systematically overestimate strengths of word representations, especially in the low-frequency range and in smaller-size vocabularies. Experience-driven differences in word familiarity prove to be faithfully captured by the subjective frequency ratings collected from responders at different experience levels. When matched on those levels, this lexical measure explains more variance than corpus-based frequencies in eye-movement and lexical decision latencies to English words, attested in populations with varied reading experience and skill. Furthermore, the use of subjective frequencies removes the widely reported (corpus) Frequency × Skill interaction, showing that more skilled readers are equally faster in processing any word than the less skilled readers, not disproportionally faster in processing lower frequency words. This finding challenges the view that the more skilled an individual is in generic mechanisms of word processing, the less reliant he or she will be on the actual lexical characteristics of that word.

The psycholinguistics of basic literacy

We review major issues in research on reading, including theories of word reading, cross-writing system comparisons, comprehension, reading difficulties, learning how to read, and cognitive neuroscience studies of reading. Each of these topics has psycholinguistic components that reflect the language foundations of reading. These foundations lie in two facts: (1) a writing system connects to a linguistic system at one or more levels, meaning that word reading is partially a psycholinguistic process; and (2): reading comprehension shares processes (e.g., parsing) with general language comprehension. One trend of recent research is the development of models of word identification that rely on single rather than dual mechanisms and their extension to explain reading difficulties. Another is the conclusion that phonology plays a role in reading that cuts across writing systems. Reading comprehension research continues to reflect two different traditions, sentence parsing and text comprehension. Both show increasing influence of general cognitive explanations, as opposed to strictly linguistic explanations, for comprehension phenomena. Studies of brain activation bring converging evidence on reading processes and provide neuroanatomical correlates of reading problems. In one area, the acquisition and teaching of reading, advances come from increasing consolidation and practical use of previous research gains.

Neural basis of semantic and syntactic interference in sentence comprehension

According to the cue-based parsing approach (Lewis, Vasishth, & Van Dyke, 2006), sentence comprehension difficulty derives from interference from material that partially matches syntactic and semantic retrieval cues. In a 2 (low vs. high semantic interference)×2 (low vs. high syntactic interference) fMRI study, greater activation was observed in left BA44/45 for high versus low syntactic interference conditions following sentences and in left BA45/47 for high versus low semantic interference conditions following comprehension questions. A conjunction analysis showed BA45 associated with both types of interference, while BA47 was associated with only semantic interference. Greater activation was also observed in the left STG in the high interference conditions. Importantly, the results for the LIFG could not be attributed to greater working memory capacity demands for high interference conditions. The results favor a fractionation of the LIFG wherein BA45 is associated with post-retrieval selection and BA47 with controlled retrieval of semantic information.

Eye-Movement Control in RAN and Reading.

ABSTRACT The present study examined the visual scanning hypothesis, which suggests that fluent oculomotor control is an important component underlying the predictive relationship between Rapid Automatized Naming (RAN) tasks and reading ability. Our approach was to isolate components of saccadic planning, articulation, and lexical retrieval in 3 modified RAN tasks. We analyzed 2 samples of undergraduate readers (ages 17–27). We evaluated the incremental contributions of these components and found that saccadic planning to nonlinguistic stimuli alone explained roughly one third of the variance that conventional RAN tasks explained in eye movements registered during text reading for comprehension. We conclude that the well-established predictive role of RAN for reading performance is in part due to the individual ability to coordinate rapid sequential eye movements to visual nonlinguistic stimuli.

The Random Forests statistical technique: An examination of its value for the study of reading

ABSTRACT Studies investigating individual differences in reading ability often involve data sets containing a large number of collinear predictors and a small number of observations. In this article, we discuss the method of Random Forests and demonstrate its suitability for addressing the statistical concerns raised by such data sets. The method is contrasted with other methods of estimating relative variable importance, especially Dominance Analysis and Multimodel Inference. All methods were applied to a data set that gauged eye-movements during reading and offline comprehension in the context of multiple ability measures with high collinearity due to their shared verbal core. We demonstrate that the Random Forests method surpasses other methods in its ability to handle model overfitting and accounts for a comparable or larger amount of variance in reading measures relative to other methods.