Semantic interpretation as computation in nonmonotonic logic: the real meaning of the suppression task

Interpretation is the process whereby a hearer reasons to an interpretation of a speaker's discourse. The hearer normally adopts a credulous attitude to the discourse, at least for the purposes of interpreting it. That is to say the hearer tries to accommodate the truth of all the speaker's utterances in deriving an intended model. We present a nonmonotonic logical model of this process which defines unique minimal preferred models and efficiently simulates a kind of closed-world reasoning of particular interest for human cognition. Byrne's "suppression" data (Byrne, 1989) are used to illustrate how variants on this logic can capture and motivate subtly different interpretative stances which different subjects adopt, thus indicating where more fine-grained empirical data are required to understand what subjects are doing in this task. We then show that this logical competence model can be implemented in spreading activation network models. A one pass process interprets the textual input by constructing a network which then computes minimal preferred models for (3-valued) valuations of the set of propositions of the text. The neural implementation distinguishes easy forward reasoning from more complex backward reasoning in a way that may be useful in explaining directionality in human reasoning.     

The large-scale structure of semantic networks: statistical analyses and a model of semantic growth

We present statistical analyses of the large-scale structure of 3 types of semantic networks: word associations, WordNet, and Roget's Thesaurus. We show that they have a small-world structure, characterized by sparse connectivity, short average path lengths between words, and strong local clustering. In addition, the distributions of the number of connections follow power laws that indicate a scale-free pattern of connectivity, with most nodes having relatively few connections joined together through a small number of hubs with many connections. These regularities have also been found in certain other complex natural networks, such as the World Wide Web, but they are not consistent with many conventional models of semantic organization, based on inheritance hierarchies, arbitrarily structured networks, or high-dimensional vector spaces. We propose that these structures reflect the mechanisms by which semantic networks grow. We describe a simple model for semantic growth, in which each new word or concept is connected to an existing network by differentiating the connectivity pattern of an existing node. This model generates appropriate small-world statistics and power-law connectivity distributions, and it also suggests one possible mechanistic basis for the effects of learning history variables (age of acquisition, usage frequency) on behavioral performance in semantic processing tasks.     

Double-decision lexical tasks in thought-disordered schizophrenic patients: a path towards cognitive remediation?

It has been shown that schizophrenics have certain difficulties in the processing of semantic context. These difficulties have usually been evaluated using lexical decision tasks with semantic priming. In this study, we chose to examine the idea of an abnormality in the early stages of semantic context processing in thought-disordered schizophrenics using two double lexical decision tasks: one with a high (25%) and one with a low (15%) proportion of related words to assess the participants' competency in controlled and possibly also more automatic context processing. The results obtained in 40 control participants and 40 schizophrenic patients revealed no significant differences in the amplitude of semantic priming between the two groups. These results suggest that, in the disorganized schizophrenic subjects evaluated in this study, the context processing processes mobilized by the employed tasks were unimpaired.     

Unconscious semantic priming extends to novel unseen stimuli

Many subliminal priming experiments are thought to demonstrate unconscious access to semantics. However, most of them can be reinterpreted in a non-semantic framework that supposes only that subjects learn to map non-semantic visual features of the subliminal stimuli onto motor responses. In order to clarify this issue, we engaged subjects in a number comparison task in which the target number was preceded by another invisible masked number. We show that unconscious semantic priming occurs even for prime stimuli that are never presented as target stimuli, and for which no stimulus–response learning could conceivably occur. We also report analyses of the impact of the numerical relation between prime and target, and of the impact of learning on priming, all of which confirm that unconscious utilization of semantic information is indeed possible.     

The analysis of intuition: Processing fluency and affect in judgements of semantic coherence

In semantic coherence judgements individuals are able to intuitively discriminate whether a word triad has a common remote associate (coherent) or not (incoherent) without consciously retrieving the common associate. A processing-fluency account for these intuitions is proposed, which assumes that (a) coherent triads are processed more fluently than incoherent triads, (b) this high fluency triggers a subtle positive affect, and (c) this affect may be experienced as a cognitive feeling and used in explicit judgement. In line with this account, it was shown that coherent triads (a) are processed faster than incoherent triads (Study 1), (b) serve as positive affective primes (Study 2), and (c) are liked more than incoherent triads (Study 3). When participants were provided with an irrelevant source of their affective reactions, they lost the ability to intuitively discriminate between coherent and incoherent triads (Study 4). Finally, an item-based analysis found that triads that are processed faster are liked more and are more likely to be judged coherent, irrespective of their actual coherence (Study 5).     

Category length produces an inverted-U discriminability function in episodic recognition memory

Four experiments used signal detection analyses to assess recognition memory for lists of words consisting of differing numbers of exemplars from different semantic categories. The results showed that recognition memory performance, measured by d _a , (a) increased as category length (CL, the number of study-list items selected from the same semantic category) increased from 1 to 8 but then decreased as CL further increased from 8 to 14, (b) was greater when 2 studied items from the same category occurred back to back, rather than being separated by 5–11 items from other categories, and (c) was greater for the first studied 2 exemplars than for the last studied 2 exemplars from blocked categories with CLs of 8 or 14. For all CLs, all z-coordinate receiver operating characteristic (z-ROC) functions were linear with slopes significantly less than 1.0, and none had a significant quadratic component. These results pose a challenge for three major classes of recognition memory models: item-noise, context-noise, and dual-process models.     

Exploring the time course of semantic interference and associative priming in the picture–word interference task

The picture–word interference (PWI) task is a widely used technique for exploring effects of semantic context on lexical access. In this task, printed words are superimposed over pictures to be named, with the timing of the interfering word relative to the picture systematically manipulated. Two experiments (N = 24 adults in each) explored the time course of effects of associates (e.g., CARROT superimposed on a picture of a rabbit) versus coordinates (e.g., CHIPMUNK superimposed on a picture of a rabbit) on naming latencies. Associates led to faster picture naming than did unrelated words, with facilitative effects occurring at stimulus onset asynchronies (SOAs, in ms) ranging from –450 to 0. Coordinates led to slower naming latencies, with the interference effect restricted to SOAs of –150 and 0. The overlapping time course of associative priming and coordinate interference provides important constraints on models of lexical access in speech production.     

Short- and long-term memory contributions to immediate serial recognition: Evidence from serial position effects

A long-standing body of research supports the existence of separable short- and long-term memory systems, relying on phonological and semantic codes, respectively. The aim of the current study was to measure the contribution of long-term knowledge to short-term memory performance by looking for evidence of phonologically and semantically coded storage within a short-term recognition task, among developmental samples. Each experimental trial presented 4-item lists. In Experiment 1 typically developing children aged 5 to 6 years old showed evidence of phonologically coded storage across all 4 serial positions, but evidence of semantically coded storage at Serial Positions 1 and 2. In a further experiment, a group of individuals with Down syndrome was investigated as a test case that might be expected to use semantic coding to support short-term storage, but these participants showed no evidence of semantically coded storage and evidenced phonologically coded storage only at Serial Position 4, suggesting that individuals with Down syndrome have a verbal short-term memory capacity of 1 item. Our results suggest that previous evidence of semantic effects on “short-term memory performance” does not reflect semantic coding in short-term memory itself, and provide an experimental method for researchers wishing to take a relatively pure measure of verbal short-term memory capacity, in cases where rehearsal is unlikely.     

Semantic ambiguity within and across languages: An integrative review

Semantic ambiguity often occurs within a language (e.g., the word “organ” in English means both a body part and a musical instrument), but it can also cross a language boundary, such that a given word form is shared in two languages, but its meanings are different (e.g., the word “angel” means “sting” in Dutch). Bilingual individuals are therefore faced not only with ambiguity in each of their languages, but also with ambiguity across languages. The current review focuses on studies that explored such cross-language ambiguity and examines how the results from these studies can be integrated with what we have learned about within-language ambiguity resolution. In particular, this review examines how interactions of frequency and context manifest themselves in ambiguity that crosses a language boundary and call for the inclusion of language context as a contributing factor. An extension of the monolingual reordered access model (Duffy, Morris, & Rayner, 1988 Duffy, S. A., Morris, R. K. and Rayner, K. 1988. Lexical ambiguity and fixation times in reading. Journal of Memory and Language, 27: 429–446. [Crossref], [Web of Science ®] , [Google Scholar]) is outlined to discuss the interactions between these factors. Furthermore, the effects of the similarity between the two meanings, task differences, and individual differences are explored. This review highlights the need for studies that test within- and cross-language ambiguity in the same individuals before strong conclusions can be made about the nature of interactions between frequency, semantic context, and language context.     

The effect of semantic relatedness on immediate serial recall and serial recognition

Several researchers have reported that learning a particular categorization leads to compatible changes in the similarity structure of the categorized stimuli. The purpose of this study is to examine whether different category structures may lead to greater or less corresponding similarity change. We created six category structures and examined changes in similarity within categories or between categories, as a result of categorization, in between-participant conditions. The best supported hypothesis was that the ease of learning a categorization affects change in within-categories similarity, so that greater (within-categories) similarity change was observed for category structures that were harder to learn.