Imagery effects and semantic similarity in sentence recognition memory

The idea that subjects often use imagery to discriminate semantically similar sentences was tested in three experiments. In the first experiment, subjects heard subject-verb-object sentences in the context of either a comprehension task or an image-generation task. Their memory for the sentences was tested using a two-alternative forced-choice recognition test in which different types of distractor sentence were used. A sentence semantically similar to the target sentence was one type; a sentence with the same subject and object nouns as the target sentence, but dissimilar in meaning, was another type; and a sentence similar in meaning to one of the stimulus sentences, but not to the target sentence, was a third type. The results showed that the image-generation instructions enhanced later recognition performance, but only for semantically similar test items. A second experiment showed that this finding only holds for high-imagery sentences containing concrete noun concepts. A third experiment demonstrated that the enhanced recognition performance could not be accounted for in terms of a semantic model of test-item discrimination. Collectively, the results were interpreted as providing evidence for the notion that subjects discriminate the semantically similar test items by elaborating the sentence encoding through image processing.     

The effects of acoustic and semantic similarity on concept identification

Since both acoustic and semantic similarity exert an influence on memory, the role of memory in concept identification (CI) was investigated by varying the acoustic and semantic similarity of the stimuli used in the CI task. Varying acoustic similarity had no effect on CI, but CI was significantly impaired when the dimensions of a CI task were semantically similar.     

The influence of acoustic and semantic similarity on long-term memory for word sequences

It has been shown that short-term memory (STM) for word sequences is grossly impaired when acoustically similar words are used, but is relatively unaffected by semantic similarity. This study tests the hypothesis that long-term memory (LTM) will be similarly affected. In Experiment I subjects attempted to learn one of four lists of 10 words. The lists comprised either acoustically or semantically similar words (A and C) or control words of equal frequency (B and D). Lists were learned for four trials, after which subjects spent 20 min. on a task involving immediate memory for digits. They were then asked to recall the word list. The acoustically similar list was learned relatively slowly, but unlike the other three lists showed no forgetting. Experiment II showed that this latter paradox can be explained by assuming the learning score to depend on both LTM and STM, whereas the subsequent retest depends only on LTM. Experiment III repeats Experiment I but attempts to minimize the effects of STM during learning by interposing a task to prevent rehearsal between the presentation and testing of the word sequences. Unlike STM, LTM proved to be impaired by semantic similarity but not by acoustic similarity. It is concluded that STM and LTM employ different coding systems.     

Distribution effects of orthographic similarity and priming by masked repetition]

This research centres on the effect that the orthographic neighbourhood has in the visual recognition of words. Specifically, we studied to what extent orthographic neighbourhood distribution, that is, the number of letter positions allowing formation of at least one neighbour (Pugh, Rexer, Peter, & Katz, 1994), influences the masked repetition priming effect. In a previous study (Mathey, Robert, & Zagar, 2004), interaction between neighbourhood distribution and orthographic priming was obtained in the lexical decision task. The Interactive Activation Model (IA; McClelland & Rumelhart, 1981) simulated this interaction. With the orthographic priming effect modified for distribution of the neighbourhood of target words, it was necessary to study whether the repetition priming effect also varied as a function of this indicator. Studying this interaction presents a major theoretical issue in specifying the activating and inhibiting processes presented in the IA model. Simulations were produced to obtain precise model predictions regarding the neighbourhood distribution effect in a repetitive priming situation for our experimental material. Target words all had two neighbours that were most frequent. These neighbours were isolated, that is, distributed over two letter positions (e.g.: TAUX/faux-toux), or associated, i.e., concentrated on one single position (e.g., SEAU/beau-peau). Targets were preceded by an identical priming (repetitive priming; e.g.: seau-SEAU) or by controlled priming (e.g., &-SEAU). The simulation results obtained using the IA model show the facilitating effects of neighbourhood distribution and repetitive priming, but no interaction between these factors. The experimental results obtained in a lexical decision task confirm these predictions. Thus, the empirical data replicate the neighbourhood distribution's facilitating effect (Mathey & Zagar, 2000) as well as the facilitating effect of masked repetition (Forster & Davis, 1984). Finally, the most interesting result is that the facilitating effect of repetition is comparable for target words with associated neighbours and target words with isolated neighbours. An explanation of the combined effects of the orthographic neighbourhood and orthographic masked repetition priming, integrating data from literature as well as from the current study, is proposed within the framework of the IA model.     

More is not always better: paradoxical effects of repetition on semantic accessibility

Repetition normally enhances memory. While in some cases the benefit of added repetition may be incremental, few would expect that massed repetition could actually reverse the benefits of brief repetition. Here we report two experiments that document a clear example of a paradoxical effect of massed repetition. Subjects first repeated words (e.g., "sheep") aloud one at a time for 0, 5, 10, 20, or 40 s. A free association phase followed in which cues could be completed with repeated words (e.g., "herd s___" for "sheep") or with semantically associated words (e.g., "fabric w___" for "wool"). Brief periods of repetition (5-10 s) resulted in priming, as would be expected based on research on repetition priming and spreading activation. Longer periods of repetition (20-40 s), however, abolished priming. Interestingly, this massed-repetition decrement was particularly robust for semantic associates of repeated words, and was evident after a 15-min delay. These findings document a paradoxical feature of the effects of rehearsal on memory: When ideas are repeated often enough, the benefits of rehearsal can actually be reversed.     

Optimal sequences of trials for balancing practice and repetition effects

This paper describes procedures for generating trial sequences to balance out practice effects and intertrial repetition effects in experiments consisting of repeated trials. In the sequences presented, each stimulus appears an equal number of times, is preceded equally often by itself and by each other stimulus, and is distributed in a “balanced” manner throughout the block of trials. Two criteria for balance are employed. One criterion aims to equalize the average position of each stimulus in the sequence. The second criterion maintains, as much as possible, a uniform interval between appearances of each stimulus in the sequence. For each criterion, optimal or near-optimal sequences are presented for experiments involving from three to nine different stimulus conditions. Suggestions are included for extending (e.g., doubling or tripling) the length of the sequences.     

Increased semantic and repetition priming in schizophrenic patients

Positive and negative priming (PP and NP) in schizophrenia were studied with a lexical-decision task. Probe words, presented 800 ms after the response to the prime (containing a word and a nonword), were either identical to, semantically related to, or unrelated to the prime target word (PP) or to the prime distractor word (NP). Schizophrenic patients displayed stronger semantic and repetition PP than controls after controlling for their slower responses. Significant NP was observed in both groups for word repetition only. The PP findings contrast with results from studies with similar prime-probe intervals but without prime responses. It is proposed that schizophrenic patients, because of impaired (controlled) processes of response selection, strongly benefit from (or rely on) the automatic retrieval of processing episodes containing response information. Related findings indicating automatic response facilitation in schizophrenia are discussed.     

Categorizing sounds by pitch: effects of stimulus similarity and response repetition

Two experiments were performed to determine whether categorization of the pitch of a probe tone is influenced by the pitch of, and response made to, a preceding prime tone. The prime and the probe could be drawn either from a pool of low-frequency sounds or from a pool of high-frequency sounds. The results of both experiments indicated that the performance obtained was best when the prime and the probe were the same pitch (and therefore required the same response), intermediate when the two sounds differed in pitch and required different responses, and slowest when the prime and the probe differed in pitch but required the same response (i.e., they were drawn from the same frequency pool). The results of Experiment 2 revealed in addition that when a repeated response was required, performance declined as the magnitude of the frequency change increased and that responses were made more quickly and accurately if the direction of the frequency change was away from the alternative category than if it was toward the alternative category. The results demonstrate that categorization of sounds by pitch is accomplished with reference to a previous processing episode.     

Category similarity effects in children's semantic memory retrieval

Fourth-, sixth-, and eighth-grade students were required to indicate whether or not a stimulus word belonged in either of two semantic categories that were held in memory. Each category pair was either semantically similar or semantically dissimilar. The results indicated that even for the youngest children, similar categories required less search time than dissimilar categories. It was suggested that while dissimilar categories had to be accessed successively prior to search, subjects were able to consolidate similar categories into a superordinate group which eliminated the time to shift from one category to another during search. The results were contrasted with clustering studies which have demonstrated a minimal amount of memory facilitation in children for conceptually related items over unrelated items.     

The effect of repetition and similarity on sequence learning

Repetition is a pervasive feature of children’s environments, and may be an important contributor to learning such complex sequential structures as language. Endress, Dehaene-Lambertz, and Mehler (2007) found that repeated tone sequences were learned more easily than sequences containing ordinal relations, but there have been no direct comparisons of repeating sequences versus sequences that contain similar, but not identical, stimuli. In Experiment 1, we compared learning from repeating tone sequences to learning from tones that varied in similarity, and confirmed that repetition is a special case for learning. In Experiment 2 we showed that the learning distinction between repeated and similar elements is not affected by whether similarity is variable. We conclude by indicating that repetition provides an important constraint on learning, and we discuss the extent to which such constraints are consistent with general-purpose statistical learning mechanisms.     

Enemies and friends in the neighborhood: orthographic similarity effects in semantic categorization

Studies investigating orthographic similarity effects in semantic tasks have produced inconsistent results. The authors investigated orthographic similarity effects in animacy decision and in contrast with previous studies, they took semantic congruency into account. In Experiments 1 and 2, performance to a target (cat) was better if a previously studied neighbor (rat) was congruent (i.e., belonged to the same animate-inanimate category) than it was if it was incongruent (e.g., mat). In Experiments 3 and 4, performance was better for targets with more preexisting congruent neighbors than for targets with more preexisting incongruent neighbors. These results demonstrate that orthographic similarity effects in semantic categorization are conditional on semantic congruency. This strongly suggests that semantic information becomes available before orthographic processing has been completed.     

Dissociation of semantic priming, word and non-word repetition effects by event-related potentials

Two experiments investigated the modulation of event-related potentials (ERPs) by semantic priming and item repetition. In Experiment 1, subjects silently counted occasional non-words against a background of words, a proportion of which were either semantic associates or repetitions of a preceding word. Compared to control items, ERPs to repeated words were distinguished by an early (ca. 200 msec) transient negative-going deflection and a later, topographically widespread and temporally sustained positive-going shift. In contrast, semantically primed words showed a relatively small, topographically and temporally limited positive-going modulation peaking around 500 msec. These data were interpreted as evidence against models of priming and repetition which postulate similar loci for these effects. In Experiment 2, subjects counted occasional words against a background of non-words, some of which were repeated. ERPs to repetitions showed a similar early ERP modulation to that in Experiment 1, and also displayed a later slow positive shift. This latter effect was smaller in magnitude and had a delayed onset in comparison to Experiment 1. It was concluded that the effects of repetition differ as a consequence of whether, prior to their first presentation, items possess a representation in lexical memory.     

Geometric and semantic similarity in visual masking

Semantic and geometric or physical similarity were manipulated separately in a backward-masking situation. When the target was a word to be read aloud, formal similarity between the letters of target and mask facilitated target recognition, as did associative similarity. Masking a target word by its own anagram also facilitated whole word report. In contrast, formal similarity was inhibitory rather than facilitatory of report when the target was spelled letter-by-letter, rather than read whole. This was true even for the same target words whose whole report was facilitated by formal similarity. A model to account for this reversal in the broader context of the neural substrate of reading is advanced. It is proposed that letter and word processing are fundamentally different in that letters are recognized by hierarchical feature analysis while words are stored and recognized wholistically by diffuse and redundant networks. Implications of the results for the study of reading are discussed.     

Interaction of semantic and perceptual processes in repetition blindness

Two experiments examined the effect of activation of higher-level semantic representations on lower-level perceptual representations. A forced-choice discrimination paradigm was used, a method known to produce repetition blindness (RB) for words unconfounded by memory demands or response bias. In Experiment 1, equivalent reductions in RB (as measured by omission error rate and by d') occurred when successive word pairs were identical in: (1) form, pronunciation, and meaning (both uppercase versions of the same word); (2) pronunciation and meaning but not form (lowercase versus uppercase; lexical identity); and (3) pronunciation, but not form or meaning (homonyms; phonological identity), relative to when the words were unrelated on all dimensions. The RB effect was markedly attenuated, but not eliminated, when the words were semantically related. Similar results were obtained in Experiment 2 using a larger group of subjects. These findings show that higher-order semantic representations can have a top-down influence on judgements based on lower-order perceptual representations. The results are discussed within the framework of a cascade model of object processing in the human brain.     

Contrasting graded effects of semantic similarity and association across the concreteness spectrum

Evidence from healthy individuals and neuropsychological patients with deep dyslexia or semantic refractory access dysphasia suggests that abstract and concrete concepts have different dependencies upon associative and similarity-based representational frameworks. However, the importance of information about semantic similarity for concepts that lie across the full concreteness spectrum has not been investigated. Here we report the performance of healthy individuals on an odd-one-out task involving semantically related word triplets that showed continuous variation for the key variables of concreteness, similarity strength, and association strength. In addition, data from a stroke aphasic patient tested on a matching-to-sample task based on the same abstract, middle-concreteness, and concrete stimuli are also presented. The effects of similarity and association strength upon performance were both shown to interact significantly with concreteness, but in opposite directions. The effect of semantic similarity increased with concreteness but the effect of association decreased with concreteness. This research provides further evidence for the proposal that abstract and concrete words have different dependencies upon associative and similarity-based information. It also develops the proposal by providing data that are consistent with a graded and not binary or ungraded model of the relationship between concreteness and these two forms of semantic relationship.     

Contrasting graded effects of semantic similarity and association across the concreteness spectrum

Evidence from healthy individuals and neuropsychological patients with deep dyslexia or semantic refractory access dysphasia suggests that abstract and concrete concepts have different dependencies upon associative and similarity-based representational frameworks. However, the importance of information about semantic similarity for concepts that lie across the full concreteness spectrum has not been investigated. Here we report the performance of healthy individuals on an odd-one-out task involving semantically related word triplets that showed continuous variation for the key variables of concreteness, similarity strength, and association strength. In addition, data from a stroke aphasic patient tested on a matching-to-sample task based on the same abstract, middle-concreteness, and concrete stimuli are also presented. The effects of similarity and association strength upon performance were both shown to interact significantly with concreteness, but in opposite directions. The effect of semantic similarity increased with concreteness but the effect of association decreased with concreteness. This research provides further evidence for the proposal that abstract and concrete words have different dependencies upon associative and similarity-based information. It also develops the proposal by providing data that are consistent with a graded and not binary or ungraded model of the relationship between concreteness and these two forms of semantic relationship.     

The similarity and diversity of semantic relations

There is a rich variety of semantic relations in natural languages. Subjects’ perceptions of similarities among relations were studied for a wider variety of relations than had been used in previous studies. Forty subjects sorted 31 cards bearing five example pairs of each of 31 semantic relations. Subjects were able both to distinguish the relations and to perceive their similarities. A hierarchical clustering analysis of the sorting data indicated that the subjects perceived five families of semantic relations (contrasts, class inclusion, similars, case relations, and part-wholes). The five families were distinguished in terms of three properties of semantic relations: contrasting/noncontrasting, logical/pragmatic, and inclusion/noninclusion. Within each family, relations also were sorted in ways consistent with their defining properties. Relations were therefore viewed not as unanalyzable primitives, but in terms of the relational properties that distinguished them.     

Identity and similarity in repetition blindness: no cross-over interaction

The difficulty in reporting both occurrences of a repeated item is a phenomenon referred to as repetition blindness (RB). RB has been proposed to result from temporal limitations in creating separate episodic tokens for a twice-activated type. Recently, Chialant and Caramazza (Cognition 63 (1997) 79-119) disputed the conventional view that RB for non-identical words (orthographic RB, as in lice and lick) results from the same mechanism as identity RB, and proposed that orthographic RB arises from competition for lexical selection. Supporting evidence was that identical and merely similar words showed different amounts of RB as a function of stimulus onset asynchrony (lag). Four experiments failed to replicate Chialant and Caramazza's finding that identity RB decreases, but orthographic RB increases, as a function of lag. Instead, RB for all stimuli, including homonym pairs, declined monotonically with lag. These results are consistent with a common mechanism underlying RB for identical and orthographically similar words and with prior research suggesting that RB in similar words occurs at a sublexical level.     

Phonological, semantic, and repetition priming with homophones

The current study investigated semantic, repetition, and phonological priming using heterographic homonyms (homophones) as stimuli in a lexical decision task. As in previous research, reliable semantic priming and repetition effects were found. In addition, the statistical additivity of these two effects was replicated using homophones as stimuli. Using homophones, a reliable phonological priming effect was found when 10-16 intervening trials were used. This effect was also statistically independent of semantic priming indicating the independence of the mechanisms responsible for semantic and phonological priming. The magnitude of the phonological priming effect was also significantly less than the magnitude of the repetition effect. This difference in effect magnitude was taken to indicate that the repetition of phonological information is not a primary contributor to the repetition effect.