The effects of word co-occurrence on short-term memory: associative links in long-term memory affect short-term memory performance

In immediate serial recall tasks, high-frequency words are recalled better than low-frequency words. This has been attributed to high-frequency words' being better represented and providing more effective support to a redintegration process at retrieval (C. Hulme et al., 1997). In studies of free recall, there is evidence that frequency of word co-occurrence, rather than word frequency per se, may explain the recall advantage enjoyed by high-frequency words (J. Deese, 1960). The authors present evidence that preexposing pairs of low-frequency words, so as to create associative links between them, has substantial beneficial effects on immediate serial recall performance. These benefits, which are not attributable to simple familiarization with the words per se, do not occur for high-frequency words. These findings indicate that associative links between items in long-term memory have important effects on short-term memory performance and suggest that the effects of word frequency in short-term memory tasks are related to differences in interitem associations in long-term memory.     

Think before you speak: pauses, memory search, and trace redintegration processes in verbal memory span

Immediate memory span and speed of memory search were assessed for words and nonwords of short and long spoken duration. Memory span was substantially greater for words than for nonwords and for short than for long items, though speed of memory search was unaffected by either length or lexicality. An analysis of the temporal pattern of responses in the memory span task indicated that inter-item pauses were longer between nonwords than words but that these pause durations were unaffected by item length. A model of verbal short-term memory span is described in which trace selection from a short-term store and the redintegration (restoration) of degraded phonological traces both occur in the pauses between saying successive items. Both trace selection and trace redintegration appear to play important roles in accounting for individual differences in memory span.     

Dissociable lexical and phonological influences on serial recognition and serial recall

The impact of the lexicality of memory items on memory performance was compared in two paradigms, serial recall and serial recognition. Experiments 1 to 3 tested 7- and 8-year-old children. Memory accuracy was only mildly impaired in lists containing nonwords compared with words in a serial recognition task involving judgements of whether the items in two sequences were in the same order (Experiment 1), although a substantial advantage for word over nonword items from the same stimulus pool was found in serial recall (Experiment 2). A stronger influence of lexicality on serial recall than serial recognition was further demonstrated in Experiments 3A and 3B, and in 4A and 4B using adult participants. These experiments also established comparable degrees of sensitivity to the phonological similarity of the memory sequences in the two paradigms. The phonological similarity effect in serial recall was found to arise from increased phoneme order errors, whereas the lexicality effect was due principally to the greater frequency of phoneme identity errors for nonwords. It is proposed that the lexicality effect originates in the redintegration of item information just prior to recall, and that this process is largely bypassed in serial recognition.     

Global processing speed as a mediator of developmental changes in children's auditory memory span

This study examined the role of global processing speed in mediating age increases in auditory memory span in 5- to 13-year-olds. Children were tested on measures of memory span, processing speed, single-word speech rate, phonological sensitivity, and vocabulary. Structural equation modeling supported a model in which age-associated increases in processing speed predicted the availability of long-term memory phonological representations for redintegration processes. The availability of long-term phonological representations, in turn, explained variance in memory span. Maximum speech rate did not predict independent variance in memory span.     

Redintegration and the benefits of long-term knowledge in verbal short-term memory: an evaluation of Schweickert's (1993) multinomial processing tree model

The impact of four long-term knowledge variables on serial recall accuracy was investigated. Serial recall was tested for high and low frequency words and high and low phonotactic frequency nonwords in 2 groups: monolingual English speakers and French-English bilinguals. For both groups the recall advantage for words over nonwords reflected more fully correct recalls with fewer recall attempts that consisted of fragments of the target memory items (one or two of the three target phonemes recalled correctly); completely incorrect recalls were equivalent for the 2 list types. However, word frequency (for both groups), nonword phonotactic frequency (for the monolingual group), and language familiarity all influenced the proportions of completely incorrect recalls that were made. These results are not consistent with the view that long-term knowledge influences on immediate recall accuracy can be exclusively attributed to a redintegration process of the type specified in multinomial processing tree model of immediate recall. The finding of a differential influence on completely incorrect recalls of these four long-term knowledge variables suggests instead that the beneficial effects of long-term knowledge on short-term recall accuracy are mediated by more than one mechanism.     

Modeling the effects of irrelevant speech on memory

The feature model (Nairne, 1990) is extended to account for the effects of irrelevant speech and concomitant interactions in immediate serial recall. In the feature model, both articulatory suppression and irrelevant speech are seen as adding noise to the memory representation, the difference being that articulatory suppression diverts more resources than does irrelevant speech. The addition of noise impairs recall because it reduces the probability of successful redintegration. When a competitor is incorrectly recalled, rather than the correct item, this competitor is recalled out of order, producing an increase in order errors. Six simulations are reported that show that the model accounts for (1) the impairment by both irrelevant speech and articulatory suppression, (2) the irrelevance of the phonological and semantic composition of the irrelevant speech, (3) greater disruption when the irrelevant speech tokens vary, (4) the abolition of the phonological similarity effect for visual, but not for auditory, items, (5) the abolition of the word length effect for both visual and auditory items, and (6) the abolition of the irrelevant speech effect under articulatory suppression for both visual and auditory items. The feature model is compared with the two other major views of irrelevant speech, the phonological store hypothesis and the changing state hypothesis.     

Irrelevant sound disrupts speech production: Exploring the relationship between short-term memory and experimentally induced slips of the tongue

To explore the relationship between short-term memory and speech production, we developed a speech error induction technique. The technique, which was adapted from a Japanese word game, exposed participants to an auditory distractor word immediately before the utterance of a target word. In Experiment 1, the distractor words that were phonologically similar to the target word led to a greater number of errors in speaking the target than did the dissimilar distractor words. Furthermore, the speech error scores were significantly correlated with memory span scores. In Experiment 2, memory span scores were again correlated with the rate of the speech errors that were induced from the task-irrelevant speech sounds. Experiment 3 showed a strong irrelevant-sound effect in the serial recall of nonwords. The magnitude of the irrelevant-sound effects was not affected by phonological similarity between the to-be-remembered nonwords and the irrelevant-sound materials. Analysis of recall errors in Experiment 3 also suggested that there were no essential differences in recall error patterns between the dissimilar and similar irrelevant-sound conditions. Weproposed two different underlying mechanisms in immediate memory, one operating via the phonological short-term memory store and the other via the processes underpinning speech production.     

Effects of word frequency and age of acquisition on short-term memory span

The experiments reported examine the effects of two highly related variables, word frequency and age of acquisition, on short-term memory span. Short-term memory span and speech rate were measured for sets of words which independently manipulated frequency and age of acquisition. It was found that frequency had a considerable effect on short-term memory span, which was not mediated by speech rate differences—although frequency did affect speech rate in one experiment. For age of acquisition, this situation was reversed; there was a small but significant effect of age of acquisition on speech rate, but no effect on memory span. This occurred despite results confirming that the stimuli used in the experiments produce an effect of age of acquisition on word naming. The results are discussed in terms of a two-component view of performance on short-term memory tasks.     

Lexical coherence in short-term memory: Strategic reconstruction or “semantic glue”?

This study explores the foundation of lexical/semantic phoneme binding effects in verbal short-term memory (STM). The immediate serial recall of pure lists of words and nonwords was compared with the recall of mixed lists that had either a predictable, alternating structure (e.g., wnwnwn) or an unpredictable structure (i.e., the serial positions of the words/nonwords could not be known in advance). The study provides evidence for two separate mechanisms by which long-term linguistic knowledge contributes to STM. First, there was evidence for automatic lexical/semantic binding effects that were independent of knowledge of lexical status. The nonwords in both types of mixed list damaged word recall and encouraged the phonological elements of words to migrate. In both alternating and unpredictable mixed lists, the phonemes of words were more likely than the phonemes of nonwords to be recalled together as a coherent item, suggesting that lexical/semantic knowledge encourages the phonological elements of words to emerge together in immediate serial recall, even when lexical status is unknown. Secondly, there was evidence for “strategic redintegration”, which was dependent on prior knowledge of the lexical status of the items in mixed lists. When participants recalled items that they knew to be words in advance, they were able to use this knowledge to constrain their responses so that they were more likely to be lexically appropriate. These findings motivate modifications to current theories of the interaction between linguistic knowledge and verbal short-term memory.     

Phonological similarity neighborhoods and children’s short-term memory: Typical development and dyslexia

In this article, we explore whether structural characteristics of the phonological lexicon affect serial recall in typically developing and dyslexic children. Recent work has emphasized the importance of long-term phonological representations in supporting short-term memory performance. This occurs via redintegration (reconstruction) processes, which show significant neighborhood density effects in adults. We assessed whether serial recall in children was affected by neighborhood density in word and nonword tasks. Furthermore, we compared dyslexic children with typically developing children of the same age or reading level. Dyslexic children are held to have impaired phonological representations of lexical items. These impaired representations may impair or prevent the use of long-term phonological representations to redintegrate short-term memory traces. We report significant rime neighborhood density effects for serial recall of both words and nonwords, for both dyslexic and typically developing children.     

Word-length effects in immediate memory: Overwriting trace decay theory

Memory is worse for items that take longer to pronounce, even when the items are equated for frequency, number of syllables, and number of phonemes. Current explanations of the word-length effect rely on a time-based decay process within the articulatory loop structure in working memory. Using an extension of Nairne’s (1990) feature model, we demonstrate that the approximately linear relationship between span and pronunciation rate can be observed in a model that does not use the concept of decay. Moreover, the feature model also correctly predicts the effects of modality, phonological similarity, articulatory suppression, and serial position on memory for items of different lengths. We argue that word-length effects do not offer sufficient justification for including time-based decay components in theories of memory.     

Redintegration and Response Suppression in Serial Recall: A Dynamic Network Model

This article presents a dynamic network model of redintegration and response suppression. Redintegration is the process that disambiguates partially retrieved memorial information into an overt response, and response suppression renders retrieved items temporarily unavailable for further report. Most models of serial recall assume the presence of both processes, but few explain how they are performed. Exploration of the network revealed that it can predict recency in serial recall of lists of varying lengths, the pattern of the associated transposition, omission, intrusion, and repetition errors, and the temporal dynamics of retrieval. The network thus augments any existing model of serial recall that assumes a distributed vector representation but does not specify the processes underlying redintegration and response suppression.     

Speech rate and the development of short-term memory span

Two experiments are reported which examine the effects of word duration on memory span in subjects of different ages. The same linear function relating recall to speech rate (assessed by the speed of repeating words) fits the results of subjects ranging in age from 4 years old to adulthood. It is concluded that developmental increases in short-term memory span can be explained in terms of increases in speech rate and that there is no evidence for an increase in short-term memory capacity. Analyses of the children's speech suggests that increases in speech rate with age reflect increases in the speed of articulation of individual words, rather than any change in the duration of pauses between successive words or changes in coarticulation between words.     

<Emphasis Type="Bold">Keeping it together: Semantic coherence stabilizes phonological sequences in short-term memory</Emphasis>

Our ability to hold a sequence of speech sounds in mind, in the correct configuration, supports many aspects of communication, but the contribution of conceptual information to this basic phonological capacity remains controversial. Previous research has shown modest and inconsistent benefits of meaning on phonological stability in short-term memory, but these studies were based on sets of unrelated words. Using a novel design, we examined the immediate recall of sentence-like sequences with coherent meaning, alongside both standard word lists and mixed lists containing words and nonwords. We found, and replicated, substantial effects of coherent meaning on phoneme-level accuracy: The phonemes of both words and nonwords within conceptually coherent sequences were more likely to be produced together and in the correct order. Since nonwords do not exist as items in long-term memory, the semantic enhancement of phoneme-level recall for both item types cannot be explained by a lexically based item reconstruction process employed at the point of retrieval (“redintegration”). Instead, our data show, for naturalistic input, that when meaning emerges from the combination of words, the phonological traces that support language are reinforced by a semantic-binding process that has been largely overlooked by past short-term memory research.     

Memory Span for Arabic Numerals and Digit Words: Evidence for a Limited-capacity, Visuo-spatial Storage System

Six experiments examined the determinants of the numeral advantage effect: the finding that memory span for Arabic numerals (1, 2, 3, etc.) is greater than for digit words (one, two, three, etc.). The speed of item identification for numeral and digit words was unrelated to memory span for the same items and a larger memory span for numerals persisted under concurrent random generation (Experiment 1). The numeral advantage, however, was abolished when the items were presented in random locations within an invisible 3x3 grid (Experiment 2) and in locations on a horizontal plane that ran contrary to the natural direction of reading (Experiment 3). When the items were presented in the same location, a disruption of the spatial component of visuo-spatial working memory eliminated the numeral advantage (Experiment 4), whereas interference with the visual component of the system did not (Experiment 5). When the items were spatially distributed in a 3x3 matrix, however, neither visual nor spatial interference abolished the effect (Experiment 6). Taken together, these findings suggest that the numeral advantage effect is mediated by discrete components in visuo-spatial working memory dedicated to the temporary storage and renewal of visual codes and questions the assumption that the underlying mechanisms in immediate, visual serial recall are equivalent between stimulus categories.     

The Word-length Effect and Disyllabic Words

Three experiments compared immediate serial recall of disyllabic words that differed on spoken duration. Two sets of long- and short-duration words were selected, in each case maximizing duration differences but matching for frequency, familiarity, phonological similarity, and number of phonemes, and controlling for semantic associations. Serial recall measures were obtained using auditory and visual presentation and spoken and picture-pointing recall. In Experiments 1a and 1b, using the first set of items, long words were better recalled than short words. In Experiments 2a and 2b, using the second set of items, no difference was found between long and short disyllabic words. Experiment 3 confirmed the large advantage for short-duration words in the word set originally selected by Baddeley, Thomson, and Buchanan (1975). These findings suggest that there is no reliable advantage for short-duration disyllables in span tasks, and that previous accounts of a word-length effect in disyllables are based on accidental differences between list items. The failure to find an effect of word duration casts doubt on theories that propose that the capacity of memory span is determined by the duration of list items or the decay rate of phonological information in short-term memory.     

A multinomial processing tree model for degradation and redintegration in immediate recall

When items are presented for immediate recall, a verbal trace is formed and degrades quickly, becoming useless after about 2 sec. The span for items such as digits equals the number of items that canbe pronounced in the available time. The length of the items affects span by affecting pronunciation rate. Other properties, such as phonological similarity and lexicality, can affect span without affecting pronunciation rate. These properties change the trace's useful lifetime by affecting redintegration. An analogy is drawn between trace reconstruction and repair of errors in speech. When a trace is degraded, one process attempts to form a phoneme string, and another process attempts to form a word. The two processes are autonomous and can be selectively influenced by lexicality and phonological similarity. The resulting processing tree models make simple predictions that depend on whether or not the influenced processes are sequential. The results are illustrated with data from experiments by Besner and Davelaar (1982).     

All parts of an item are not equal: Effects of phonological redundancy on immediate recall

The process of redintegration is thought to use top-down knowledge to repair partly damaged memory traces. We explored redintegration in the immediate recall of lists from a limited pool of partly phonologically redundant pseudowords. In Experiment 1, four kinds of stimuli were created by adding the syllable /ne/ to two-syllable pseudowords, either to the middle (/tepa/ vs. /tenepa/) or to the end (/tepane/), or adding a different syllable to each item (/tepalo/, /vuropi/). The repeated syllable was thought to be available for redintegration. Lists of two-syllable pseudowords were recalled best, items with a redundant end were intermediate, and items with a redundant middle-syllable were as hard as nonredundant three-syllable items. In Experiment 2, the last syllable was predictable from context but not shared between all stimuli, reducing phonological similarity between items. Performance did not differ from the situation with identicallast syllables. In Experiment 3, a shared first syllable had a detrimental effecton memory. An error analysis showed that beneficial redundancy effects were accompanied by harmful similarity effects, impairing memory for nonredundant syllables. The balance between the two effects depended on syllable position.     

The association between auditory memory span and speech rate in children from kindergarten to sixth grade.

This study reexamined the association between speech rate and memory span in children from kindergarten to sixth grade (N = 152) in order to potentially account for the inconsistencies within the published literature on this topic. Some of the inconsistencies in past research may reflect the different methods adopted in assessing speech rate. In particular, repeating word triples may itself involve memory demands, contaminating the correlation between speech rate and memory span in younger children. Analyses using composite speech rate and memory span measures showed that speech rate for word triples shared variance with memory span that was independent of speech rate for single words. Moreover, speech rate for word triples was largely redundant with age in explaining additional variation in memory span once the effects of speech rate for single words were controlled.     

Implications of aging,lexicality, and item length for the mechanisms underlying memory span

Hulme, Maughan, and Brown (1991) provided evidence that the contribution of long-term memory to memory span performance was additive to the contribution of rehearsal rate (e.g., Baddeley, 1986). The present study further explored the relationship between these two contributions in younger and older adults. Speech rates and spans for short, medium, and long words and nonwords were obtained from subjects. Older adults had slower speech rates and smaller spans than did younger adults. Both groups’ data were fit well by linear functions relating speech rates to spans. However, the slope of the function that relates speech rate to memory span was greater for words than for nonwords. This finding supports the idea that long-term memory, as well as rehearsal rate, contributes to span performance, and that this contribution is not simply additive.