Clive Frankish

Phonotactic influences on short-term memory

The impact of phonotactic probabilities on serial recall was investigated in a series of experiments. In Experiments 1A and 1B, 7 and 8 year olds were tested on their serial recall of monosyllabic words and of nonwords varying in phonotactic frequencies. A recall advantage to words over nonwords remained when stimuli were balanced for phonotactic probability, but nonword recall showed superior accuracy for high over low probability nonwords, as in Experiment 2. The nonword frequency effect appears to reflect the frequency of constituent syllables rather than biphones. Both lexicality and high phonotactic frequency led to increased proportions of full over partial recall of the memory stimuli. These findings indicate that decayed memory traces in phonological short-term memory can be reconstructed using either lexical or phonotactic knowledge.

Modality-specific grouping effects in short-term memory

Abstract Previous investigations of immediate serial recall have attributed the beneficial effect of temporal grouping to the increased rehearsal time afforded by intralist pauses. The three experiments described in the present article, together with previously reported data on auditory grouping, indicate that there is a substantial recall advantage for grouped lists with auditory but not with visual presentation. It is argued that temporal grouping of auditory lists enables greater use to be made of precategorical acoustic storage (PAS). Implications of these findings for theories concerning utilization of PAS information are briefly discussed.

Recognition accuracy and user acceptance of pen interfaces

The accuracy of handwriting recognition is often seen as a key factor in determining the acceptability of hand-held computers that employ a pen for user interaction. We report the results of a study in which the relationship between user satisfaction and recogniser performance was examined in the context of different types of target application. Subjects with no prior experience of pen computing evaluated the appropriateness of the pen interface for performing three different tasks that required translation of handwritten text. The results indicate that the influence of recogniser performance on user satisfaction depends on the task context. These findings are interpreted in terms of the task-related costs and benefits associated with handwriting recognition. Further analysis of recognition data showed that accuracy did not improve as subjects became more practised. However, substantial gains in accuracy could be achieved by selectively adapting the recogniser to deal with a small, user-specific subset of characters.

Long-term knowledge effects on serial recall of nonwords are not exclusively lexical

S. Roodenrys and M. Hinton (2002) reported superior recall for nonwords with large rather than small lexical neighborhoods when constituent biphone frequency was controlled, but comparable recall of high and low biphone frequency nonwords when neighborhood size was controlled, suggesting that long-term knowledge effects on nonword recall are lexically based. We report two experiments in which the same manipulations were made, but with neighborhood size controlled at the level of neighbor type. In Experiment 1, biphone frequency significantly influenced nonword recall when neighborhood size was controlled in this way. In Experiment 2, neighborhood size significantly influenced nonword recall when biphone frequency was controlled. These findings suggest that long-term knowledge contributions to nonword recall are not exclusively lexical but are based instead on both lexical and phonotactic knowledge of a language.

Perceptual organization and precategorical acoustic storage.

Current views of precategorical acoustic storage (PAS) have been largely based on differences in the level of recall of terminal list items as a function of input modality and on experiments in which various types of suffixes are added to unstructured auditory lists. Experiments with grouped lists reveal that PAS can make a far more extensive contribution to serial recall. A series of four experiments investigated grouping effects in relation to existing accounts of consolidation, attentional selection, and auditory masking in PAS. Grouping effects obtained with very brief intralist pauses were inconsistent with the consolidation and masking hypotheses. Contrary to the attentional hypothesis, nontemporal grouping by voice or by spatial location was found to be as effective as grouping by extended pauses. When nontemporal methods of grouping were combined with intralist pauses, the two sets of grouping cues were no better than one, suggesting that list segmentation by pauses and by item attributes must be explained in terms of a single process. These results are discussed in the context of previous research that implies the existence of an auditory store with a capacity greater than previously attributed to PAS. Existing data on modality and suffix effects are seen as specific instances of a more general relation between the structure of spoken sequences and their subsequent recall.

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.

Speech recognition technology for individuals with disabilities

There are estimated to be about 15 million people in the United States alone who are disabled to some degree, and it has been hypothesized that at least 10% of the worlds population experience some sort of physical impairment. Although such statistics are always open to debate, this does not detract from the suggestion that one of the most promising application areas for Automatic Speech Recognition (ASR) is in helping people with disabilities. It is within this context that the advantages of speech recognition are discussed, both in general and with specific relation to this user group. During the last decade, the development of more sophisticated techniques for analyzing incoming speech combined with the increased processing power of microcomputers has resulted in improved recognition performance. Consequently, speech recognizers are now either operational or being considered in a variety of industrial tasks and in office and aerospace environments. ASR applications specific to disabled users are revie...

Language familiarity effects in short-term memory: the role of output delay and long-term knowledge.

Four experiments examined the origins of language familiarity effects in bilingual short-term recall. In Experiments 1A and 1B, bilingual adults were tested on serial recall and probed serial recall of words and nonwords in their first and second languages. A first-language advantage was obtained on both measures, indicating that the beneficial effects of language familiarity are not exclusively attributable to lesser output delay during overt recall. In Experiments 2A and 2B, the same group of bilinguals was tested on serial recall and serial recognition of word lists in both languages. Although a sizeable first-language advantage was obtained on the serial recall measure, recognition performance was comparable in the two languages. On the basis of these results it is suggested that language differences in bilingual immediate memory arise in large part as a consequence of the differential availability of language-specific long-term knowledge with which to support retrieval processes in serial recall.

Design guidelines for speech recognition interfaces

Seven classes of design guidelines are described for interfaces which use speech recognition. The guidelines concern: (i) allocation of function within complex systems; (ii) parallel processing of speech with other modalities; (iii) design of command vocabulary; (iv) choice of command syntax; (v) use of feedback; (vi) template and user training; and (vii) choice of evaluation criteria.

Lexical and semantic influences on item and order memory in immediate serial recognition: Evidence from a novel task

Previous studies have reported that, in contrast to the effect on immediate serial recall, lexical/semantic factors have little effect on immediate serial recognition. This has been taken as evidence that linguistic knowledge contributes to verbal short-term memory in a redintegrative process at recall. Contrary to this view, we found that lexicality, frequency, and imageability all influenced matching span. The standard matching span task, requiring changes in item order to be detected, was less susceptible to lexical/semantic factors than was a novel task involving the detection of phoneme order and hence item identity changes. Therefore, in both immediate recognition and immediate serial recall, lexical/semantic knowledge makes a greater contribution to item identity than to item order memory. Task sensitivity, and not the absence of overt recall, may have underpinned previous failures to show effects of these variables in immediate recognition. We also compared matching span for pure and unpredictable mixed lists of words and nonwords. Lexicality had a larger impact on immediate recognition for pure than for mixed lists, in line with findings for immediate serial recall. List composition affected the detection of phoneme but not item order changes in matching span; similarly, in recall, mixed lists produce more frequent word phoneme migrations but not migrations of entire items. These results point to strong similarities between immediate serial recall and recognition. Lexical/semantic knowledge may contribute to phonological stability in both tasks.