Representations Of Spoken Words Research Articles

Evidence for a Spoken Word Lexicon in the Auditory Ventral Stream.

The existence of a neural representation for whole words (i.e., a lexicon) is a common feature of many models of speech processing. Prior studies have provided evidence for a visual lexicon containing representations of whole written words in an area of the ventral visual stream known as the visual word form area. Similar experimental support for an auditory lexicon containing representations of spoken words has yet to be shown. Using functional magnetic resonance imaging rapid adaptation techniques, we provide evidence for an auditory lexicon in the auditory word form area in the human left anterior superior temporal gyrus that contains representations highly selective for individual spoken words. Furthermore, we show that familiarization with novel auditory words sharpens the selectivity of their representations in the auditory word form area. These findings reveal strong parallels in how the brain represents written and spoken words, showing convergent processing strategies across modalities in the visual and auditory ventral streams.

Neural Response Suppression Predicts Repetition Priming of Spoken Words and Pseudowords

An important method for studying how the brain processes familiar stimuli is to present the same item on more than one occasion and measure how responses change with repetition. Here we use repetition priming in a sparse functional magnetic resonance imaging (fMRI) study to probe the neuroanatomical basis of spoken word recognition and the representations of spoken words that mediate repetition priming effects. Participants made lexical decisions to words and pseudowords spoken by a male or female voice that were presented twice, with half of the repetitions in a different voice. Behavioral and neural priming was observed for both words and pseudowords and was not affected by voice changes. The fMRI data revealed an elevated response to words compared to pseudowords in both posterior and anterior temporal regions, suggesting that both contribute to word recognition. Both reduced and elevated activation for second presentations (repetition suppression and enhancement) were observed in frontal and posterior regions. Correlations between behavioral priming and neural repetition suppression were observed in frontal regions, suggesting that repetition priming effects for spoken words reflect changes within systems involved in generating behavioral responses. Based on the current results, these processes are sufficiently abstract to display priming despite changes in the physical form of the stimulus and operate equivalently for words and pseudowords.

Recent developments in language acquisition and reading research:The phonological basis of children’s reading difficulties

AbstractThis review examines the convergence of recent developments in the fields of language and literacy development and, in particular, developments relating phonological development to both language and reading development. It begins by examining the issue of how children represent spoken words. In particular, it presents recent work arguing that, throughout early and even middle childhood, children’s representations of spoken words are reorganised as sequences of phonemes. The second section examines poor readers’ phonologicol recoding difficulties and, in particular, the contribution of phonological awareness to early reading success. This section includes an overview of phonologicol awareness training studies in “at-risk” preschool and kindergarten children. The final section examines phonologicol processing difficulties as a common underlying cause of reading dificulties.This section provides a theoretical context for practitioners to understand diverse findings relating performance on a wide range of tasks to children’s reading achievement.

Specificity of memory representations for spoken words.

Many theories of spoken word recognition assume that lexical items are stored in memory as abstract representations. However, recent research (e.g., Goldinger, 1996) has suggested that representations of spoken words in memory are veridical exemplars that encode specific information, such as characteristics of the talker's voice. If representations are exemplar based, effects of stimulus variation such as that arising from changes in the identity of the talker may have an effect on identification of and memory for spoken words. This prediction was examined for an implicit and explicit task (lexical decision and recognition, respectively). Comparable amounts of repetition priming in lexical decision were found for repeated words, regardless of whether the repetitions were in the same or in different voices. However, reaction times in the recognition task were faster if the repetition was in the same voice. These results suggest a role for both abstract and specific representations in models of spoken word recognition.

Possible word boundary constraints on multiple activation of form‐based representations of spoken words.

Most current theories of spoken word perception (e.g., Cohort theory, TRACE, Shortlist, Neighborhood Activation Model) propose that multiple form‐based representations of words are activated in memory during recognition. Few of these theories, however, specify the precise nature of multiple activation of words in fluent speech or the constraints on activation imposed by information within the speech signal. The degree to which word boundary information in two‐word sequences may limit lexical hypotheses during recognition was examined. Using a cross‐modal priming technique, spoken two‐word utterances, such as NOTE‐RAIL, were presented and the activation of possible lexical items spanning the word boundary (e.g., TRAIL) was investigated. The implications of these results for current theories of spoken word recognition will be discussed. [Work supported by NIDCD.]

Talker variability and token variability in spoken word recognition: Effects on repetition priming in lexical decision and an explicit recognition memory task

Recent research (e.g., Goldinger, 1993) has suggested that representations of spoken words in memory may be veridical exemplars that encode instance‐specific information, such as characteristics of the talker’s voice. This account of spoken word recognition was examined in work presented at a recent meeting of the Acoustical Society (Luce and Lyons, 1994). Evidence was presented in support of the claim that exemplar‐type representations are used in explicit recognition memory tasks, but that more abstract representations play a role in tasks such as lexical decision. However, the previous work did not make clear the explicitness of the information encoded in exemplar‐type representations; specifically whether only gross factors such as talker variation are encoded, or whether within speaker variation is also encoded. The present research expands on these previous results by comparing the effects of talker variation and same‐speaker token variation on a repetition priming and an explicit recognition memory task. [Work supported by NIDCD.]

What happened to the “n” of sink? Children's spellings of final consonant clusters

Children's spellings provide a window on their representations of spoken words. These representations may not always match those assumed by the conventional orthography. We examined one specific case in which this may be true, that of CVCC (consonant-vowel-consonant-consonant) syllables. In three experiments, first graders often produced spellings such as “sach” for the nonword /sænt∫/ and “pit” for the nonword /pılt/. We suggest that many children consider these syllables to contain three phonemes instead of four phonemes. Rather than treating the postvocalic nasal or liquid as a separate phoneme, children consider it an attribute of the vowel. This suggestion was supported by the results of a phoneme counting test in which children often used three tokens for syllables such as /sænt∫/ and /pılt/. The results suggest that, in addition to difficulties in gaining access to phonemes, nonstandard phonemic representations may be a stumbling block in the acquisition of alphabetic literacy.

Effects of stimulus variability on perception and representation of spoken words in memory.

A series of experiments was conducted to investigate the effects of stimulus variability on the memory representations for spoken words. A serial recall task was used to study the effects of changes in speaking rate, talker variability, and overall amplitude on the initial encoding, rehearsal, and recall of lists of spoken words. Interstimulus interval (ISI) was manipulated to determine the time course and nature of processing. The results indicated that at short ISIs, variations in both talker and speaking rate imposed a processing cost that was reflected in poorer serial recall for the primary portion of word lists. At longer ISIs, however, variation in talker characteristics resulted in improved recall in initial list positions, whereas variation in speaking rate had no effect on recall performance. Amplitude variability had no effect on serial recall across all ISIs. Taken together, these results suggest that encoding of stimulus dimensions such as talker characteristics, speaking rate, and overall amplitude may be the result of distinct perceptual operations. The effects of these sources of stimulus variability in speech are discussed with regard to perceptual saliency, processing demands, and memory representation for spoken words.

The representation of voice information in spoken word recognition: Differential effects of repetition in lexical decision and recognition

Much recent research [e.g., Goldinger (1993)] has suggested that representations of spoken words in memory are veridical exemplars that encode specific information, such as characteristics of the talker’s voice. This exemplar account of phonetic form representations was examined in an experiment in which subjects heard two lists of words. The words in the first list were spoken by both a male and a female speaker. The words in the second list were (1) repetitions of words in the first list spoken by the same talker, (2) repetitions of words spoken by a different talker, or (3) new words. Subjects always made lexical decision responses to the stimuli in the first list. For the second list, subjects in one condition again made lexical decision responses; subjects in another condition performed a recognition task. Results indicated that changing the voice of the talker across lists had no effect on the repetition priming effect when subjects made lexical decisions to the second list. However, in the recognition task, reaction times were faster if the voice of the talker was preserved. These results are consistent with an exemplar-based account for recognition but also implicate the role of more abstract representations in identification. [Work supported by NIDCD.]

Do the Beginnings of Spoken Words Have a Special Status in Auditory Word Recognition?

Six cross-modal priming experiments were conducted that investigated whether representations of spoken words in memory (base word) may be activated by similar-sounding nonwords. The experiments demonstrated that nonwords that differed in one or two linguistic features from a base word resulted in significant priming effects for semantic associates to the base word. Nonwords that deviated by more linguistic features from a base word showed no priming effects. Finally, nonwords in which either the medial or initial portions were altered showed comparable priming effects. This result held for a set of two-syllable words and for a group of longer words (at least three syllables). A model of auditory word recognition is discussed in which partial acoustic-phonetic information in a spoken word is mapped onto a lexical representation in memory based on goodness-of-fit. It is argued that this mapping process affords no particular status to word-initial phonemes.

Speeded comparisons of spoken words

This research evaluated the claim that spoken words are recognized as soon as they diverge from all other words in memory. Pairs of consonant‐vowel‐consonant (CVC) words were presented for a speeded SAME‐DIFFERENT response. Subjects were instructed to respond SAME or DIFFERENT as soon as they could determine if the second word of the pair was the same as or different from the first word. Reaction times were measured from the onset of the second word. The overlap between the two members of the pairs of spoken words was systematically manipulated to determine if reaction times to say DIFFERENT were a simple function of the point at which the two words differed or whether overlapping segments after the divergence point would affect reaction times. The results are consistent with the notion that the temporal window for mapping information onto representations of spoken words in memory spans more than a feature or segment and that spoken words are not necessarily recognized as soon as they diverge from other words in memory.

Object‐naming deficits in developmental dyslexia

ABSTRACTThis paper reports two experiments which focus on the object naming deficits of dyslexic readers. In Experiment 1, dyslexic and normal readers were asked to name objects depicted by pictures or following their spoken definition. Ten‐year‐old dyslexics named fewer objects correctly than other children of a similar age, performing only as well as a younger group of 8‐year‐old normal readers. This was true irrespective of the modality through which they were tested. In terms of naming latency, however, they were similar to comparison groups. In Experiment 2, nine‐year‐old dyslexic and normal readers performed as well as each other in a receptive vocabulary test in which pictures had to be matched to spoken words. However, once again, on a picture naming test, the dyslexics did less well than controls. We argue that dyslexic children are subject to verbal naming difficulties which cannot be accounted for by generally low levels of vocabulary knowledge. Their problems are attributable not to difficulties in semantic representation but to difficulties with the lexical‐phonological representation of spoken words they know. We propose that, in turn, these difficulties are related to their memory and reading problems.

Synthetic voices for computers

The two methods described for giving voices to computers recognize the importance of economical storage of speech information and extensive vocabularies, and consequently are based on principles of speech synthesis. The first, formant synthesis, generates connected speech from low-bit-rate representations of spoken words. The second, text synthesis, produces connected speech solely from printed English text. For both methods the machine must contain stored knowledge of fundamental rules of language and acoustic constraints of human speech. Formant synthesis from an input information rate of about 1000 bits per second is demonstrated, as is text synthesis from a rate of about 75 bits per second. To give the reader an opportunity to evaluate some of the results described, a sample recording is available; see Appendix A for details.