Paul A. Luce

Recognizing Spoken Words: The Neighborhood Activation Model

Objective: A fundamental problem in the study of human spoken word recognition concerns the structural relations among the sound patterns of words in memory and the effects these relations have on spoken word recognition. In the present investigation, computational and experimental methods were employed to address a number of fundamental issues related to the representation and structural organization of spoken words in the mental lexicon and to lay the groundwork for a model of spoken word recognition. Design: Using a computerized lexicon consisting of transcriptions of 20,000 words, similarity neighborhoods for each of the transcriptions were computed. Among the variables of interest in the computation of the similarity neighborhoods were: 1) the number of words occurring in a neighborhood, 2) the degree of phonetic similarity among the words, and 3) the frequencies of occurrence of the words in the language. The effects of these variables on auditory word recognition were examined in a series of behavioral experiments employing three experimental paradigms: perceptual identification of words in noise, auditory lexical decision, and auditory word naming. Results: The results of each of these experiments demonstrated that the number and nature of words in a similarity neighborhood affect the speed and accuracy of word recognition. A neighborhood probability rule was developed that adequately predicted identification performance. This rule, based onLuces (1959) choice rule, combines stimulus word intelligibility, neighborhood confusability, and frequency into a single expression. Based on this rule, a model of auditory word recognition, the neighborhood activation model, was proposed. This model describes the effects of similarity neighborhood structure on the process of discriminating among the acoustic‐phonetic representations of words in memory. The results of these experiments have important implications for current conceptions of auditory word recognition in normal and hearing impaired populations of children and adults.

When Words Compete: Levels of Processing in Perception of Spoken Words:

Current theories of spoken-word recognition posit two levels of representation and process: lexical and sublexical. By manipulating probabilistic phonotactics and similarity-neighborhood density, we attempted to determine if these two levels of representation have dissociable effects on processing. Whereas probabilistic phonotactics have been associated with facilitatory effects on recognition, increases in similarity-neighborhood density typically result in inhibitory effects on recognition arising from lexical competition. Our results demonstrated that when the lexical level is invoked using real words, competitive effects of neighborhood density are observed. However, when strong lexical effects are removed by the use of nonsense word stimuli, facilitatory effects of phonotactics emerge. These results are consistent with a two-level framework of process and representation embodied in certain current models of spoken-word recognition.

A web-based interface to calculate phonotactic probability for words and nonwords in English.

Phonotactic probability refers to the frequency with which phonological segments and sequences of phonological segments occur in words in a given language. We describe one method of estimating phonotactic probabilities based on words in American English. These estimates of phonotactic probability have been used in a number of previous studies and are now being made available to other researchers via a Web-based interface. Instructions for using the interface, as well as details regarding how the measures were derived, are provided in the present article. The Phonotactic Probability Calculator can be accessed athttp://www.people.ku.edu/~mvitevit/PhonoProbHome.html.

Priming lexical neighbors of spoken words: effects of competition and inhibition

Two experiments employing an auditory priming paradigm were conducted to test predictions of the Neighborhood Activation Model of spoken word recognition (Luce & Pisoni, 1989, Neighborhoods of words in the mental lexicon. Manuscript under review). Acoustic-phonetic similarity, neighborhood densities, and frequencies of prime and target words were manipulated. In Experiment 1, priming with low frequency, phonetically related spoken words inhibited target recognition, as predicted by the Neighborhood Activation Model. In Experiment 2, the same prime-target pairs were presented with a longer inter-stimulus interval and the effects of priming were eliminated. In both experiments, predictions derived from the Neighborhood Activation Model regarding the effects of neighborhood density and word frequency were supported. The results are discussed in terms of competing activation of lexical neighbors and the dissociation of activation and frequency in spoken word recognition.

Phonetic priming, neighborhood activation, and PARSYN

Perceptual identification of spoken words in noise is less accurate when the target words are preceded by spoken phonetically related primes (Goldinger, Luce, & Pisoni, 1989). The present investigation replicated and extended this finding. Subjects shadowed target words presented in the clear that were preceded by phonetically related or unrelated primes. In addition, primes were either higher or lower in frequency than the target words. Shadowing latencies were significantly longer for target words preceded by phonetically related primes, but only when the prime-target interstimulus interval was short (50 vs. 500 msec). These results demonstrate that phonetic priming does not depend on target degradation and that it affects processing time. We further demonstrated that PARSYN—a connectionist instantiation of the neighborhood activation model—accurately simulates the observed pattern of priming.

Speech Perception, Word Recognition and the Structure of the Lexicon

This paper reports the results of three projects concerned with auditory word recognition and the structure of the lexicon. The first project was designed to experimentally test several specific predictions derived from MACS, a simulation model of the Cohort Theory of word recognition. Using a priming paradigm, evidence was obtained for acoustic-phonetic activation in word recognition in three experiments. The second project describes the results of analyses of the structure and distribution of words in the lexicon using a large lexical database. Statistics about similarity spaces for high and low frequency words were applied to previously published data on the intelligibility of words presented in noise. Differences in identification were shown to be related to structural factors about the specific words and the distribution of similar words in their neighborhoods. Finally, the third project describes efforts at developing a new theory of word recognition known as Phonetic Refinement Theory. The theory is based on findings from human listeners and was designed to incorporate some of the detailed acoustic-phonetic and phonotactic knowledge that human listeners have about the internal structure of words and the organization of words in the lexicon, and how, they use this knowledge in word recognition. Taken together, the results of these projects demonstrate a number of new and important findings about the relation between speech perception and auditory word recognition, two areas of research that have traditionally been approached from quite different perspectives in the past.

Phonotactics, Neighborhood Activation, and Lexical Access for Spoken Words

Probabilistic phonotactics refers to the relative frequencies of segments and sequences of segments in spoken words. Neighborhood density refers to the number of words that are phonologically similar to a given word. Despite a positive correlation between phonotactic probability and neighborhood density, nonsense words with high probability segments and sequences are responded to more quickly than nonsense words with low probability segments and sequences, whereas real words occurring in dense similarity neighborhoods are responded to more slowly than real words occurring in sparse similarity neighborhoods. This contradiction may be resolved by hypothesizing that effects of probabilistic phonotactics have a sublexical focus and that effects of similarity neighborhood density have a lexical focus. The implications of this hypothesis for models of spoken word recognition are discussed.

Acoustic-phonetic representations in word recognition*

This paper reviews what is currently known about the sensory and perceptual input that is made available to the word recognition system by processes typically assumed to be related to speech sound perception. In the first section, we discuss several of the major problems that speech researchers have tried to deal with over the last thirty years. In the second section, we consider one attempt to conceptualize the speech perception process within a theoretical framework that equates processing stages with levels of linguistic analysis. This framework assumes that speech is processed through a series of analytic stages ranging from peripheral auditory processing, acoustic-phonetic and phonological analysis, to word recognition and lexical access. Finally, in the last section, we consider several recent approaches to spoken word recognition and lexical access. We examine a number of claims surrounding the nature of the bottom-up input assumed by these models, postulated perceptual units, and the interaction of different knowledge sources in auditory word recognition. An additional goal of this paper was to establish the need to employ segmental representations in spoken word recognition.

A computational analysis of uniqueness points in auditory word recognition

To determine the extent to which words in isolation may be recognized prior to their offsets,uniqueness points oroptimal discrimination points were computed for all words in a 20,000-word computerized lexicon with more than two phonemes. The results of this analysis revealed that the frequency-weighted probability of a word’s diverging from all other words in the lexicon prior to the last phoneme was only 39. This finding suggests that an optimally efficient strategy of word recognition may be severely limited in scope due to structural properties of the mental lexicon.

Capacity demands in short-term memory for synthetic and natural speech

Three experiments were performed that compared recall for synthetic and natural lists of monosyllabic words. In the first experiment, presentation intervals of 1, 2, and 5 s per word were used. Although free recall was consistently poorer overall for the synthetic lists at all presentation rates, the decrement for synthetic stimuli did not increase differentially with faster rates. In a second experiment, zero, three, and six digits were presented visually for retention prior to free recall of each spoken word list in a preload paradigm. Fewer subjects were able to correctly recall all of the digits for the six-digit list than the three-digit list when the following word lists were synthetic. The third experiment required ordered recall of lists of natural and synthetic words. Differences in ordered recall between the synthetic and natural word lists were substantially larger for the primacy portion of the serial position curve than the recency portion. These results indicate that difficulties observed in the perception and comprehension of synthetic speech are due, in part, to increased processing demands in short-term memory.