Connectionist modelling of the separable processing of consonants and vowels

describe two aphasic patients, with impaired processing of vowels and consonants, respectively. The impairments could not be captured according to the sonority hierarchy or in terms of a feature level analysis. Caramazza et al. claim that this dissociation demonstrates separate representation of the categories of vowels and consonants in speech processing. We present two connectionist models of the management of phonological representations. The models spontaneously develop separable processing of vowels and consonants. The models have two hidden layers and are given as input vowels and consonants represented in terms of their phonological distinctive features. The first model is presented with feature bundles one at a time and the hidden layers have to combine their output to reproduce a unified copy of the feature bundle. In the second model a "fine-coded" layer receives information about feature bundles in isolation, and a "coarse-coded" layer receives information about each feature bundle in the context of the prior and subsequent feature bundle. Coarse-coding facilitated processing of vowels and fine-coding processing of consonants. These models show that separable processing of vowels and consonants is an emergent effect of modular processors operating on feature-based representations. We argue that it is not necessary to postulate an independent level of representation for the consonant/vowel distinction, separate from phonological distinctive features.     

Effects of phoneme repertoire

In three experiments, listeners detected vowel or consonant targets in lists of CV syllables constructed from five vowels and five consonants. Responses were faster in a predictable context (e.g., listening for a vowel target in a list of syllables all beginning with the same consonant) than in an unpredictable context (e.g., listening for a vowel target in a list of syllables beginning with different consonants). In Experiment 1, the listeners’ native language was Dutch, in which vowel and consonant repertoires are similar in size. The difference between predictable and unpredictable contexts was comparable for vowel and consonant targets. In Experiments 2 and 3, the listeners’ native language was Spanish, which has four times as many consonants as vowels; here effects of an unpredictable consonant context on vowel detection were significantly greater than effects of an unpredictable vowel context on consonant detection. This finding suggests that listeners’ processing of phonemes takes into account the constitution of their language’s phonemic repertoire and the implications that this has for contextual variability.     

Call me Alix,not Elix: vowels are more important than consonants in own‐name recognition at 5 months

Consonants and vowels differ acoustically and articulatorily, but also functionally: Consonants are more relevant for lexical processing, and vowels for prosodic/syntactic processing. These functional biases could be powerful bootstrapping mechanisms for learning language, but their developmental origin remains unclear. The relative importance of consonants and vowels at the onset of lexical acquisition was assessed in French‐learning 5‐month‐olds by testing sensitivity to minimal phonetic changes in their own name. Infants’ reactions to mispronunciations revealed sensitivity to vowel but not consonant changes. Vowels were also more salient (on duration and intensity) but less distinct (on spectrally based measures) than consonants. Lastly, vowel (but not consonant) mispronunciation detection was modulated by acoustic factors, in particular spectrally based distance. These results establish that consonant changes do not affect lexical recognition at 5 months, while vowel changes do; the consonant bias observed later in development does not emerge until after 5 months through additional language exposure.     

Word learning and phonetic processing in preschool-age children

Consonants and vowels have been shown to play different relative roles in different processes, including retrieving known words from pseudowords during adulthood or simultaneously learning two phonetically similar pseudowords during infancy or toddlerhood. The current study explores the extent to which French-speaking 3- to 5-year-olds exhibit a so-called “consonant bias” in a task simulating word acquisition, that is, when learning new words for unfamiliar objects. In Experiment 1, the to-be-learned words differed both by a consonant and a vowel (e.g., /byf/-/duf/), and children needed to choose which of the two objects to associate with a third one whose name differed from both objects by either a consonant or a vowel (e.g., /dyf/). In such a conflict condition, children needed to favor (or neglect) either consonant information or vowel information. The results show that only 3-year-olds preferentially chose the consonant identity, thereby neglecting the vowel change. The older children (and adults) did not exhibit any response bias. In Experiment 2, children needed to pick up one of two objects whose names differed on either consonant information or vowel information. Whereas 3-year-olds performed better with pairs of pseudowords contrasting on consonants, the pattern of asymmetry was reversed in 4-year-olds, and 5-year-olds did not exhibit any significant response bias. Interestingly, girls showed overall better performance and exhibited earlier changes in performance than boys. The changes in consonant/vowel asymmetry in preschoolers are discussed in relation with developments in linguistic (lexical and morphosyntactic) and cognitive processing.     

The structure of graphemic representations

The analysis of the spelling performance of a brain-damaged dysgraphic subject is reported. The subject's spelling performance was affected by various graphotactic factors, such as the distinction between consonant and vowel and graphosyllabic structure. For example, while the subject produced many consonant and vowel deletion errors when these were part of consonant and vowel clusters, respectively (e.g., sfondo----sondo; giunta----gunta), deletions were virtually never produced for single consonants flanked by two vowels (e.g., onesto----oesto) or for single vowels flanked by two consonants (e.g., tirare----trare). The demonstration that graphosyllabic factors affect spelling performance disconfirms the hypothesis that graphemic representations consist simply of linearly ordered sets of graphemes. It is concluded that graphemic representations are multidimensional structures: one dimension specifies the grapheme identities that comprise the spelling of a word; a second dimension specifies the consonant/vowel status of the graphemes; a third dimension represents the graphosyllabic structure of the grapheme string; and, a fourth dimension provides information about geminate features.     

The mental representation of lexical form: a phonological approach to the recognition lexicon

We propose a psycholinguistic model of lexical processing which incorporates both process and representation. The view of lexical access and selection that we advocate claims that these processes are conducted with respect to abstract underspecified phonological representations of lexical form. The abstract form of a given item in the recognition lexicon is an integrated segmental-featural representation, where all predictable and non-distinctive information is withheld. This means that listeners do not have available to them, as they process the speech input, a representation of the surface phonetic realisation of a given word-form. What determines performance is the abstract, underspecified representation with respect to which this surface string is being interpreted. These claims were tested by studying the interpretation of the same phonological feature, vowel nasality, in two languages, English and Bengali. The underlying status of this feature differs in the two languages; nasality is distinctive only in consonants in English, while both vowels and consonants contrast in nasality in Bengali. Both languages have an assimilation process which spreads nasality from a nasal consonant to the preceding vowel. A cross-linguistic gating study was conducted to investigate whether listeners would interpret nasal and oral vowels differently in two languages. The results show that surface phonetic nasality in the vowel in VN sequences is used by English listeners to anticipate the upcoming nasal consonant. In Bengali, however, nasality is initially interpreted as an underlying nasal vowel. Bengali listeners respond to CVN stimuli with words containing a nasal vowel, until they get information about the nasal consonant. In contrast, oral vowels in both languages are unspecified for nasality and are interpreted accordingly. Listeners in both languages respond with CVN words (which have phonetic nasality on the surface) as well as with CVC words while hearing an oral vowel. The results of this cross-linguistic study support, in detail, the hypothesis that the listener's interpretation of the speech input is in terms of an abstract underspecified representation of lexical form.     

Does letter position coding depend on consonant/vowel status? Evidence with the masked priming technique

Recently, a number of input coding schemes (e.g., SOLAR model, SERIOL model, open-bigram model, overlap model) have been proposed that capture the transposed-letter priming effect (i.e., faster response times for jugde-JUDGE than for jupte-JUDGE). In their current version, these coding schemes do not assume any processing differences between vowels and consonants. However, in a lexical decision task, Perea and Lupker (2004, JML; Lupker, Perea, & Davis, 2008, L&CP) reported that transposed-letter priming effects occurred for consonant transpositions but not for vowel transpositions. This finding poses a challenge for these recently proposed coding schemes. Here, we report four masked priming experiments that examine whether this consonant/vowel dissociation in transposed-letter priming is task-specific. In Experiment 1, we used a lexical decision task and found a transposed-letter priming effect only for consonant transpositions. In Experiments 2-4, we employed a same-different task - a task which taps early perceptual processes - and found a robust transposed-letter priming effect that did not interact with consonant/vowel status. We examine the implications of these findings for the front-end of the models of visual word recognition.     

Forgetting and redintegration of consonants and vowels in pseudoword lists

Immediate recall of phonemes was studied in a pseudoword span task. Finnish participants recalled lists of increasing length, consisting of C(consonant)V(vowel)CVCV pseudowords. The lists were made up from pools of 12 pseudowords. There were three types of lists. In the non‐redundant lists the items were unpredictable combinations of consonants and vowels. In consonant‐redundant lists, all items had the consonant frame /t/_/s/_/l/. In vowel‐redundant lists, all items had the vowel frame _/u/_/e/_/o/. Unlike redundant last syllables in a previous experiment, neither consonant nor vowel redundancy helped list recall. Instead, a harmful phonological similarity effect was apparent in the vowel‐redundant case but not the consonant‐redundant case. A phoneme‐level analysis of recall showed that consonants were recalled better in consonant‐redundant lists and vowels were recalled better in vowel‐redundant lists compared to non‐redundant lists. Vowels appeared to be more important for discrimination between items, with redundancy resulting in confusions. The consequences of phoneme‐level forgetting and redintegration for item‐ and list‐level recall are discussed.     

Misreadings by beginning readers of Serbo-Croatian

Errors in reading aloud by the beginning reader have been interpreted as reflecting the difficulty and the importance of phonemic segmentation for the acquisition of reading skills. Results from previous studies on English words patterned as consonant-vowel-consonant showed: (1) more errors on vowels than on consonants; (2) more errors on word final consonants than on word initial consonants; and suggested that (3) consonant errors were based on phonetic confusions while vowel errors were not. In contrast to their English counterparts, the beginning readers of Serbo-Croatian tested in the present study committed proportionally fewer errors on their reading of vowels than of consonants but in common with their English counterparts, their reading of final consonants was more vulnerable to error than their reading of initial consonants. This pattern of errors was found for both word and pseudoword consonant-vowel-consonant structures and the pattern of vowel confusion, like the pattern of consonant confusions, was rationalized by speech-related factors. The differences between the patterns of confusions for Serbo-Croatian and for English could be due to the difference between the two orthographies in the precision with which they represent the phonology or to the fact that the vowels of English are qualitatively less distinct phonologically than the vowels of Serbo-Croatian.     

Beyond stop consonants: Consonantal specificity in early lexical acquisition

Previous research has shown that 20-month-old infants can simultaneously learn two words that only differ by one of their consonants, but fail to do so when the words differ only by one of their vowels. This asymmetry was interpreted as developmental evidence for the proposal that consonants play a more important role than vowels in lexical specification. However, the consonant/vowel distinction was confounded with another distinction, that of the continuous status of the phonemes used (discontinuous stop consonants versus continuous vowels). The present study investigated 20-month-olds’ use of phonetic specificity while simultaneously learning two words that differ by a continuous consonant. The results obtained parallel those previously found for stop consonants, confirming the original claim of an asymmetry between the roles of consonants and vowels at the lexical level.     

The processing of consonants and vowels in reading: evidence from the fast priming paradigm

We assessed the early encoding of consonant and vowel information in the reading of English, using the fast priming paradigm. With 30-msec prime durations, gaze durations on target words were shorter when preceded by high-frequency consonant-same primes (which shared consonant information with the target word; e.g., lake-like) than when preceded by vowel-same primes (which shared vowel information with the target word; e.g., line-like), but there were no priming effects for low-frequency primes. With 45-msec prime durations, however, there was no effect of prime frequency and gaze durations on target words were shortened equally when they were preceded by consonant-same primes and vowelsame primes, as compared with control primes (e.g., late-like). The results suggest that the processing of consonants is more rapid than that of vowels, providing further evidence for the distinction between consonant and vowel processing in the reading of English.     

Syllable similarity: The effects of differences in vowels, consonants, and order

In five experiments with synthetic and natural speech syllables, a rating task we used to study the effects of differences in vowels, consonants, and segment order on judged syllable similarity. The results of Experiments I-IV support neither a purely phonemic model of speech representation, in which vowel, consonant, and order are represented independently, nor a purely syllabic model, in which the three factors are integrated. Instead, the data indicate that subjects compare representations in which adjacent vowel and consonant are independent of one another but are not independent of their positions in the syllable. Experiment V provided no support for the hypothesis that this position-sensitive coding is due to acoustic differences in formant transitions.     

The role of distinctive features in the dichotic perception of vowels

A dichotic listening experiment was conducted to determine if vowel perception is based on phonetic feature extraction as is consonant perception. Twenty normal right-handed subjects were given dichotic CV syllables contrasting in final vowels. It was found that, unlike consonants, the perception of dichotic vowels was not significantly lateralized, that the dichotic perception of vowels was not significantly enhanced by the number of phonetic features shared, and that the occurrence of double-blend errors was not greater than chance. However, there was strong evidence for the use of phonetic features at the level of response organization. It is suggested that the differences between vowel and consonant perception reflect the differential availability of the underlying acoustic information from auditory store, rather than differences in processing mechanisms.     

Testing the two-cycle model by different word types

The two-cycle model assumes that consonants in words are processed more quickly than vowels. This study tested the two-cycle model with different word types using a priming task which presented consonants or vowels before the target. Analysis showed presenting consonants before the target was beneficial in processing the target for the words with the letter compositions of CWC and CVCV. In contrast, presenting vowels before the target was beneficial for the words with the letter composition of VCVC. The words with the letter composition of VCCV showed no difference between the consonant prime and the vowel prime. The two-cycle model was not supported across all types of words.     

Developmental differences in short-lived auditory memory for various classes of speech sounds

Spoken serial recall by second-grade children of aurally presented lists of digits, synthetic stop consonants, and synthetic vowels showed a significant suffix effect (selective debilitation of recall at the final position under the stimulus suffix condition) only for the lists of digits and not for either consonants or vowels. Making the synthetic syllables more distinctive by simultaneously covarying the consonant and vowel failed to produce a suffix effect under a strict scoring criterion which required both consonant and vowel to be recalled correctly; however, when subjects were given credit for partially correct answers the suffix effect emerged. Adults given the redundant consonant-vowel syllables showed a significant suffix effect with the strict scoring criterion. However, when consonants and vowels varied orthogonally, the adults' performance showed the suffix effect only under the lenient scoring criterion. An argument is made for equivalence of basic memorial processing between children and adults, the difference being in the number of features needed to disambiguate the target items and in the ability to integrate these features to exploit interstimulus redundancy.     

Mechanisms of aural encoding: VII. Differences in consonant and vowel recall in a Peterson and Peterson short-term memory paradigm

Ss either saw or heard lists of three syllables which differed by their initial consonant phoneme or their final vowel phoneme. After 5 or 15 sec of mental arithmetic, Ss were required to recall the syllables. Following auditory presentation, vowels were recalled more accurately than consonants in all serial positions and at both delays. In addition, spoken consonants and vowels showed primacy and recency effects. Following visual presentation, consonants and vowels were recalled with equal accuracy at both delays, and no recency effects were observed. These data suggest that superior recall of vowels over consonants results from differential decay of these stimuli in an acoustic storage. These data are consistent with previous experiments showing that, during serial recall, the final vowels in a sequence are recalled more accurately than the final consonants.     

Forgetting and redintegration of consonants and vowels in pseudoword lists

Immediate recall of phonemes was studied in a pseudoword span task. Finnish participants recalled lists of increasing length, consisting of C(consonant)V(vowel)CVCV pseudowords. The lists were made up from pools of 12 pseudowords. There were three types of lists. In the non-redundant lists the items were unpredictable combinations of consonants and vowels. In consonant-redundant lists, all items had the consonant frame /t/_/s/_/l/. In vowel-redundant lists, all items had the vowel frame _/u/_/e/_/o/. Unlike redundant last syllables in a previous experiment, neither consonant nor vowel redundancy helped list recall. Instead, a harmful phonological similarity effect was apparent in the vowel-redundant case but not the consonant-redundant case. A phoneme-level analysis of recall showed that consonants were recalled better in consonant-redundant lists and vowels were recalled better in vowel-redundant lists compared to non-redundant lists. Vowels appeared to be more important for discrimination between items, with redundancy resulting in confusions. The consequences of phoneme-level forgetting and redintegration for item- and list-level recall are discussed.     

The relation between discriminability and memory for vowels,consonants, and silent-center vowels

People remember lists of vowel-contrasting syllables better than lists that vary only in stop consonant identity. Most views suggest that this difference is due to the structure of immediate memory and the greater discriminability of vowels compared with consonants. In all of these views, there is a presumed systematic relationship between discriminability and recall so that the more discriminable an item, the better that item should be recalled. The 11 experiments reported here measured the relative discriminability of and compared serial recall for (1) intact syllables that varied only in the medial vowel, (2) intact syllables that varied only in the initial consonant, and (3) syllables with the center vowel replaced by silence (so-called silent-center vowels). When item discriminability, as measured by identification, was equated for consonant-contrasting and silent-center lists, serial recall performance was also equal. However, even when the vowels were less discriminable than the consonants or silentcenter vowels, serial recall performance for the vowels was still better. These results are problematic for theories based on acoustic discriminability but can be explained parsimoniously by Nairne’s (1990) feature model.     

Converging sources of evidence on spoken and perceived rhythms of speech: cyclic production of vowels in monosyllabic stress feet

The article reviews the literature from psychology, phonetics, and phonology bearing on production and perception of syllable timing in speech. A review of the psychological and phonetics literature suggests that production of vowels and consonants are interleaved in syllable sequences in such a way that vowel production is continuous or nearly so. Based on that literature, a hypothesis is developed concerning the perception of syllable timing assuming that vowel production is continuous. The hypothesis is that perceived syllable timing corresponds to the times sequencing of the vowels as produced and not to the timing either of vowel onsets as conventionally measured or of syllable-initial consonants. Three experiments support the hypothesis. One shows that information present during the portion of an acoustic signal in which a syllable-initial consonant predominates is used by listeners to identify the vowel. Compatibly, this information for the vowel contributes to the vowel's perceived duration. Finally, a measure of the perceived timing of a syllable correlates significantly with the time required to identify syllable-medial vowels but not with time to identify the syllable-initial consonants. Further support for the proposed mode of vowel-consonant production and perception is derived from the literature on phonology. Language-specific phonological conventions can be identified that may reflect exaggerations and conventionalizations of the articulatory tendency for vowels to be produced continuously in speech.