Producing and recognizing words with two pronunciation variants: Evidence from novel schwa words

This study examined the lexical representations and psycholinguistic mechanisms underlying the production and recognition of novel words with two pronunciation variants in French. Participants first learned novel schwa words (e.g., /??nyk/), which varied in their alternating status (i.e., whether these words were learned with one or two variants) and, for alternating words, in the frequency of their variants. They were then tested in picture-naming (free or induced) and recognition memory tasks (i.e., deciding whether spoken items were learned during the experiment or not). Results for free naming show an influence of variant frequency on responses, more frequent variants being produced more often. Moreover, our data show an effect of the alternating status of the novel words on naming latencies, with longer latencies for alternating than for nonalternating novel words. These induced naming results suggest that both variants are stored as lexical entries and compete during the lexeme selection process. Results for recognition show an effect of variant frequency on reaction times and no effect of variant type (i.e., schwa versus reduced variant). Taken together, our findings suggest that participants both comprehend and produce novel French schwa words using two lexical representations, one for each variant.     

Producing and recognizing words with two pronunciation variants: evidence from novel schwa words

This study examined the lexical representations and psycholinguistic mechanisms underlying the production and recognition of novel words with two pronunciation variants in French. Participants first learned novel schwa words (e.g., /??nyk/), which varied in their alternating status (i.e., whether these words were learned with one or two variants) and, for alternating words, in the frequency of their variants. They were then tested in picture-naming (free or induced) and recognition memory tasks (i.e., deciding whether spoken items were learned during the experiment or not). Results for free naming show an influence of variant frequency on responses, more frequent variants being produced more often. Moreover, our data show an effect of the alternating status of the novel words on naming latencies, with longer latencies for alternating than for nonalternating novel words. These induced naming results suggest that both variants are stored as lexical entries and compete during the lexeme selection process. Results for recognition show an effect of variant frequency on reaction times and no effect of variant type (i.e., schwa versus reduced variant). Taken together, our findings suggest that participants both comprehend and produce novel French schwa words using two lexical representations, one for each variant.     

Sex differences in phonological processes: Speeded matching and word reading

Sex differences in phonological processing were investigated in four experiments. Two experiments required college students to decide whether two five-letter strings matched. Same-case (AA) pairs of letter strings could be matched using physical features, whereas mixed-case (Aa) pairs of letter strings required the mediation of a speech-based code (letter name) for the comparison. Women were significantly faster than men when the comparisons required the speech-based codes. In another experiment, college students read lists of words and lists of pseudohomophones to determine whether there was a sex difference in the computation of phonology for unfamiliar words (assembled phonology). In a final experiment, students read lists of words with phonologically inconsistent spelling patterns to determine whether there was a sex difference in accessing pronunciations of familiar words (addressed phonology). Women were more proficient than men under both of these conditions. Results were interpreted in terms of a female advantage in both prelexical and lexical processing, an advantage that may stem from a sex difference in the quality of the phonological representations.     

The pseudohomophone effect: The role of visual similarity in non-word decisions

The pseudohomophone effect, that is, the finding that non-words that are pronounced like words (e.g. MEEN) take longer to reject in a lexical decision task than other pronounceable non-words (e.g. NEEN), has been used to support the hypothesis of phonological receding in lexical access. The lexical decision experiments reported here matched pseudohomophones to control non-words that were as visually similar to words as were the pseudohomophones. For both normal and aphasic subjects, reaction times to reject the pseudohomophones were no longer than those for the visual controls. However, the pseudohomophones did take longer to reject than other pronounceable non-words which were not as visually similar to words. The results suggest that most previous findings of the pseudohomophone effect result from the greater visual rather than phonological similarity of the pseudohomophones to words. The absence of a phonological effect on the non-words in the present study implies that phonological coding is an optional rather than a slow, but obligatory process.     

Exposure to print and word recognition processes

The effect of exposure to print on the efficiency of phonological and orthographic word recognition processes was examined by comparing two groups of university students having similar reading comprehension scores but different levels of exposure to print. Participants with a high level of exposure to print were faster and more accurate in naming pseudowords, in choosing the correct member of a homophone pair, and in making lexical decisions when nonwords were pseudohomophones. In the lexical decision task, low-print-exposure participants were more sensitive to the frequency of the orthographic patterns in the stimuli. The results of a form priming task demonstrated that high-print-exposure participants more quickly and strongly activated the orthographic representations of common words and subsequently more strongly activated the corresponding phonological representations. Even among successful students, differences in exposure to print produce large differences in the efficiency of both orthographic and phonological word recognition processes.     

Base word Frequency and Pseudohomophone Naming

The advantage of naming pseudohomophones over non-pseudohomophones has been interpreted as reflecting the contribution of whole-word lexical representations in phonological coding. A whole-word interpretation was further supported by Taft and Russell (1992), who reported a pseudohomophone frequency effect such that pseudohomophones were named faster if they corresponded to high- than to low-frequency base-words (e.g. poast vs. hoast ). Experiment 1 replicated this pseudohomophone frequency effect using the Taft and Russell items. Further analyses showed, however, that the pseudohomophones in Taft and Russell's high-frequency group were more orthographically similar to words than the pseudohomophones in the low-frequency group. These differences in orthography may have been the cause of the 'frequency' effects. In Experiment 2, a new set of high- and low-frequency pseudohomophones was constructed that were matched on orthographic factors (i.e. SPBF and N). With these items, a standard pseudohomophone advantage was found such that pseudohomophones were named faster and more accurately than non-pseudohomophones. However, in contrast to Taft and Russell's results, pseudohomophone naming was not related to base-word frequency. We conclude that the pseudohomophone advantage occurs at a postlexical stage in non-word naming.     

Processing variant forms in spoken word recognition: the role of variant frequency

Recognition of a spoken word phonological variant--schwa vowel deletion (e.g., corporate --> corp'rate)--was investigated in vowel detection (absent/present) and syllable number judgment (two or three syllables) tasks. Variant frequency corpus analyses (Patterson, LoCasto, & Connine, 2003) were used to select words with either high or low schwa vowel deletion rates. Speech continua were created for each word in which schwa vowel length was manipulated (unambiguous schwa-present and schwa-absent endpoints, along with intermediate ambiguous tokens). Matched control nonwords were created with identical schwa vowel continua and surrounding segments. The low-deletion-rate words showed more three-syllable judgments than did the high-deletion-rate words. Matched control nonwords did not differ as a function of deletion rate. Experiments 2 and 3 showed a lexical decision reaction time advantage for more frequent surface forms, as compared with infrequent ones, for schwa-deleted (Experiment 2) and schwa-present (Experiment 3) stimuli. The results are discussed in terms of representations of variant forms of words based on variant frequency.     

Phonological assembly in reading: Lexical contribution leads to violation of graphophonological rules

According to current models of reading, the phonological representation-of a-letter string could be generated by means of two different procedures. The first consists in the looking up or addressing of the phonological representation of the appropriate word in the mental lexicon. The second consists in the assignment of a phonological code to the various orthographic units that occur in the letter string and the assembly of them into a unitary phonological representation. However, the various models of phonological assembly differ in the nature of the knowledge that such assembly requires. On the one hand, dual-route theories assert that the assembly process operates by means of extralexical graphophonological rules and, therefore, without reference to lexical knowledge. On the other hand, in analogy theories and interactive models of phonological assembly, the lexical phonological representations of words orthographically close to the letter string (the orthographic neighborhood) are supposed to contribute to the-translation process. The aim of the experiments reported here was to empirically distinguish between these two categories of models. In Experiment 1, subjects were asked to pronounce pseudowords containing the letterg. Results indicate that the phonological representation assigned to the letterg depends on the pronunciation favored by the lexical neighbors. Experiment 2 shows that lexical contribution to phonological assembly also takes place in lexical decision.     

Rule-governed missing information in spoken word recognition: schwa vowel deletion

Vowel deletion is a phonological process in which an unstressed /inverted e/ (schwa) vowel is deleted during pronunciation. In Experiment 1, vowel-deleted and vowel-reduced versions of two- and three-syllable words rated for pronunciation acceptability showed reduced acceptability for deleted vowel versions with a greater decrement for two-syllable words. Experiments 2 and 3 used vowel-intact and vowel-deleted productions preceded by themselves (repetition prime), their alternative production (variant prime), or a control (unrelated) prime. Lexical decisions to three-syllable vowel-intact and vowel-deleted targets, as well as to two-syllable vowel-intact targets, showed greater priming in the repetition conditions than in the variant conditions. Two-syllable vowel-deleted targets, however, showed comparable repetition and variant priming. The results are discussed in terms of lexical activation and representation of phonological variants. A model is offered in which activation based on similarity triggers utilization of phonological inferences only when required for successful recognition.     

Visual word recognition is accompanied by covert articulation: evidence for a speech-like phonological representation

Two lexical decision task (LDT) experiments examined whether visual word recognition involves the use of a speech-like phonological code that may be generated via covert articulation. In Experiment 1, each visual item was presented with an irrelevant spoken word (ISW) that was either phonologically identical, similar, or dissimilar to it. An ISW delayed classification of a visual word when the two were phonologically similar, and it delayed the classification of a pseudoword when it was identical to the base word from which the pseudoword was derived. In Experiment 2, a LDT was performed with and without articulatory suppression, and pseudowords consisted of regular pseudowords and pseudohomophones. Articulatory suppression decreased sound-specific ISW effects for words and regular pseudowords but not for pseudohomophones. These findings indicate that the processing of an orthographically legal letter sequence generally involves the specification of more than one sound code, one of which involves covert articulation.     

Looks aren’t Everything: Pseudohomophones Prime Words but Nonwords do not

After examining literature that deals with phonological and orthographic effects associated with pseudohomophones, the current effort deviates from the norm by using fewer pseudohomophones (20) and extending the lags between primes and targets (M=8). Word and pseudohomophone primes were found to facilitate lexical decision response latencies to word targets. Response latencies to word targets were not influenced by nonword primes, however. The presence of pseudohomophone effects was demonstrated by longer response latencies and higher error rates for pseudohomophones (e.g., DREEM) that were equated in orthography to nonword controls (e.g., DROAM). Despite the frequency effect observed for base words, the pseudohomophones did not exhibit an effect of base word frequency. The results suggest that phonological codes exert an influence on lexical representation but are not frequency sensitive.     

Automatic phonological coding of unattended printed words

A lexical decision response to an attended printed word can be slowed when the word is accompanied by an unattended word that is semantically related. Does this hold for an unattended word that is not related to the target word but sounds as if it is? The homophone WASTE can be shown to affect the lexical decision response to RUBBISH, but how does the incongruently related homophone WAIST affect RUBBISH? If incongruent homophones of words related to the attended word can influence processing when they are not being attended to, then it must be through automatic processing into a phonological code, either before or after lexical access. Experiment 1 reports such an effect, and it is concluded that a phonological representation is generated preattentively and influences the semantic processing of the attended word after generation of that representation. Experiment 2 confirms the effect and investigates the possibility that the effect exists because of the strategic use of phonological encoding of attended words. To discourage the use of a phonological strategy for lexical access, all nonwords were pseudohomophones. The influence persisted, however, with attended words still being affected by the presence of incongruent homophones of related words.     

Large-Scale Modeling of Wordform Learning and Representation

The forms of words as they appear in text and speech are central to theories and models of lexical processing. Nonetheless, current methods for simulating their learning and representation fail to approach the scale and heterogeneity of real wordform lexicons. A connectionist architecture termed the sequence encoder is used to learn nearly 75,000 wordform representations through exposure to strings of stress-marked phonemes or letters. First, the mechanisms and efficacy of the sequence encoder are demonstrated and shown to overcome problems with traditional slot-based codes. Then, two large-scale simulations are reported that learned to represent lexicons of either phonological or orthographic wordforms. In doing so, the models learned the statistics of their lexicons as shown by better processing of well-formed pseudowords as opposed to ill-formed (scrambled) pseudowords, and by accounting for variance in well-formedness ratings. It is discussed how the sequence encoder may be integrated into broader models of lexical processing.     

Reading aloud pseudohomophones in Italian: Always an advantage

In the present article, the lexical contribution to nonword reading was evaluated using Italian pseudohomophones that contained atypical letters or letter sequences. Pseudohomophones were read faster than orthographically matched nonwords in both mixed (Experiment 1) and pure (Experiment 2) lists; in addition, a base-word frequency effect was obtained in both conditions. The same pseudohomophone advantage was observed when nonwords without atypical letter sequences were mixed in the experimental list (Experiment 3 ), and it disappeared only in lexical decision, in which pseudohomophones were rejected as quickly as control nonwords. The pattern of results was explained by assuming that, due to their orthographic properties, the Italian pseudohomophones did not benefit from an orthographic lexical contribution and were mainly processed through the interaction system between the sublexical mechanisms and the phonological output lexicon.     

Repetition priming of words and pseudowords in divided attention and in amnesia

Repetition priming of novel stimuli (pseudowords) and stimuli with preexisting representations (words) was compared in two experiments. In one, 19 normal male subjects performed a lexical decision task with either focused or divided attention. In another, lexical decision performance was compared between 8 male Korsakoff patients and 8 alcoholic control subjects. In control conditions, repetition speeded responses to both stimulus types. Experimental conditions that minimized the contribution of episodic memory to task performance eliminated reaction time priming for pseudowords but not for words. However, in these same conditions, repetition increased the likelihood that pseudowords would be incorrectly classified. These results indicate that preserved repetition priming effects in amnesia do not solely reflect activation of representations in semantic memory.     

The role of phonological recoding in orthographic learning

The objective of this research was to explore whether orthographic learning occurs as a result of phonological recoding, as expected from the self-teaching hypothesis. The participants were 32 fourth- and fifth-graders (mean age = 10 years 0 months, SD = 7 months) who performed lexical decisions for monosyllabic real words and pseudowords under two matched experimental conditions: a read aloud condition, wherein items were named prior to lexical decision to promote phonological recoding, and a concurrent articulation condition, presumed to attenuate phonological recoding. Later, orthographic learning of the pseudowords was evaluated using orthographic choice, spelling, and naming tasks. Consistent with the self-teaching hypothesis, targets learned with phonological recoding in the read aloud condition yielded greater orthographic learning than those learned with concurrent articulation. The research confirms the critical nature of phonological recoding in the development of visual word recognition skills and an orthographic lexicon.     

Inflected words in production: Evidence for a morphologically rich lexicon

Current evidence suggests that there is a difference between the representations of multimorphemic words in production and perception. In perception, it is widely believed that both whole-word and root representations exist, while in production there is little evidence for whole-word representations. The present investigation demonstrates that whole-word and root frequency independently predict the duration of words suffixed with –ing, –ed, and –s, which reveals that both root and word representations play a role in the production of inflected English words. In a second line of analysis, we find that the number of inflected phonological neighbours independently predicts the duration of monomorphemic words, which extends these results and suggests that whole-word representations exist at the lexical level. Together these results suggest that both root and word representations of inflected words are stored in the lexicon and are relevant for the production of both monomorphemic and multimorphemic words.     

Phonological encoding in the lexical decision task

In lexical decision experiments, subjects have difficulty in responding NO to non-words which are pronounced exactly like English words (e.g. BRANE). This does not necessarily imply that access to a lexical entry ever occurs via a phonological recoding of a visually-presented word. The phonological recoding procedure might be so slow that when the letter string presented is a word, access to its lexical entry via a visual representation is always achieved before phonological recoding is completed. If prelexical phonological recodings are produced by using grapheme-phoneme correspondence rules, such recodings can only occur for words which conform to these rules (regular words), since applications of the rules to words which do not conform to the rules (exception words) produce incorrect phonological representations. In two experiments, it was found that time to achieve lexical access (as measured by YES latency in a lexical decision task) was equivalent for regular words and exception words. It was concluded that access to lexical entries in lexical decision experiments of this sort does not proceed by sometimes or always phonologically recoding visually-presented words.     

The activation of phonological representations by bilinguals while reading silently: evidence from interlingual homophones

These experiments investigated whether bilinguals activate phonological representations from both of their languages when reading silently in one. The critical stimuli were interlingual homophones (e.g., sank in English and cinq in French). French-English and English-French bilinguals completed an English lexical decision task. Decisions made by French-English bilinguals were significantly faster and more accurate for interlingual homophones than for matched English control words. In subsequent experiments, the homophone facilitation effect in the latency data disappeared when distractors were changed to pseudohomophones, when cognates and interlingual homographs were added to the experiment, and when the proportion of critical stimuli was decreased. However, the homophone effect in the error data remained. In contrast, English-French bilinguals revealed little evidence of an interlingual homophone effect. Several attempts were made to increase the saliency of the nontarget language, however these manipulations produced only a small effect in the error data. These results indicate that the activation of phonological representations can appear to be both language-specific and nonspecific depending on the proficiency of the bilinguals and whether they are reading in their weaker or stronger language.     

Interaction and representational integration: evidence from speech errors

We examine the mechanisms that support interaction between lexical, phonological and phonetic processes during language production. Studies of the phonetics of speech errors have provided evidence that partially activated lexical and phonological representations influence phonetic processing. We examine how these interactive effects are modulated by lexical frequency. Previous research has demonstrated that during lexical access, the processing of high frequency words is facilitated; in contrast, during phonetic encoding, the properties of low frequency words are enhanced. These contrasting effects provide the opportunity to distinguish two theoretical perspectives on how interaction between processing levels can be increased. A theory in which cascading activation is used to increase interaction predicts that the facilitation of high frequency words will enhance their influence on the phonetic properties of speech errors. Alternatively, if interaction is increased by integrating levels of representation, the phonetics of speech errors will reflect the retrieval of enhanced phonetic properties for low frequency words. Utilizing a novel statistical analysis method, we show that in experimentally induced speech errors low lexical frequency targets and outcomes exhibit enhanced phonetic processing. We sketch an interactive model of lexical, phonological and phonetic processing that accounts for the conflicting effects of lexical frequency on lexical access and phonetic processing.