Diminutives facilitate word segmentation in natural speech: Cross-linguistic evidence

Final-syllable invariance is characteristic of diminutives (e.g., doggie), which are a pervasive feature of the child-directed speech registers of many languages. Invariance in word endings has been shown to facilitate word segmentation (Kempe, Brooks, & Gillis, 2005) in an incidental-learning paradigm in which synthesized Dutch pseudonouns were used. To broaden the cross-linguistic evidence for this invariance effect and to increase its ecological validity, adult English speakers (n=276) were exposed to naturally spoken Dutch or Russian pseudonouns presented in sentence contexts. A forced choice test was given to assess target recognition, with foils comprising unfamiliar syllable combinations in Experiments 1 and 2 and syllable combinations straddling word boundaries in Experiment 3. A control group (n=210) received the recognition test with no prior exposure to targets. Recognition performance improved with increasing final-syllable rhyme invariance, with larger increases for the experimental group. This confirms that word ending invariance is a valid segmentation cue in artificial, as well as naturalistic, speech and that diminutives may aid segmentation in a number of languages.     

Infants' sensitivity to allophonic cues for word segmentation

A series of four experiments was conducted to determine whether English-learning infants can use allophonic cues to word boundaries to segment words from fluent speech. Infants were familiarized with a pair of two-syllable items, such as nitrates and night rates and then were tested on their ability to detect these same words in fluent speech passages. The presence of allophonic cues to word boundaries did not help 9-month-olds to distinguish one of the familiarized words from an acoustically similar foil. Infants familiarized with nitrates were just as likely to listen to a passage about night rates as they were to listen to one about nitrates. Nevertheless, when the passages contained distributional cues that favored the extraction of the familiarized targets, 9-month-olds were able to segment these items from fluent speech. By the age of 10.5 months, infants were able to rely solely on allophonic cues to locate the familiarized target words in passages. We consider what implications these findings have for understanding how word segmentation skills develop.     

Cue integration with categories: Weighting acoustic cues in speech using unsupervised learning and distributional statistics

During speech perception, listeners make judgments about the phonological category of sounds by taking advantage of multiple acoustic cues for each phonological contrast. Perceptual experiments have shown that listeners weight these cues differently. How do listeners weight and combine acoustic cues to arrive at an overall estimate of the category for a speech sound? Here, we present several simulations using a mixture of Gaussians models that learn cue weights and combine cues on the basis of their distributional statistics. We show that a cue‐weighting metric in which cues receive weight as a function of their reliability at distinguishing phonological categories provides a good fit to the perceptual data obtained from human listeners, but only when these weights emerge through the dynamics of learning. These results suggest that cue weights can be readily extracted from the speech signal through unsupervised learning processes.     

Integration of multiple speech segmentation cues: a hierarchical framework

A central question in psycholinguistic research is how listeners isolate words from connected speech despite the paucity of clear word-boundary cues in the signal. A large body of empirical evidence indicates that word segmentation is promoted by both lexical (knowledge-derived) and sublexical (signal-derived) cues. However, an account of how these cues operate in combination or in conflict is lacking. The present study fills this gap by assessing speech segmentation when cues are systematically pitted against each other. The results demonstrate that listeners do not assign the same power to all segmentation cues; rather, cues are hierarchically integrated, with descending weights allocated to lexical, segmental, and prosodic cues. Lower level cues drive segmentation when the interpretive conditions are altered by a lack of contextual and lexical information or by white noise. Taken together, the results call for an integrated, hierarchical, and signal-contingent approach to speech segmentation.     

The use of probabilistic lexicality cues for word segmentation in Chinese reading

In an eye-tracking experiment we examined whether Chinese readers were sensitive to information concerning how often a Chinese character appears as a single-character word versus the first character in a two-character word, and whether readers use this information to segment words and adjust the amount of parafoveal processing of subsequent characters during reading. Participants read sentences containing a two-character target word with its first character more or less likely to be a single-character word. The boundary paradigm was used. The boundary appeared between the first character and the second character of the target word, and we manipulated whether readers saw an identity or a pseudocharacter preview of the second character of the target. Linear mixed-effects models revealed reduced preview benefit from the second character when the first character was more likely to be a single-character word. This suggests that Chinese readers use probabilistic combinatorial information about the likelihood of a Chinese character being single-character word or a two-character word online to modulate the extent of parafoveal processing.     

Recurrent speech patterns as cues to the segmentation of multisyllabic sequences.

First and second language acquisition both require that speech be segmented into familiar, multiphonemic units (e.g., words and common phrases). The present research examines one segmentation cue that is of considerable theoretical interest: the repetition of fixed sequences of speech. On each trial, subjects heard repetitions ('pre-exposures') of two artificially-constructed, multisyllabic patterns that shared an embedded segment 1 or 2 syllables long (e.g., 2 shared syllables: [ga-li-SE] and [li-SE-stu]). There were 2 and 6, 4 and 4, or 6 and 2 repetitions of the two patterns, randomly ordered. Subjects were then to indicate the groupings they perceived within a subsequent, longer sequence containing both of the pre-exposed patterns (e.g., [ga-li-SE-stu]). Responses varied systematically with the size of the embedded segment, the repetition frequencies of the two pre-exposed patterns, and the serial position of each pre-exposure. The results illustrate how investigations of the processing of speech patterns may contribute to an understanding of some elementary aspects of language learning.     

Cues to speech segmentation: Evidence from juncture misperceptions and word spotting

The question of whether Dutch listeners rely on the rhythmic characteristics of their native language to segment speech was investigated in three experiments. In Experiment 1, listeners were induced to make missegmentations of continuous speech. The results showed that word boundaries were inserted before strong syllables and deleted before weak syllables. In Experiment 2, listeners were required to spot real CVC or CVCC words (C = consonant, V = vowel) embedded in bisyllabic nonsense strings. For CVCC words, fewer errors were made when the second syllable of the nonsense string was weak rather than strong, whereas for CVC words the effect was reversed. Experiment 3 ruled out an acoustic explanation for this effect. It is argued that these results are in line with an account in which both metrical segmentation and lexical competition play a role.     

Word segmentation with universal prosodic cues

When listening to speech from one’s native language, words seem to be well separated from one another, like beads on a string. When listening to a foreign language, in contrast, words seem almost impossible to extract, as if there was only one bead on the same string. This contrast reveals that there are language-specific cues to segmentation. The puzzle, however, is that infants must be endowed with a language-independent mechanism for segmentation, as they ultimately solve the segmentation problem for any native language. Here, we approach the acquisition problem by asking whether there are language-independent cues to segmentation that might be available to even adult learners who have already acquired a native language. We show that adult learners recognize words in connected speech when only prosodic cues to word-boundaries are given from languages unfamiliar to the participants. In both artificial and natural speech, adult English speakers, with no prior exposure to the test languages, readily recognized words in natural languages with critically different prosodic patterns, including French, Turkish and Hungarian. We suggest that, even though languages differ in their sound structures, they carry universal prosodic characteristics. Further, these language-invariant prosodic cues provide a universally accessible mechanism for finding words in connected speech. These cues may enable infants to start acquiring words in any language even before they are fine-tuned to the sound structure of their native language.     

Lexically restricted utterances in Russian, german, and english child-directed speech

This study investigates the child-directed speech (CDS) of four Russian-, six German, and six English-speaking mothers to their 2-year-old children. Typologically Russian has considerably less restricted word order than either German or English, with German showing more word-order variants than English. This could lead to the prediction that the lexical restrictiveness previously found in the initial strings of English CDS by Cameron-Faulkner, Lieven, and Tomasello (2003 ) would not be found in Russian or German CDS. However, despite differences between the three corpora that clearly derive from typological differences between the languages, the most significant finding of this study is a high degree of lexical restrictiveness at the beginnings of CDS utterances in all three languages.     

Category induction from distributional cues in an artificial language

The ability to identify the grammatical category of a word (e.g., noun, verb, adjective) is a fundamental aspect of competence in a natural language. Children show evidence of categorization by as early as 18 months, and in some cases younger. However, the mechanisms that underlie this ability are not well understood. The lexical co-occurrence patterns of words in sentences could provide information about word categories--for example, words that follow the in English often belong to the same category. As a step in understanding the role distributional mechanisms might play in language learning, the present study investigated the ability of adults to categorize words on the basis of distributional information. Forty participants listened for approximately 6 min to sentences in an artificial language and were told that they would later be tested on their memory for what they had heard. Participants were next tested on an additional set of sentences and asked to report which sentences they recognized from the first 6 min. The results suggested that learners performed a distributional analysis on the initial set of sentences and recognized sentences on the basis of their memory of sequences of categories of words. Thus, mechanisms that would be useful in natural language learning were shown to be active in adults in an artificial language learning task.     

Burst cues, transition cues, and hemispheric specialization with real speech sounds

Subjects listening to dichotically presented real speech stop and fricative consonants, with and without transitions, showed larger laterality effects in the transition-less condition. In a second study, laterality effects for burst cues and transition cues were compared; using the stop consonants /b/ and /d/. Again, burst cues produced a larger laterality effect. These results are not compatible with a lateralized speech “decoder”, and are interpreted as favoring a Semmes (1968) model of hemispheric differences, differential processing.     

A computational model of word segmentation from continuous speech using transitional probabilities of atomic acoustic events

Word segmentation from continuous speech is a difficult task that is faced by human infants when they start to learn their native language. Several studies indicate that infants might use several different cues to solve this problem, including intonation, linguistic stress, and transitional probabilities between subsequent speech sounds. In this work, a computational model for word segmentation and learning of primitive lexical items from continuous speech is presented. The model does not utilize any a priori linguistic or phonemic knowledge such as phones, phonemes or articulatory gestures, but computes transitional probabilities between atomic acoustic events in order to detect recurring patterns in speech. Experiments with the model show that word segmentation is possible without any knowledge of linguistically relevant structures, and that the learned ungrounded word models show a relatively high selectivity towards specific words or frequently co-occurring combinations of short words.     

The differential role of phonological and distributional cues in grammatical categorisation

Recognising the grammatical categories of words is a necessary skill for the acquisition of syntax and for on-line sentence processing. The syntactic and semantic context of the word contribute as cues for grammatical category assignment, but phonological cues, too, have been implicated as important sources of information. The value of phonological and distributional cues has not, with very few exceptions, been empirically assessed. This paper presents a series of analyses of phonological cues and distributional cues and their potential for distinguishing grammatical categories of words in corpus analyses. The corpus analyses indicated that phonological cues were more reliable for less frequent words, whereas distributional information was most valuable for high frequency words. We tested this prediction in an artificial language learning experiment, where the distributional and phonological cues of categories of nonsense words were varied. The results corroborated the corpus analyses. For high-frequency nonwords, distributional information was more useful, whereas for low-frequency words there was more reliance on phonological cues. The results indicate that phonological and distributional cues contribute differentially towards grammatical categorisation.     

Modeling human performance in statistical word segmentation

The ability to discover groupings in continuous stimuli on the basis of distributional information is present across species and across perceptual modalities. We investigate the nature of the computations underlying this ability using statistical word segmentation experiments in which we vary the length of sentences, the amount of exposure, and the number of words in the languages being learned. Although the results are intuitive from the perspective of a language learner (longer sentences, less training, and a larger language all make learning more difficult), standard computational proposals fail to capture several of these results. We describe how probabilistic models of segmentation can be modified to take into account some notion of memory or resource limitations in order to provide a closer match to human performance.     

The impact of attention load on the use of statistical information and coarticulation as speech segmentation cues

In two artificial language learning experiments, we investigated the impact of attention load on segmenting speech through two sublexical cues: transitional probabilities (TPs) and coarticulation. In Experiment 1, we observed that coarticulation processing was resilient to high attention load, whereas TP computation was penalized in a graded manner. In Experiment 2, we showed that encouraging participants to actively search for “word” candidates enhanced overall performance but was not sufficient to preclude the impairment of statistically driven segmentation by attention load. As long as attentional resources were depleted, independently of their intention to find these “words,” participants segmented only TP words with the highest TPs, not TP words with lower TPs. Attention load thus has a graded and differential impact on the relative weighting of the cues in speech segmentation, even when only sublexical cues are available in the signal.     

A variable that may affect individual differences in the child-directed speech of Japanese women

Sixty-one single Japanese-speaking women between the ages of 18 and 26 years were recorded as they read aloud picture books to a young child and as they conversed with another Japanese-speaking woman. When their utterances were acoustically compared between the two settings with regard to prosodic features, both the average pitch and pitch excursions exhibited a significant increase when interacting with the child in 17 of the 61 women. In 36 of the remaining 44 subjects, neither of these parameters showed such changes. This individual variability was not related to the subjects' liking for picture books, previous experience with reading or being read them, or with baby-sitting. The only variable that could explain the results was whether the subjects had grown up with one or more siblings or as only children. If they were only children, the prosodic modification was significantly less likely to occur     

Comparing word fragment completion and cued recall with letter cues

Cued recall with word stems as cues and fragment completion rely on different types of letter cues and also differ in the explicit-implicit nature of the retrieval orientation. Despite these differences, variables effective in one task may be effective in the other because both rely on letter cues. Two variables known to affect cued recall were manipulated: Lexical set size (number of words that fit the letter cue) and meaning set size (number of associates generated to the studied words). Across four experiments, subjects in each task were less likely to recover targets from larger lexical sets. However, meaning set size affected cued recall but not fragment completion. These results indicate that fragment completion and letter-cued recall are based on lexical search but that cued recall also involves a semantic search component. Furthermore, type of retrieval cue had a greater effect than type of retrieval orientation.     

Relative salience of intonation fall and pause as cues to the perceptual segmentation of speech in an unfamiliar language

The modification of subjects' attention by two prosodic features was investigated using their reaction times to a nonspeech stimulus which coincided with these features. By using a foreign language unfamiliar to the subjects (Czechoslovakian), the influence of semantic and syntactic knowledge was controlled. The results indicate that for native speakers of English intonation fall is a relatively more important cue to the perceptual segmentation of speech than is pause.     

"Frequent frames" in German child-directed speech: a limited cue to grammatical categories

Mintz (2003) found that in English child-directed speech, frequently occurring frames formed by linking the preceding (A) and succeeding (B) word (A_x_B) could accurately predict the syntactic category of the intervening word (x). This has been successfully extended to French (Chemla, Mintz, Bernal, & Christophe, 2009). In this paper, we show that, as for Dutch (Erkelens, 2009), frequent frames in German do not enable such accurate lexical categorization. This can be explained by the characteristics of German including a less restricted word order compared to English or French and the frequent use of some forms as both determiner and pronoun in colloquial German. Finally, we explore the relationship between the accuracy of frames and their potential utility and find that even some of those frames showing high token-based accuracy are of limited value because they are in fact set phrases with little or no variability in the slot position.     

New evidence for chunk-based models in word segmentation

There is large evidence that infants are able to exploit statistical cues to discover the words of their language. However, how they proceed to do so is the object of enduring debates. The prevalent position is that words are extracted from the prior computation of statistics, in particular the transitional probabilities between syllables. As an alternative, chunk-based models posit that the sensitivity to statistics results from other processes, whereby many potential chunks are considered as candidate words, then selected as a function of their relevance. These two classes of models have proven to be difficult to dissociate. We propose here a procedure, which leads to contrasted predictions regarding the influence of a first language, L1, on the segmentation of a second language, L2. Simulations run with PARSER (Perruchet & Vinter, 1998), a chunk-based model, predict that when the words of L1 become word-external transitions of L2, learning of L2 should be depleted until reaching below chance level, at least before extensive exposure to L2 reverses the effect. In the same condition, a transitional-probability based model predicts above-chance performance whatever the duration of exposure to L2. PARSER's predictions were confirmed by experimental data: Performance on a two-alternative forced choice test between words and part-words from L2 was significantly below chance even though part-words were less cohesive in terms of transitional probabilities than words.