Phonological awareness has been shown to be important for early literacy development in unilingual readers of English. We investigated its impact in English and Mandarin for 57 multilingual pupils whose language backgrounds were English, Chinese (Mandarin/dialect), or Bahasa Indonesia, using a homophone decision task, an English lexicality spelling test, and a Hanyu Pinyin (romanised Mandarin) spelling test. All three groups of pupils were studying English and Mandarin in the same school, and so, somewhat unusually, the influence of their language background (especially script exposure) could be seen in the absence of differences in teaching strategies. In English, the results showed that a relationship between script exposure and phonological awareness develops in line with the orthographic depth hypothesis (Frost, Katz, & Bentin, 1987). The Bahasa Indonesia group exhibited the highest levels of alphabetic phonological awareness, followed by the English group and then the Chinese group. In Mandarin, the pupils' performance on the Hanyu Pinyin spelling test suggested that tonal phonological awareness is relatively independent of alphabetic phonological awareness. It seems that language background can influence the nature and development of phonological awareness, and that this in turn may affect children's strategies for the subsequent acquisition of a second (or third) written language.

This study evaluated the effects of practice with a referential communication task on the form and effectiveness of elderspeak, a speech register targeted at older listeners. The task required the listener to reproduce a route drawn on a map following the speaker's instructions. Young adults were given extended practice with this task to determine if they would modify their fluency, prosody, grammatical complexity, semantic content, or discourse style. The effectiveness of the young speakers' instructions was also evaluated in terms of how accurately their older partners could reproduce the routes and in terms of the older adults' evaluations of their own communicative competence. With practice, the young adults' instructions became shorter, simpler, slower, and more repetitious; these selective changes did not affect the older adults' accuracy, but did result in lower self-ratings of communicative competence by the older partners. In a second study, a new group of young adults was given extended practice with young adults as partners. The practice effects were limited to fluency (sentence length and speech rate) and had no effect on the young partners' accuracy or selfratings of communicative competence.

Children using Augmentative and Alternative Communication (AAC) picture boards often produce sequences of symbols that do not reflect the grammatical structure of the language spoken in their environment. These irregularities may reflect an incomplete or incorrect representation of linguistic structure. Alternatively, they may simply be the result of constraints on the communication mode itself. This study examined constituent order and deletion in picture board communication in children who are known to have intact linguistic knowledge (i.e., normally developing native speakers of English). The results suggest that English syntactic knowledge is not automatically applied to picture board communication. Additional skills may be required to transpose linguistic knowledge into the visual-graphic modality used in AAC systems. Moreover, the regularity of non-English responses suggests that there is a bias for specific structures in visual communication systems.

Phonetic skills and vocabulary size were investigated in 37 toddlers (24 to 31 months) identified with specific expressive language impairment and 20 comparison toddlers matched on age, SES, and nonverbal ability. The use of consonants was highly consistent across two contexts: 20 minutes of structured testing and 5 minutes of free play. The comparison children produced almost triple the number of consonant types and five times as many consonant tokens as the late talkers. However, the most frequently present consonants were similar across groups, indicating delayed rather than deviant phonetic development in the late talkers. The late talkers used a much higher proportion of their consonants in initial position than did the normally developing toddlers. Consonant types and tokens were significantly related to reported vocabulary size on Rescorla's (1989) Language Development Survey. No significant predictive relationship was found between either phonetic repertoire or vocabulary size at intake and age 3 language outcome. However, age 3 MLU and IPSyn scores were significantly predicted by age-normed severity of expressive delay at intake (Reynell Expressive z score), with more delayed children showing poorer outcomes.

This report presents findings from the first phase of a longitudinal study examining the development of articulation quality and specific phoneme awareness during the fourth year of life. Articulation and phoneme awareness of /r/ and a control phoneme were assessed for 80 3-year-old children. Children whose articulation of /r/ was accurate for all items were contrasted with those who were accurate on none. The groups differed in their awareness of /r/, after controlling for awareness of the control phoneme on two of the three phoneme awareness tasks and after controlling for relevant cognitive variables. These findings support the hypothesis that accuracy of articulation is related to quality of phoneme awareness in young children.

Three experiments were conducted to compare the development of orthographic representations in children learning to read English, French, or Spanish. Nonsense words that shared both orthography and phonology at the level of the rhyme with real words (cake-dake, comic-bomic), phonology only (cake-daik, comic-bommick), or neither (faish, ricop) were created for each orthography. Experiment I compared English and French children's reading of nonsense words that shared rhyme orthography with real words (dake) with those that did not (daik). Significant facilitation was found for shared rhymes in English, with reduced effects in French. Experiment 2 compared English and French children's reading of nonsense words that shared rhyme phonology with real words (daik) with those that did not (faish). Significant facilitation for shared rhyme phonology was found in both languages. Experiment 3 compared English, French, and Spanish children's reading of nonsense words (dake vs. faish) and found a significant facilitatory effect of orthographic and phonological familiarity for each language. The size of the familiarity effect, however, was much greater in the less transparent orthographies (English and French). These results are interpreted in terms of the level of phonology that is represented in the orthographic recognition units being developed by children who are learning to read more and less transparent orthographies.

This study tested the hypothesis that individual differences in immediate verbal memory span would predict success in second language vocabulary acquisition. The subjects learned 56 English–Finnish translations during two sessions using a method in which they were encouraged to distribute their learning and to use semantic encoding strategies where appropriate. Verbal, but not visuo-spatial, memory span was correlated with the rate of vocabulary learning, a result that could not have occurred because of immediate retrieval from a short-term buffer. When tested one week later, the subjects were less likely to remember those words they had had difficulty learning, even though they had studied these items more often. The theoretical and practical implications of the findings for vocabulary learning are discussed.

Deficits in lexical access in children with specific language impairments (SLI) have been inferred from word naming, recall, and categorization tasks, but no evidence exists concerning these children's ability to identify words in spoken input. We presented successive auditory time gatings of unfamiliar words; familiar, phonologically related words; and familiar, phonologically unrelated words to school-age children with and without SLI. The groups did not differ significantly in the point at which they recognized familiar words, but the subjects with SLI required significantly more of the acoustic-phonetic signal than did their peers to recognize unfamiliar words. For all word types, the subjects with SLI were significantly less likely to respond with correct initial consonants at the earliest gated interval than were their peers. Our results suggest that both representational and perceptual inefficiencies may contribute to slowed lexical access in children with SLI.

This study investigated the development of automaticity in second language word recognition. In an earlier study (Segalowitz & Segalowitz, 1993) we demonstrated that a reduction in the coefficient of variation of lexical decision reaction time (CVRT) - the standard deviation of reaction time divided by mean reaction time (RT) - reflects a restructuring of underlying cognitive processing mechanisms in the direction of increased automaticity and not a simple speed-up of those mechanisms. In the current study, English speakers studying French performed multiple lexical decision tasks. Differences in CVRT were compared cross-sectionally and longitudinally. As in the earlier study, crosssectional analyses showed that CVRT correlated positively with RT for initially fast, but not initially slow, responders. CVRT also correlated positively with RT in longitudinal analyses. These results confirm that, with extended learning experience, the cognitive components underlying word recognition are restructured (automatized) and not simply speeded-up.

This study compared the language skills in a group of very low-income toddlers with those of a middle-income sample matched on age and sex. The assessment instrument was the MacArthur Communicative Development Inventory (CDI) for toddlers, a parent report form. The scores for the low-income group were strikingly lower on the three key indices evaluated: size of expressive vocabulary, age of appearance of word combinations, and complexity of utterances. The entire lowincome distribution was shifted about 30% toward the lower end of the middle-income distribution for both productive vocabulary and grammatical development. The magnitude of these income/ social class effects was larger than reported in most prior reports for children in this age range. This finding underscores the cautionary note issued by the CDI developers, which states that the published CDI norms, based on a middle-class sample, may not be directly applicable to low-income samples.

As children develop language, they also acquire strategies to utilize surface forms to interpret the meaning and functions of sentences. This study investigated how Chinese-speaking children (aged 5, 7, 9, 11 and 13) and adults utilize structural and prosodic cues to interpret the pragmatic function of given and new information intrasententially. The results revealed that: (a) children did not utilize word order cues effectively in NVN sentences with normal stress; (b) children utilized the marked grammatical structure of passives and pseudoclefts more effectively than clefts; (c) when stress converged or competed with word order or syntactic structures, children responded nonuniformly; and (d) animacy was a factor that affected children's interpretation of given versus new. Interaction effects are discussed in light of the Competition Model.

This research investigated communicative intent in 31 toddlers who were slow to talk and 32 normally developing toddlers matched on SES, age, and nonverbal cognitive ability. Communicative intent was studied during free play, both with the mother and with an unfamiliar examiner. Late talkers had lower rates of communication, initiation, and joint attention, but when total communicativeness was controlled for, they were just as likely to initiate, respond, and maintain joint attention as were their peers. As expected, the late talkers relied more on nonword vocalization, gestures, and gesture/oral combinations than their normally developing peers. Children in both groups initiated much more with the examiner (who was instructed to be passive) than with their mothers. Finally, regression analysis suggested that intake expressive language delay severity was the best predictor of age 3 MLU and IPSyn language outcome in the late talkers. However, after expressive delay severity was accounted for, late talkers who were more interested in initiating communication and sustaining joint attention had worse outcomes than late talkers whose communicative drive was weaker, suggesting that they had a more severe underlying language dysfunction.

This study is an investigation of the patterns and strength of the connections between English dative verbs and the double object dative (DOD) argument structure in native speaker production. The subjects completed three written production tasks using dative and other verbs from different semantic classes of verbs. The results show that alternating dative verbs varied in their patterns of connection to argument structures, but were consistent in the strength of their connection to the DOD argument structure across subjects and tasks. There was no support for production differences due to verb class membership, but the results do support a model of lexical representation that represents variable strengths of association between individual verbs and argument structures. With respect to the design of production and processing studies, the results do not support the treatment of alternating dative verbs as a unitary group, nor a priori assumptions about relative argument structure complexity between alternating and nonalternating verbs, but do support the use of production tasks to determine individual verb argument structure preferences.

It is well established that syntactic form and communicative function do not always correspond: for instance, a syntactic question might function as a request for information (“did you see the play?”) or a request for acknowledgment when there is no doubt about polarity of the response (“there's a tower at Blackpool, isn't there”). Using data from a corpus of 18 child–adult conversations, we distinguished adult utterances that solicited information from those soliciting acknowledgment (i.e., where the response was predictable, and the utterance served a predominantly social function). Both types of utterance were usually responded to by children, but the form of response differed according to the communicative function of the utterance. Nonverbal and prosodic responses (e.g., nods or “mmh”) were significantly more likely to occur in response to utterances soliciting acknowledgment than in response to yes/no questions that solicited information. There were consistent form–function relationships for responses as well as for soliciting utterances. Nonverbal nods and headshakes were not functionally equivalent to verbal “yes” and “no.”

Vocabulary scores, word span, nonword repetition, rhyme detection, and articulation rate measures were obtained for a group of 30 preschool children with a mean age of 4;11. After allowing for age and intelligence, it was found that the best predictors of vocabulary were memory span, rhyme detection, and nonword repetition, accounting for 19%. 15%, and 13% of the variance, respectively. Of these children, 28 were tested again on the same measures 13 months later. At the later stage, memory span and rhyme scores again predicted vocabulary to a significant extent, but nonword repetition score did not. Of the measures taken initially, memory span and rhyme detection significantly predicted later vocabulary scores, whereas nonword repetition just failed to reach significance. Cross-lagged correlations showed (hat memory span and rhyme detection on the first test predicted later vocabulary after partialing out initial vocabulary scores. However, initial vocabulary scores did not predict later memory span (or rhyme detection) after partialing out the initial span (or rhyming) scores. Articulation rate was not significantly related to vocabulary at either age. The results are interpreted as indicating that the phonological store of working memory (required by word span, rhyme detection, and nonword repetition, but independent of articulation rates) contributes to vocabulary development.

This study examined the effects of differing degrees of bilingualism on the nonverbal problemsolving abilities of children in grade 3. Three linguistic groups were compared on problem-solving tasks designed to measure control of attention or analysis of knowledge, processes previously shown to develop differently in monolingual and bilingual children solving linguistic problems (Bialystok, 1988). In this study, an English-speaking monolingual group was compared with a French–English bilingual group and a Bengali-English bilingual group. All of the children in the study were similar except for their language background. Tests of language proficiency confirmed that the French-English subjects were balanced bilinguals and that the Bengali-English subjects were partial bilinguals. The balanced French–English bilinguals showed better performance on the non-linguistic tasks requiring control of attention than both the partial bilingual group and the monolingual group. There were no differences found between the groups on the non-linguistic task requiring analysis of representational structures. These results indicate that balanced bilinguals carry over their linguistic advantage in control of attention into the non-linguistic domain.

This study examines the effect of valency (defined as the associative value that represents the number of two-kanji words containing in first or second position the first-positional kanji of the word) on the recognition of two-kanji words in Japanese. Lexical decisions were 24 ms faster for words in the high-valency condition than in the low-valency condition, but were 26 ms slower for nonwords in the high-valency condition than in the low-valency condition. While these results suggested a significant interaction between valency and wordness, a regression analysis indicated that the frequency and valency of the first constituent kanji are significant factors for word recognition, and that the frequency of the first constituent kanji is a significant factor for nonword recognition. The secondary-frequency effect hypothesis is put forth to explain the effects of valency and other variables on lexical decision time.

Variation in mothers' child-directed speech and in their children's rates of language development were examined as a function of child birth order and family socioeconomic status (SES). A total of 63 children between 18 and 29 months were recorded in dyadic interaction with their mothers on two separate occasions, 10 weeks apart. The children included first and later boms who came from high-SES and mid-SES backgrounds. Analyses of the children's speech at the second visit showed that the first-bom children were more advanced in lexical and grammatical development than the later-bom children, and that the later-bom children were more advanced in the development of conversational skill. High-SES children showed more advanced lexical development than mid-SES children. These differences are interpreted as the result of differences in language learning experience associated with birth order and SES, some of which were in evidence in the mothers' speech recorded at the first visit. With respect to theories of language acquisition, these findings suggest that language experience plays a nontrivial role in language development, and that the nature of that role is different for different components of language development. With respect to general developmental consequences of birth order and SES, the findings indicate that differences in early language experience may set the stage for later developmental differences, but that when long-term and pervasive differences are observed, as is the case for SES-related differences in achievement, it is likely that there are pervasive and continuing differences in experience.

Bilingual children face a variety of challenges that their monolingual peers do not. For instance, switching between languages requires the phonological translation of proper names, a skill that requires mapping the phonemic units of one language onto the phonemic units of the other. Proficiency of phonological awareness has been linked to reading success, but little information is available about phonological awareness across multiple phonologies. Furthermore, the relationship between this kind of phonological awareness and reading has never been addressed. The current study investigated phonological translation using a task designed to measure children's ability to map one phonological system onto another. A total of 425 kindergarten and second grade monolingual and bilingual students were evaluated. The results suggest that monolinguals generally performed poorly. Bilinguals translated real names more accurately than fictitious names, in both directions. Correlations between phonological translation and measures of reading ability were moderate, but reliable. Phonological translation is proposed as a tool with which to evaluate phonological awareness through the perspective of children who live with two languages and two attendant phonemic systems.

This article reports the longitudinal follow-up of 41 preschool children as they moved into reading. When the children were 3 years old, they participated in a detailed assessment of their language, print, and metalinguistic skills. At the end of first grade, the children received two tests of phonological awareness and three reading measures: sound–symbol knowledge, word identification, and passage comprehension. Overall language development at age 3 just as strongly, or even more strongly, correlated with reading scores at age 7 as it had with metalinguistic and print awareness scores at age 3. In addition, the overall metalinguistic skills and print awareness of 3 year olds made significant contributions to reading achievement beyond what was provided by tacit language development. Specific metalinguistic domains were also good predictors of reading, with phonological and structural awareness offering more than word awareness.