Since the publication of Nelson's (1973) monograph, a number of studies have documented negative relationships between referential style and some form of maternal directiveness. However, the precise nature of these relationships is still far from clear. The present study represents an attempt to resolve this confusion by investigating the relationship between different measures of maternal directiveness and different measures of referential style in the same group of eight mother–infant dyads. A distinction is made, first, between attentional directiveness and behavioral directiveness, and, second, between referential vocabulary measures based on a fixed number of vocabulary items (i.e., 50 words) and measures taken at particular age points (in this case, 1;4). Correlational analysis of these different measures shows not only that attentional and behavioral directiveness are not significantly related, but also that they are differentially predictive of different measures of referential style. Attentional directiveness is significantly related to referential style at 1;4, but not to referential style at 50 words; behavioral directiveness is significantly related to referential style in the child's subsequent 50 words, but not to referential style at 1;4. These findings suggest that, although the attentional regulation hypothesis may be potentially useful in explaining differences in children's actual rate of vocabulary development, it may be less valuable as a means of explaining stylistic variation in early vocabulary composition. This conclusion underlines the need to distinguish relationships between mothers' interactional behavior and stylistic variation in their children's early language from more general effects of maternal behavior on children's overall rate of language development.

Personal event narratives and fictional stories are narrative genres which emerge early and undergo further development throughout the preschool and early elementary school years. This study compares personal event and fictional narratives across two language-ability groups using episodic analysis. Thirty-six normal children (aged 4 to 8 years) were divided into high and low language-ability groups using Developmental Sentence Scoring (DSS). Three fictional stories and three personal event narratives were gathered from each subject and were scored for length in communication units, total types of structures found within the narrative, and structure of the whole narrative. Narrative genre differences significantly influenced narrative structure for both language-ability groups and narrative length for the high language-ability group. Personal events were told with more reactive sequences and complete episodes than fictional stories, while fictional stories were told with more action sequences and multiple-episode structures. Compared to the episodic story structure of fictional stories, where a prototypical ‘good” story is a multiple-episode structure, a reactive sequence and/or a single complete episode structure may be an alternate, involving mature narrative forms for relating personal events. These findings suggest that narrative structures for personal event narratives and fictional stories may follow different developmental paths. Finally, differences in productive language abilities contributed to the distinctions in narrative structure between fictional stories and personal event narratives. As compared to children in the low group, children in the high group told narratives with greater numbers of complete and multiple episodes, and their fictional stories were longer than their personal event narratives.

Name agreement in Spanish and English in response to 264 pictures was assessed in monolinguals and in bilinguals, who varied in rated skill in the two languages. Most of the pictures were adapted from a standardized set of line drawings of common objects (Snodgrass & Vanderwart, 1980). Name agreement decreased as language skill decreased, and agreement was lower when labels were given in Spanish rather than in English. The relationship between name agreement and word frequency, word length, and (in the case of English) age of acquisition was assessed; both word frequency and word length were found to be related to agreement. Modal responses given by monolingual subjects were nearly identical in the two languages, and the types of non-modal responses were affected by both naming language and language skill.

This research examines the influence of bilingualism on third language learning in a bilingual community, the Basque Country. The English-language achievement of students instructed through the majority (Spanish) and the minority (Basque) languages in the Basque Country was measured. In addition, other cognitive, sociostructural, social psychological, and educational variables were also included in the study. Several sets of regression analyses were carried out to analyze the role of bilingual education. The results indicated that bilingualism and several other variables (intelligence, motivation, age, and exposure) were good predictors of English-language achievement. This research suggests, then, that immersion in the minority language for Spanish-speaking students and school reinforcement of the native language for Basque-speaking students have positive linguistic outcomes.

The relationship of gesticulation to speech has received considerable theoretical and empirical attention from investigators interested in the verbal status of gesture, its use in prelinguistic children, and the role of gestures in social and pragmatic communication. The relationship of gesticulation to sign language, in contrast, has received less attention. Although the gestures of deaf children have been investigated in the contexts of language acquisition and linguistic flexibility, the functions of gestures used by deaf versus hearing individuals have not been examined. One difficulty for such a study stems from the fact that gesture and sign language occur in the same modality. Gesture and sign are considered here with an eye toward determining those aspects of manual communication that are specific to users of signed languages and those in common with users of oral languages. This examination reveals that gestures produced by deaf individuals can be distinguished from the sign language in which they are embedded, both in terms of their privilege of occurrence and their semantic and pragmatic functions.

Studies of sentence comprehension in different languages have enhanced our understanding of how syntactic and semantic information are used in interpreting sentences. However, few studies have attempted to assess how such information affects sentence-processing speed. The present study explored effects of case inflections, gender agreement, noun animacy, and word position on decisions and reaction times for, agent-object assignment in Croatian, a case-marking language with a rich grammatical morphological system. In accordance with Bates and MacWhinney's Competition Model, results provided evidence that the various cues to agent-object relations converge and compete to affect not only agent-object decisions, but also the speed with which those decisions are made. However, several findings from the present study did not support predictions made by the Competition Model. For example, cue strength (i.e., the relative weights that listeners assign to cues) was not a good predictor of reaction time for agent-object assignment in the present study. In addition, reaction times were not always shortened by converging information or inhibited by competing information – both of which were predicted by the model. Rather, our results support the view that, in a given language, there may be constraints on the particular combinations of cues which may converge to facilitate, or compete to inhibit, agent-object assignment. The pattern of cue interactions, which was obtained for the response latencies of normal adults in the present study, is similar to that revealed by off-line choice data from a group of Croatian aphasics (Smith & Bates, 1987), another situation in which ceiling effects of the dominant case cue were removed; thus, these results may tap into a basic aspect of the human sentence processor.