Catharine H. Echols - US grants

The proposed program of research investigates the process by which children identify words in speech and associate words to referents. More specifically, it concerns the extent to which very young children may be sensitive to cues in the speech signal which (a) indicate words and other linguistic units and (b) provide information concerning associations between words and referents. This research will contribute significantly to an understanding of how children succeed at the fundamental tasks of identifying linguistically-relevant units in the stream of speech and mapping those units to appropriate real-world referents. The potential role of cues to linguistic units and their referents as initial wedges into language learning will be investigated in four interrelated series of experiments: 1. Infants' sensitivity to cues to word-level units in speech. Tendencies of young infants to perceive certain syllables as especially salient and to treat as a single sequences indicated by particular acoustically based cues will be evaluated using a modified head turn preference procedure, in which preferences for more salient versus less salient changes can be assessed by recording infants' orientation to speakers from which each type of stimulus is presented. 2. Infants' sensitivity to cues to metrical structure. Metrical (rhythmic) structure may' provide cues not only to words, but also to larger linguistic units in speech. The possibility that infants may become tuned to rhythmic properties of their native language will be assessed in experiments using the headturn preference procedure to identify infants' responses to native and nonnative rhythmic patterns across the first year of life. Preferences of Spanish-hearing and Chinese-hearing infants will also be assessed. 3. Interactions between perceptual, representational and production factors. imitation and word-learning studies will be used with two- and three-year olds to disentangle the effects of representational and production factors in contributing to the form of early words. 4. Infants' sensitivity to cues that assist with the mapping of words to referents. Infants' sensitivity to information in the linguistic input concerning the semantic or grammatical categories of novel words will be investigated using a familiarization procedure. Infants will see events involving moving objects. The events will be paired with labeling sequences. If linguistics cues direct 9-16 months-old infants attention toward certain potential referents, then infants should be more likely to notice changes in those particular elements than in other elements of the event. The proposed research may provide insights into the process by which young children identify structure in their native language. By enhancing our understanding of normal language development, valuable information may be provided for early identification and remediation in cases of delayed language acquisition.

Learning the meanings of words might appear to be a simple task: An adult names an object; the child, hearing the name in association with the object, determines that the name refers to the object. But consider, the following situation: An individual says, "Where is the mandrel?" She picks up one object, examines it with a frown, then puts it down. She picks up a second object, her face brightens, and she says "Here it is!" If word learning requires nothing more than associations between words and objects, people should interpret "mandrel" as a label for the first object. However, neither adults nor young children make this error. Instead, they make inferences about the speaker's referential intent, using cues such as facial expression and eye gaze, to identify which entity the speaker wishes to label. With funding from the Linguistics Program, the Developmental and Learning Sciences Program and the Office of International Science and Engineering, Dr. Catharine Echols will examine the role of referential intent in word learning. Studies will assess variability in the availability and form of referential intent cues, children's use of the cues, and the possible consequences of such variability for language learning. Naturalistic observations of conversations between mothers and children and experimental procedures will be used. The studies will be carried out across two languages and cultures, English in the U.S. and Brazilian Portuguese in Brazil, and will examine the effects of socio-economic status within each country. Additional studies will test similar questions with deaf and hearing children whose parents are either deaf or hearing. Dr. Richard P. Meier, an expert on the acquisition of American Sign Language, is a collaborator on the project. The project will provide important new knowledge about the language learning process, and it has valuable applications. Children from economically deprived backgrounds are at risk for delays in language learning, and delays also are evident in the language development of deaf children of hearing parents. A reduced availability of cues to referential intent-or a reduced capacity to use any available cues-could contribute to these delays. Consequently, the findings of this project could be used to improve early intervention programs.

Research has shown that accent ? defined as the manner of pronunciation that is particular to an individual or group of individuals ? can trigger a variety of negative impressions and judgments about a speaker. Like adults, young children treat foreign-accented individuals less positively than those with native accents. These negative reactions to foreign-accented individuals have often been interpreted as evidence of social biases: Because a foreign- accented speaker is viewed as a member of an ?out group,? negative perceptions and attitudes are elicited. An additional possibility, however, is that the impressions are linked to the fluency with which foreign-accented speech is processed: Difficulties in processing lead to negative affect that is misattributed as an unfavorable impression of the speaker. The proposed research will evaluate the role of processing fluency in children?s accent-related biases, and examine implications for modifying those biases The proposed research assesses 4- to 7-year-old children?s responses to foreign-accented speech, and factors influencing those responses, by asking children to judge the credibility of statements: If a foreign accent results in a negative bias, listeners should be more likely to judge a statement produced by speaker with a foreign accent as false than one produced by a native speaker. These ages are of interest because changes in language representation and processing skills are occurring that could affect fluency of processing. A first study tests the basic question of whether a foreign accent elicits lowered credibility judgments from children. A second experiment evaluates the processing fluency account by independently varying accentedness and comprehensibility in the speech presented to the children. A third study examines the degree to which specific factors?including age, vocabulary size, processing skills, and exposure to accented speech?predict children?s credibility judgements. A fourth study determines whether any initial biases favoring native-accented speech will be overridden if children receive evidence that the speaker is credible. A fifth study evaluates whether observed biases are malleable, by providing children with experience with the specific accents. Finally, a sixth study tests whether general exposure to foreign-accented speech can lead to reductions in accent-related biases. Non-native speakers of English now account for about 12.5% of the United States? population; many of these individuals have foreign accents. An understanding of the degree to which children exhibit biases against non-native speakers, and the bases for those biases, can lead to strategies for promoting positive interactions between native-accented children and their non-native accented teachers, physicians, peers and other community members.