Conventionality lies at the center of human communication. For communication to be effective there must be symbols that are shared among speakers. In this proposal, Dr. Koenig investigates the possibility that understanding the conventional nature of words serves as a powerful tool in language acquisition. While speakers are often consistent in the particular language they use, the presence of different languages across speakers raises important questions about the scope of children's conventional understanding when more than one language system is involved. Can monolingual toddlers learn foreign terms from foreign speakers? Do they treat foreign terms as distinct from inaccurate terms? This project will compare 24-month-olds' ability to learn new words from an accurate English speaker versus an inaccurate English speaker and a native Dutch speaker. Given promising evidence that children can learn foreign terms presented in Dutch, an interesting question arises of how broadly they generalize what they have learned across speakers. This study will examine toddler's generalization across different types of speakers, native and foreign. Broad generalization would indicate that infants have accepted the word, but not demarcated the boundaries of their own system versus a foreign language. Narrow generalization -- generalizing English but not foreign terms to users of your own language -- would suggest that infants have begun to understand that different systems can exist. In the final component, this project will directly examine the possibility that observation of consistent word use across speakers is more powerful evidence in favor of a word than is consistency that is demonstrated by only one speaker.
The major goal of this research is to provide insight into different aspects of conventionality in language: (1) knowing that conventional systems can, in principle, vary across speakers; (2) knowing that word knowledge generalizes to other speakers, but not all speakers; and (3) knowing that the basis of conventionality lies in consistent word use across speakers. Each aspect reflects an important conceptual achievement in understanding language as a system, and a central aim of this project is to identify the emergence of each component. As a result, it will contribute to scientific knowledge about the development of conventional understanding, and advance current theories of language development. Beyond elucidating a critical aspect of language learning, these studies will promote future research in bilingual populations and shed light on the more general question of how children extract and generalize cultural and social norms.
Most children acquire their first language with a wizard-like ease. However, once children have learned their first language, a second language (that comes later) presents a different set of problems. By the year 2030, an estimated 40% of children in US schools will be non-native speakers of English (Wirt et al., 2003). The projects funded by this NSF award (PI: M. Koenig) have been dedicated to investigating children's ability to learn new terms from foreign speakers as well as their willingness to generalize newly learned terms to new speakers. Through this research, we have a much better understanding of how young children treat foreign speakers and foreign language boundaries. To sum across the findings listed below, important aspects of learning words in a second language recruit the child’s existing native language knowledge. By supporting the mastery of language skills in their native language, children are better able to acquire new skills and information in a new language. The following summarizes some of our primary findings: Terminological note: L1 = ‘native language’ and L2 = ‘foreign language’ (1) Monolingual English-learning two-year-olds with higher L1 vocabularies successfully learned words for objects in a foreign language, unlike two-year-olds with lower vocabularies. (2) In a training study, we found age-related changes in the benefits of L1 word knowledge. Two-year-olds' long-term memory was best for foreign terms that labeled an object previously labeled in L1, while 3-year-olds’ long term foreign-term recall was strong both with and without prior L1 training. (3) When learning words in their native language, English-learning 2- and 3-year-old children showed no evidence of difficulty learning from a nonnative speaker with a heavy accent, suggesting that social preferences and acoustic unfamiliarity do not interfere with word learning from comprehensible, nonnative speech. (4) However, when learning words in a foreign language (Spanish), two- and three-year-old children remembered words more easily when the foreign speaker used English phonology, suggesting that acoustic familiarity supports the segmentation of foreign speech, and word-object mapping. (5) Similarly, children learn foreign words which contain highly frequent sounds in their L1 (e.g., "fen") easier than foreign words that contain infrequent sounds ("fetch"). Thus, in a phonological sense as well, L1 learning might be leveraged to promote L2 learning. (6) 3- to 4-year-olds need 3-6 presentations of a word in L2 to learn it versus 1 in their native language (i.e., they are not "fast-mapping" in L2 initially). (7) In a training study that presented both familiar and novel object labels, children learn foreign words for objects with which they are familiar more easily than for objects that are completely novel. (8) Knowing translation equivalents in L1 does NOT INTERFERE with L2 word learning (which many people are afraid of & follows from the EF literature Children learn foreign words better if they already have a word for the object in their native language;). (9) Five-year-olds learn foreign words better than 4-year-olds who learn better than 3-year-olds (i.e., they require fewer presentations to learn and can remember words better over a few days). So it might be optimal to wait until children are a little older, or support their vocabulary growth in L1, before teaching them words in a second language.
