Mowgli in the jungle of words: Comprehension and later lexical development

Mowgli in the jungle of words: Comprehension and later lexical development Eef Ameel (eef.ameel@psy.kuleuven.be) and Gert Storms (gert.storms@psy.kuleuven.be) Department of Psychology, University of Leuven, 3000 Leuven, Belgium Barbara C. Malt (barbara.malt@lehigh.edu) Department of Psychology, Lehigh University, Bethlehem, PA 18015 USA Abstract Production data suggest that meanings of common nouns continue to change well past the early years of language acquisition (Andersen, 1975; Ameel, Malt, & Storms, 2008). Here we used two comprehension tasks to further evaluate the nature of later lexical learning. In a name applicability task, seven- to 13-year old Dutch-speaking children judged whether each of three names applied to common household containers. In a typicality judgment task, participants judged how good an example the containers were of the three names. Both tasks revealed continued evolution of word use up to age 13. However, all names were overextended by the children, in contrast to the production data in which both over- and underextension were found. Moreover, the child lexical categories showed considerable overlap, indicating strong inter-category relatedness. With age these overextensions as well as the overlap gradually disappeared and the lexical categories became more distinct over time. Keywords: word learning; later lexical development; lexical categories; typicality judgments Introduction The speed with which young children add words to their vocabulary seems to suggest that word learning is a relatively easy process. By age two, they are able to produce about 600 words. In the period from age two to six, they are estimated to acquire around 14.000 words, at a rate of ten words a day (Carey, 1978). The meanings assigned to these words, however, only partially correspond to the meanings adults attach to the words. The incomplete word meanings inevitably cause children to commit errors. Some words are used too broadly, a phenomenon called overextension, for example, when the word cow is applied to all four-legged animals. A category is underextended when its name is assigned to too few items, for example, when the word cow is only applied to the child’s cuddly animal. These errors disappear when the full conventional meaning is grasped, but this may take months, or even years. A number of studies have shown learning periods for verb meanings extending to age 8 or 9 (e.g., for pour and fill, Gropen, Pinker, Hollander, & Goldberg, 1991; see also Bowerman, 1974; Pye, Loeb, & Pao, 1996). But even for common nouns, the time frame for fully acquiring their meaning and use seems to extend beyond the early years of language acquisition. This reflects the broad and complex extension of common nouns referring to concrete, simple objects. Ball for English speakers, for example, often refers to smooth, bouncy, spherical, deflatable things made for play, but it can also refer to non-smooth balls of paper, non- bouncy beanbag balls, non-spherical footballs, non- deflatable baseballs and billiard balls, and things not for play such as balls of yarn, balls of string, and tea balls. A few studies clearly demonstrated the extended learning trajectory for nouns that are used to refer to common household objects. Andersen (1975), for example, asked English-speaking children aged 3 through 12 to name drinking vessels, and she found that children’s naming did not match adult naming until the age of 12. Ameel et al. (2008) had Dutch- speaking Belgian children of ages 5, 8, 10, 12 and 14 name common household containers, which were mostly called fles, bus, and pot by adults (Ameel, Storms, Malt, Sloman, 2005). They found that early use of fles and pot was overextended and gradually narrowed from 5-year-olds to adults, while the opposite pattern was found for bus. Gradually, the children converged toward the adult naming, but even the 14-year-olds still slightly differed in their word use from the adults. The results of Ameel et al. (2008) were obtained using a production task in which participants named the object in each picture however they chose. A production task offers several advantages as opposed to a comprehension task in which participants are generally asked to make a forced choice decision between two or more options. A first advantage is that it does not constrain responses to be from a set pre-determined by the experimenter to be suitable. Production is also a purer measure of what participants know because they have to come up with a name without having any suggested to them. Given a forced choice between fles and bus (or any other word pair), participants will pick one regardless of how little knowledge they may have of the words. Finally, production has a certain ecological validity in that outside the laboratory, children and adults frequently select names for objects for purposes of communication, and they do so without having options presented to them. However, there are also some disadvantages related to production tasks. First, production tasks might be cognitively more demanding than comprehension tasks, since they require not only a sense of familiarity with the presented material but also retrieval from memory. Children have had less language exposure than adults, so even if they have some knowledge of a word and of appropriate word-