We report three eyetracking experiments that examine the learning procedure used by adults as they pair novel words and visually presented referents over a sequence of referentially ambiguous trials. Successful learning under such conditions has been argued to be the product of a learning procedure in which participants provisionally pair each novel word with several possible referents and use a statistical-associative learning mechanism to gradually converge on a single mapping across learning instances. We argue here that successful learning in this setting is instead the product of a one-trial procedure in which a single hypothesized word-referent pairing is retained across learning instances, abandoned only if the subsequent instance fails to confirm the pairing – more a ‘fast mapping’ procedure than a gradual statistical one. We provide experimental evidence for this Propose-but-Verify learning procedure via three experiments in which adult participants attempted to learn the meanings of nonce words cross-situationally under varying degrees of referential uncertainty. The findings, using both explicit (referent selection) and implicit (eye movement) measures, show that even in these artificial learning contexts, which are far simpler than those encountered by a language learner in a natural environment, participants do not retain multiple meaning hypotheses across learning instances. As we discuss, these findings challenge ‘gradualist’ accounts of word learning and are consistent with the known rapid course of vocabulary learning in a first language.

Three experiments explored how words are learned from hearing them across contexts. Adults watched 40-s videotaped vignettes of parents uttering target words (in sentences) to their infants. Videos were muted except for a beep or nonsense word inserted where each "mystery word" was uttered. Participants were to identify the word. Exp. 1 demonstrated that most (90%) of these natural learning instances are quite uninformative, whereas a small minority (7%) are highly informative, as indexed by participants' identification accuracy. Preschoolers showed similar information sensitivity in a shorter experimental version. Two further experiments explored how cross-situational information helps, by manipulating the serial ordering of highly informative vignettes in five contexts. Response patterns revealed a learning procedure in which only a single meaning is hypothesized and retained across learning instances, unless disconfirmed. Neither alternative hypothesized meanings nor details of past learning situations were retained. These findings challenge current models of cross-situational learning which assert that multiple meaning hypotheses are stored and cross-tabulated via statistical procedures. Learners appear to use a one-trial "fast-mapping" procedure, even under conditions of referential uncertainty.acquisition | induction | language | vocabulary F undamental for each child entering the human community is the acquisition of word meanings: discovering which language sounds map onto which interpretations. Because these mappings are arbitrary and vary cross-linguistically, growing a vocabulary poses a classic learning problem for humans, both infant learners of a first language and second-language learners who must replace the original mappings with a new set. This experiencedependent learning problem for humans contrasts with animal communication systems in which the interpretations of speciesspecific barks, chirps, and growls are largely given for free by nature. The present article provides experimental evidence concerning the primitive initial procedure by which humans acquire vocabulary items.A common assumption is that form-to-meaning mappings are discovered in a process mediated by observation of extralinguistic events: The learner matches recurrent speech events to recurrent aspects of the observed world. For example, when an English speaker says "dog" or a French speaker says "chien," there is likely to be a co-occurring dog sighting. Young children often acquire a word's meaning after a single such exposure to its use in context (1), particularly if there is strong pragmatic support (2) or a restrictive syntactic environment (3). The sheer size of the average vocabulary at age 6 y [estimated at 6,000-8,000 words (1)] suggests that this "fast mapping" of a sound segment onto its interpretation must happen very often, as is also attested in many laboratory studies (4,5).Yet the world of words and their contexts is enormously complex. Few words are taught systematically, even in middleclass environments with...

Children vary greatly in the number of words they know when they enter school, a major factor influencing subsequent school and workplace success. This variability is partially explained by the differential quantity of parental speech to preschoolers. However, the contexts in which young learners hear new words are also likely to vary in referential transparency; that is, in how clearly word meaning can be inferred from the immediate extralinguistic context, an aspect of input quality. To examine this aspect, we asked 218 adult participants to guess 50 parents' words from (muted) videos of their interactions with their 14-to 18-mo-old children. We found systematic differences in how easily individual parents' words could be identified purely from this socio-visual context. Differences in this kind of input quality correlated with the size of the children's vocabulary 3 y later, even after controlling for differences in input quantity. Although input quantity differed as a function of socioeconomic status, input quality (as here measured) did not, suggesting that the quality of nonverbal cues to word meaning that parents offer to their children is an individual matter, widely distributed across the population of parents.language acquisition | word learning | SES C hildren's vocabularies vary greatly in size by the time they enter school (1, 2). Because preschool vocabulary is a major predictor of subsequent school success (3), this variability must be taken seriously and its sources understood. Some of this variability resides in the individual capacities and temperament that infants bring to the word learning task (4, 5). However, environmental influences are also bound to play instrumental roles. Accordingly, we examined the contextualized speech input parents provide to infants during the second year of life as a potential source of the massive vocabulary differences found at school entry.It is already known that the sheer quantity of linguistic input is an important determinant of vocabulary size; overall, the more words children hear early in development, the larger their subsequent vocabularies. This relationship holds true both for types (different words) and tokens (number of words heard, including repetitions) (6, 7). These quantity differences are correlated with socioeconomic status (SES). Children from low SES homes are typically exposed to fewer words early in development (8, 9) and have smaller vocabularies at school entry than children from high SES homes (10).Taken alone, the correlation of vocabulary size with amount of input is puzzling because as a general rule language learners do not seem to require a large number of exposures to a word to acquire its meaning (11). In experimental settings, for example, children have been shown to acquire and retain a new word heard only once or a very few times (12-14). The likelihood, then, is that certain exposures to a new word are especially informative, supporting secure and rapid inferences to meaning. For example, common sense insists that it will be ...

No abstract

Three experiments explored how words are learned from hearing them across contexts. Adults watched 40-s videotaped vignettes of parents uttering target words (in sentences) to their infants. Videos were muted except for a beep or nonsense word inserted where each “mystery word” was uttered. Participants were to identify the word. Exp 1 demonstrated that most (90%) of these natural learning instances are quite uninformative, whereas a small minority (7%) are highly informative, as indexed by participants’ identification accuracy. Preschoolers showed similar information sensitivity in a shorter experimental version. Two further experiments explored how cross-situational information helps, by manipulating the serial ordering of highly informative vignettes in five contexts. Response patterns revealed a learning procedure in which only a single meaning is hypothesized and retained across learning instances, unless disconfirmed. Neither alternative hypothesized meanings nor details of past learning situations were retained. These findings challenge current models of cross-situational learning. Learners appear to use a one-trial fast mapping procedure, even under conditions of referential uncertainty.