Infant word learning has become a popular field of study over the past decade.Research during this time has shown that infants can learn, in a short period of time, to attach words to objects. Two experiments on the role of social cues in early word learning are reported using tightly controlled conditions. Fourteen-and 18-month-old infants were trained by viewing a video of an adult pointing and nodding towards one of two different novel objects appearing on a screen simultaneously, while novel labels were emitted through a speaker. Infants' looking times to each object were recorded both during training and test trials. Our analyses indicated that both 14-and 18-month-olds looked significantly longer at the object that the adult pointed to in the training trials.However, only 18-month-olds showed any evidence of looking longer at the target object during the test in the consistent condition than in the inconsistent (control) condition.These studies are important because they show, in a controlled laboratory study of infant word learning, that different types of social cues are available at different ages. Fourteenmonth-olds are aware of adult pointing and head turning and can follow those cues to an object during training. However, it isn't until 18 months of age that infants seem able to use those cues in the service of actual word learning. Infant word learning has become a popular field of study over the past decade.Research during this time has shown that infants as young as 14 months can learn, in a short period of time, to attach words to objects (e.g., Werker, Cohen, Lloyd, Casasola, and Stager, 1998;Shafer & Plunkett, 1998). By 18 months of age, infants are using a number of assumptions to aid them in the process of linking words to objects. Three of the assumptions infants use to organize their linguistic world are the "whole object" assumption, "mutual exclusivity," and the "taxonomic" assumption. According to the whole object assumption, early word learners will attach a new label to a whole object, rather than to a part of that object (Markman, 1990); mutual exclusivity is the assumption that a child will assign a novel label to an object that the child does not already have a label for, rather than to an object for which the child has a label (Markman & Wachtel, 1988). Lastly, the taxonomic assumption is the notion that children will extend a new label to objects of the same kind, rather than objects that are thematically related to the known object (Markman, 1990). Children use these constraints as a way of minimizing the number of possible meanings for a given word. As Quine (1960) has pointed out in his gavagai example, there is an indefinite number of possible interpretations for any one item given the presentation of a label and it's referent. In order for children to learn language, they must be able to pick out the correct interpretations for new words that they encounter when listening to others' speech.However, these cognitive constraints thought to be used by early word learners are not ...
