Acquiring relational meaning from the situational context: What linguists can learn from analyzing videotaped interaction

The acquisition of linguistic expressions of relational meaning is typically regarded as a difficult task (Gentner 1978; Gleitman 1990).2 What hampers this process, as compared to the acquisition of expressions of object reference, is not so much the complexity of relational concepts per se, which children seem to understand well before their first birthdays for many types of relational concepts (Casasola, Bhagwat, and Ferguson 2006; Mandler 2006; Sootman-Buresh, Woodward, and Brune 2006), but the question which subset of the large set of relations perceivable in situational context should be mapped to a linguistic expression. Furthermore, as especially Gleitman (1990) argues, not all relations expressed are directly perceivable, as some pertain to mental states of intentional agents and others are simply not present in the here-and-now of the speech situation. As Gleitman succinctly puts it (p. 5): “[T]here is not enough information in the whole world to learn the meaning of even simple verbs, or . . . there is too much information in the world to learn the meaning of . . . verbs.” In this chapter, we evaluate Gleitman’s double problem. Is the context indeed too poor or too rich for the learner to find situational, contextual correlates for the linguistic items he perceives that can help him bootstrap the meaning of