Infants Interpret Ambiguous Requests for Absent Objects Megan M. Saylor Vanderbilt University Patricia Ganea University of Virginia The current studies investigated 2 skills involved in 14- to 20- month-olds’ ability to interpret ambiguous requests for absent objects: tracking others’ experiences (Study 1) and representing links between speakers and object features across present and absent reference episodes (Study 2). In the basic task, 2 experimenters played separately with a different ball. The balls were placed in opaque containers. One experimenter asked infants to retrieve her ball using an ambiguous request (“Where’s the ball?”). In Study 1, infants used the experimenter’s prior verbal and physical contact with the ball to interpret the request. A control condition demonstrated that infants were interpreting the request and not responding to the mere presence of the experimenter. Study 2 revealed that only infants who were given stable cues to the ball’s spatial location appropriately interpreted the request: When spatial information was put in conflict with a color cue, infants did not select the correct ball. Links to infants’ spatial memory skills and emerging pragmatic understanding are discussed. Keywords: language development, theory of mind, spatial memory, pragmatic competence Absent reference comprehension is a pivotal accomplishment of early life. Its emergence supports development in a variety of areas including concept formation and memory, as it takes language beyond the concrete present to entities that are abstract and hypo- thetical. Previous research has demonstrated that such comprehen- sion emerges across infants’ second year (Ganea, 2005; Hutten- locher, 1974; Miller, Chapman, Branston, & Reichle, 1980; Sachs, 1983; Saylor, 2001, 2004; Saylor & Baldwin, 2004). However, it is still not clear what skills support its emergence. The current research addresses this gap. Before detailing the current studies, we describe a task that is central to understanding absent reference to set the stage for a discussion of two skills infants may recruit when interpreting such references. Decoding References to the Absent Oftentimes when speakers refer to absent things, the referent is quite clear. For example, if a speaker uses a term that picks out a unique individual (e.g., “Todd” or “the moon”) or if there is only one referent available, then identifying the referent may be rela- tively unproblematic. Previous research has revealed that infants from 13 months can interpret such unambiguous absent references (e.g., Ganea, 2005; Saylor, 2004). However, absent reference will often be less transparent because there may be multiple possible exemplars that an expression can refer to. For example, if a father refers to “the dog,” an infant only understands the reference if she identifies the particular dog that is specific to her shared experi- ence with a dog but not to dogs in general or dogs that she has seen with other people. In this case, the infant’s task is to identify the particular category member a speaker is referring to. To do so, infants must track others’ experiences vis-a `-vis referent objects and maintain links between people and objects across present and absent reference episodes. We discuss each skill and its relation to absent reference understanding below. Tracking Others’ Experiences To track others’ experiences vis-a `-vis referent objects, infants must recognize which category member the speaker is familiar with. Research on adults’ pragmatic competence suggests that the endpoint of this skill is recognizing what knowledge is shared between speakers (i.e., mutual knowledge; Clark & Marshall, 1981; Pickering & Garrod, 2004). Inferences about mutual knowl- edge involve recognizing (a) how experiences affect knowledge states (e.g., seeing = knowing) and (b) iterative relationships between the knowledge states of self and other (e.g., “I know that you know that I know that you know . . .”). In other words, to determine whether something is mutually known, speakers need to recognize that knowledge of something exists in the mind of their interlocutor and to recognize the relationship between their own knowledge states and others’ knowledge states. These skills may be out of infants’ reach for two reasons. First, the recognition of causal relationships between experiences and knowledge states does not emerge until the preschool years (e.g., Montgomery, 1992). Next, infants’ limited processing resources may make it difficult to represent iterative relationships. Infants may track others’ experiences using the more modest strategy of noting which objects and events their speech partners Megan M. Saylor, Department of Psychology and Human Development, Vanderbilt University; Patricia Ganea, Department of Psychology, Univer- sity of Virginia. Portions of these data were presented at the Jean Piaget Society Con- ference, Vancouver, British Columbia, Canada, June 2005. We thank C. Brooke Carroll, Diana Carver, Elisabeth Galle, Alexandra Fortuna, Emilie Peloubet, Jennifer Peterson, Katherine Singleton, Tywanquilla Walker, and Jennifer Wright for assistance with collecting and coding the data as well as John Rieser and Bethany Rittle-Johnson for comments on the manu- script. Correspondence concerning this article should be addressed to Megan M. Saylor, Department of Psychology and Human Development, Vander- bilt University, Nashville, TN 37203. E-mail: m.saylor@vanderbilt.edu Developmental Psychology Copyright 2007 by the American Psychological Association 2007, Vol. 43, No. 3, 696 –704 0012-1649/07/$12.00 DOI: 10.1037/0012-1649.43.3.696 696