Can Infants Use Transitive Inference in Attribution of Goals to Others?

Abstract

Transitive inference refers to the ability to use knowledge of pre-existing relationships to infer relationships between entities that have not been directly compared. This form of logical inference is an important skill for many social species, and has been thought to arise in an immature form in humans between the ages of four and six years. The experimental methods used to test this ability in humans often require some verbal skill, gross or fine motor coordination, a memory capable of containing numerous relationships, and often a great deal of time and repetition in testing. These methods of testing may have been too demanding on other physical and cognitive abilities to be successfully completed by children under four years of age, regardless of their ability to make transitive inferences. The present study used methods sensitive to infant cognition to test the current theory that the ability to make transitive inferences does not develop until the age of four. Nine-month-old infants were tested in three separate experiments using a visual habituation paradigm similar to that used by Woodward (1998) and through investigation of infants’ own imitative actions. Experiment 1 verified that infants can track the goals of others in a habituation paradigm when the goal object changes position throughout habituation trials, both through looking time measures and imitative action. Experiment 2 used an extension of this paradigm to examine the ability to make transitive inferences across a three item chain, serially ordered by the actor’s object preference, and no evidence of transitive inference was observed. Experiment 3 tested infants’ ability to habituate to and recall multiple goals using context as a cue to actor choice. Infants were able to consistently track only one of the pairings, suggesting that avoidance, in addition to selection, may play a role in infant performance in the visual habituation paradigm.