Acquiring a Proper Name by Correcting for the Speaker's False-belief About Identity

• In Event: 3-224 - The Nature of Infants' and Preschoolers' Sensitivity to Other People's Object-representations

Integrative Statement

Is early theory of mind (ToM) fundamentally different from the later ToM, in particular, does it lack the ability to compute false-beliefs about object identities? We explore this issue with young preschoolers and toddlers, through proper names acquisition. Proper names pick out particular individuals and are learned through specific acts of labeling. Suppose Jane is mistaken about the identity of a dog currently in a box and mislabels it with her pet dog’s name. Will a young child map the name with the dog actually in the box or will she correct for Jane’s false belief and learn the name of the intended referent?
We presented three-year-olds with a narrated, illustrated story in which Jane puts one of two identical dogs in a box and leaves. In her absence, this dog jumps out of the box and leaves the scene; then the other dog jumps into the box. Jane returns with her brother, points to the box, and says, “That’s Fido in the box. I put Fido in the box for sleep”; they then leave. In the test scene, the first dog returns and the two dogs are presented side by side. Subjects are asked to point to Fido. Out of 20 three-year-olds (mean = 3 years 5 months, SD = 3.41 months; range = 36-47), 15 pointed to the correct dog—the speaker’s intended referent (Binomial p=.04, two-tailed; BF=3.2 in favor of H1). Could children, however, have tracked the referent through Jane’s description, “I put Fido in the box …”, rather than through the identity false-belief? To rule out this possibility, we conducted a new study on a new group of three-year-olds, omitting this description from the script. Thus, in the naming-scene, Jane points and says simply, “That’s Fido in the box!” In addition, we included a true-belief (TB) condition in which Jane returns before the dogs change their original places, so her intended referent is the same as the actual referent. Out of 21 three-year-olds in the FB condition (mean = 3 years 5 months, SD = 3.89 months; range = 36-47), 16 pointed to the correct dog—the speaker’s intended referent (Binomial p=.027, two-tailed; BF=4.68 in favor of H1). In the TB condition, 16 out of 19 three-year olds (mean = 3 years 5 months, SD = 3.77 months; range = 36-47 months) pointed to the correct dog—the actual and the speaker’s intended referent (Binomial p=.004, two-tailed, BF=27.05 in favor of H1). Hence, three-year-olds appear to correct for the speaker’s false-belief about identity in a labeling task. Figure 1 summarizes the results of studies 1 and 2.
These results show that a Speaker’s referential intention may override stimulus-world associations in early world learning. Also, they suggest that early theory of mind (ToM) may sometimes be able to represent false beliefs about object identities, indicating its developmental continuity with later ToM.

Authors

• Gala Stojnic, Rutgers University, NJ
  Presenting Author

• Alan Miller Leslie, Rutgers University, NJ
  Non-Presenting Author