The Developmental Origins of Logical Inference: Deduction and Domain-Generality
MetadataShow full item record
CitationMody, Shilpa. 2016. The Developmental Origins of Logical Inference: Deduction and Domain-Generality. Doctoral dissertation, Harvard University, Graduate School of Arts & Sciences.
AbstractIs there a fundamental divide between the types of thoughts that human adults can entertain and those available to infants and nonhuman animals? The research in this dissertation explores the developmental origins of abstract, combinatorial, propositional thought. As a case study, we examined infants’ and children’s ability to make a logical inference, the disjunctive syllogism: A or B, not A, therefore B.
In Paper 1, we asked when infants begin to recruit negation in reasoning. When shown that a toy was hidden in one of two buckets, and that one of those buckets was empty, 17-month-olds (but not 15-month-olds) used that negative information to preferentially approach the other bucket. When shown that two blocks placed on a toy together activated the toy, but that one of those blocks did not activate the toy by itself, 17-month-olds (but not 15-month-olds) used that negative information to preferentially attempt to activate the toy with the other block. This parallel onset, particularly when combined with a similar pattern of findings in the word learning literature, suggests that a common underlying factor led to the change in infants’ performance – likely, the ability to use negation flexibly in reasoning.
In Papers 2 and 3, we looked for evidence that children’s performance on these tasks actually reflected a deductive inference, rather than a simpler non-deductive strategy that did not require representing the disjunctive relation between the options. In particular, we looked for evidence that excluding one option led children to be deductively certain of the other option. In Paper 2, we used an adaption of the search task described above to assess whether seeing that one location was empty made children certain that the object was hidden in the other location. We found that 3-, 4-, and 5-year-olds showed evidence of making this inference, while 2.5-year-olds did not. Paper 3 used an analogous design in a version of the causal reasoning paradigm described above, and found that seeing that one block was inert led 3- and 3.5-year-olds to be certain that the other block could cause the toy to activate, while 2.5-year-olds showed no evidence of making this inference.
Together, these studies suggest that children begin to be able to recruit an abstract, combinatorial, domain-general representation of negation at around 1.5 years of age. However, it is not until 3 years of age that there is evidence that children actually make the disjunctive syllogism inference, recruiting representations of negation and disjunction to come to a deductively certain, logical conclusion.
Citable link to this pagehttp://nrs.harvard.edu/urn-3:HUL.InstRepos:33840745
- FAS Theses and Dissertations