Relational similarity

Relational similarity

1. A trademark of analogical reasoning

   • Objects compared in an analogy should be linked by the same relations

2. Carey (1978): children move from first order to higher order generalisations

3. Keil & Batterman (1984) UNCLE STUDY: characteristic + defining features -> shift in representational stucture does not happen for everything at the same time

4. 2 views

   • Conceptual difficulties: similarity judgements difficult because they are not constant
     • Keil & Carroll (1980): children aged 3 to 6 where shown pictures of houses and arrows
       • They could choose the tallest for houses, but not for arrows (which usually appear on a horizontal axis)
   • Perceptual difficulties: objects vary in several dimensions, hard to inhibit the irrelevant one
     • Garner: some dimensions are easy to separate (shape, colour, size) while others cannot be judged independently (brightness and saturation)
     • Gentner et al (1996): within + between domains, 4yo
       • randomized trials: performance at chance
       • within-dimensional trials, then cross-dimensional trials = improve in performance

5. Language

   • Ratterman & Gentner (1998): 3 experiments with 3-5 yo to find out whether relational labels improve performance
     • 1st exp: cross-mapping task for rich + sparse objects
       • 2nd exp: "mummy", "daddy", "baby" improved performance
         • 3rd exp: "big", "little", "tiny": children trained with relational labels produced sig more relational responses than children in the no-label condition
   • children start to group things into categories a week before "naming spurt" - present in normally developed children + children with Down syndrome