Schwartz, D. L. & Black, J. B. (1996) Shuttling between depictive models and abstract rules: Induction and fallback. Cognitive Science. 20. 457--497.

@Article{,
  author = 	 {Daniel L. Schwartz and John B. Black},
  title = 	 {Shuttling between depictive models and abstract
rules: Induction and fallback},
  journal = 	 {Cognitive Science},
  year = 	 {1996},
  OPTvolume = 	 {20},
  OPTpages = 	 {457--497},
}

Author of the summary: Jim Davies, 2002, jim@jimdavies.org

Cite this paper for:

• depictive model
• depictive representations turning into numerical, abstract representations. [461]

Two claims fall from this: People will use imagery when they would but can't use empirical evidence, and people can learn abstact rules from this thought experimenting.

This study explores these by looking at how people switch from non-imagery to imagery in thinking about gears. In gears there is the "parity rule:" if there are an odd number of gears connected, then the first will go in the same direction as the last. The prediction is that no one would need empirical evidence if there were sufficient confidence in this rule, and would thus not care to use mental imagery.

Their sense of a "depictive model" is the same as that of "analog imagery" (Shepard and Cooper 1982) plus some physical rules. [460]

The hypotheses were supported: As they gained experience with gear problems, answers were quicker, explanations became more quantitative, and gestures disappeared. These evidence going from simulation to rule following.

Summary author's notes:

• The authors don't use the term "thought experimenting;" I do.