[ CogSci Summaries home | UP | email ]

Glasgow, J., Narayanan, N. H., Chandrasekaran, B. (1995). Diagrammatic Reasoning: Cognitive and Computational Perspectives. AAAI Press/MIT Press: Cambridge, MA.

  ALTauthor = 	 {},
  ALTeditor = 	 {Janice Glasgow, N. Hari Narayanan, B. Chandrasekaran},
  title = 	 {Diagrammatic Reasoning: Cognitive and Computational Perspectives},
  chapter = 	 {},
  publisher = 	 {AAAI Press/MIT Press},
  year = 	 {1995},
  OPTaddress = 	 {Cambridge, MA},

Author of the summary: Jim Davies, 2001, jim@jimdavies.org

Cite this paper for:


Some distinctions:

External world: the world.
External diagrammatic representations: constructed part of the external world, meant to represent something else in the external world.
Internal diagrams or Images: internal representations with pictoral properties

I: Phenomenal image: the experience of an image
R: "Pattern of activation in the neural structure in long-term memory that contained information which gave rise to I."
P: Pattern of neural structure that is in the brain when agent experiences I.

R is a neural pattern-- it makes no sense to talk about whether or not it is propositional or imagistic. Whether it can be consistently described propositionally is independent of whether it can be consistently described imagistically. It's a false distinction. [p xvii]

visual information: "information that humans can extract by inspection from an image or from the world by directing visual attention to it." e.g. shapes, certain simple spatial relations, color, texture, etc.

Diagram, in this book, refers both to mental images and externally drawn images. [p xxii]

why diagrams are helpful: they preserve locality information (neighborhood information, relative intersections, etc.)

When diagrams are helpful: When you can visualize a problem (perhaps using culturally learned visual analogs) such that the solution is easily gettable from perceptual processes on the visualization.

[p xxiii] R: representation (non-visual)
P: Predicate we are interested in computing
VR: Visualization of R
PR: Predicate we can get from VR (we can get it efficiently from the visual architecture)
Also suppose we can get P from PR by using the mapping used to get from R to VR. This describes how a visualization could be useful.

[p xxiv] In our culture we often represent time as a line, and can reason about time by reasoning about line lengths.

Summary author's notes:

Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies (jim@jimdavies.org)
Last modified: Thu Apr 15 11:07:19 EDT 1999