|InterJournal Complex Systems, 482
|Manuscript Number: |
Submission Date: 10318
Comment on manuscript revision number 50812|
ICCS Review for paper 397 by L. Sacks.
Subject(s): CX.67, CX.66, CX.35
Category: Brief Article
This is an interesting paper. My personal interest is in the Modularity aspect, especially in its relation to the Karmiloff-Smity (1992) work. The 'punctuated evolution' side is also interesting; but this should be compared with - related - Simulated Annealing techniques rather than issues of biological evolution. The former improves convergence - as it does in this manuscript. The latter is more of an emergent property... Beyond Modularity is specifically about the relation between modules in the mind - rather than their specific existence. In particular there is a critically important issue of how modules convert information for representation both back into their own domains and into other modules (Representational Rediscription RR). Although the role of modules is extensively discussed. This issue has been shown to be critically important in human evolution - see The Prehistory of the Mind, Steven Mithen 1998 and other authors in evolutionary psychology. This manuscript shows that modular structures can evolve in a neural network. It would be interesting to get a more intuitive feel for what each module was handling i.e. in which domains they operated. In the Karmiloff-Smith book the mind handles the Domains like: Language, Physics, Maths, Psychology etc. More importantly it would be interesting to know how or if the individual modules aided each other through some RR like mechanism. Overall a good attack on these issues which are critical to understanding the human mind.
|Retrieve Manuscript Abstract|
|Submit referee report/comment|