Select All
Christopher Caldwell,
The New Yorker,
Feb 24, 2004
In yesterday's OLDaily I talked a bit about having a non-cognitivist theory of cognition. Let me expand on that a bit with this short item. Connectionism and neural networks become cognition, in my view, through similarity. While similarity is usually caricatured in cognitivist texts, a proper analysis may be found in the work of Amos Tversky, who, as this article notes, proposed "a feature-based ‘contrast model’ of similarity, in which common features tend to increase the perceived similarity of two concepts, and where feature differences tend to diminish perceived similarity." My own refinement of this position, summarized here, involves the introduction of context to similarity measures, thus producing "relevant similarity". Anyhow, this article, a review of Barry Schwartz's "The Paradox of Choice," draws on on Tversky's work, in part, to explain why people make choices that are 'good enough' rather than optimal. A lot follows from this, including most of my own theoretical perspective regarding online learning; well worth adding to your reading list is not only this volume but also Tversky's Preference, Belief and Similarity, released last December.
Today: 4 Total: 4 [Direct link] [Share]
