Herbert A. Simon
Professors Grahoff and May (henceforth, G&M) have proposed an interesting model of scientific discovery, and have applied it to the discovery of the reaction cycle for urea synthesis, a problem solved in 1932 by Hans Krebs. As they compare their proposed explanation of Krebs’ discovery to one earlier published by Deepak Kulkarni and myself (Henceforth, K~S), and as their judgments on our program are rather negative, I have drafted a few remarks to explain why we proceeded as we did. 1 I will leave to the forthcoming meeting the discussion, which I expect will be lively and productive, of broader issues of the relations of both of these programs to general theories of scientific discovery: both theories that seek to explain the actual behaviors of scientists and theories that have the normative intent of demonstrating efficient discovery procedures. Thus, I will only address some of the specific criticisms G&M made of our simulation.