Robbit: "... seems eminently pertinent to this starnge notion that LW would have anything much to contribute to proto-scientific modeling of "expert judgement". The last thing he was interested in is armchair scienctific model building and speculation." " Rob: I'm unsure if you missed this point or if I am misreading you. They guy said, in a moment of curiosity, something like: "I wonder what expert judgment really is?" (for purposes of wondering whether AI could replicate it). I followed by directing him to Wittgenstein's concern with the nature of aesthetical judgments. This concern surely implicates his question and would have a great deal to offer. Especially the parts about: seeing as, non-universality, appreciation is shown not proved, "getting it," non-formalistic style of judgment, etc. etc. So, before he can ask whether AI can simulate connoisseur judgment, he needs to know what that is. And if he comes to believe that a magical kind of technology in the future might replicate it, he would need further help to see that this has no bearing on his current belief that statistical and formal methods stand distinct from and outperform connoisseurs. One confusions arises from not knowing what connoisseur judgment is, the other from not seeing what happens to his language game when the conditions that create it get messed up in the future. Did I misread your quote? SW _______________________________________________ Wittrs mailing list Wittrs@xxxxxxxxxxxxxxxxxxx http://undergroundwiki.org/mailman/listinfo/wittrs_undergroundwiki.org