M.D. Terekhov - Post-graduate Student, Philosophy Faculty, Philosophy of Science Department, Lomonosov Moscow State University. E-mail: stereodrome@yandex.ru

Frame problem may be understood technically as a problem of representing features of the world, which actions of a cognitive agent don-t affect; philosophically it-s a problem of epistemic relevance. In order to solve the frame problem it was proposed to associate information to make big chunks, «frameworks for representing knowledge» (M. Minsky). H. Dreyfus criticized this approach, because the reduction of complexity actually was not done: a frame-based reasoning system would face the choice from millions of frames for recognizing frames etc. Cognition must be embodied and embedded: this condition was captured in Heidegger-s «reaniness-to-hand» concept and in phenomenological philosophy of Merleau-Ponty. But there may be some objections to the Heideggerian AI: it seems that in particular it only transforms the general problem of relevance. There are also general objections to the connectionist incarnation of Heideggerian AI expressed by J. Fodor: he claims connectionism to fail to comply with systematiс and productive character of human thought.

 

1. Smysl, relevantnost i predely tekhnologii. Beseda G. Krejjslera s KH. Drejjfusom // KHORA. 2008. № 1.
2. Minskijj M. Frejjmy dlja predstavlenija znanijj. M.: Mir. 1979.
3. KHajjdegger M. Bytie i Vremja. M.: Ad Marginem. 1997.
4. ChowS.J.What-s the Problem with the Frame Problem - // Review of Philosophy and Psychology. 2013. №4.
5. Churchland P.On the Nature of Intelligence. Turing, Church, von Neumann and the Brain. Neurophilosophy at Work, Cambridge University Press. 2007.
6. Dreyfus H. Why Heideggerian AI failed and how fixing it would require making it more Heideggerian // Artificial Intelligence. 2007. P. 171.
7. Fodor J., Pylyshyn Z.Connectionism and cognitive architecture // Cognition. 1988. № 28.