From owner-cse575-fa08-list@LISTSERV.BUFFALO.EDU Fri Aug 29 09:02:29 2008 Date: Fri, 29 Aug 2008 09:01:25 -0400 From: "William J. Rapaport" Subject: 3 Key Ideas in (Computational) Cognitive Science To: CSE575-FA08-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: 3 Key Ideas in (Computational) Cognitive Science ------------------------------------------------------------------------ Here are a few more details on 3 key ideas underlying (computational) cognitive science, which I discussed briefly at the end of Thursday's lecture. 1. Newell & Simon's Physical Symbol System Hypothesis 2. The representational theory of mind & Fodor's language of thought 3. Dennett's theory of the intentional stance ------------------------------------------------------------------------ 1. Newell & Simon's Physical Symbol System Hypothesis ------------------------------------------------------------------------ Allen Newell & Herbert Simon were two of the pioneers of AI and of cognitive science (Simon won the Nobel prize in economics.) In their 1976 paper, "Computer Science as Empirical Inquiry", http://www.cse.buffalo.edu/~rapaport/575/F07/cgsmiscrdgs.html#newellsimon76 they proposed the "Physical Symbol System Hypothesis": A physical system can *exhibit intelligent behavior* iff it is a physical *symbol system*. A "symbol system" is essentially a mathematical model of computation (what is know more technically as a Turing machine--more on this later in the semester). And a "physical" symbol system is a physical implementation of a symbol system, i.e., a computer. ------------------------------------------------------------------------ 2. The representational theory of mind (RTM) & Fodor's language of thought (LOT) ------------------------------------------------------------------------ The RTM (or the CRUM--computational representational understanding of mind--as Thagard 2005(*) calls it :-) says: Mental processes are computations (or "rules") that operate on symbols that are mental representations (of things in the world). The philosopher Jerry Fodor's LOT hypothesis is: Mental representations are a language, with a syntax (i.e., a grammar) (and *maybe* with a semantics). You can read more about these in MITECS, and in Fodor's 1975 book, _The Language of Thought_: http://www.cse.buffalo.edu/~rapaport/575/F07/cgsmiscrdgs.html#fodor75 The reason I think that Fodor's LOT might not come with a semantics is because of Fodor's related theory of "methodological solipsism", essentially the view that the proper study of the mind does *not* need to pay any attention to the way the world is, roughly because everything that the mind believes about the world has first been "internalized"; see his 1980 paper, "Methodological Solipsism Considered as a Research Strategy in Cognitive Psychology": http://www.cse.buffalo.edu/~rapaport/575/F07/cgsmiscrdgs.html#fodor80 ------------------------------------------------------------------------ 3. Dennett's theory of the intentional stance ------------------------------------------------------------------------ In his 1971 article, "Intentional Systems", http://www.cse.buffalo.edu/~rapaport/575/F07/cgsmiscrdgs.html#dennett71 the philosopher Daniel C. Dennett argued that it makes sense to treat certain complex systems *as if* they had beliefs and other "intentional" states, even if they don't. (By "intentional", Dennett merely means "mental"; this is a term of art in philosophy that I'll say more about later in the semester.) There is a sort of hierarchy of "stances" that one can take towards such a complex system: a) The physical stance, in which you explain and predict the system's behavior in terms of its physical implementation: This is best for explaining and predicting malfunctions. E.g., you could say that a chess-playing computer crashed because transistor #7 misfired (or something like that). b) The design stance, in which you explain and predict the system's behavior in terms of its design. In the case of a computer program (which, by the way, is what Simon believes is the prime example of a complex system), this would be its program. E.g., you could say that a chess-playing computer castled because of a certain IF-statement on line 7 of its program. c) The intentional stance, in which you explain and predict the system's behavior by treating it like an intelligent entity, i.e., in terms of beliefs, desires, and intentions. E.g., you might say that the chess-playing computer castled because it believed that that was the best response to its opponent's Armenian defense (or whatever). When Kasparov lost to Deep Blue, he described his computer opponent in intentional terms! ======================================================================== (*) http://cogsci.uwaterloo.ca/courses/resources.html