Folk Psychology and the Language of Thought: Understanding Human Thought Processes, Study notes of Art

Download Folk Psychology and the Language of Thought: Understanding Human Thought Processes and more Study notes Art in PDF only on Docsity! Folk Psychology and the Language of Thought (Ch. 1 of Psychosemantics) Commonsense belief/desire psychology, aka “Folk Psychology” (FP) An ordinary example of Folk-psychological explanation: “Why did A deliberately bump into B?” “Because she likes him and was hoping to meet him.” In more detail, the explanation says something (very roughly) like: 1. A is attracted to B. 2. A believes that B might have or develop corresponding feelings toward A. 3. Ceteris paribus, if X is attracted to Y, and X believes that Y might have or develop corresponding feelings toward X, then X desires to form a relationship with Y. 4. So A desires to form a relationship with B. 5. A believes that she won’t form a relationship with B unless she meets him. 6. A believes that she won’t meet B unless she deliberately bumps into him. 7. Ceteris paribus, if X believes that P won’t happen unless Q does, and that Q won’t happen unless R does, then X believes that P won’t happen unless R does. 8. So A believes that she won’t form a relationship with B unless she bumps into him. 9. A believes that it is within her power to bump into B. 10. (BD  A) If X desires P, and X believes that P won't happen unless Q does, and X believes it is within X's power to bring it about that Q, then ceteris paribus, X tries to bring it about that Q. 11. So A tries to bring it about that she bumps into B. Three claims about FP: First claim: FP works very well very often.  It works so well, that we often ignore it and only attend to the cases where it fails. o Arranging meetings, etc. 1  “the theory from which we get this extraordinary predictive power is just good old commonsense belief/desire psychology” (3) o Note: there are alternatives to FP.  FP allows us to infer intentions from sounds (i.e. speech), and behavior from intentions. “Philosophical” objection: All these supposed “generalizations” are hedged with ceteris paribus clauses.  Does it make the generalizations trivially true? Do they have the form “If XXX happens, then YYY is sure to happen, unless it doesn’t”: o If someone says “I will do A”, then she intends to do A, unless she doesn’t. Reply 1: The ceteris paribus clauses cannot be trivial, because the generalizations work (and others like them don’t)  If you really want something, you’ll try to think of a way to get it (ceteris paribus) VS. If you really want something, you’ll avoid it (ceteris paribus).  Even though the generalizations have exceptions (and even though we may not be able to list them all), they are still informative. Reply 2: Explaining how ceteris paribus clauses work is a problem for every special science (psychology, economics, anthropology, geology, biology). It’s a substantial issue in the philosophy of science.  The fact that many scientific laws hold only ceteris paribus is a big part of why statistics and probability are so central to the sciences.  compare: “A meandering river erodes its outside bank” (vs. “rivers catch fire”) Conclusion: FP’s use of ceteris paribus clauses is as legitimate as any special sciences’ use of them.  Any hope for completely spelling out the contents of these ceteris paribus clauses will follow the same strategy that the special sciences uses: looking for better generalizations “down a level”. Second claim: FP is a deep theory of human cognition. (p. 7) FP’s version of psychological explanation is not just a bunch of platitudes. Instead, it exhibits the ‘deductive structure’ that is so characteristic of explanation in real science. 2 o causation of mental states by mental states (chains of thought)  It is characteristic of FP that it attributes contents and causal powers to the very same mental things that it takes to be semantically evaluable. o In fact, causal relations among propositional attitudes typically respect their relations of content, and belief/desire explanations turn on this.  You believe that Iris will clean the barn, because you believe that either she or Ed will, and you believe that Ed won’t.  Your mind is organized so as to be presently entertaining the belief that either Iris or Ed will clean the barn, and the belief that Ed won’t. The presence of these two occurrent states (plus lots of other background) physically causes you to have the belief that that Ed will clean the barn.  The causal connection between beliefs also correlates with a rational connection. More importantly, it correlates with the kinds of inferences humans actually draw. o FP explanations become quite general, because they quantify over agents, propositions, etc. This makes such generalizations empirically predictive (and potentially false). o Our explanations track the semantic relations holding between sentences, but the semantic relations correlate with the causal relations holding between them, so we end up correctly predicting behavior, mental states, etc. Thus, in the simplest cases, our trains of thought are like rational arguments.  Many interesting issues here; e.g., why did we develop so that the semantic and causal relations among our thoughts are so well correlated? (iii) Generalizations preserved  The causal powers of the attitudes must be roughly what common sense says they are. Things like modus ponens, (BD  A), etc. are okay, but (**) is not: o (**) If X desires that P and X believes that P or Q, then X believes that P.  Can be lots of exceptions, but the thesis of FP shouldn’t be made trivial. The Representational Theory of Mind (RTM) RTM is the only going theory that can explain FP 5  since FP must be explained, we have a quick argument in support of RTM. o For a view of cognition that denies RTM, cf. van Gelder 1995 At the heart of RTM is the postulation of a language of thought: an infinite set of ‘mental representations’ which function both as the immediate objects of propositional attitudes and as the domains of mental processes. Definition of RTM: 1. For any organism O and any attitude A toward the proposition P, there is a psychological (computational/functional) relation R and a mental representation MR such that: MR means that P, and O has A iff O bears R to MR. To believe that P is to have a mental representation that means that P tokened in your head in the believing kind of way. 2. Mental processes are causal sequences of tokenings of mental representations. Two reasons for accepting RTM: First reason: “RTM underlies practically all current psychological research on mentation”. EXAMPLES:  SOAR, ACT-R  ‘infants extract algebra-like rules that represent relationships between placeholders (variables), such as “the first item X is the same as the third item Y”, or more generally, that “item I is the same as item J”.’ G. Marcus et al. ‘Rule learning by seven-month old infants’ Science 283 (1999).  “Physical contact between source and target results in the transfer of some effect or quality, which we call essence, from source to target”. P. Rozin and C. Nemeroff 2000, ‘Sympathetic magical thinking: the contagion and similarity “heuristics”’ Heuristics and biases  Y. Rottenstreich and A. Tversky, ‘Unpacking, repacking and anchoring: advances in support theory’ Psychological review 104 (1997). s(X) = the support for hypothesis X P(A, B) = the probability that A rather than B holds 6 )()( )( ),( BsAs As BAP   s(A)  s(A1  A2) = s(A1) + s(A2) Second reason: RTM also suggests why causal relations among propositional attitudes often respect semantic ones. Spelling out of second reason. The short story: RTM + computational theory of mind: Cognitive activity isn’t merely a causal sequence of mental representations, but that it’s the computational processing of symbols. The longer story:  Pretend that symbols in the language of thought have a syntax like the English sentences used to express them. Consider the causal powers of e.g., ‘Sally became happy’ so that we mirror its semantic properties: o “Sally became happy” has the general form of “[Person-name] [VP become+[Past tense] [Adjective]]” Some possible laws of thought: [note: These laws can be viewed as causal laws about physical changes in a person.] o [Person-name] [VP become+[Past tense] [Adjective]]  Somebody [VP become+ [Past tense] [Adjective]] o [Noun Phrase] [VP become+[Past tense] [Adjective]]  [Noun Phrase] [VP is+[Past tense] not [Adjective]] o [Noun Phrase] [VP become+[not-future tense] happy]  [Noun Phrase] [VP is+[Past tense] sad] Not laws of thought: o [Noun Phrase] [VP become+[Past tense] [Adjective]]  [Noun Phrase] [VP become+[Past tense] [a Republican]] o [Noun Phrase] [VP become+[Past tense] [Adjective]]  Somebody [VP become+ [Past tense] [Adjective]] Thus, the (relevant part of the) mind is hypothesized to be a sort of machine M such that:  M’s operations consist entirely of transformations of symbols o The operations in question only alter the shapes of various symbols. o M is sensitive only to the syntactic properties of these symbols. 7  e.g., the mental representation that I am good and you are bad contains as a subcomponent the mental representation that I am good.  Without LOT, there may be causal connections between the mental representations of P and (P & Q), but the one needn’t be a constituent of the other. (This is a crucial issue.) Three Arguments for LOT Background Assumption: intentionality should not figure into the causal laws: I suppose that sooner or later the physicists will complete the catalogue they’ve been compiling of the ultimate and irreducible properties of things. When they do, the likes of spin, charm, and charge will perhaps appear on their list. But aboutness surely won’t; intentionality simply doesn’t go that deep….If aboutness is real, it must be really something else. (Psychosemantics, 98) Argument 1 (methodological) (P) Suppose: c1 causes e1, c2 causes e2, and c3 causes e1 and e2, Then it is ceteris paribus reasonable to hold that c3 is constituted by c1 and c2. (P) minimizes unexplained coincidences  e.g., the way to get an e1 type effect is either by c1 or by c3 (or by c4, which causes e1 and e4, or by c5, etc.). So let: c1 = having the state of believing that corn is good. c2 = having the state of believing that cabbage is bad. c3 = having the state of believing that corn is good and cabbage is bad. e1 = the cognitive and behavioral effects of believing that corn is good. e2 = the cognitive and behavioral effects of believing that cabbage is bad. Then by (P) it is ceteris paribus reasonable to hold that c3 is constituted by c1 and c2. Argument 2: Psychological processes Psychological theories frequently advert to mental representations.  this way of talking can’t just be translated away, because mental representations not only used to specify the content of intentional states, they’re also used to specify how various mental processes work. 10 o Various aspects and intermediate states of these processes have no intentional interpretation, so the process can’t be reduced to a sequence of intentional states.  The appeal to mental representations is part of a computational theory of how one mental representation is generated from another. o Mary sang a song in the church loudly o Sue kissed Ben in the barn Stephanie o The horse raced past the barn fell o While Mary was mending the socks fell off her lap. o The rat the cat the dog chased bit died. Argument 3: Productivity and systematicity Productivity: You can believe that John is happy John and Jane are happy John and Jane and Bill are either happy or silly, etc. You can believe infinitely many distinct things.  This is easily explained if our abilities to believe have a combinatorial semantics of the sort LOT proposes.  But productivity is an empirical claim that requires a performance/competence distinction. Systematicity: If you understand the sentence “John loves Mary”, then you also understand the sentence “Mary loves John”, etc.  Systematicity, like productivity, can be easily explained if LOT is true.  Systematicity does not require a performance/competence distinction. But how do we know cognitive capacities are systematic?  They must be at least as systematic as language, because language functions to express thoughts. 11