Three Kinds of Intentional Psychology (1975)
Daniel Dennett
Heil, pp. 298-320
I. Folk Psychology
(298)
A. Two kinds of explanation
1. Suppose you are asked, “What do all magnets have in common?” There are two ways to answer this question:
a. definitional, conceptual: All attract iron (folk physics)
b. theoretical, causal: All have a certain microstructure that explains their ability to attract iron
2. or again, “What do all elements with the same valence have in common?” (298-299)
a. they combine with other elements in same integral ratios
b. they have a particular sort of distribution of electrons in their orbits
3. similarly, if one asks, “What feature do we both have in common if both of us believe that cats eat fish?” there are two ways to answer this question, as well (298)
B. Logical Behaviorism (299)
1. Ryle favors purely conceptual answers to such questions (299)
2. Now although he was wrong to think that such answers were in conflict with causal explanations, he was right in thinking that conceptual answers were not to be given at the micro-reductive, theoretical level (q.v.)
3. One could argue that some concepts, such as “genuine Winston Churchill autograph” or “intelligent action,” have an essential causal component
4. Nevertheless, Ryle could argue that even if every instance of intelligent action has a causal explanation, exactly how it is caused is irrelevant to what makes it intelligent
5. Dennett compares behaviorism in psychology to the use of the concept of valence in chemistry before it could be explained by physics (300)
C. Beliefs
1. Getting back to the question about what two people with the same belief have in common, many feel that this will ultimately be explained in terms of neurophysiology, much as physics explained valence
2. However, while valence is a technical term in chemistry, terms like “belief” and “desire” are part of our ordinary language
3. Thus, we should turn to “folk psychology” to see what sorts of things we are expected to explain (300)
a. When we learn to use words like “belief” and “desire,” no one really tells us what these things are
b. We don’t learn folk psychology as an explicit theory; there are no textbooks in it (300-301)
c. rather, we learn to use it the way we learn to speak our native language (300-301, q.v.)
4. For folk-psychology, roughly, beliefs are information-bearing states that arise from perception and that, together with desires, lead to action (301, q.v.)
5. But that still leaves many questions unanswered, such as whether animals have beliefs, whether beliefs must be expressed in language, etc. (q.v.)
D. Dennett gives two reasons for being interested in detailed answers to these questions about folk psychology
1. Folk psychology exists as a phenomenon like religions, myths, or styles of clothing that make up part of the world in which we live and which sciences like anthropology may investigate
2. But also, like the folk physics of magnets and unlike astrology, folk psychology seems to be true and thus a candidate to be incorporated somehow in the body of scientific knowledge (301)
a. The first step here would be to critically examine folk psychology, separating the true from the false
b. To do this, we need to determine which parts of it are needed for successful explanations and predictions
c. If it turned out to be a bad theory, we could discard it along with the beliefs and desires it postulates, but we’d have to replace it with something else (301-302)
d. Dennett adds that the new theory would have to explain the predictive successes of the old, but that’s going a bit too far (302)
E. How good is folk psychology? (302)
1. Sometimes it is unable to explain some behaviors in terms of beliefs and desires
2. But it also has many strengths
a. Use it in many areas of daily living. E.g. driving.
b. Can be extended into many areas – gives example of interpreting a movie. (For that matter, the writers, directors, and actors use it to produce the movie!)
c. Children pick it up easily
d. We are able to use it without knowing anything about how the brain works (302)
F. Normative Character of Folk Psychology
1. Unlike the physical sciences, when folk psychology explains an action it defends it as reasonable under the circumstances (q.v.)
2. That is, it treats the agent as rational
3. For this reason, Dennett regards folk psychology as an idealizing, abstract, instrumentalist method of interpreting and predicting behavior
G. Folk psychology as an idealizing method
1. That is, in folk psychology we treat each other as “intentional systems” whose behavior can be predicted by attributing beliefs, desires, and rationality to them in the following way: (302-303, q.v.)
a. An agent’s beliefs are the ones it should have given its perceptual abilities, environment, etc. That is, they are largely true, and any false beliefs it may have require an explanation. (303)
b. Its desires are the ones it ought to have, given its biology, etc. Abnormal desires require an explanation.
c. Its behavior consists of actions that would be rational given its beliefs and desires
2. Of course, our perceptual systems are not perfect (303)
a. evolution provides “only a passable jury-rig”
b. these short-cuts may yield false beliefs in abnormal situations
3. similarly, we’re not perfectly rational in our thinking, either (303-304)
a. But folk psychology works very well because we are rational enough (304)
b. thus we treat each other as if we were rational agents
c. This assumption that we are rational allows us to interpret ourselves as intentionally acting on beliefs and desires and to predict a lot of behavior
4. Dennett then is claiming that folk psychology can be seen as a kind of logical behaviorism (q.v.)
a. that is, that it allows us to answer the question what it means to say that someone has a certain belief by saying that it means that a person will act certain ways under certain conditions
b. What ways and what conditions? Those that would be rational given its beliefs and desires
5. how rational are we? (305)
a. if we are designed by natural selection, we are nothing more than “a bag of tricks, patched together by a satisficing Nature,” that is, no better than we need to be to get by (305, q.v.)
b. sometimes Nature demands the opposite of what logic dictates, such as that we jump to conclusions
c. some psychological research shows us to be very irrational
1.) not only jumping to conclusions, but biased by all sorts of irrelevant factors.
2.) However, Dennett argues that this behavior is elicited under artificial conditions.
d. a better indicator of how rational we are may be the fact that even a child can do better than Schank’s scripts at interpreting restaurant stories
H. Folk psychology as an abstract method (306)
1. It’s abstract in the sense that beliefs and desires need not be considered actual intervening psychological states
2. Rather, the role of the concept of belief is like that of the concept of a center of gravity in physics, allowing you to make certain calculations
I. Folk psychology as an instrumentalist method
1. That is, useful for predictions, etc.
2. Beliefs and desires are real in the same sense that centers of gravity or the Earth’s equator are real
3. Reichenbach distinguishes two kinds of referents for theoretical terms: (306)
a. Illata: postulated theoretical entities.
b. Abstracta: conventions that allow calculations, etc.
4. For Dennett, beliefs and desires are abstracta
J. Dennett concedes that many philosophers, such as Jerry Fodor, hold just the opposite point of view, according to which beliefs and desires are real things governed by causal laws that we can learn from experience (quotation, 306 q.v.)
K. But Dennett makes two responses:
1. First, he admits that when we interpret other people’s behaviors, we do rely on rough generalizations based on experience (306-307)
a. We don’t simply logically deduce what their beliefs and desires ought to be from the goal of survival
b. For example, we learn by experience that some people desire things that are not good for them, like cigarettes (307)
c. It is because people draw on their experiences with people like themselves that it is hard for them to interpret the behavior of foreigners
d. Nevertheless, Dennett believes that this empirical knowledge is used in a way that assumes a normative foundation (307)
2. Second, the issue for Dennett is not what folk psychology is but what it ought to be
a. Hence, it would be beside the point to argue against Dennett that people take beliefs and desires to be concrete, causally interacting illata; the question is whether they should
b. Dennett also argues that there is nothing concrete about the way in which we attribute beliefs to people
1.) Example of people learning in different ways that Jacques shot his uncle dead in Trafalgar Square – the detective who caught him, the guys who read it in the paper, etc.
2.) We would say that they all believe a Frenchman committed murder in Trafalgar Square, but that does not imply that they said this proposition to themselves, or thought it
3.) but more importantly, by imputing this common belief to some people, we are not thereby postulating that there is some similarly structured object in the heads of all of them (308)
L. Dennett affirms that our ordinary folk psychology notion of belief is actually somewhere between being an illata and being an abstracta. (308)
M. Also argues that “belief” is unappealing as a scientific concept
1. Compares it to Anaxagoras’s ancient notion of “seeds,” according to which flour, water, and yeast contain bread seeds, bread contains flesh, blood, and bone seeds, and so on to infinity
2. Similarly, there seems to be no way to avoid attributing an infinity of beliefs to people
3. One way to solve this problem might be to say that we have
a. only a finite number of “core” beliefs that are explicitly stored. These are illata.
b. The rest are all virtual, implicit beliefs, merely abstracta
4. But Dennett argues:
a. We could very well be wrong about what our core beliefs are. It could be that all of our beliefs are virtual and that what was explicitly stored were procedures, memory addresses, etc. (q.v.)
b. Although we could say that it is the core elements that actually play the causal role, there seems to be good reasons to hold on to the notion that beliefs can cause actions, blushes, verbal slips, etc. (309)
E.g. “Her belief that John knew her secret made her blush.”
N. So Dennett makes a suggestion:
1. We can avoid debate over whether intentional explanations are causal explanations by assigning causal role to core elements and regarding beliefs as virtual
2. We split apart the two parts of folk psychology and create two separate theories: (310)
a. One abstract, idealizing, instrumentalistic: pure intentional system theory
b. The other a concrete, micro-theoretical account of the actual realization of these intentional systems: sub-personal cognitive psychology
3. We can then ask whether the former can be reduced to the latter
II. Intentional System Theory
A. similar to and overlapping with disciplines like game theory, decision theory
B. Is holistic as well as abstract and normative: that is, it deals with behavior of the whole organism or system
1. Beliefs and desires are attributed to the whole system, not the parts (310)
2. The individual realizations of these systems are “black boxed”
C. The theory explains the production of new beliefs and desires from old ones
1. It might seem that it is dealing with actual natural processes
2. But this would be like mistaking a parallelogram of forces with actual mechanical links
D. Intentional system theory deals with performance specifications of believers (311)
1. That is, it asks such questions as what an organism’s capabilities must be to survive in a certain environment (q.v.)
2. But is silent on how those abilities are realized
3. For example,
a. it might say that in a certain environment, birds and bats will need to recognize and avoid a certain poisonous insect
b. Given the difference between bats and birds, there is no reason to think that the biological realization will be the same in both
4. So Dennett does not think that intentional systems theory will reduce to the neurophysiological level (311)
E. One may object that the attribution of beliefs and desires seems to work and it can’t just be magic (312)
F. Dennett concedes that only physical systems with a great deal of complexity could realize these intentional systems, but that given the differences in evolutionary history among animals and personal history among people, there is no reason to think we all do it the same way.
G. N.B. Dennett’s philosophy thus appears to be a kind of functionalism
III. Sub-Personal Cognitive Psychology
(313)
A. At first, the job of sub-personal cognitive psychology is to explain something that seems inexplicable
1. The brain, from the point of view of Intentional system theory and evolutionary biology, appears to be a semantic engine, that is, to discriminate among its inputs in accordance with what they mean and act accordingly
2. But from the point of view of neurophysiology – and plain common sense – the brain appears to be a syntactic engine
3. How does the brain get semantics out of syntax?
B. Dennett suggests that the brain is designed to approximate a solution to the problem, to mimic a semantic engine (313)
1. For instance, an animal may need to know when it has had enough food, but it will manage with a friction-in-the-throat-followed-by-stretched-stomach detector
2. This will work well enough in normal environments
3. That is, living things will settle for a system that seems to discriminate meanings by discriminating things that co-vary with meanings (314)
4. Evolution has provided us the means not only to do this but to get better at it through learning (314-315)
C. For Dennett, it is then the task of sub-personal cognitive psychology to propose and test models of the kinds of brain activity that allow us to do these things (315, q.v.)
1. Should have same strengths and weaknesses (315)
2. It is here that we will find good theoretical entities or illata
a. They may or may not resemble things like beliefs and desires
b. But we can sure that their labels will carry intentionality, that is, that they will have content or meaning
c. However, in order for these illata to carry these labels, the theorist will have to look outside the system to see what effects these produce in the environment (315)
d. thus progress in sub-personal cognitive psychology will blur the boundaries between it and intentional system theory
e. For Dennett, to study internal processing in total isolation from interaction with the environment is not really to do psychology at all (315-316)
D. Dennett quotes Friedman to the effect that cognitive psychology should begin with the social instead of the individual level (316)
E. Dennett agrees but emphasizes that the social is only a part of the macro-level, which has to do with organism-environment interaction generally
IV.
Conclusion (317)
A. Given these three kinds of psychology, what will reduce to what?
1. Will not find type-type identities of the sort that reductive materialists used to talk about
2. But can folk psychology be reduced to intentional system theory?
3. here, too, does not expect identities
4. Something like the belief that cats eat fish is a functional state that can be physically implemented in more than one way (317)
B. Turing Machine functionalism
1. On this philosophy, for two things to believe that cats eat fish is for both of them to share a Turing machine description according to which they are in the same logical state (q.v.)
2. But Dennett thinks this is too strong: there is no more reason to think that everyone who believes the same thing has the same program than to think that they all have the same kind of brain state
3. Nevertheless, machine functionalism suggests an analogy to Dennett (318)
a. Just as we may ask what makes two people have the same belief, we can ask what do two embodiments of a Turing machine have in common when they are in the same logical state?
b. In the latter case, we say that according to this particular Turing machine description, they are both in the same state (q.v.)
c. Similarly, we can say that two organisms or systems can be characterized as intentional systems and that according to this description, they are in the same state (318)
C. So, for Dennett, the claim that every mental phenomenon that is talked about in folk psychology is intentional-system-characterizable provides a sort of reduction of the mental
D. But then the final question is not whether intentional systems theory reduces to neurophysiology, but whether a physiological system could be legitimately interpreted as an intentional system (319, q.v.)
V. A realist’s worry : Whether you have certain beliefs, on Dennett’s account, seems to depend on someone else’s interpretation of your actions
Dennett would consider your beliefs as real in the same sense that your center of gravity is real.