What do you think?
Rate this book
Explaining Behavior: Reasons in a World of Causes
Why do human beings move? In this lucid portrayal of human behavior, Fred Dretske provides an original account of the way reasons function in the causal explanation of behavior. Biological science investigates what makes our bodies move in the way they do. Psychology is interested in why persons—agents with reasons—move in the way they do. Dretske attempts to reconcile these different points of view by showing how reasons operate in a world of causes. He reveals in detail how the character of our inner states—what we believe, desire, and intend—determines what we do.
180 pages, Paperback
First published January 1, 1988
5 people are currently reading
101 people want to read
About the author
Fred I. Dretske
11 books18 followersFrederick Irwin Dretske is a philosopher noted for his contributions to epistemology and the philosophy of mind. Recent work centers on conscious experience and self-knowledge. He was awarded the Jean Nicod Prize in 1994. Dretske received his Ph.D from The University of Minnesota and taught for a number of years at the University of Wisconsin-Madison before moving to Stanford University. After retiring from Stanford, he moved to Duke University where he is now research professor of Philosophy.
Dretske holds externalist views about the mind, and thus he tries in various writings to show that by means of mere introspection one actually learns about his own mind less than might be expected.
Dretske holds externalist views about the mind, and thus he tries in various writings to show that by means of mere introspection one actually learns about his own mind less than might be expected.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 2 of 2 reviews
January 25, 2016
(I forgot to mark this book as read, despite having finished it a while ago.)
I was debating for a while whether to rate this book or not, but in the end I decided to do it and to give it one star. The book has way too much philosophy and speculation and way too little (/neuro)science. Especially absent from the book is any sort of indication of how the theory being developed might be used to predict anything empirically, rather than just being a sort of explanatory model that is really when you come right down to it just about what you might decide to call different aspects of behaviour and motivation. The author notes on page 126 that "There is, of course, no point in arguing about words", but when I read that comment I couldn't help thinking that if he really thinks that, he must also think that his book is completely pointless, because that's really pretty much all this book is doing - arguing about words. When is something a belief, and when is it only properly to be considered a desire? Who gives a crap? Why is this distinction of any relevance to anything? I'm sure the author would argue that if only I'd read the book more carefully I would be able to better appreciate why this distinction is important - but I figure a lot of what he's doing in this book is just a bunch of useless word games and definitions. Some of the implicit assumptions in the book used to justify the distinctions made seemed to me both arbitrary and problematic.
The book contains a lot of speculation, a lot of what seemed to me but more or less irrelevant distinctions and subtleties of no relevance to anything. Most of the book is about how you might decide to classify different types of behaviours and behavioural motivations, but why you would want to use the classification system on offer by the author is basically blowing in the wind. It felt to me a bit like reading an economist who hasn't learned any math yet - let's define x this way, let's define y that way, but let's not develop any sort of testable model but instead just talk about these things (and z, and q, and...) until the reader gets bored enough to close the book. I did read the book to the end, but that was mostly because I was deeply confused as to where the author was heading. When I finished I concluded that he probably didn't know either.
Not all the observations, distinctions, and details included in the book are worthless, but all in all I thought this was a weak book which didn't really go anywhere. A disappointing read.
I was debating for a while whether to rate this book or not, but in the end I decided to do it and to give it one star. The book has way too much philosophy and speculation and way too little (/neuro)science. Especially absent from the book is any sort of indication of how the theory being developed might be used to predict anything empirically, rather than just being a sort of explanatory model that is really when you come right down to it just about what you might decide to call different aspects of behaviour and motivation. The author notes on page 126 that "There is, of course, no point in arguing about words", but when I read that comment I couldn't help thinking that if he really thinks that, he must also think that his book is completely pointless, because that's really pretty much all this book is doing - arguing about words. When is something a belief, and when is it only properly to be considered a desire? Who gives a crap? Why is this distinction of any relevance to anything? I'm sure the author would argue that if only I'd read the book more carefully I would be able to better appreciate why this distinction is important - but I figure a lot of what he's doing in this book is just a bunch of useless word games and definitions. Some of the implicit assumptions in the book used to justify the distinctions made seemed to me both arbitrary and problematic.
The book contains a lot of speculation, a lot of what seemed to me but more or less irrelevant distinctions and subtleties of no relevance to anything. Most of the book is about how you might decide to classify different types of behaviours and behavioural motivations, but why you would want to use the classification system on offer by the author is basically blowing in the wind. It felt to me a bit like reading an economist who hasn't learned any math yet - let's define x this way, let's define y that way, but let's not develop any sort of testable model but instead just talk about these things (and z, and q, and...) until the reader gets bored enough to close the book. I did read the book to the end, but that was mostly because I was deeply confused as to where the author was heading. When I finished I concluded that he probably didn't know either.
Not all the observations, distinctions, and details included in the book are worthless, but all in all I thought this was a weak book which didn't really go anywhere. A disappointing read.
February 14, 2026
The reading experience is: a dumb guy wrestling to work out the implications of one interesting idea he has overheard in a discussion with the cool kids. The book is an odd mix of disguised ordinary language philosophy (lots of distinctions between things that are not clearly different, like 'goal-directed' vs 'goal-intended', state with meaning in and to an organism, etc.) and examples of learning theory, interspersed with poorly assimilated concepts from actual theoretical biology (Lewontin's distinction between selectional and developmental explanations is barely recognizable). It's repetitive, often painfully boring to read. And the central claim, that learning provides organisms with reasons for acting in a way that evolution does not, is not argued for, although Dretske clearly thinks it is. I have to side with Dennett on this by the way. Again.
Some notes:
Ch. 1, pp. 23-24, the problem of singling out ‘the’ primary cause of behaviour. There are a number of background conditions that must be present for something regarded as ‘the’ cause to have its effect. Dretske writes: “I have nothing particularly original to say about how one identifies the cause of something from among the many events and conditions on which it depends. It seems fairly clear that this selection is often responsive to the purposes and interests of the one doing the describing… (A few examples later). Causal conditions may be out there in the world, independent of our purposes and interests, but something’s status as the cause is, it seems, in the eye of the beholder.” (p. 24).
First, I don’t think that the selection is in the eye of the beholder: it has to do with the kind of relation the organism under study has with those external background and supporting conditions. Second, I don’t think “something’s status as the cause is… in the eye of the beholder”, because we can have explanatory models that single out (‘the’) causes as difference-makers in a set of non-arbitrarily defined conditions, depending, for instance, on how often they co-occur, so as to rank-order them by causal relevance, and the explanations appealing to them according to their corresponding explanatory depth (see Woodward 2005, 2021, and Strevens 2008).
An example of background conditions given in ch. 2 is: the thermostat’s being wired in such and such ways causes, or is part of what causes, its indicating such and such temperatures. Distinction between structuring and triggering causes.
Structuring cause/triggering cause
Ch. 4: The fact that C indicates F explains why C causes M: this is how we make meaning matter. Then we make nature select C for indicating F and causing M in virtue thereof. But that doesn’t explain how organisms come to have the machinery required. Only once the machinery, or the variation, exists, is selection possible. Dretske says that a creature has such machinery because its ancestors had it. Selectional vs developmental explanations (Lewontin 1983): I have only martini-imbibed friends, but is it a fact about each of my friends (development), or a fact about me (selection) (Dretske’s example)? But then, doesn’t the machinery constrain the range or kind of indicatings available for selection to choose from? Don’t we have to explain how it comes to be that the creature has the very resources to indicate F? This was exactly Lewontin’s point by the way: development generates the variation on which selection acts. Mother Nature doesn’t just wake up one day and say: why don’t I ‘hard-wire’ the indicatings of F by Cs to the causings of Ms (by Cs)? It’s like waking up one morning: why don’t I buy a house in Florida? Well, because I don’t have the money.
But then, on p. 95, Dretske goes on to say that even though natural selection can make it so that Cs indicate Fs and cause Ms, it cannot make it so that Cs cause Ms in virtue of their indicating Fs. I don’t understand why. Dretske thinks that only individual learning can make meaning matter in this way, but he does not say why (Dennett 1992 is also puzzled by this, in case you think I’m misreading Dretske). Dretske, p. 101, writes about the organism learning a capacity to “tell” that F occurs, but that sounds just like having F-indicators. I don’t see what difference it makes whether they are onto- or phylogenetically acquired. I agree, however, that organisms can gain something in having the capacity to assess circumstances as affording the release of a certain action pattern, or the use of a certain representation to promote a certain goal, but, roughly, the assessment must either be one indicator inhibiting another, or something different from indication. Probably if the circumstances are partly engineered by the organism itself, that helps the organism ‘knowing’ what the circumstances afford—but precisely because it doesn’t have to learn it.
Dretske writes that, prior to learning, the organism may do M when something happens, but not because it happens. Since the ‘when’ case is meant as designating a chance event (“the bird was just poking around”, p. 102), I don’t see the difference between the ‘because’ case and selection. Dretske later writes, in the conclusion of the chapter, of the difference between structures that mean something tout court and those that mean something to the organism (p. 107). Still not sure what this means. It looks like what Dretske wants is for the organism to reflect on his state, take a step back, and make a judgement: ‘Ooh, my F-detector just lit up! But wait, I am downtown, and there are no Fs here, only F*s—I always get confused. Phew, that was close, but no need to pay attention.’
In the same spirit, the difference between the case in which the organism takes appropriate action to the presence of F by having F-detectors, and the case in which it does so by having G-detectors, where Gs and Fs co-occur reliably enough, is not clear to me. Dretske writes that, in the latter case, the organism’s behaviour is dependent on the correlation’s continued existence. That’s true, of course, but likewise, in the former case, the organism’s behaviour is dependent on the continued existence of the correlation between the activation of its F-detectors and the presence of Fs, rather than, say, F*s—which look a lot like Fs, but ain’t. I genuinely see no difference at all between using world-world correlations and mind-world correlations, between using Gs to indicate Fs and using one’s own brain states as indicators of Gs or Fs, or one via the other, or what you like. ‘Ooh, but everyone knows you don’t use your own thoughts! What about my precious aboriginal intentionality?’ Yeah right
Some notes:
Ch. 1, pp. 23-24, the problem of singling out ‘the’ primary cause of behaviour. There are a number of background conditions that must be present for something regarded as ‘the’ cause to have its effect. Dretske writes: “I have nothing particularly original to say about how one identifies the cause of something from among the many events and conditions on which it depends. It seems fairly clear that this selection is often responsive to the purposes and interests of the one doing the describing… (A few examples later). Causal conditions may be out there in the world, independent of our purposes and interests, but something’s status as the cause is, it seems, in the eye of the beholder.” (p. 24).
First, I don’t think that the selection is in the eye of the beholder: it has to do with the kind of relation the organism under study has with those external background and supporting conditions. Second, I don’t think “something’s status as the cause is… in the eye of the beholder”, because we can have explanatory models that single out (‘the’) causes as difference-makers in a set of non-arbitrarily defined conditions, depending, for instance, on how often they co-occur, so as to rank-order them by causal relevance, and the explanations appealing to them according to their corresponding explanatory depth (see Woodward 2005, 2021, and Strevens 2008).
An example of background conditions given in ch. 2 is: the thermostat’s being wired in such and such ways causes, or is part of what causes, its indicating such and such temperatures. Distinction between structuring and triggering causes.
Structuring cause/triggering cause
Ch. 4: The fact that C indicates F explains why C causes M: this is how we make meaning matter. Then we make nature select C for indicating F and causing M in virtue thereof. But that doesn’t explain how organisms come to have the machinery required. Only once the machinery, or the variation, exists, is selection possible. Dretske says that a creature has such machinery because its ancestors had it. Selectional vs developmental explanations (Lewontin 1983): I have only martini-imbibed friends, but is it a fact about each of my friends (development), or a fact about me (selection) (Dretske’s example)? But then, doesn’t the machinery constrain the range or kind of indicatings available for selection to choose from? Don’t we have to explain how it comes to be that the creature has the very resources to indicate F? This was exactly Lewontin’s point by the way: development generates the variation on which selection acts. Mother Nature doesn’t just wake up one day and say: why don’t I ‘hard-wire’ the indicatings of F by Cs to the causings of Ms (by Cs)? It’s like waking up one morning: why don’t I buy a house in Florida? Well, because I don’t have the money.
But then, on p. 95, Dretske goes on to say that even though natural selection can make it so that Cs indicate Fs and cause Ms, it cannot make it so that Cs cause Ms in virtue of their indicating Fs. I don’t understand why. Dretske thinks that only individual learning can make meaning matter in this way, but he does not say why (Dennett 1992 is also puzzled by this, in case you think I’m misreading Dretske). Dretske, p. 101, writes about the organism learning a capacity to “tell” that F occurs, but that sounds just like having F-indicators. I don’t see what difference it makes whether they are onto- or phylogenetically acquired. I agree, however, that organisms can gain something in having the capacity to assess circumstances as affording the release of a certain action pattern, or the use of a certain representation to promote a certain goal, but, roughly, the assessment must either be one indicator inhibiting another, or something different from indication. Probably if the circumstances are partly engineered by the organism itself, that helps the organism ‘knowing’ what the circumstances afford—but precisely because it doesn’t have to learn it.
Dretske writes that, prior to learning, the organism may do M when something happens, but not because it happens. Since the ‘when’ case is meant as designating a chance event (“the bird was just poking around”, p. 102), I don’t see the difference between the ‘because’ case and selection. Dretske later writes, in the conclusion of the chapter, of the difference between structures that mean something tout court and those that mean something to the organism (p. 107). Still not sure what this means. It looks like what Dretske wants is for the organism to reflect on his state, take a step back, and make a judgement: ‘Ooh, my F-detector just lit up! But wait, I am downtown, and there are no Fs here, only F*s—I always get confused. Phew, that was close, but no need to pay attention.’
In the same spirit, the difference between the case in which the organism takes appropriate action to the presence of F by having F-detectors, and the case in which it does so by having G-detectors, where Gs and Fs co-occur reliably enough, is not clear to me. Dretske writes that, in the latter case, the organism’s behaviour is dependent on the correlation’s continued existence. That’s true, of course, but likewise, in the former case, the organism’s behaviour is dependent on the continued existence of the correlation between the activation of its F-detectors and the presence of Fs, rather than, say, F*s—which look a lot like Fs, but ain’t. I genuinely see no difference at all between using world-world correlations and mind-world correlations, between using Gs to indicate Fs and using one’s own brain states as indicators of Gs or Fs, or one via the other, or what you like. ‘Ooh, but everyone knows you don’t use your own thoughts! What about my precious aboriginal intentionality?’ Yeah right
Displaying 1 - 2 of 2 reviews