A blog about problems in the field of psychology and attempts to fix them.

Thursday, August 18, 2011

The Myth of Knowledge

Spurred by Sabrina's comment on my first post, and some of the things I have been writing about on Gary's blog "Minds and Brains", I wanted to talk a little bit about the Myth of Knowledge. This is an intense vestige of dualism, and of enlightenment philosophy. The modern notion of Knowledge is a brilliant 18th and 19th century idea, which is just plain wrong. I'm a big fan of anachronistic ideas like this - the great chain of being, intelligence, etc. - but they are hypotheses that are false, and they are now interfering with our building a more coherent psychology.

The Myth of Knowledge 

Once upon a time it was believed that one of the most basic psychological kinds was 'knowledge', i.e., a person 'knowing' something. Well, not a person, but a mind. The body sits there, but the mind/soul/spirit knows things. Several hundred years of philosophy (from at least Descartes on) started with epistemology, i.e., started with a knower and with things known. But what is it to know something?

I'm not sure how to explain the traditional understanding of knowledge except by example. So, for an easy one, let's take driving a car. What do we mean when we say that someone 'knows how to drive a car'? We mean a) that the person can, in fact, drive a car, b) that the person could describe to us how to drive a car, c) that the person could teach someone how to drive a car, d) that the person could tell us what they are doing while they drive a car, e) that the person can mental represent the driving process while not driving, e) that the person has an accurate conceptual schema of driving. Probably I could list a few other things, but that's enough for now. The 'myth' of knowledge is that those different abilities are necessarily combined together. It is an insidious myth.

What's Wrong With This?
Just to point out a few problems with this myth: Does a person with no legs, who successfully coaches football punters 'know' how to kick a football? Does a person who is a successful therapist, but cannot accurately explain what about his sessions helps patients 'know' how to do therapy? Does a person who expertly drives a car, but cannot explain the optical invariants that they use to brake properly 'know' how to drive? Does a person who wins fencing championships 'know' how to fence, even if they cannot teach others? Does a successful couples therapist who never manages to keep a healthy relationship 'know' how to have a good relationship? Does a person who successfully imagines punching through a board, but never manages to actually do so 'know' how to punch through a board? Does a car mechanic in a coma 'know' how to fix cars? Does an Alzheimer's patient who often forgets his children's still 'know' who they are during the bad spells?

The problems are so deep that they cannot be solved by mumbling about 'implicit' and 'explicit' probes knowledge, 'implicit' and 'explicit' types of knowledge, or even 'implicit' and 'explicit' knowledge systems in the brain. The notion of 'knowledge' that reached its peak 200 years ago is simply wrong, it was a perfectly good hypothesis, but it is wrong. There is now plenty of evidence that the ability to do something, the ability to talk about doing something, and the ability to think about doing something, have no need to be intertwined.

Because I am a behaviorist-leaning kind of guy, I would additionally point out that when behavior, talking, and thinking come into conflict, behavior wins. In my article trying to connect ecological and social psychology, I used an example out of Aikido, the martial art that prefers not to hurt people unnecessarily. Indulging in horrible generalizations: In the Western cultures - steeped in dualism and the myth of knowledge - we thinking that 'knowing' is about 'thinking', but in Eastern cultures this is not so. In Aikido, one of your goals is to blend with your opponent's movements so you inflict minimal harm. Your goal is not to think about blending, not be able to explain how to blend, nor to be able to accurately imagine blending, rather your goal is to actually blend when the time comes. A person 'knows' how to blend when they do it without thinking, and regardless of whether they can teach how to blend or explain what they did after the fact. (By the way, that article is part of a 7 article discussion, including my latest addition now available online.)

Another Example
In parting, one of the reasons this was on my mind is because the issue of 'discrimination' was brought up on one of the other blogs. My dissertation was on infant object-permanence abilities from an ecological perspective. Unintentionally, I stumbled into the mess that is the infant-looking-time literature. One of my conclusions while trying to make sense of my data is that we get into trouble when we say that an animal 'discriminates', but we do not say the behavior with which they discriminate. We often don't feel the need to say the behavior because we have bought the Myth of Knowledge, we believe that 'discriminating' is a mental activity that can be connected with any behavior... but it is not. Discriminating is a bodily act, and even if it never manifests in what we would call 'overt' behavior, all discrimination comes in the form of a particular discriminating response on the part of the body. When we fail to specify the behavior, we have had an incomplete thought. Much confusion in psychology would be avoided if instead of saying that 'a rat' discriminated between two tones, we said that the rat's 'bar pressing' discriminated between two tones. Similarly, we should not say that 'an infant' failed to discriminate two ways in which objects went out of sight, but that their 'reaching behavior' failed to discriminate ways of going out of sight.

The question of whether an infant 'knows' an object is still present after it has gone out of sight will continue to produce no end of the worst kinds of arm-chair philosophizing, and is impossible to investigate. Determining the conditions under which infants 'behaviorally discriminate situations in which an out of sight object is present from situations in which an out of sight object is absent' presents challenges, but is entirely tractable.

Going Back
To go back to the beginning, we would do well to avoid using the word 'knowledge' or to use it sparingly and carefully. Instead, we should talk about the observable abilities that make us think someone 'knows' something. The legless coach can explain how to kick a football. The successful therapist can help other people build stronger relationships. The expert driver can drive a car. The fencing champion can outmatch a wide variety of opponents. Same observable happenings described, but with none of the nasty problems created by the implication of 'knowledge.'


  1. This is a good post, Eric. For me, in sociology of science and technology studies, there is a big question: What would STS look like without "knowledge" as a crutch during analysis?

    Can to hazard a guess/comment?


  2. Cross posted at: http://www.installingorder.org/what-does-the-knowledge-myth-mean-for-skatsts

  3. Interesting. It seems though it isn't knowledge that is core to your discussion but our 'knowing' - the human experience. What makes some things 'compiled' knowledge (in our experience, like riding a bike) and some things expressible knowledge (which we can teach, like nuclear physics)? Is it a physiological problem between muscle memory, the spinal cord, and the thing we call the brain?

  4. Nicholas,
    I'm not sure how "knowledge" is used as a crutch in STS, but I'm willing to hazard a guess...

    The first thing you would need to do is to stop talking about what people know, and start talking about what people know based on some measure. So instead of talking about 'knowledge of how to fix a car,' you talk about 'knowledge - as measured by how long it takes someone to actually fix a car.' Of course, at that point you realize that the first part is superfluous. You can just say that you are interested in 'how long it takes someone to build a car.'

    Now, when you do that, people will start to object, they will say "But that's not what I'm interested in, I'm interested in what the person KNOWS!" At that point, you have a problem, because either they are confused about what they are interested in, or they are interested in something that does not exist.

  5. Robert (Anonymous),
    Well, here is the core of the problem. Your final question packs a mighty assumption: That if there is a difference between the things we can do and the things we can explain how to do, there must be some underlying physiological problem... as if it would be the natural state for us to be able to do everything we can explain and visa versa. But I would suggest that such a fortuitous convergence of ability is exceptional, rather than typical.

    Let's say you have one person who can ride a bike, but can't explain or teach how to ride a bike, and another person who can explain and teach how to ride a bike, but can't actually ride herself. We are obviously accurate if we say that one successfully rides, the other successfully explains. What do we add to that obviously accurate statement when we talk about them as having two types of 'knowledge', or of expressing their 'knowledge' in two different ways? All we add is a porting in of the Knowledge Myth. We add the assumption that all those abilities mentioned in my first post are most naturally found in a tightly connected fashion. One consequence of this, is that we are sent looking for the abnormality responsible for this particular disconnect.

    Without the Myth of Knowledge, there is no mystery, no need to posit an abnormality. Without the Myth, we are free to simply talk about what people do and how they learn to do those things.

  6. One way of making the distinction between "knowing-how" and "knowing-that" suggested by "anonymous" is to note that the behavior in justifying them can be different. There's no better way for me to justify my claim that "I know a little Aikido" than to accept an invitation onto the mat. However, the best I can do in justifying my claim "I know the capital of Australia is Canberra" is to say that a good Australian friend told me it is, and to point out the interesting fact that if you take a map of Australia and flip it up into the northern hemisphere overlapping the US, Canberra, ACT (Australian Capital Territory) falls more-or-less on Washington, DC.

    Nonetheless, I agree with Eric that both types of knowledge are justified by behavior. It's just that one type of behavior - using language - gets a lot of special attention. And I agree that in technical conversation it's probably a good idea to be careful in attributing the ability to "know".

    Editorial note: despite Sabrina's often being totally justified in doing so, she has never "spurned" me in her comments, although they have often "spurred" me to further reflection. (:>)

  7. Hmmm... I think things get much more complicated when you are making claims about your own behavior. Tradition has it that the question of knowledge starts with the question of self-knowledge. I suspect, however, that psychology would have made better progressing by starting with evaluating the knowledge of others.

    In that context, you can test my knowledge by asking if I can tell me what the capital of Australia is, and having me reply. Note that in so doing, you don't need to talk about what I 'know', you can juts worry about what I can 'tell you' under various conditions.

    Trying to figure out how to justify the things that you tell yourself when you ask yourself questions, layers much complexity onto the problem.


    As for the editorial note.... I have no idea what you are talking about, you should check again ;- ) (Odd that I can edit the original post, but not the comments.)

  8. My view of "knowing-that" is Sellarsian:

    [I]n characterizing ... a state as that of knowing ... we are placing it in the logical space of reasons, of ... being able to justify what one says.

    (Sect 36 of "Empiricism and Phil of Mind")

    Ie, in asserting "I know-that X" I am asserting the ability to give reasons to my peers - not to myself - justifying my belief that "X" is true. Ie, just as you suggest, I am inviting my peers to "evaluat[e] the knowledge [claims] of others" - in this case "the others" being me.

    I was just extending that idea to "knowing-how", where justification by performance presumably beats justification by logical argument, ie, the "giving of reasons".

    I think this is consistent with your position and just thought a little different perspective might help in addressing anonymous's queries.

    As for "spurred", I apparently misread it at first encounter as "spurned" but neglected to go back and reread it. If you didn't change it, that must be the case because it's OK now. Sorry.

  9. Two brief comments:

    I used to really enjoy talking to Esther Thelen about things like this. She was always fairly measured in her writing, but the thrust of that writing was also driven by the very strong view she was happy to tell you in person: that she thought the performance-competence distinction was utterly incoherent. Frankly, most of the developmental work out of IU was about proving this, in some way or another. Esther was working on habituation when she got sick, and was busy demonstrating that what infants 'know' is entirely a function of what you are asking them to do :)

    Second, a minor note: I read the abstracts for the paper you mentioned, Eric, and looked at the Castro abstract: dude, they were way harsh.

  10. I had one or two amazing conversations with Ester before her death. You are write, as cutting as her works were, she was even more exciting in person. I am still traumatized by a talk I heard at SRCD not long after her death, where one of her modeling collaborators had added a variable to the model that stood in for "memory strength". Uhg! The whole point was that you could explain the phenomenon as a system-level result without needing cognitive variables.... but he was more interested in the modeling itself than the implications for psychology.

    Yeah, Castro and Lafuente were bizarrely harsh. My apparent heresies were many and varied. While I tried to stick to substance in my reply, I couldn't resist a couple digs of my own.