Re: Introspection

From: HARNAD Stevan (
Date: Tue May 21 1996 - 18:16:27 BST

> From: "Whitehouse Chantal" <>
> Date: Mon, 20 May 1996 15:15:03 GMT
> Introspection is basically thinking about thinking. It's looking
> inside our own minds to try and work out how they, and their
> processes, work. For example, if asked how you remembered somebody's
> name you would probably try to work backwards- "I remembered the name
> because I pictured his face, I got his face from memory- how did I do
> that?" There are two main problems with using introspection to study
> the mind. Firstly we'd never know for sure if we knew about everthing
> that was going on inside our heads. There may be some vital
> unconscious process going on that we couldn't ever grasp hold of or
> comprehend. We may feel that we understand it all but this could just
> be a subjective feeling of having the question answered which would
> come from the mind.

That's right. Introspection gives the impression that there IS no
problem of explaining how the mind works. It gives you the impression
that you know how it works -- or that there is nothing to explain.

What is the critical test of whether anything you learn from
introspection REALLY explains, and REALLY explains how your mind works?
See if you can use it to get a model to do what you can do. A model is just
something that gets the same input that your mind does, does with it
what your introspection tells you your mind does, and then comes out with
the same output your mind comes out with. Your example of remembering
someone's name is as good as any: How would you use introspection to
design a model that could remember what you can remember, given the same
experience and the same inputs you get?

The answer is that the introspection doesn't help at all. SOMETHING in
there is successfully remembering the name, but introspection doesn't
give us a clue as to what it is or how it works: It just takes it for

It's incorrect, however, to say that these real structure and processes,
the ones we are not conscious of, that are doing the work for us, are
things we cannot grasp hold of or comprehend: We just can't grasp hold
of or comprehend them BY INTROSPECTION. Any more than we can introspect
the laws of physics or the principles of engineering. We need to do
experiments and modeling (we have to do cognitive psychology!), and,
mainly, we have to come up with good THEORIES of what those unconscious
processes might be. But once we find them, and confirm that they work,
there's no reason we can't grasp or comprehend them.

As to where theories come from... Here we must turn to the question of
creativity. But the answer is certainly not: introspection. Once you
do get a good theory, it will of course be in your mind. But it won't
have gotten there by your simply introspecting for it. You will have had to
think, and see things, and learn things, and a lot of that unknown,
unconscious processing will have to have gone on, before it hands to you
on a platter the winning theory (much the way it hands to you on a
platter the name of the person you were trying to recall). But
introspection can hardly take the credit for any of this.

> The second problem is that you don't know if what
> you, yourself, feel or the processes your mind uses are the same as
> everyone elses (or if anyone thinks anything at all). Every single
> person could have different thought mechanisms. It seems unlikely as
> our biological systems are all similar, but the point is that we
> can't be positive. Nothing can be proven with introspection.

That is the other-minds problem. Insofar as introspection is concerned,
the relevant point is that I cannot observe what is going on in anyone's
mind but my own. That means that my introspective "observations" are
unlike any other observations in science, because all other observations
can be repeated and confirmed by anyone. But I cannot repeat and confirm
YOUR experiences. I can't even repeat or confirm my own (e.g., think of
false-memory syndrome); I can only say, truly, that it FEELS as if this
is the same thing I felt yesterday. But how could I test and confirm it?
(See Dennett & Kinsbourne on Orwellian and Stalinesque memory.)

So if the first problem with introspection is that it does not reveal
how the mind works, hence it does not provide a useable THEORY of the
mind, then the second problem is that introspection doesn't even provide
useable DATA (otherwise known as observations), because no one else has
access to the data, so they cannot be confirmed or disconfirmed.

It's incorrect, however, that this raises the problem that the
unconscious structures and processes that ACTUALLY generate both our
behavioural capacities and our introspective experiences may all be
different in each of us. On the contrary, they are probably mostly the
same. It's just that introspection is no way to find out what they are.

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:41 GMT