Re: How to get HOW from WHEN and WHERE

From: HARNAD Stevan (harnad@cogsci.soton.ac.uk)
Date: Fri May 03 1996 - 17:10:30 BST


> From: "Herheim Aaste" <AH595@psy.soton.ac.uk>
> Date: Thu, 2 May 1996 17:32:29 GMT
>
> "How can you find out HOW from WHEN and WHERE?"
>
> Can you?

Perhaps. This is a question on which there is more than one viewpoint.
The reaction time ("cognitive chronometry") and brain imaging studies
described by Kosslyn (Overview of Image and Mind) and by Posner & Raichle
http://cogsci.soton.ac.uk/~bbs/Archive/bbs.posner.html
certainly seem to be answering some "how" questions (concerning, for
example, whether or not imagery is involved in mental rotation tasks).
The computationalists (who think the hardware details are irrelevant,
and think the "how" questions are all answered by the finding out what
algorithm is being computed) don't think information on brain
localisation can help: Are they right? This is a question on which you
can make a case for both sides.

But the rest of what you are asking doesn't seem to be about the "how"
question at all, but about the mind/body problem. Remember that even if
the answer to the question of whether or not you can find out HOW from
WHEN/WHERE is "yes," it does not follow that from finding out HOW you
will solve the mind/body problem! To solve the mind/body problem you
need the answer to yet another "how," namely: if the explanation of how
the brain DOES something (remembers, recognises, calculates, etc.) turns
out to be THIS (substitute for "THIS" whatever you like: a computed
algorithm, the connections of a neural net, activity in the inferior
temporal gyrus), how could THIS feel like something?

Don't hold your breath while you wait for an answer for that one...

> Even if I know that a certain area in the brain is active when I
> experience a certain feeling or try to solve a certain task,- this
> tells me absolutely nothing about HOW THAT BRAIN ACTIVITY TURNED INTO
> MENTAL ACTIVITY!

Correct. But, as I said, that's the mind/body problem, whereas the "how"
question that was intended here was, in a way, a more answerable one:
What is the causal explanation of what I DO (not of what I FEEL whilst
I'm doing it)? First sort out whether or not when/where can answer that
how question. Leave the bigger "How" question ("How can ANYTHING
physical be mental?") for later, for after you have an answer to the
easier "little" how question.

> Is it possible to answer the question comparing the mind with a
> computer (like we've done) ,- then how?

It is possible to answer the little how question, the one intended by
the exam question, with a computational answer: There ARE things that we
can do that, in and of themselves, are mysterious: We can do them, but
we haven't the faintest idea HOW we do them. (Think of today's lecture:
We can recognise birds, and call them "birds," but we have no idea how
we do it: we can't even list the features that something has to have in
order to belong to the category "bird".) A computational answer (or a
neural net answer, or an analog processing answer) could answer how; the
question here is whether information about WHEN and WHERE can help us find
the HOW.

That's the little how. Having found the little how, we can turn to the
big How: If this causal explanation for how I am able to do so and so is
correct (be it a net or a computation or a shadow-manipulation), How can
we explain that whilst the causal mechansim does its stuff (the net
connects, the algorithm computes, the shadow transforms) we are FEELING
something: conscious experience is going on.

That big How question will not receive an answer in this course; it
will merely be raised. But don't keep returning to it in place of the
little how questions, because most of the course is about the little
how (how do we do it? what is the causal mechanism for the doing??),
not the big How (How does the causal mechanism of doing also generate
or become feeling?).

> Would this mean that we hypothesise that feelings are just cognition,-
> I still can't get around the question how increased activity in the
> prefrontal cortex (along with other typical indicators of depression)
> translates into a certain mental state,- namely a feeling of sadness,
> decreased interest in the external environment etc. Is it the specific
> "decoding" of brain events to mental events we are looking for?

Set that aside for now, and focus on what kind of brain activity is
needed to get something DONE, to give you a CAPACITY (say, to remember a
name, to do a calculation, to recognise a face, to assign a category,
to understand a sentence). Focus on that little how when you take on the
when/where question. This is not the place for the big How.



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:39 GMT