Re: Aaron Sloman's "The Irrelevance of Turing Machines to AI" article

From: Wolf Kirchmeir (wwolfkir_at_sympatico.ca)
Date: 08/04/04 

Date: Wed, 04 Aug 2004 09:58:00 -0400

Eray Ozkural exa wrote:

[...]
> What's more interesting is how little attention is being paid to help
> students develop these cognitive skills in their early education.
> Instead they are being taught that they ought to concentrate on
> unnecessary details, which can always be found in a book.
[...]

Again, I used to think so, until this very attitude infected curriculum
revisions at the primary/intermediate levels here in Ontario (and
elsewhere in Canada.) The problem is, that if you don't know a pretty
large amount of "unnecessary detail", it's very difficult to know where
to start looking, and worse, it's even more difficult to judge the
utility and relevance of what you do find.

I had already discovered that you can't learn "cognitive skills" without
those messy details. It takes a lot of practice in sorting lots of
details to learn how to build a classification tree, for example. And
then, transferring those "cognitive skills" from one domain to another
is surprisingly difficult. As someone said, "The devil is in the
details." You may be an excellent classifier in domain A, but if you
don't know enough details in domain B, your classification will be
wrong. Reminder 1: Much scientific argument is about how to classify
some object or phenomenon - IOW, it's an argument about details.
Reminder 2: If you don't know enough details, you may not be able to
tell that your classification may be wrong, let alone how to fix it.

Academics, who from a fairly early age have little contact with "average
people" don't seem to understand this, because they tend to be better
than average at "transfer of learning," nad by the time they hit grad
school, they have trandferred a lot "cognitive skills" from one domain
to another, so it seems easy to do. (Eray, I found your expression of
pity for my students offensive, BTW.) One technique that works at least
some of the time is explicit instruction and practice in applying skills
learned in domain A to domain B -- but both must be dealt with in all
their messy detail.