Re: Mental Imaging

From: HARNAD Stevan (
Date: Thu Mar 07 1996 - 12:49:57 GMT

cn> From: "Nelson Claire" <>
cn> Date: Thu, 7 Mar 1996 09:39:10 GMT
cn> Does mental imaging incorporate the notion of ' imagination'?

Yes it definitely does. Mental images (which need not be visual, but can
be in any sense: hearing, touch, smell, or even just mood) can either be
everyday ones, as in picturing people or objects you've seen lately; or
they can be rare insights, such as Kekule's image of the snake eating
its tail, which gave him the idea of the structure of the Benzene Ring
in organic chemistry.

cn> What is imagination ?

Imagination is everything going on in your mind that is not directly
caused by what you are looking at (or hearing, etc.) from the outside
world at the time. So when you look at a real apple, what you see (which
is of course also something that is just going on in your head) is not
usually called an "image" (although it is); what you see is the apple.

But when the apple is not there, and you picture it in your mind, then
you are imaging. Normally we reserve the word "imagination" for an even
more restricted kind of imaging: when you are imaging things that don't
exist (or exist only in your "imagination"), for example, unicorns (or
pink elephants).

So you have seeing (1) real objects, (2) picturing real objects in your
head, and (3) picturing imaginary objects in your head. In a way, ALL
of these involve imagery, because in all cases you are seeing
something, and "seeing" is a mental event, and it happens in your head,
not outside your head. Sometimes you can't even tell whether you are
seeing a real object or just imagining you are seeing it (that's called
a hallucination).

As you see, most of the questions about the mind are already there in
these questions about imagery: What IS a mental image? We all know what
a real picture is, out there in the world. But what is a picture in your
mind? We also know who is looking at a picture when it is in the outside
world. It is you. So what is going on in your mind when you see and
recognise an apple is the kind of thing we need to explain in
explaining the mind.

But who is looking at the picture when it is an image IN your mind,
rather than an object in front of you, in the world? Another mind?
A mind in your mind? And what's going on in THAT mind when it is
looking at the picture in YOUR mind?

(This is the problem of the homunculus: You cannot explain what is going
on in the mind by saying there are pictures in there, because then you
have to explain what is going on in the mind of the viewer of those
pictures, and so on [more pictures? inside more minds?]. An explanation
of the mind has to "discharge" the homunculus, get rid of any reliance
on it, replace it with some kind of a mindless process that can work
on its own, and not only in the mind of yet another mind!

Kosslyn's and Posner's and Pinker's papers are about the evidence, both
behavioural evidence and brain evidence, that there really are pictures
in our heads; they also give some computational evidence that "pictures"
can do things on their own, without needing anyone to look at them.

[Of course, on the deeper question, of why all the mindless mechanical
processes, freed of the need for homunculi, don't simply do it all --
why, in other words, there should be minds at all -- Kosslyn and Posner
and Pinker can unfortunately offer nothing at all, because that is
again the mind/body problem.]

cn> Does mental imaging explain in anyway
cn> the means by which people can ' mentally ' come up with
cn> ideas / images for inventions etc.... ?

Eventually, we hope. Creativity will be the subject of the last two
lectures in this course. But for now, before we leap to creative
cognition, we first have to try to understand ordinary, "noncreative"

sh> From: "HOLMES Sharon" <>
sh> Date: Thu, 7 Mar 1996 11:02:41 +0000
sh> Are the mental images that you and I produce when we come up with an
sh> invention a by-product of computational processes or the processes
sh> themselves?

This comment shrewdly raises a question that came up in lecture:
Why is it only images in the mind that suffer from a homunculus problem
(the problem of who looks at the images, and what's going on in HIS
mind)? Why don't "propositions" (words, sentences) face the same
problem? If what's going on in my mind is more like sentences
(descriptions) than pictures (depictions), who is "reading" and
understanding the sentences, and what is going on in HIS mind?

Fortunately, this question has been definitively answered by a couple of
decades of findings in artificial intelligence: Sentences are
strings of symbols, like the symbols on the computer screen you are
looking at right now. Computer programs are likewise strings of symbols,
and rules for manipulating those symbols.

I will give an example in Friday's lecture, but think of some of the
"recipes" you learned in maths, such as the rule for factoring a
quadratic equation of the form:

aX + bX**2 + y = 0

by using the formula:

X = -b +/- (SQRT(b**2-4ac)/2a).

By applying the second rule/recipe to any string of symbols
like the one first one, you can factor any quadratic equation and find
X. That kind of mechanical manipulation of symbols is what a computer
does. Moreover, if you have forgotten what all these symbols MEAN it
doesn't matter, because the computer doesn't know what they mean
either, and doesn't need to! It is merely mindlessly following a rule,
and the correctness of the rule (sometimes called an "algorithm," but
there's no need to remember that unless you already know what an
algorithm is) is what ensures that the computer's output will be

So we KNOW that computation (propositions, sentences) can work -- can
produce behaviour that normally requires a mind to produce it -- without
the need for a homunculus. Can the same be said for images? Or is what's
going on in your head really ALL just computations, and the images are
just pretty decorations that aren't really doing any of the work?

This is what the imagery debate is really about. And the answer is that
there ARE physical processes -- they are also called "analog" processes
-- such as rotating objects and superimposing one object on another (the
way I did in lecture with the two transparencies, to see whether the
little x did or did not overlap with the large F) that can be shown
(again by computers, but this times the computers are just SIMULATING
the physical process: we could have used the real physical process to
show the same thing, and that would not involve any symbols or
computation) to be able to produce certain kinds of behaviour that,
again, normally requires a mind to produce it -- again, without the
need for a homunculus.

In other words, midnless, mechanical manipulation of images, just like
mindless, mechanical manipulation of symbols, can produce behaviour
that normally requires a mind to produce it. So, to a reverse engineer
of the mind, both symbol-manipulation (computation) and
image-manipulation (imagery) are potential explanations of the mind.

This will all become clearer as the course continues. And it will come
back again and again. So don't worry -- but DO do the readings and
Skyreadings, and do them now, so it starts to make sense sooner, rather
than at the end, when you may be too rushed to understand it. And don't
be like the passive kitten in the Held/Hein experiment I described in
the first lecture: Don't just attend the lectures and tutorials and read
and Skyread: Skywrite too. Ask questions or try out your kid-sibling
explanatory skills to try to explain or comment on what you have
understood so far...

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:38 GMT