Kim Sterelny, "The Imagery Debate"
Ned Block, "Mental Pictures and Cognitive Science"
It is said that we know that some of the representations in the brain/mind are not structured linguistically: mental images.
Pick on a mental image that you can conjure up before your mind's eye.
QUESTION: Does anybody have difficulties with this? It is suggested that some of the pioneers of behaviourism had a difficulty over conjuring up images.
We have discussed how a proposition may represent. We discussed it. I didn't say we understood it.
How does this mental image that we have represent the thing it is an image of?
My image is of the Grand Canal. How does this image represent the Grand Canal?
Does it represent the Grand Canal in the way that a painting of the Grand Canal represents the Grand Canal?
There are said to be two accounts of how mental images represent.
One is that they represent in the way in which pictures represent. This is pictorialism.
The other is that they represent in the way in which propositions represent. This is descriptivism.
This is the view that is attributed by Sterelny to Pylyshyn:
'...Images are networks of quasi-linguistic representations in a computational workspace.' (Sterelny, in Lycan, p.616.)
_____________________________________________
But what are we talking about when we talk about a mental image?
It may mean the experience of imagery, eg the picture of Venice I have when I day-dream of a holiday.
Or it may mean the internal representations involved in mental imagery - whatever in the physical brain underpins the experience of seeing the Grand Canal 'with the mind's eye' when I day-dream.
For the physicalist this will be a distinction without any significance.
I will call the experiencing a day-dream Grand Canal a phenomenological image.
_______________________________________________
Let me point to some facts which have been cited as supporting the idea that there must be images in the brain.
3. The process of imagination
Rotation
Display diagram 1.
Appoint two timers.
I'm going to display two images and I want you to work out whether they are different views of the same object or different objects.
When you have worked it out, raise a hand. The timers will see how long it takes.
Now I will display another pair of images. Work out please whether they are images of the same object. Raise a hand when you have the answer. The timers will see how long it takes.
(Ask one timer to time Eli. Ask the other to take the time when 6 members of row 3 say register.)
[Ask people how they go about answering the question.]
When this is done properly, the result is that the second exercise takes longer.
In each case, what people say they are doing is rotating the first image to see if it coincides with the second.
Supposing this is what they are doing. Then it should take longer dealing with the second pair than the first because you have to rotate image A more in the second case.
This is what happens.
It is argued that this is evidence that there is a sort of image in the brain which we rotate. This supports the pictorial account of what an image is.
Scanning
You are asked to memorize a map.
The map is withdrawn and you are then asked to recall it in consciousness.
You are asked to focus on a particular feature:
And then to confirm whether feature A is on the map.
And then to confirm whether feature B is on the map.
If B is further away from the focus than A is, confirming whether B is present should take longer than it takes to confirm that A is present.
This is taken as evidence that what you are doing in checking out whether A (and then B) is present is scanning an image.
_____________________________________
But it is said there can't be images in the brain:
Here are some reasons. Block calls them 'flat-footed':
Open up the skull and you don't actually see any pictures.
Phenomenological images are eg irredescant blue but bits of the brain do not have the same properties.
(Leibniz Law: 'If a is the very same thing as b, whatever is true of a is true of b.')
We leave this to the reading.
So maybe we conclude from these objections that there are no pictures in the brain.
Is there anything left of pictorialism?
Even though we may conclude that there are no images in the brain, perhaps there are distinctive types of representation - not images, but 'pictorial' nonetheless.
Let me explore this.
When we asked whether there were propositions in the brain we said: there may not be propositions but there may be representations of propositions, quasi-propositions.
Can we not make the same move in relation to pictures? Granted there are no pictures in the brain. But may there not be representatioons of pictures, quasi-pictures?
If we pose this question, the tempting answer is surely Why not?
If I do have a phenomenological image of the Grand Canal, anyone with the least leanings towards physicalism will think this must be unperpinned by a representation of a picture in the brain.
When I actually see the Grand Canal, if imaging is involved, there again you presumably have a representation of that image in the brain.
This does not mean that the representation must be pictorial however.
The representation of a picture does not need to be itself 'pictorial'.
It could be (so it is claimed) propositional.
Take a simple case. A picture of a rectangle. A propositional represention of a rectangle might begin:
A straight line runs from one point to another, say A and B.
From both A and B straight lines of equal length drop down vertically. their ends are joined by a fourth straight line.
This is a description of a rectangle. A representation of a rectangle in propositions.
It is claimed that we can represent any pictures in this kind of way: via a long list of descriptive propositions.
(And this is how sophisticated graphics on von Neumann machines is done. You have a long list of instructions describing what lines in what colours at what lengths and at what angles are to be drawn on the screen.)
So we can see how a picture might be represented propositionally.
The question is, is there an alternative? Might a picture be represented pictorially?
_______________________________________
Attempts to say what the difference is between a representation being pictorial and a representation being 'discursive' or propositional:
1. To say images are pictorial is to say the brain uses them in the same way as maps and pictures are used.
2. To say images are pictorial is to say they have the same semantic properties as pictures have. Eg spatial closeness in what is represented is represented by closeness. In a picture of the Grand Canal, the picture of the Rialto bridge is above the pictured water. In the world, the bridge is above the water.
3. To say images are pictorial is to say they are analogue representations.
Two differnt ideas may be involved here. One is that every difference in the representation is reflected in a difference in the object represented. Eg every change in position in a radar blip represents a change of distance of the tracked plane. Call this denseness.
The second idea is that the relation between representation and what is represented is not conventional. Eg
A parallel: a mercury thermometer represents the temperature of the environment analogically and not conventionally.