[ExI] Do digital computers feel?

[Ben]

Brent Allsop wrote, in various responses:

 > in summary, a "red" property is something that reflects 650 NM light, 
and a "redness" quality is a quality of our knowledge of such.
For all we know, my redness could be more like your greenness - both of  
which we only say represents the same "red".

 > Would you agree with that, and does that help?

Yes, and No.

I agree, but it doesn't help. As various other people keep pointing out, 
a difference that makes no difference is not actually a difference at all.

You are not addressing my assertion that there is no such thing as 
'plain and simple redness', that can somehow exist separate from 
memories and perception. 'Redness' /is/ a perception. It makes no sense 
to take it away and still have it exist.

I will ask again: Can you show that 'redness' exists outside a human 
mind? You claim that it's a fundamental thing, so it can't be dependent 
on, or composed of, something else. Please show us what it is, if not 
patterns of information in a mind.



 > for Ben's sake: in this simplified world there are "red and green 
signals in the optic nerve" that can be easily inverted.

Then this simplified world is so far away from reality as to be useless. 
You can't just wave away the way something actually works in order to 
present a 'simplified version' of how it works. That's just making 
things up.

We look at complex systems and extract simplified models so that we can 
cope with thinking about them, and make testable predictions. No matter 
how simplified they are, though, they still have to be /based on 
reality/. If not, they're worse than useless, they're misleading and a 
total waste of time.

Disregarding the fact that there are no inherent red and green signals 
in the optic nerve is not a simplification, it's a falsification.

Here is my (very) simplified version of sensory processing of a 'red' 
object:

A red ball causes receptors in the retina to produce three types of 
spatially-mapped signals in the optic nerve (red-green, luminance and 
blue-yellow, according to colour-opponency theory) that are sent 
(mainly) to specific locations in the thalamus, which routes them to a 
set of different visual maps in the cortex. (note that these signals 
have no intrinsic meaning on their own. Their meaning is determined by 
their origin and destination. 'Line Labelling').

What we know as the concepts 'colour', 'shape', 'brightness', 'size', 
'location', etc., are created by a complex set of cross-associations 
between these and other sensory maps, combined with previously-stored 
memories and modified by the specific architecture of the individual's 
brain. Only at this point, after literally thousands of neuronal events, 
does the concept of 'big red ball right in front of me' come into being. 
Up to then, something like 'redness' simply does not exist, any more 
than Keanu Reeves' nose exists at any point between a movie camera and 
your television screen. In fact, less so, due to the positional coding 
used in the visual nervous sytem.

At all the intermediate points, between the rods and cones of the eyes 
and the conscious perception, all that exists is a set of spike-trains 
in a bunch of axons, setting off another bunch of spike trains (by what 
means is irrelevant), in a great big network. They represent all kinds 
of things, but the only differences between them is the timing of the 
spikes, where they come from and where they are going. There is no 
redness, greenness, bigness, leftness, roundness, etc., etc., in any of 
the individual signals. This is what the science tells us. This is what 
we have to work with when coming up with philosophical theories of how 
we perceive things.

It's a bit like numbers. There is no "Pi quality" in the number 2, is 
there? Neither is there any 'redness quality' in ||||||  ||| |||| 
(that's supposed to be a simplified representation of a spike-train. Not 
a real one).


The key point, I think, is that perception, despite how it appears to 
us, is not a passive process of transmitting existing signals (or 
'properties') to an internal observer, but a dynamic process of 
creation. We don't so much see the red ball as create it in our minds. 
It's that very process of creation that we call 'conscious perception'. 
And this explains a ton of perceptual illusions, like for instance 
seeing a giant monster on the horizon for a second, before adjusting our 
experience to our memories, and seeing a tiny spider on the window. Or 
feeling the non-existent backwards motion of your car in stationary 
traffic for a second, when the cars next to you start moving forwards. 
For that second, the false backwards motion is as real to you as 
anything else. You may even move forward in reaction to the backward 
movement of the car. Then your mind updates itself, and your reality 
changes.

So, there is no 'essence of red' which winds its way from your eyes to 
your conscious experience. Instead, the experience of a red thing is the 
endpoint of a complex series of events, none of which can be said to be 
'red' in any meaningful and consistent way. The Taj Mahal is made from 
bricks, but not a single brick has the least bit 'Taj Mahal quality' 
about it. They're all just ordinary bricks.

When you stare at a cyan square for about 30 seconds, then look at a 
blank white area, what do you see? Why? Do you understand that that 
experience is 'conjured up' by our visual system? How could that be the 
result of an 'elemental red quality'?



 > But it may not help if you believe there are not elemental qualities 
out of which our brain builds or paints composite qualitative 
experiences with

Belief has nothing to do with it. There is simply no support for the 
idea of sensory experiences being 'elemental qualities', and no need for 
it. It's sheer fantasy. Our brains build experiences by processing and 
combining information from many different cortical and sub-cortical 
regions, in the form of patterns of action potentials travelling down 
millions of interconnected axons. These patterns of information have 
meaning in relation to our interactions with our internal and external 
environments.

Here's another argument against the concept of 'elemental redness': If 
redness is elemental, what about purpleness? Bluey-greyness? 
yellowish-brownness? How many colours can we percieve? Can they all be 
'elemental'? And what about all the other things we are capable of 
perceiving? Does it really make sense that you can call something that 
there must be potentially more of than there are particles in the 
universe, 'elemental'? Things just don't work that way. Enormous numbers 
of different things are built up from smaller numbers of simpler things. 
There's no reason why our perceptions shouldn't be the same.



 > Obviously, an abstracted word like 'red' does not have a redness 
quality we can experience.

What? are you saying that the word "red" can't trigger the production of 
the experience of redness? I don't know about you, but I can produce 
that experience whenever I want, including on hearing the words "red", 
"scarlet", and many more. In fact words like "postbox", "Matador's 
cape", "sunset", and so-on can do it.

If you're saying something as obvious as "the word red is not red", I 
don't know why that's worth saying. The word Buffalo is not a Buffalo, 
we all know that. It's just as true that 650nm light is not a particular 
set of action potentials in a particular set of neurons, nor is it the 
RGB colour code #FF0000.

The qualitative difference in how knowledge (or any information) is 
represented is irrelevant. I really don't care if my copy of Beethoven's 
9th symphony is on a CD or a .wav file or grooves in a plastic disc, 
it's the same symphony. A book with references to 'red' can have all 
those words replaced with a little red square, and the book still has 
the same meaning (as long as the reader has colour vision). So I don't 
actually think that it matters whether an information-processing system 
represents red as #FF0000 in a register or as a specific set of voltages 
on a specific set of wires, or a certain pattern of neural spike trains 
in a certain set of axons, it can all mean the same thing. What matters 
is how those representations relate to the system as a whole, not what 
form they take.


Ben Zaiboc