Part 1 : TITLE PAGE | Preface | What is Consciousness? | Outline of the system
Part 2 : Building bricks | Layer-1 | Layer-2 | Layer-3 | Layer-4 | Layer-5
Part 3 : Discussion | Arguments | Conclusions | Addenda
Tartan Hen Publications : Home | more books | Contact : feedback@tartanhen.co.uk
DISCUSSION
What Computers can doI have now completed the description of all five layers of the mechanism. More detail can be found in the addendum. These details include how visual scenes might be analysed, how objects and shapes can be recognised and how physical objects can be represented using the structures I have defined. I have not dealt only with easy cases. I have tackled physical objects which have no easily defined dimensions - like a cloud, a river and an ocean, plural objects, and plural objects with no defined number. I have tackled abstract concepts like "duty" and "responsibility". I have even included something (which I believe has never been attempted before) the representation of the concept "representation". All of these techniques are offered, not as practical proposals, but as plausible accounts of how these things might be done. This is not a prescription. It is just an argument designed to convince the reader that a solution is, in principle, not beyond the bounds of what is feasible.
I don't think, however, that I really need to convince readers that computers are capable of doing some very clever thing. Computers can now recognise text, the spoken word and identify complex visual images. They can control robotic devices as they carry out complex operations. In every case the performance is less than perfect, and there are still some tasks which are beyond them. But research in artificial intelligence continues to advance. And anyway, we humans are not perfect. So perhaps we should be tolerant of computers which still fall a bit short of full human-like performance.
Evolution and Saving Space
I have shown that a mechanism, which would have a reasonable prospect of a human-like performance, could evolve from a very simple system design and do so in small steps, each step being associated with some survival advantage. I hope the reader is convinced of that, at least.
The plausibility of consciousness having an evolutionary origin, is an important aspect of my thesis. Various writers - Searle and Block being two prominent examples - have tried to set up "thought experiments" which feature a stylised mechanism. In each case we are invited to suppose that the mechanism is capable of carrying out human-like activities. At the same time, however, it is obviously lacking in any conscious understanding of what the mechanism is actually doing.
Not only are these systems, quite incapable of doing what the author suggests, but in order to make the system appear feasible in principle, each author has also tried to release himself from practical problems of storage limitation. This was achieved by supposing an infinite or near-infinite storage capacity for his system. We are invited to regard this as a convenience, a mere technical fix which does not affect the logical principles.
I find that ironic. A crucial aspect of the mechanism which I describe here, is the suggestion that the evolution of consciousness, was brought about by the survival advantage of having a practical way for the storage of ever greater amounts of information, without exceeding the available storage capactity. Consciousness, in this view, is not only a mechanism for making intelligent predictions, it is a brain-space-saving device.
What does it feel like to be a Robot?
I am not overly protective of the technical details of my brain mechanism. I don't care if some bits get changed this way or that. The only thing I want to preserve is that interpretation procedure, which constructs and consults the self-MIND model. So let us brush aside technical objections and assume that an artificial brain like the one I have described, has been constructed. The question we must now address is this -
I ask the question in that form, because many years ago, Thomas Nagel in a famous paper (Philosiphical Review, 1974) asked the question - "What does it feel like to be a bat?". Nagel's paper, and that question, has achieved icon status among the phenomenalists, or non-physicalists or whatever we want to call those who hold that consciousness stands apart from physical mechanisms [*]. There is a widely held belief, among that group, that that "bat-question" presents any physical or reductionist attempt to explain consciousness, with an insurmountable obstacle.
[*] NOTE: there are serveral different philosophical positions under this roof, some of which would object to being called 'non-physicalist'. What unites them, however, is the view that a straight forward physical account of a mechanism is not sufficient to explain what they regard as the phenomenon of consciousness.
Consider the subtle way in which the form of the 'bat-question', mitigates against a physicalist response. By stressing the word "feel", the questioner seems to be placing a barrier between the physical mechanism and the topic of discussion. There is, in that word, an implication which pre-supposes the result the questioner desires. We are not talking about mechanisms. We are talking about feelings. And that is different - isn't it? Or is it? Isn't that the question which we should be asking?
According to the non-physicalists, to be the robotic mechanism I have described, would be to experience nothing at all. If you were that robot, and assuming the mechanism is able to operate in an ideal kind of way, you would go about your business without feelings, without being aware of anything. You would walk, sit, discuss the weather, tell jokes, go to work, carry out your normal activities, describe what you had done the previous day, sleep, dream, make love, and even discuss the finer points of consciousness, all without having the least knowledge or understanding of what you were doing. Above all, you would, if asked, CLAIM to have feelings although you would not have anything of the kind.
This idea, that a brain mechanism can operate like that without the aid of feelings, is very strange indeed. Suppose we were able to manufacture an electronic component to match every biological component of the human brain in terms of its functional performance. We can already do this for certain biological components. Damaged nerve fibres can be replaced by electronics to allow paraplegics to do some limb movements. So it is not too fanciful to suggest we might extend that capability across the board.
Assuming, then, that we also have a wiring diagram for the human brain, we could construct an exact functional replica of the human brain. According to the non-physicalists, that robotic mechanism would perform in a way that was indistinguishable from a human, but it would have no feelings at all. It would experience nothing. But the strange (and to my mind) quite alarming implication of that, is that if the two brains were wired the same way, and the artificial system was able to operate as the non-physicalists say it can, then your brain and mine is also capable of operating in the same mechanistic way, without enlisting the aid of feelings or of consciousness.
That seems to imply that consciousness is nothing more than a spectator. When YOU discuss the finer points of your consciousness and your feelings, it is your brain mechanism, which is formulating these utterances. Your consciousness plays no part. When Professor Roger Penrosse wrote his book "The Emperor's New MInd, and penned the following words ....
I more or less know when I am conscious myself, and I take it that other people experiences something corresponding to what I experience."
.... that was not the conscious Roger Penrose who was responsible. It was the unconscious mechanism of his brain which wrote thus. Meanwhile, conscious Roger Penrose observed these actions, but was unable to play any part. So the question we must ask is - was that mechanism telling the truth or was it not? Did it utter a true statement without having the relevant information with which which it could justify that statement?
And that's the problem. When the mechanism claims to have feelings, it is basing its assertions on the information it has stored within its self-representation. But that, according to the non-physicalists, is not a sound reason for making any such assertion. In the case of a robotic mechanism, it would be telling a lie. In the case of Roger Penrose, however, it is telling the truth, but only because, unknown to it, the consciousness of Roger Penrose happens to inhabit the same brain.
This situation is so bizarre it reads like a script for Red Dwarf.
"Wasn't me Gov. It was my brain mechanism that stole those gems and gave you the false alibi."
But then that wouldn't be YOU saying that either. Your unconscious brain mechanism would be saying those words, and it would be claiming, falsely, to be YOU, while it was, at the same time, trying to pass the buck (to itself?).
Your conscious self would be like the pilot of an airliner, lying bound and gagged in the cockpit, watching the plane being flown by the autopilot. Not just flying the plane - switching on the tanoid and saying to the passengers "This is your captain speaking."
The reply to this, which some non-physicalists offer, is to say that although consciousness is not part of the mechanism, it is linked to it by a causal connection. It is 'generated' by the mechanism rather as steam is generated by a steam engine. Consciousness, in this view, becomes a by-product of the mechanism. A conscious experience can, therefore, never express a feeling which is different from the one being enacted (but not felt) by the mechanism. All it does is to reflect what the mechanism is doing, (for reasons which are totally determined by the mechanism itself without reference to consciousness). When I spoke above of the mechanism telling a lie and consciousness being aware of the lie, but also being unable to do anything about it, I was creating a circumstance which can never arise. If the mechanism sees fit to tell a lie, the conscious experience which will accompany that, will be a feeling that lying is the appropriate thing to do at that moment. So consciousness is not in control of itself. The feelings which consciousness experiences are totally under the direction of the unconscious mechanism.
And that is a problem too. When a conscious person experiences a true kind of consciousness, that feeling is being created by an unconscious mechanism which believes itself to be conscious but is telling lies when it tells us that. The unconscious mechanism and the conscious person both believe themselves to be conscious, and both base that belief on the information stored in the brain, and available to the unconscious mechanism. But one of them is telling the truth and the other is lying.
So if consciousness is completely separated from the mechanism, we have a problem. And if consciousness is generated by the mechanism, we have a problem. I think the non-physicalists have just painted themselves into a corner.
What is never explained by the non-physicalists, is how the unconscious mechanism could cause the associated conscious experience. And that, I suspect, is how it must remain. If anyone did manage to produce an explanation of how the mechanism does that, that explanation would take the form of a mechanism. It has to do that, because that is what an explanation is - it's is a causal mechanism. And the non-physicalists would then immediately take the view that that secondary mechanism was a 'mere' mechanism and so could not be the 'real' explanation of consciousness.
One gets the impression that trying to find an explanation which will satisfy a non-physicalist is like trying to find the end of a rainbow. As you approach, it recedes. It is always out of reach. It has to be so, because in the non-physicalist belief system, that is how consciousness is defined - that which is always beyond explanation. The 'Hard Problem' must remain hard, otherwise it would not be the 'Hard problem' any more.
But we do not need to get into this absurd position. We have an alternative explanation. Why do we need to introduce an unfathomable thing which is caused by the mechanism when we have that mechanism which is fully in control of itself? Instead of saying one causes the other, we could say that they are the same thing. It is rather like the difference between the agitation of the molecules in a body of gas, and the temperature of the same mass of gas. If we insist that they are separate things, with the molecular activity causing the temperature, we will get into all kinds of nonsensical bother. Temperature is, in fact, just another way of looking at the degree of activity of the molecules. In like manner, consciousness is just another way of looking at what the mechanism is doing. Your MECHANISM and YOU are one and the same thing. You are the whole of that mechanism. Yes ... but you SEEM TO BE just one part of it. You SEEM TO BE not the whole thing, just that interpretation procedure which constructs and then refers to the self-MIND. And when you say you are FEELING, all you are really saying is that your interpretation procedure is processing a particular kind of data. When you say you are seeing something which is RED, what you really mean, is that your interpretation procedure is processing sensory data (encapsulated in concept structures) which contain a particular kind of notation. The colour RED (the experience of REDNESS) is just the notation used by your brain to record these experiences. The concept structure corresponding to RED, has all kinds of remembered associations. But RED itself, is just a notation. Even non-physicalists have to admit that the brain must have some kind of notation in terms of which experiences will be recorded. It could just be the disposition of certain brain synapses. It is unlikely to be anything which could be placed on a sheet of paper.
PAIN is also just part of the notation. It is also a sensory experience which is labelled NASTY, and it causes a wide-spread system interrupt. Readers with a knowledge of computer technology will recognise the concept of a system interrupt. When some external events impinge on a computer (a key being struck on the keyboard for example), the system reacts by stopping all other on-going processes, recording the data needed to restart those other processes, and then deals with the new event. It can also assess the importance of the new interrupting event, and decide whether to deal with it immediately, or assign it to a queue of jobs pending. PAIN will usually be one of those which are dealt with immediately. Because it is NASTY it must be avoided.
"But does it actually FEEL pain?," asks the non-physicalist.
The answer is "Yes".
"IT" is the interpretation process.
"IT" is the process which is interrupted.
"IT" is the process which has to deal with the new experience.
"IT" records the experience as NASTY.
So "IT" feels the pain.
And when we say that, all we are saying is that "IT" does all these things in response to the input of sensory data (and the associated interrupt) which we normally associate with pain.
And if YOU are that mechanism, then "IT" and "YOU" are the same thing.
And what then does it feel like to be that robot? Much the same as being a human mechanism with the same functionality.
And what does it feel like to be a bat? I guess a bat does not form very many concepts and does not interpret the world in a complex way. But it must have a three dimensional model of the world, and it must be able to plot its own trajectory, and those of the insects it is hunting. So why does it actually see? My guess is that "seeing" is much more likely than "hearing" even although it is a sonic form of sensing. But seeing is a more precise way of plotting positions. I don't know what kind of notation it will use to record that map of positions. If it comes to that, I don't know what notation other people use. But I do know they have a similar three dimensional map in their heads. So I guess that being a bat, is much like being a human, in respect only of the three dimensional plot of positions. My guess is it is rather like my layer-2 system with, perhaps, a hint of layer-3.
Other Issues: Philosophical Implications
That concludes the main thrust of my case. There are a number of other issues which are raised by this discussion. These I have transferred from this section to a number of others which the reader may visit by clicking on the links below. Beside each link I have included a short note which gives a gist of the content.
REALITY
The mechanism builds a representation of external reality. It has no access to an external world except what comes to it through its sensory apparatus. For it, therefore, that internal reality is its REAL reality. It is a structure, a store of information, which I will denote {REALITY}.
PROBABILITY and DETERMINISM
Probability is a measure of human expectation of outcomes. Determinism is the illusion of absolute certainty of outcome. Two factors combine to destroy absolute certainty. One is Quantum Mechanics which prevents certainty at the atomic level in the physical world. The other is Chaos Theory which ensures that in some systems, small errors propagate to produce large errors, destroying our ability to predict with accuracy.
FREE WILL
The concept of free will is not challenged by a mechanistic description of brain function. If the concept of SELF is identified with the mechanism itself, rather than with some disembodied non-physical entity which accompanies the mechanism, then choices made by that mechanism, are being taken by SELF. The SELF therefore carries full responsibility for those decisions.
TRUTH
In my system, truth (or falsehood) is the result obtained by comparing two structures. If one of the structures is called {REFERENCE} and the other is called {TEST}, the comparison is uni-directional and can be denoted this way -
ADDENDA
For a complete list of additional topics.
Part 1 : TITLE PAGE | Preface | What is Consciousness? | Outline of the system
Part 2 : Building bricks | Layer-1 | Layer-2 | Layer-3 | Layer-4 | Layer-5
Part 3 : Discussion | Arguments | Conclusions | Addenda
Tartan Hen Publications : Home | more books | Contact : feedback@tartanhen.co.uk
Copyright © Hugh Noble (Nov 2006)