Brains working towards machine consciousness

 "The World in my Mind, My Mind in the World", Igor Aleksander, 2005, imprint academic, Exeter.

Subtitled "Key mechanisms of consciousness in people, animals and machines"

OK, I like Igor Aleksander.  Not to name drop (because I am guessing most readers of this may not have heard of him anyway), I met Prof. Aleksander at COGRIC, and admire his work greatly.

So, upfront, I want to get the bad part of this review out of the way.  This book is riddled with typing errors.  I don't think it was proof-read.  This vexes me.

However, the content is much more important.  It is a good read, with good explanations.  But I do have some issues with parts of it.  For instance the assertion that "In waking life, the perceptual is always there" - is it?  Is it really?  Or are we just defining 'waking' differently?  Or, perhaps, perceptual. 

"The perceptual is the 'here and now'.", so, we agree about that. I wonder why I am talking to a book.  But if it is always present in waking life, we are presumably not talking about a conscious awareness of the perceptual, although that is what I would assume was meant, because there are those times when we are most definately awake, but not conscious of what is happening around us.  Indeed, we need not be conscious of what we are doing to undertake quite complex tasks.  Worryingly, driving is an example.  Conversation appears to be another.  I might argue that education is too, but that might get me into trouble.

More importantly, Aleksander outlines his Five Tests for Being Conscious.  Personally, I think these are Five Tests of Being Conscious, but that is another matter - he describes 5 axioms of consciousness.

Axiom 1 : Me in a real world out there.  Feeling that you are part of, but separate to, the 'out there' world

Axiom 2 : My experience of the world out there.  Also described as imagination.  The ability to persist a model of the world when not directly perceptually aware of it.

Axiom 3 : Out to get experience. Whilst this sounds like it is about intention, it is about attention.  Attentional mechanisms.

Axiom 4: Thinking Ahead. Part of the cognitive package, of course, and I would agree it is necessary for pragmatic consciousness, but am not entirely sure it is actually necessary for consciousness.  But that is a minor issue, and I am more interested in pragmatic consciousness anyway.

Axiom 5 : Emotions - the guardians of thought.  Now, I know there are others who think emotions are necessary components of even good AI, let alone consciousness.  I certainly think they are an interesting aspect, and an emotionless artificial system is much less likely to pass muster than one with them.  But are they necessary?  Really?  Some of the problem here is again in definitions.  Some practitioners refer to things like 'hungry' as an emotion.  Mind you, some philosophers do too.  Hungry is not an emotion.  Sorry, it just isn't.  And a small child flinching from a large object is not, to me, a sign of fear, although it is, apparently, to Aleksander as he uses it as an example.  It is an autonomic response designed to remove said child from danger.  Fear is the emotion which you get when your sympathetic nervous system kicks in, adrenaline starts to be released and have its effect, but you cannot perceive the cause of the physiological effects.  Fear can only be of the unknown.  You can respond to the known, and the physiological response can be the same as that of fear, but fear itself, I strongly contend, is always of the unknown.  Fear is psychological, not physiological.  Emotions are psychological not physiological.

Another point of contention I have with Aleksander is that he seems to suggest that subconscious 'brain' activity is almost an afterthought, although still an essential component of a functioning conscious system.  I do not think that this can be a correct view (and rather hope I have just misinterpreted what he is saying).  However, it is brought up in the same part of the book which covers sleep, an area which I think Aleksander has got things spot on. 

He even has examples from the marvellous model they have developed and use, a brain-like automata.  And this is worth reading the book for, on its own.  Well, at least the appendix, which describes the concepts.

There is much more in the book, which has a high content to page ratio.  And, to be honest, I like a book I can argue with.  It would be better if I had the author to argue with, because books don't answer back, but I am content for the moment with the book.