Impossible Minds: Ny Neurons, My Consciousness [Hardcover]

It is easy to see why Aleksander is attacked by both philosophers and neuroscientists. He seems to say hes working on a consicous machine, and that there is no problem in saying computers can/are conscious. This book will clarify things a little, however. Aleksander holds compiters and machines, can be artificially consicous, which is very different from normal human consicousness. How different, or in what way? Well, thats what hes trying to find out. And Aleksander is workind on machines so that he can gain understanding on the human kind of consciousness, not to be known as the first researcher to create a conscious computer. IT seems then, that he is on the naturalist and the neuroscientists side, but is approaching the problem from an engineering angle, to be more specific, from the viewpoint of automata theory.
And so, this book is about consciousness, and how it arises out of neurons. He first gives a basic guess, or hypothesis, and then defends it and expans it to see how far it goes. Consciousness is the product of neurons and their functioning, with the crucial function of iconic representations and learning, according to Aleksander. Most of the book deals with diagrams of automata theory, flow charts and neural units, the tools of connectionism. It all looks good at first reading, the points seem obvious. Sure, there is no real talk of actual brain mechanisms, but Aleksander goes to the basics, and it seems to hold up.
Aleksander then brings in important insights into the field of the science of consicousness. If consciousness is the product of neurons, and neurons process information, it does not seem absurd to hold that information processing in something other than neurons can be conscious too.
I do have doubts on how much Aleksander can explain with his current methods. For example, he theorizes on how qualia can emerge from the automata theory diagrams, but this seems to only say that some units, because of previous learning and their interactions, represtent "blue" or "red" or whatever. This is not what philosophers think of the mystery of qualia, however. Neuroscientists talk of "color coding cells" of "color areas", etc., but the problem is explaining not that the cells represent color, but how by virtue of doing so, color presents itself to our experience the way it does. We are pretty sure neurons can represent blue, but we have very little idea of how this neuron makes the blues "blueness" so striking in our subjective consicous experience. In this sense, philosophers will not be satisfied with Aleksanders account. But then again, no other scientific theory seems to quench philosophers thirst for explanations in the qualia issue, so it can hardly be taken against Aleksander.
This book is what contributions to the consicousness debate should look like. It brings in new insights, is quite simple to read and follow, it argues for a better understanding of consicousness through multidiciplinary approaches (in this case, automata theory), and genuinely makes some advancement. Aleksander however leaves out so many details, that most will find his proposals wanting.

NOT PERFECT - BUT OUTSTANDING
==============================

Writing a review in 2009 about a book of 1996 needs some justification, especially when I state that the book of Aleksander is not perfect, but outstanding!

1. A STILL OPEN QUESTION

Aleksander states that close to the 1990s consciousness has become a highly popular topic as a subject for conferences (p.291) and John G.Taylor, one of the major contributors to an engineering approach of consciousness, says in his book 'The Race for Consciousness' (1999)[p.178] that the 'race for consciousness has started in earnest'. These opinions are further confirmed by a simple statistics restricted to the field of engineering given by the IEEE electronic library which indicates an increase from 18 publications about consciousness in 1980-89 to 82 in 1990-99 and to more than 304 in 2000-2009. The real numbers including numerous different disciplines are much larger. But despite the growing amount of publications in many fields no unified solution has been reached. In a recent paper 'Can consciousness ever be modeled?'(2009) G.Taylor states it as follows '... up to now a satisfactory answer has not been forthcoming, at least one acceptable to all who hear it.'[p.175]. What is missing besides several unsolved questions in the details of the neural machinery as well as in the psycho dynamics of human persons is the general lack of a commonly accepted framework integrating the main views. This is why I think that the book of Aleksander is still worth reading and even outstanding despite of being non perfect.

2. ENGINEERING

Although today our modern world cannot any longer exist without the contributions of engineering, engineering as such is still not very well established within the sciences. While science usually (only) wants to 'understand', engineering goes beyond understanding, engineering wants to 'construct' real working systems. To do this it is necessary in engineering to form large teams of experts (several hundreds, several thousands, or even more) to transform a problem into a working solution. This requires highly sophisticated processes of communications and interactions (human swarm intelligence) which goes far beyond forms of cooperations which are practiced in science. Furthermore in engineering it is not enough to 'mimick' nature but to abstract from nature underlying principles which can be translated into efficient working systems. In the case of 'consciousness' engineering is not satisfied with the scientific findings of biology, psychology, neural sciences etc. but it tries to 'translate' these biological structures into technical structures of maximal simplicity as working systems. Although these technical systems can structurally and functionally be the 'same' as the biological givens, they can 'look' quite different. The book of Aleksander about consciousness is a book written from the point of view of engineering

3. ENGINEERING MEETS PHILOSOPHY

A long time --perhaps until the 1950s-- the topic of consciousness was primarily a topic solely of philosophy, psychology and some related disciplines. It was nearly inconceivable that this topic could become a serious subject of engineering (despite some examples of building artificial machines mimicking some human properties in the past). But finally this happened. Aleksander gives an explicit picture of the view of many important philosophers about topics like 'qualia', 'consciousness', 'mind', 'language' etc. and explains then with sufficient detail how these concepts can be translated into so called 'neural state machines'. This does not mean that this is the only possible translation (many more are conceivable), but it is at least one translation, which works and which gives an engineer the possibility to built machines which can mimick central properties before attributed only to human persons, including 'consciousness'.

4. ENGINEERING SUPPORTS SCIENCE

Aleksander does no limit his book to philosophy but extends his scope to many more disciplines including biology, linguistics, neurosciences etc. While the sciences until today are divorced in multiplicity and a unifying framework is not at the horizon the approach of Aleksander shows how it could be possible to unify all the findings of the different disciplines by translating them into one technical framework provided by the paradigm of neural state machines.

5. OUTSTANDING

What makes the book of Aleksander outstanding, even 13 years later, is the fact that he has gathered so many disciplines --including extensively philosophy-- and bringing them together to one working system. I do not know of any other book trying such a holistic approach from the point of engineering. Even large international funded projects of today do not yet follow such an integrated radical approach.

6. NOT PERFECT

Clearly, a book with such a broad scope of reflected disciplines, viewpoints, and methods can not be perfect in all details. But this does not matter. Details can be found in hundreds of specialized papers. What is seriously missing today are books (and projects) which have an open and general engineering approach from the very beginning which are looking to the phenomena of consciousness and mind from all the important points of view. I am recommending this book to all my students and I myself use it as a source of inspiration (also because of its broad scale of elaborated references).

7. SMALL POINTs OF CRITICISM

7.1 TURING MACHINES

On several occasions, especially pp.181-185 Aleksander criticizes the limits of the universal turing machine, introduced 1936/7 by Alan Matthew Turing, and 'conventional computers' as mere practical forms of turing machines and offers the neural state machines as an improvement (This argument can also be found very often outside of the book of Aleksander, especially used by philosophers). This argument is dangerous, if not completely wrong. The universal turing machine (UTM) as such is not machine at all (!), but only a mathematical concept to represent the most general idea of computing, we know today. A real conventional computer is one (!) possible instance of this mathematical concept and in the same sense is a real neural state machine another possible instance of the mathematical concept of an UTM too (this is the reason why we can simulate neural state machines with the aid of conventional computers).

7.2 REDUCTIONISM/ EPIHENOMENALISM

It seems to me that Aleksander underestimates the problem of the 1st person view as described by Nagel and the 3rd person view as exemplified by the behavioral or neural sciences. While the 'content' of a 1st person view is epistemological different from the 'content' of a 3rd person view it is logically completely impossible to 'identify' these epistemologically different contents as 'ontologically the same'. If we 'correlate' some neural measurements with some descriptions of 1st person communications then we can hypothesize that a certain 1st person assembly of qualia seems to correlate with some neural activities (neural state changes), but we can not identify 'assemblies of qualia' ontologically with these neural patterns and pattern changes. Qualia as such have no ontological state. They are pure logical entities which can logically be mapped to neural entities and thereby gain some possible meaning (or vice versa). This disallows also any kind of 'dualism'. It would be nice to discuss this further with the author.