The chatter of tiny computers

We humans are unique, as Plato observed long ago. We are the only featherless biped having broad nails. This is not enough for James Trefil, veteran science communicator, who seeks to defend the uniqueness of human traits he holds dear: love, consciousness and the seven different inflections of the Czech language. This is his cue for a whirlwind tour of animal psychology and machine learning, the two fronts on which he sees the enemy approaching. The mentality of animals concerns him less than the prospect that we may one day build artificial brains that will replace us. This preoccupation stems from the day he got the heebie-jeebies watching some state of the art computer software solve a few modest problems.

What follows is a pastiche of topics - language acquisition, mirror self-recognition, neural nets, the Turing test - all rather well worn and treated better elsewhere. I was interested to learn that Carl Linnaeus originally classified the rhinoceros as a rodent, but learnt surprisingly little else.

Perhaps then Trefil offers some synthesis of this material with a view to resolving the genuinely preoccupying question that his book poses? Sadly, no. Having reviewed the status of these disparate research areas he then renders most of them obsolete by suddenly introducing complexity theory. It is on this fledgling science that Trefil pins his hopes. As he observes, we do not yet know whether complexity will prove to respect a central set of principles or whether there is no theory of complexity, just the plain fact of it.

It therefore comes as an acute disappointment to discover finally that the author's conclusion that, yes, we are unique, rests on laws of complexity as yet undiscovered. Trefil conjectures that these may dictate that a) the brain is too difficult to understand and therefore replicate, or b) the brain may be too difficult to replicate even if we understand it. Well, who knows what they will dictate and as a thesis this is pretty feeble.

It is also unclear to whom this book is addressed; clearly the scientifically minded, but who will gain from such familiar material? Perhaps a certain kind of adolescent, though such already tend painfully to judge themselves unique.

The author also seems unclear about his readership; it is surely absurd to interrupt an exposition of Kurt Godel's proof with an explanation of what is a prime number. Trefil's style is pleasant enough, though rather consciously self-indulgent with frequent soapbox rhetoric and too many rather flat anecdotes.

Finally, Are We Unique? exemplifies a persistent myopia among computational scientists in their new-found capacity as commentators on the human condition. Their vision rarely extends beyond the individual (brain).

What truly characterises humans and surely makes them unique is that the content of their minds and the quality of their experience is determined by a cultural milieu to which they contribute but which is in part independent of them. I am not so sure that an unschooled solitary human naked on the savannah would appear so different from a chimpanzee or that a human brain suspended in some vat would not resemble the computer it faced over a chessboard. Humans are made persons through being socially situated with respect to one another in a way that computers are not. You will not find the source of poetry or humour or any other of those most human traits in the wiring of the brain but in the wiring of society. The omission of this perspective in Are We Unique? betrays a tacit assumption that a sufficiently large number of interfaced computers would somehow constitute a society. But the complexity of electrical impulses and chemicals shuffling around the brain may be not greater than society's everyday exchange of information, misinformations, insinuation, allusion and coercion. It is only in the unlikely event that this should ever become clear theat we might embark on the dubious enterprise of designing machines to enter the fray.

Thomas Sambrook is honorary research fellow in anthropology, University of Durham.