"To make computers which are like brains one would indeed have to make them of analogous units [realistic neurons], connected in a similar [realistic specialized brain region inter-connectivity] way, capable of analogous plastic modifications [complex mechanisms for synaptic strength modification] and with analogous responses to incoming data, both from outside the computer [sensory input] and from within its own works [capacity for introspective thinking]."

Although Rose had the guts to put the word "consciousness" in his title, discusion of consciousness is poetically brief in this book, but he hits the high points. For example, in dealing with the popular question of what subsets of the universe have conciousness he clearly explains that complex brains allow consciousness. We view human consciousness as being special because, ".....we have defined consciousness not as a static [Platonic?] form but as a process involving interaction between individual and environment."

This theme of "individual and environment" is greatly expanded upon in Rose's 1992 bookThe Making of Memory: From Molecules to Mind. Go here to see a review by Robert P. Pula. (a backup of this link). Another review by Danny Yee. Pula deals well with Rose's emphasis on the "collective". I would only add, that as painful as it is to listen to scientists try to reconcile their personal beliefs with their research, this is what the world needs more of. Rose should be congratulated for inviting the public into his life in such a personal way. Most scientists are too tight-assed to even admit that Rose has done the right thing in writing a book like this.

I share Rose's view that an understanding of the mechanisms of biological memory can be a Rosetta stone for translating between mind-speak and brain-speak. I also share Rose's intuition that all people can get something useful out of learning about how the brain works. Epistemology is a central issue in all human cultures. Can anyone have a rational epistemology without understanding the brain?

It is great fun to watch Rose deconstruct everything from Descartes to Eric Kandel. What a joy ride: Rose then constructs a contrafactual world in which biology develops before physics, sparing the world from reductionism! I wonder if the mathematics of complex systems might have been developed before calculus in such a universe? It is mind-warping to watch Rose juggle his anti-reductionism (for example, he suggests poker playing for the Turing test, and says that poker involves "non-cognitive inputs" that no machine can ever deal with) with his reductionistic attempt to explain human memory in terms of molecules.

It is instructive to follow Edelman in his confrontation with connectionists; at least Edelman has been involved with using computer models to study intelligence. I will (try to) be the first to agree with Rose that most "neural" network models are not very neural, but I think he is wrong to condemn Artificial Intelligence research by claiming that AI workers think that they are explaining how biological brains work. Sure, some folks have gotten overly enthusiastic about their computer models, but Rose is wrong to condemn the entire field.

In fact, it is Rose's penchant for condemning almost everyone and everything that is the biggest problem with his book. We can all welcome constructive criticism, but Rose too often just following his personal biases onto the battle field and presents warped caricatures of the ideas he wants to demolish. The book is full of Rose's authoritarian statements such as, "the neuronal systems that comprise the brain, unlike a computer, are radically indeterminant". Rose is certain of this because. "brains do not work with information in the computer sense, but with meaning". Rose is adament: "AI needs to know its place", he lectures that AI needs to, "show some humility in confronting that marvellous object of study, the brain." Rose's saving grace is that he can wax poetic: he calls the idea that a machine could be intelligent "an unholy Hegelian synthesis" of Cartesian Dualism and Behaviorism. Surprising that Rose knows the word humility, maybe he could practice what he preaches. It is not hard to forgive Rose for his limited view of AI, physics, philosophy, and the study of complex systems. It gets harder to forgive him for his hatchet job on fellow neuroscientists. I guess we have to take pity that his system, the chick, has not gotten as much press as Aplysia or Drosophila or even learning-defective gene knockout mice.

The heart of the book is okay...Rose describing the story of his work. Unfortunately, his odd mixture of reductionism/anti-reductionism paints him into a corner. In the end he is left to lament that he can not imagine a biochemical mechanism for memory recall. Gee, Steve, what if memory recall is not built on the same kind of biochemical processes that account for memory storage? What if complex neural nets are able to respond to new imputs with what we call "memory recall" in ways that only depend on quick receptor and ion channel-mediated electrical activity? But no, that is not possible for Rose since he has condemned the idea of neural networks as being to deterministic to hold any human meaning. One hopes Rose is joking when he throws around ideas such as the possiblity that procedural memory is more stable because it might be partly in muscle and sinue. Maybe we should just take this as a warning that if scientists are going to try to write books for the general public, they should do so either before getting too old or else only after finding a good editor! 

Go to John's Book Page.

Go to John's Home Page.