If nothing else, this book should provide motivation for a few more all-nighters by the artificial intelligensia. Dyson suggests that since we are all bozos when it comes to knowing how to construct human-like intelligence in a box, me might as well just sit back and wait for it to spontaneously evolve on the internet. Thanks, George, I'll put that on my "To Do" list once I have immortality and can afford to wait around a few millenia.

Read this book for some little known facts about the origins of the computer age. Go here for some online material from Dyson.

Dyson's book starts with Thomas Hobbes' 17th century exploration of the "metaphysical vacuum" between the intelligence of God and the intelligence of man. There is room here for expanded treatment of how other intellectuals have tried to deal with this "vacuum". For example, a good source of information is Birth of a Worldview by Robert Doran. Doran deals with some of the contortions that early Christians went through in trying to decide on a theory of intermediates between God and man. By the time Hobbes arrived on the scene, Western Europe had in place a 1,500 year history of conflict over this issue. There were certain philosophers from Plato to Descates who the Church had learned how to deal with.....the establishment did not need someone like Hobbes coming late onto the scene and pointing out that Plato's Ideal Forms and Descates' mind/body dualism are silly ideas.

Unfortunately, Dyson pollutes his first chapter by inviting in a modern silly idea: emergence as an alternative to reductionism. Dyson starts well,  correctly stating that "emergent behavior" simply means behavior that is hard to predict, that is, talk of "emergence" is talk motivated by one's  epistemological deficits. However, Dyson quickly turns to the pop philosophy of Danny Hillis as an authoritative source to support the bogus claim that emergent phenomena imply "the impossibility of a reductionistic explanation of thought". Thus are we set on the course of the book by which Dyson maons reapeatedly: Gee, the mind and intelligence are mysterious, maybe they will just emerge spontaneously out of a wild and crazy global computer network. I ask: who really wants a fog within which we can hide the soul? Answer: the same vested interests that persecuted Hobbes.

Dyson's Chapter 2, "Darwin Among the Machines" centers on Samul Butler, a man who most people will know of either through his novel, Erewhon, or through the conflict Butler had with Darwin over the theory of evolution as mentioned in the Appendix to Darwin's autobiography. Dyson gives a one line tip of the hat to Aristotle and Lucretius before continuing on from Hobbes to Butler and extending the theme of the idea of spontaneous emergence of intelligence within machines.

After dealing with the history of Butler, Dyson jumps ahead a century to Freeman Dyson's Origins of Life. There is a long history of physical scientists crossing-over disciplinary boundaries in order to try to tell biologists some nugget of wisdom.....a classic example is the attempt of Lord Kelvin's to prove Darwinism wrong by calculating that the temperature of the Earth would be much lower than it is if the Earth were as old as Darwin suggested. Of course, Darwin published underestimates of the age of the Earth because he feared the ridicule of such physical scientists who were themselves wrong. The movement of ideas and methods between the various specialized subdisciplines of science is very important. It is natural that methodologies developed by physical scientists for studying nonliving systems should be applied to the study of life. However, it is amazing that most physical scientists find it so hard to realize and admit that they do not know their biology.

Another well-known example is Erwin Schroedinger's little book, "What is Life?", which contains nothing in the way of biological insight and is a tough read for a biologist (the book is an interesting exposition of the "discovery" by a physicist that while physicists were totally unable to explain "genetic memory" using the statistical molecular dynamics of gasses, the quantum theory of covalent bonds was just what is needed to make sense of the stability of genes.....put into English, this just means that stable molecular structures are required for life, an "insight" that did not need to be explained to biologists who were already intent on explaining life in terms of the complex molecules found inside living organisms), but it was influential in attracting physical scientists like Francis Crick to biology (see Crick's autobiographical What Mad Pursuit.)

What was the origin of life on Earth? Freeman Dyson's theory is that metabolism came first and replicating genes came second. Anyone interested in the issue of the role of molecular replication in life should read Stu Kauffman's The Origins of Order and Dug Hofstadter's Godel, Escher, Bach. Kauffman correctly argues that life is not as improbable as most physical scientists (including F. Dyson) have always feared. Most theorists of the origin of life have centered their attention on the idea of an "RNA world" as the most likely chemical origin of life. RNA molecules can function both as genetic molecules and as enzymes. Both protein and DNA may have been later additions that "fine tune" life. There is much interest in the possibility that RNA might be able to achieve its own molecular self-replication. Complementary nucleotide base-pairing is the solution to the problem of gene replication that is used by all existing forms of life on Earth. Of course, it is easy to imagine forms of life that would not use this trick. In any case, at some point this trick was developed, so life as we know it required the "inventions" of nucleic acid replication. Most biologists interested in the origin of life are content to work on the problem of how nucleic acid replication arose. This is a key problem in the origin of life. People like Freeman Dyson should not be surprised when most biologists ignore other issues such as the possibility of a "protein metabolism world". It is the intuition of scientists who are familiar with life that the origin of nucleic acid replication is an important and approachable problem. And sure, much of what gets done in biology is done just because it is trendy and because it can be done. It is much easier to work with nucleic acid replication than protein "statistical reproduction." You really cannot fault people who live in a publish or perish world for working on a trendy theory for which there are accessible experiements to be done. Yes, biology has become replication-centric in this century. Yes there is much to say about life that does not involve replication. We would be fools to balk at calling a robot with humal-level intelligence "alive" just because it could not reproduce itself.  Many biological organisms are "drones" that cannot themselves reproduce, yet we call them "alive".

This book is full of odds and ends like Dyson telling us about his imagining of thoughts in trees and the larks of early computer scientists such as driving home backwards after parties. It seems that ever since The Double Helix, science historians know what the key to popular history of science is. Dyson does not seem satisfied with the debate between mind/brain monists and dualist.....introducing his trialism of life, nature, and machine. I fail to see the point of such a world view, particularly since Dyson himself spends so much effort pointing out the similarities between complex adaptive systems that fall into each of these three "categories".

As you near the end of this book, Dyson keeps repeating this mantra: "intelligence arises from interacting systems." What is the alternative to this "nugget of wisdom"? That intelligence arises from thin air or from non-physical "mind-stuff" or from systems with no moving parts? Those philosophers and theologians who are anti-materialists and anti-reductionists and prefer mind/brain dualism are unlikely to be converted by computer scientists who continue to promise machine intelligence and never quite deliver.

A key issue is: how can we build an intelligent machine? Dyson seems to hope that machine intelligence will simply emerge spontaneously. Dyson seems to fear that we are too stupid to figure out how biological brains produce minds. My prediction is that biologists will figure out how brain molecules and cells produce mind and we will then be able to apply that understanding of biological intelligence to the problem of making machine intelligence. Computer scientists who ignore and fail to cooperate with neuroscientists are simply making their task (the construction of machine intelligence) harder than it should be. Dyson seems satisfied with the idea that all living organisms on Earth, particularly humans, evolved within an evironment, the "natural" world, that biological brains "never quite comprehend". According to Dyson, it is fitting that we should now replace the "natural" world with an "artificial" world of man-made machines which we will never quite comprehend. "We have traded one jungle for another." Dyson tells us that this "trade" is a source of sanity and hope. Hey, if we evolved as ignorant savages, why shouldn't we be happy to live in what ever information age jungle the likes of folks like Bill Gates manage to subject us to? This image of the future may satisfy Dyson, but I reject it. I prefer the humanitarian vision of people like E. O. Wilson, Carl Sagan, and Francis Crick; not coincidentally, people more in tune with the realities of biology than our computer scientists who are overly dominated by the limited mechanical concepts of physical science. Of course, as we reach the end of this century (still 1999 as I write this) some physical scientists are starting to face the complexity of life: a good example is Murray Gell-Mann. About 1,000,000 years ago the human brain went down an evolutionary path that made it devoted to being a universal modeling machine, capable of allowing humans to survive by learning complex behaviors that are suited to a complex and changing environment. The human intellect, aided by extrasomatic knowledge bases and our mechanical computing devices will be able to understand itself and through that understanding we will be able to rationally construct more intelligent machines. Some branches of humanity may decide to stick with the bliss of ignorance, but the rest of us will move on to bigger and better things.

One last issue: Dyson is impressed by the neuroevolutionary ideas of Bill Calvin. I fear that this is a case of one enthusiastic poularizer of science putting words and ideas into another enthusiastic popularizer of science. By the time all of the metaphors and over-simplifications settle into place, a complex set of issues has become lost in feel-good slogans. Human intelligence is not the result of simply getting rid of "bad connections" by way of a mechanism like natural selection to get rid of "bad mutations". Exaclty what is the basis of human intelligence is still being worked out, but biologists know that there are "instructionistic" as well as "selectionistic" processes at work. As an antidote to Dyson's imbalanced perspective, go here. For a more solid statement of evolutionary processes in mind, see Gerald Edelman's work; it is more difficult reading than Calvin's popularizations of the subject, but more true to the real complexities of the human brain.

Go to John's Book Page.

Go to John's Home Page.