book: Kevin Kelly 14 Apr 1995

Out of Control

The New Biology of Machines, Social Systems, and the Economic World.

An accessible and entertaining explanation of why the coming years will probably be the Age of Biology -- particularly evolution and ethology -- and what this will mean to almost every aspect of our society.

Out of Control is a book about possibilities. A book about the future. Wired editor Kevin Kelly sketches a future in which the differences between natural and artificial systems have ceased to be fundamental. A future in which the natural has become more artificial and the artificial more natural: neo-biological civilisation.

Kelly's vision of the future is based on the concept of parallel systems. Parallel systems are formed by a large number of relatively simple units, each connected with a large number of other units in the system. There is no central control mechanism: each unit is autonomous, as it were. Because they process information in parallel, such systems might be said to possess a non-linear causality. Examples of parallel systems are an ant colony, a flock of birds, the human brain, neural networks, the market economy and (indeed) Internet. The counterpart of the parallel system is the serial system. These systems process information serially, giving rise to a linear causality. A well-known example is the computer with its Von Neumann architecture. Most mechanical systems are also serial systems.

Parallel systems (Kelly calls them distributed systems and, more suggestively, swarm systems, hive minds and vivisystems) have a number of advantages over serial systems. The large number of simple units and parallel information processing makes these systems (reasonably) resistant to disruption and damage. Damage to a (limited) number of units does not necessarily influence the system's functioning. The death of a certain number of ants in an ant colony or of a certain number of brain cells need not influence the functioning of those systems as a whole. Parallel processing also creates a high level of adaptability to sharply changing circumstances or even completely new ones. A good example might be the capabilities of neural networks. It is precisely their serial construction that makes serial systems much more vulnerable and less flexible. The removal of one link in the serial chain is usually enough to stop the entire system from functioning. A disadvantage of parallel systems is that their functioning is much less transparent, due to the intensive contact between the units and parallel information processing. Another disadvantage of great importance to Kelly is that there is no central control in parallel systems. Parallel systems can be guided, but only as a shepherd guides his sheep. To use Kelly's term, they are swarm systems. Think of the flock of birds that suddenly changes direction. It is not possible, as in serial systems, to send each unit of the system in a certain direction; it is impossible to have direct control of all units. Kelly calls the system 'out of control'.

According to Kelly, lack of central control is not a disadvantage. Parallel systems are not unguided systems. Their capacity for adaptation makes them capable of surviving in all sort of different situations. While there is no 'pre-meditated plan' for solving a certain problem, parallel systems are capable of finding 'solutions' to new situations. There is no central control or guidance that creates the solution. No single unit of the system has control, but the system as a whole is capable of 'self-control'. Kelly mentions the emergence of control. Just as intelligence is an emergent characteristic of the functioning of the brain, so control can be an emergent characteristic of different kinds of parallel systems. Control then becomes the same as (successful) adaptation, though Kelly does not make this claim so explicitly. To Kelly, 'out of control' actually means 'without a central control mechanism'.

Kelly is a materialist. While 'life' is traditionally regarded as a characteristic with which only natural organisms are endowed, Kelly takes quite another view. I take the view that life is a nonspiritual, almost mathematical property that can emerge from networklike arrangements of matter, to which he then adds life can be copied from living bodies as a delicate structure of information (spirit or gene?) and implanted in new lifeless bodies, whether they are of organic parts or machine parts. If both organic, natural systems and artificial systems can be similar parallel systems, then both can possess the same emergent characteristics, he argues. According to Kelly, the differences between the natural and the artificial are not fundamental. He foresees a blurring of the borders between the two areas. Machines can and will live; natural systems can and will become more artificial.

Kelly's future vision takes its final form when he treats on the mechanism he believes will determine the development of both natural and artificial systems in the future: evolution. Darwin has provided us with the key to the concept of the development of life forms out of other life forms. Evolution also provides an explanation for the creation of life from dead matter. Kelly emphasises repeatedly that this is a step-by-step process of construction from the bottom up. Evolution is a parallel process. Unlike many traditional creation myths, there is no master plan that is carried out during the course of evolution, no design, no top--down structure.

Evolution is out of control. In light of his materialistic viewpoint, Kelly's view that evolution does not apply exclusively to biological systems is not surprising. He believes that the multiformity of nature will develop among artificial systems eventually. According to Kelly, the first steps have already been taken -- Tom Ray's evolving computer programs (including the development of parasites), Danny Hillis' evolution simulation on parallel computers, Gerald Joyce's experiments with rna strings; natural evolution mechanisms seem also to recur in an artificially constructed environment. This does not surprise Kelly: nature contains the best solutions to complex problems. We continue to be astonished by the adaptive capacity and flexibility of systems structured from the bottom up (molecules, organisms, colonies, eco-systems, etc.). He quotes Danny Hillis, with whom he agrees: There are only two ways we know of to make extremely complicated things. One is by engineering, and the other is evolution. And of the two, evolution will take the more complex. According to Kelly, nature teaches the best way to deal with complex problems. Thus, we must structure our artificial systems from the bottom up and allow them to evolve to attain an optimal solution. Natural processes, especially evolution, are exemplary. He believes that the natural and the artificial world will follow the laws of nature. The border between both worlds will blur. The future holds a neo-biological civilisation.

The consequence of the creation of a neo-biological civilisation is the loss of control. When artificial systems acquire a bottom up structure and the ability to develop themselves with increasing independence, these systems will become increasingly independent from us. That is the price we must pay for progress. Kelly considers it a fair one. Not only will the best solutions for complex problems emerge, new forms will also emerge, just as nature continually produces new species and solutions. He envisions mutating buildings, living polymers, evolving software, swarms of small, cheap robots, like the ants in an ant colony... The increase in diversity, the possibility of the creation of new forms, artificial or natural, the open-ended universe: Kelly views the our future with hope.

We are gods. But not traditional gods. We are capable of creating much, perhaps even our own successors. But, according to Kelly, we mustn't attempt to maintain control over that which we create. We will have to learn to accept our new role.

Like every vision of the future, Kelly's is speculative. Not all of his claims are supported by fact. Some are debatable. For example, one wonders if evolution is indeed the best mechanism for solving complex problems. That may be true for the moment, but does not mean that no better mechanism will emerge in the future, considering that evolution does not produce (optimal) results. This would undermine Kelly's conviction that natural processes form the guideline for future civilisation.

The engaging thing about Kelly is that he combines a clear style with openness of thought. He continually poses questions and is not content with accepted theories. It is also an erudite book. Kelly is at home in a variety of areas. His vision is inspiring, certainly for those who share his materialistic view of the world. And hopeful.

Enlarge

Cover of "Out of Control" -

English

hardcover

0-201-57793-3

$28