In recent years there have been a growing number of attempts to integrate theories of self-organisation with evolutionary theory. Stuart Kauffman, professor at the Santa Fe Institute, argues that the laws of physics generate general principles of self-organisation which must be considered when studying the evolution of life (although he believes that science does not yet have an adequate conceptual or mathematical framework to deal with such issues) [Kauffman 93]. He says on the subject:
``The building blocks of life at a variety of levels from molecules to cells to tissues to organisms are precisely the robust, self-organized, and emergent properties of the way the world works. If selection merely molds further the stable properties of its building blocks, the emergent lawful order exhibited by such systems will persist in organisms. The spontaneous order will shine through, whatever selection's further siftings.
Can selection have reached beyond the spontaneous order of its building blocks? Perhaps. But we do not know how far. The more rare and improbable the forms that selection seeks, the less typical and robust they are and the stronger will be the pressure of mutations to revert to what is typical and robust. That natural order, we may suspect, will indeed shine through.'' [Kauffman 95] (p.189).
Many have argued that the developmental processes of organisms also impose significant constraints upon organismal design (e.g. [Wolpert 93], [Goodwin 94], [Raff 97]). Note that the nature of these constraints is subtly different to the general principles of self-organisation proposed by Kauffman and others; developmental constraints are somewhat arbitrary with respect to underlying physical laws, but they are nevertheless real constraints for organisms employing any kind of developmental process. Maynard Smith says on the issue:
``For 50 years to my knowledge, and maybe for much longer than that, people have been saying that our ideas about evolution will be transformed when we have an adequate theory of development.'' [Maynard Smith 96].
The distinguished geneticist C.H. Waddington remarked that:
``The hereditary differences which arise in animals are not quite random, like the differences between two heaps of bricks. They result from changes in orderly systems of development, and each new variety has an order of its own, maybe less, but sometimes more, complex than that of the original form from which it was developed.'' [Waddington 57] (p.7).
Developmental biologist Lewis Wolpert argues that the self-organisational and feedback mechanisms operating during the development of an embryo to an adult organism radically restrict the space of morphologies that evolution is free to explore. In his book on animal development, The Triumph of the Embryo, Wolpert says:
``Developmental mechanisms, together with their genetic control, put a severe constraint on the evolution of animal form. It is not selective pressures that have kept the basic pattern of the vertebrate arm the same, but the fact that altering the basic pattern is almost impossible. Therefore, not all imaginable animals are possible. Any theory of evolution must incorporate an appreciation of developmental mechanisms.'' [Wolpert 93] (p.195).
There are many arguments over the relative importance of the spontaneous order inherent in the physical world, of the constraints imposed by developmental processes, and of natural selection in determining morphology (e.g. [Maynard Smith 86] p.43), but all these factors clearly have some part to play in the process. These considerations indicate that we cannot understand life purely through studying the genes, but we must also think about their interactions with the physical medium in which they are embedded. Richard Belew and Melanie Mitchell warn that ``common metaphors for the genes (e.g., programs, blueprints, etc.) as the information-carrying component in evolution all rest on an inappropriate `preformationist' view of information, as if information `... exists before its utilization or expression.' '' [Belew & Mitchell 96] (p.14, original emphasis).2.4
Another important point about evolution taking place in a medium with some inherent self-organisational properties is that, although the possible forms achievable by evolution may be restricted, the `instructions' for making those forms that are achievable can be compactly described on the organism's genome. In other words, the genome does not have to encode information about every aspect of the adult organism's design, because some features will just fall into place `for free' as the developmental process unfolds, due to the self-organisational properties of the constituent matter.