Major Contributions

In the following list I summarise what I regard as the major contributions of this work.

1.

The design and implementation of Cosmos, a Tierra-like artificial life platform.

2.

The description and analysis of a wide variety of experiments with Cosmos, constituting the fullest and most systematic investigation of the behaviour of such a system to have been reported. An important part of this investigation concerns the role of contingency in determining the course of evolution.

3.

The identification of a number of problems, some theoretical and some methodological, that limit the utility of Tierra-like platforms as tools for the scientific investigation of open-ended evolution in general, and of the evolution of life in particular.

4.

An analysis of the logic of reproduction in terms of more specific issues, with particular emphasis on processes of self-reproduction in the context of evolution. This includes topics such as: the superiority of genetic reproduction over reproduction by self-inspection; discussion of self-reproducing programs in Tierra-like platforms in terms of von Neumann's analysis of the logic of self-reproduction; issues relating to the explicit versus implicit encoding of the various components of von Neumann's genetic architecture; an emphasis on the role of phenotypes; the desirability of allowing unrestricted interactions between individuals to promote open-ended evolution; and discussion of the evolution of symbolic information and of fundamentally new measuring instruments, and how these may be achieved.

5.

Suggestions for how the approach to modelling open-ended evolution, and the evolution of life, may be improved. In particular, it is emphasised that careful consideration must be given not only to modelling individuals, but also to modelling the environments in which they exist, and the sorts of interactions allowed. Waddington's paradigm for an evolutionary process [Waddington 69] is suggested as a possible unifying framework within which to develop a better approach to the synthetic modelling of evolution and life.