An Alternative Approach to the Synthesis of Life

Tim Taylor
Institute of Perception, Action and Behaviour
Division of Informatics
University of Edinburgh

tim.taylor [at] ed.ac.uk

Abstract

This work concerns the creation of a computational environment that could support the origin and open-ended evolution of artificial life. It is argued that most previous attempts in this area, from John von Neumann to Tom Ray, and others (many of whom have doubtlessly been inspired by the work of Richard Dawkins) have placed too much emphasis on the careful engineering of individual organisms to act as seeds for an evolutionary process. In particular, these attempts often regard the process of self- replication as a major design issue; at the same time, they tend to pay little attention to the design of the environment and the way in which organisms interact with it.

An alternative, and very different, approach to the topic is presented, inspired in part by the theoretical work of Howard Pattee. In this approach, the ability of genotypes to self- replicate is taken for granted. Rather, the emphasis is much more on the environment and the processes that it supports; it is assumed that the environment has its own dynamics and self- organisational properties, and that genomes can "sculpt" these dynamics by supplying constraints. Genomes that generate local dynamics that promote their own stability and reproductive success will, by definition, survive, and thereby serve as the seeds of an evolutionary process. A simple model is introduced to exemplify these ideas. Initial results suggest that the appearance of genome-regulated self-stabilising dynamics (a major aspect of the origin of life) is a robust feature of the model. The model also has desirable properties for long-term open-ended evolution.


[Tim Taylor's publications] [Tim Taylor's homepage]