Random and better-than-random processes in biological and social evolution
University of Glasgow
Last modified: June 30, 2007
Since the work of Darwin, evolution has traditionally been considered as acting upon organisms through processes of random mutation and natural selection. With the development of genetics and neo-darwinism, biological evolution has been conceived of as acting through mutations and the introduction of variation at the level of the genome (DNA), and natural selection at the level of the phenome (macroscopic characteristics). Variation has been considered to be introduced randomly with respect to function, with only a very small proportion of mutations leading to increased fitness.
In contrast, human actions have traditionally been considered to be highly goal-directed, and thus better than random. This contrast has led to concerns about the validity of modelling social systems using biological evolution, an issue of relevance to complex systems approaches to understanding social and biological systems. This paper presents recent theories and evidence from biology that suggest that the introduction of variation in biological systems may be better than random, as well as from psychology that indicate that human actions are often less goal-directed than they often appear. Several alternative conceptions that may help us to reconcile the two types of system are discussed. These include the differences between the two as quantitative rather than qualitative, the notion that both types of system involve random and non-random processes, and alternative understandings of prediction and learning. The implications of these conceptions, including the importance of the level of randomness in determining the efficiency of the evolutionary process, are also discussed.