Review of our approach to modelling evolution

Evolution is seen as involving 2 or more populations with many individuals. There are the mechanisms of copying, adding variation, and selection that generate the second population from the first. Selection in this context means that some individuals in the original or parent population are copied more often than others, so that the second population is a 'biased' copy of the first population, preserving most of the characteristics, but slowly moving in the direction indicated by the selection criteria.

Evolution can be seen as a method for copying information from one population to the next. Variation and selection introduce a bias in the copying, so that the population 'improves' or 'learns' over successive generations.

DNA is the main type of information encoding typically associated with the copying process in evolution. The representation of innate skills has not yet been decoded, so we have to proceed from general constraints, with feasibility models, and engage in a lot of speculation.

For skills, the mechanics of the process of evolution, i.e. copying, variability, and selection, may also apply to other forms of information encoding and information copying.

• We shall look at mimicry with a parent to child apprenticeship as a second process that may involve similar mechanics of evolution to pass skills from the parent population to the child population.

We shall also speculate about within-individual evolution. For this theory and model, different skills are the individuals in the population. This theory uses the processes of evolution to account for learning from experimentation (using variations of the skilled actions).

We speculate that the brain contains a population of skills at any given time. This inventory of skills changes over time. In this theory, the population of skills at one age (say 3 years old) is the parent to the population of skills at a future age (say at 4 years old).

We assume that skills can be forgotten, i.e. that the information representation for the skill disappears from our brain or where ever it may be stored, if it is not used (like dynamic memory in computers).

• Skills are copied forward, i.e. from the present population to the next population.
• Skills have a higher probability of being copied when they have been used and when they have been successful (selection leading to learning).
• Skills, as stored in the brain, are represented in some compressed form. The higher the level of compression, the more likely it is that variation occurs in interpreting the stored description and generating the full detail of muscle control. Successful variations may well be noted and stored.

For each one of the above applications of the abstract theory of evolution, the information representation of skills is at the heart of the theory. In the next part we shall focus on this issue with a theory of an 'inner language' that represents skills, allows further compression, is useful for building on and reusing skill components. I see this 'inner language' as a programming language for the information processing required for applying skills to generate action.