Weasel Words, Genetic Algorithms and Coarse Acknowledgements

The ‘weasel words’ example by Richard Dawkins is based on a simple genetic algorithm and is aimed at demonstrating the likelihood of evolution due to natural selection and mutations. The algorithm rewards fine changes in single letters of a random sequence towards the desired phrase. Many traits in living things, however, exhibit irreducible complexities, which make fine changes such as substitutions of single amino-acids in protein chains unlikely to be rewarded by improved fitness. Improved fitness is often achievable only via coarse changes in the overall structure, and in many cases such changes will require moving through inferior individuals. A more realistic genetic algorithm demonstrates that the impact of coarse acknowledgments on the convergence speed is so significant that the process does not converge even within a million generations.