Probing the Persistent Question Marks

The puzzle of the persistent question marks (a phrase coined by Harvey in 1993) refers to the occurrence of residual learnable alleles in simulations of the Baldwin effect, which concerns interactions between learning and evolution. Explanations of the puzzle have focused on the role of fitness proportional selection pressures and of drift. We replicated the original 1987 model by Hinton and Nolan, extending the number of generations simulated until populations converged on stable genotypes, thus quantifying the number of residual question marks and the time to homogeneity (only one allele remaining in the population for each gene in a chromosome). Previous explanations of the residual question marks imply that algorithms that maintain strong selection pressure should not result in residual question marks. We tested this implication by simulating Hinton and Nolan’s Baldwin model replacing frequency proportional selection by tournament selection. Fewer residual question marks remained, but they were still present in many populations. We analyzed the relative factors that contribute to these effects in both fitness proportional and tournament selection runs. We conclude that homogeneous question marks are a significant factor in populations exhibiting the Baldwin effect in all the types of selection strategies tested in this study.