Daniel Saunders

Computer science only became established as a field in the 1950s, growing out of theoretical and practical research begun in the previous two decades. The field has exhibited immense creativity, ranging from innovative hardware such as the early mainframes to software breakthroughs such as programming languages and the Internet. Martin Gardner worried that "it would be a sad day if human beings, adjusting to the Computer Revolution, became so intellectually lazy that they lost their power of creative thinking" (Gardner, 1978, p. vi-viii). On the contrary, computers and the theory of computation have provided great opportunities for creative work. This chapter examines several key aspects of creativity in computer science, beginning with the question of how problems arise in computer science. We then discuss the use of analogies in solving key problems in the history of computer science. Our discussion in these sections is based on historical examples, but the following sections discuss the nature of creativity using information from a contemporary source, a set of interviews with practicing computer scientists collected by the Association of Computing Machinery’s on-line student magazine, Crossroads. We then provide a general comparison of creativity in computer science and in the natural sciences.

Some cultural evolution theorists claim that humans tend to imitate the traits of people who are similar to themselves—men tend to imitate men, women tend to imitate women, and so on. These theorists further suggest that selection has shaped human psychology to attend to similarities and weigh them when learning from others. The argument typically works like this: If similar people face similar problems, then learning from those people can ensure humans learn the most relevant information to solve problems they will face. Little formal evolutionary modelling has explored the conditions under which this argument is valid. This article develops a series of models to answer this question. The general insight is that the viability of the evolutionary argument depends largely on what we assume the function of social roles to be. If, as is the default view in the cultural evolution literature, social roles facilitate coordination, then the model is not very robust with respect to the initial conditions, parameter settings, or population structure. However, if social roles facilitate the division of labour, then similarity-biased learning evolves under a wide range of conditions. These results can improve our understanding of the origins of inequality. Some philosophers have proposed evolutionary bargaining models as potential explanations for inequality. These models make frequent use of similarity-biased learning assumptions. I suggest some ways to improve the research programme on bargaining models in light of these results.