The fuzzy notion of “complexity” is pervasive in various areas of psychology. It is unsurprisingly a pillar of the new (Bayesian) paradigm approach of randomness perception, but also appears in the study of language, reasoning, in neuropsychology, or even in social psychology. The formal notion of complexity is made precise within com- puter science under the name of algorithmic (or Kolmogorov-Chaitin) complexity. Unfortunately, algorithmic complexity is uncomputable. However, recent developments in computer science have given rise to a reliable approximation of algorithmic complexity for short strings of symbols. We will first present a brief overview of the theoretical basis of this algorithmic complexity for short strings (ACSS), now implemented as an R-package, and describe several recent applications of ACSS across various subject matter within psychology. More specifically, we will ar- gue that algorithmic complexity fills in a gap in the Bayesian account of randomness perception.


Nicolas Gauvrit

Integrative Individual Cognitive Science e-session

Photos by : David Rytell