Is creativity computable?

P. N. Johnson-Laird
DOI: https://doi.org/10.1080/20445911.2024.2313354
2024-02-16
Journal of Cognitive Psychology
Abstract:Is creativity based on the brain's computations? Some scientists say, No, because of Gödel's argument that humans can grasp the truth of sentences that formal algorithms cannot. A taxonomy implies that only three sorts of algorithm could be creative: neo-Darwinian algorithms, neo-Lamarckian ones and Multi-level ones. These algorithms accommodate "breaking the rules", and they have the computational power—the working memory for intermediate results—needed for natural language and for the creation of informal algorithms. A programme illustrates how jazz musicians can improvise with the minimum of power—no use of working memory—because they have access to a long-term memory for chord sequences created using a similar computational power to natural language. A programme that reverse engineers Boolean electrical circuits uses a verification algorithm to check that they are correct. And programmes that verify sentences obviate Gödel's argument. Hence, whether creativity is computable is an open question.
psychology, experimental
What problem does this paper attempt to address?