Reviewing, Creativity, and Algorithmic Information Theory

We connect critical review and analysis of creative objects to a recent domain-independent creativity assessment framework by Mondol and Brown (2021a; 2021b). Reviewing is interesting for at least three reasons. Reviews are time- and space-limited, unlike other tasks. Reviews are a creative task about creative tasks, and that meta-creativity is interesting to consider theoretically. And reviews cause communication and learning; the various actors (the primary creator, the reviewer, and the reader) interact in complex ways. We show how Mondol and Brown’s framework connects to the process of review, and show how topics like summarization, contextualization and learning fit within an algorithmic information theory frame. We also give some interesting examples, such as analysis of conceptual art and concert reviews, as computation tasks. We finish by showing that (as is often true of algorithmic information theory ideas) it is hard to fulfill our objectives with practical systems, due to uncomputability or intractibility issues