Computable Aesthetics for Dance

Abstract

We discuss development of a universal conception of what (or if any) part of dance aesthetics is amenable to computation and if there is any, then how it can be deployed for new exciting development of computational resources for dance. Claims to universality, while objectively verifiable with data, have an underlying theme of positing universals of aesthetics in the practice of a classical discipline. It is natural to think that classical dance forms evolved with conscious expression of the initially unconscious, universal tenets of aesthetic experience of performers and audience. It is commonly claimed that such a universality binds the performers and audience together, as well as practitioners of various classical and folk dance forms across geography, time, and cultures. Classical forms derive their classicity from a conscious expression of such universals. However, the expression of universals itself is not so explicit as to be amenable to computation easily. If we want to harness the computing power unleashed by the recent developments in machine learning for helping experts in artistic creativity in dance, then we need such explicit expression and a certain clarity of separation between what is and what is not so expressible. In this work, we raise three questions:

• 1. Are there universals of aesthetics? We argue that there are.

• 2. Can such universals be identified? How? From particular to general or vice versa?

• 3. Once identified, can we make one or several computational models to classify and synthesize aesthetically attributed artwork and performances?

We discuss plausible answers to each of these questions and indicate beginnings of practical pursuit of the same.

Sangeeta Chakrabarty is a.k.a Sangeeta Jadhav.