Exploring the potential of concept associations for the creative generation of linguistic artifacts: A case study with riddles and rhetorical figures

Abstract

Automatic generation of linguistic artifacts is a problem that has been sporadically tackled over the years. The main goal of this paper is to explore how concept associations can be useful from a computational creativity point of view to generate some of these artifacts. We present an approach where finding associations between concepts that would not usually be considered as related (for example life and politics or diamond and concrete) could be the seed for the generation of creative and surprising linguistic artifacts such as rhetorical figures (life is like politics) and riddles (what is as hard as concrete?). Human volunteers evaluated the quality and appropriateness of the generated figures and riddles, and the results show that the concept associations obtained are useful for producing these kinds of creative artifacts.