Understanding and Quantifying Creativity in Lexical Composition

Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1246–1258, Seattle, Washington, USA, 18-21 October 2013. c⃝2013 Association for Computational Linguistics Understanding and Quantifying Creativity in Lexical Composition Polina Kuznetsova Jianfu Chen Department of Computer ScienceYejin Choi Stony Brook University Stony Brook, NY 11794-4400 {pkuznetsova,jianchen,ychoi}@cs.stonybrook.edu Abstract Why do certain combinations of words such as “disadvantageous peace” or “metal to the petal” appeal to our minds as interesting ex- pressions with a sense of creativity, while other phrases such as “quiet teenager”, or “geometrical base” not as much? We present statistical explorations to understand the char- acteristics of lexical compositions that give rise to the perception of being original, inter- esting, and at times even artistic. We first ex- amine various correlates of perceived creativ- ity based on information theoretic measures and the connotation of words, then present ex- periments based on supervised learning that give us further insights on how different as- pects of lexical composition collectively con- tribute to the perceived creativity. 1 Introduction An essential property of natural language is the gen- erative capacity that makes it possible for people to express indefinitely many thoughts through indef- initely many different ways of composing phrases and sentences (Chomsky, 1965). The possibility of novel, creative expressions never seems to exhaust. Various types of writers, such as novelists, journal- ists, movie script writers, and creatives in adver- tising, continue creating novel phrases and expres- sions that are original while befitting in expressing the desired meaning in the given situation. Consider unique phrases such as “geological split personal- ity”, or “intoxicating Shangri-La of shoes”,1 that 1Examples from New York Times articles in 2013. continue flowing into the online text drawing atten- tion from readers. Writers put significant effort in choosing the per- fect words in completing their compositions, as a well-chosen combination of words is impactful in readers’ minds for rendering the precise intended meaning, as well as stimulating an increased level of cognitive responses and attention. Metaphors in particular, one of the quintessential forms of lin- guistic creativity, have been discussed extensively by studies across multiple disciplines, e.g., Cog- nitive Science, Psychology, Linguistics, and Liter- ature (e.g., Lakoff and Johnson (1980), McCurry and Hayes (1992), Goatly (1997)). Moreover, re- cent studies based on fMRI begin to discover bio- logical evidences that support the impact of creative phrases on people’s minds. These studies report that unconventional metaphoric expressions elicit signif- icantly increased involvement of brain processing when compared against the effect of conventional metaphors or literal expressions (e.g., Mashal et al. (2007), Mashal et al. (2009)). Several linguistic elements, e.g., syntax, seman- tics, and pragmatics, are likely to be working to- gether in order to lead to the perception of creativ- ity. However, their underlying mechanisms by and large are yet to be investigated. In this paper, as a small step toward quantitative understanding of lin- guistic creativity, we present a focused study on lex- ical composition two content words. Being creative, by definition, implies qualities such as …

Content truncated to 500 words