Effects of Discourse Goals on the Process of Metaphor Production

Effects of Discourse Goals on the Process of Metaphor Production Akira Utsumi (utsumi@inf.uec.ac.jp) Kota Nakamura (kota@utm.inf.uec.ac.jp) Maki Sakamoto (sakamoto@inf.uec.ac.jp) Department of Informatics, The University of Electro-Communications 1-5-1, Chofugaoka, Chofushi, Tokyo 182-8585, Japan Abstract Only a few attempts have so far been made at exploring the process of metaphor production, although a large number of studies have addressed metaphor comprehension. Therefore, in this paper, we address the problem of how people generate metaphors or identify an apt vehicle for a given topic of me- taphors. Specifically, we examine how the process and prod- uct of metaphor production differ between two discourse goals of metaphor, namely an explanatory purpose (e.g., to clarify) and a literary purpose (e.g., to aesthetically pleasing). Exper- iment 1 analysed the metaphors (or vehicles) generated in the metaphor production task, and demonstrated that people iden- tified more prototypical exemplars of the property attributed to the topic as a vehicle for explanatory metaphors than for liter- ary metaphors. In addition, it was found that explanatory me- taphors were more apt and conventional, and had high topic- vehicle similarity than literary metaphors, while literary meta- phors were more familiar and imageable than explanatory me- taphors. Experiment 2 used a priming paradigm to assess the online availability of prototypical and less prototypical mem- bers of the topic property during metaphor production. The re- sult was that both prototypical and less prototypical members were activated in producing literary metaphors, while neither members were activated in the production of explanatory me- taphors. These findings indicate that the process of metaphor production is affected by discourse goals of metaphor, and sug- gest that only prototypical members of the category are rapidly searched for a vehicle during the production of explanatory metaphors, while both prototypical and less prototypical mem- bers are searched to generate literary metaphors. Keywords: Metaphor production; Discourse goal; Priming Introduction Metaphor has been a main topic of research in cognitive sci- ence, because metaphorical expressions are frequently ob- served in our everyday use of language. Hence, a large number of studies have been made on how people compre- hend metaphors (e.g., Bowdle & Gentner, 2005; Gibbs, 2008; Glucksberg, 2001; Utsumi, 2011). In contrast, only a few studies have addressed the process of metaphor production (for a notable exception, see, for example, Chiappe & Chi- appe, 2007; Katz, 1989). This paucity of research on me- taphor production is especially problematic, given that by its very nature a metaphor is an intentional, linguistic device em- ployed to convey ideas that might be otherwise inexpressible. To ameliorate this situation, in this paper, we empirically ex- plore the process of metaphor production. Previous studies on metaphor production analysed the products of metaphor production (i.e., metaphorical expres- sions produced) in terms of the qualitative dimensions of me- taphors and/or individual differences. Concerning the qualita- tive dimensions of metaphor products, Fainsilber and Ortony (1987) demonstrated that descriptions of emotional states contained more metaphorical language than did descriptions of actions. Katz (1989) examined the properties of metaphor vehicles by asking participants to choose, from a set of al- ternatives, a vehicle that completes a given sentence frame of science) as comprehensi- (e.g., Chemistry is the ble and apt metaphors. The result was that participants were likely to choose the vehicles that were moderately distant from the topic and referred to concrete domains. Concern- ing individual differences, it was demonstrated that the quan- tity and quality of metaphor products were affected by indi- vidual differences such as writing experience (experienced or novice) (Williams-Whitney, Mio, & Whitney, 1992), reason- ing and imagery ability (Katz, 1989), gender (male or female) (Hussey & Katz, 2006), and working memory capacity (Chi- appe & Chiappe, 2007). For example, Chiappe and Chiappe (2007) demonstrated that people with high working memory capacity produced more apt metaphors than did low capacity individuals. Although shedding light on the specific aspects of meta- phor production, these studies did not address one important aspect of metaphor production, namely discourse goals of metaphor. Because metaphors (and other figurative language) are intentionally used to accomplish certain communication goals (Roberts & Kreuz, 1994), it is obviously crucial to ex- plore the effects of discourse goals on the process of meta- phor production. Discourse goals that are accomplished by the use of metaphorical expressions can be classified broadly into two classes: explanatory and literary purposes (Steen, 1994; Utsumi, 2005). These two goals are quite different and sometimes incompatible with each other; explanatory meta- phors are used to clarify certain properties of the topic, while literary metaphors are used to evoke an aesthetically pleas- ing feeling by enriching the meanings conveyed by the meta- phors. It naturally follows that discourse goals are likely to affect the process of generating metaphors, or more specifi- cally choosing the vehicles of metaphors. This paper aims at examining how the process and product of metaphor produc- tion differ between these two discourse goals. In metaphor production, people often have in mind a topic that they want to express and some properties that they intend to attribute to the topic. They must identify an appropriate or apt vehicle to convey the intended meaning (i.e., the in- formation that the topic has the property). Hence, metaphor production essentially involves the process of searching for or retrieving an apt vehicle. According to the categorization (or attributive category) theory of metaphor (Glucksberg, 2001; Glucksberg & Keysar, 1990), apt vehicles must not only have the intended property but also be a prototypical exemplar of