C REATING M USICAL C ADENCES VIA C ONCEPTUAL B LENDING : E MPIRICAL E VALUATION AND E NHANCEMENT OF A F ORMAL M ODEL A STERIOS Z ACHARAKIS , MAXIMOS K ALIAKATSOS -PAPAKOSTAS , C OSTAS T SOUGRAS ,&E MILIOS C AMBOUROPOULOS Aristotle University of Thessaloniki, Thessaloniki, Greece THE COGNITIVE THEORY OF CONCEPTUAL BLENDING may be employed to understand the way music becomes meaningful and, at the same time, it may form a basis for musical creativity per se. This work constitutes a case study whereby conceptual blending is used as a creative tool for inventing musical cadences. Specifically, the per- fect and the renaissance Phrygian cadential sequences are used as input spaces to a cadence blending system that produces various cadential blends based on musicolog- ical and blending optimality criteria. A selection of ‘‘novel’’ cadences is subject to empirical evaluation in order to gain a better understanding of perceptual relationships between cadences. Pairwise dissimilarity ratings between cadences are transformed into a percep- tual space and a verbal attribute magnitude estimation method on six descriptive axes (preference, originality, tension, closure, expectancy, and fit) is used to associate the dimensions of this space with descriptive qualities (closure and tension emerged as the most prominent qual- ities). The novel cadences generated by the computational blending system are mainly perceived as single-scope blends (i.e., blends where one input space is dominant), since categorical perception seems to play a significant role (especially in relation to the upward leading note move- ment). Insights into perceptual aspects of conceptual bending are presented and ramifications for developing sophisticated creative systems are discussed. Received: May 16, 2016, accepted May 1, 2017. Key words: conceptual blending, musical cadence, com- putational creativity, empirical evaluation, harmony perception N EW CONCEPTS MAY BE CREATED EITHER BY ‘‘exploring’’ previously unexplored regions of a given conceptual space (exploratory creativ- ity), transforming established concepts in novel ways (transformational creativity), or making associations between different conceptual spaces that share some structural relations (combinational creativity). Boden (2004) maintains that the latter (i.e., combinational creativity) has proved to be the hardest to describe for- mally. This paper explores aspects of combinational cre- ativity in the domain of music, and more specifically, the harmonic structure of music. Conceptual blending is a cognitive theory developed by Fauconnier and Turner (2003) whereby elements from diverse but structurally related mental spaces are ‘‘blended,’’ giving rise to new conceptual spaces that often possess new powerful interpretative properties and allowing better understanding of known concepts or the emergence of novel concepts altogether. Concep- tual blending theory is useful for explaining the cogni- tive process that humans undergo when engaged in creative acts, and is akin to Boden’s notion of combi- national creativity. A computational framework that extends Goguen’s formal approach (Goguen, 2006) has been developed in the context of the COINVENT (Con- cept Invention Theory) project (http://www.coinvent- project.eu) (Schorlemmer et al., 2014). According to this framework, two input spaces are described as sets of weighted properties and relations, and after their generic space is computed, the amalgamation process (Eppe et al., 2015) leads to the creation of consistent blends that are optimal according to some criteria relating to the blending process and to the knowledge domain of the modeled spaces (the amalgamation process poten- tially includes multiple ‘‘generalization paths,’’ leading to many different blends). With regard to music, conceptual blending has been predominantly theorized as the cross-domain integra- tion of music structural and extramusical domains such as text or image (e.g., Cook, 2001; Moore, 2012; Zbi- kowski, 2002, 2008). Additionally, it has been studied in the context of ‘‘musicogenic’’ meaning (Koelsch, 2013), which refers to physical, embodied, emotional, and personality-related responses to music; such studies include work on music and motion by Johnson and Lar- son (2003) or empirical studies on pitch perception and image schemata in children by Antovic ´ (2009, 2011). Finally, there have been studies that touch upon issues of structural mappings/blending between different spaces Music Perception, VOLUME 35, ISSUE 2, PP. 211–234, ISSN 0730-7829, ELECTRONIC ISSN 1533-8312.  2017 BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA ALL RIGHTS RESERVED. PLEASE DIRECT ALL REQUESTS FOR PERMISSION TO PHOTOCOPY OR REPRODUCE ARTICLE CONTENT THROUGH THE UNIVERSITY OF CALIFORNIA PRESS ’ S REPRINTS AND PERMISSIONS WEB PAGE, HTTP:// WWW. UCPRESS . EDU/ JOURNALS . PHP? P¼REPRINTS . DOI: https://doi.org/10.1525/ MP.2017.35.2.211 Cadence blending 211