The Effect of Repetition and Expertise on Liking and Complexity in Contemporary Music Valerio Velardo, *1 Mauro Vallati #2 * School of Music, Humanities and Media, University of Huddersfield, UK # School of Computing and Engineering, University of Huddersfield, UK 1 valerio.velardo@hud.ac.uk, 2 m.vallati@hud.ac.uk ABSTRACT Aesthetic perception of music has been extensively researched in the last decades. Numerous studies suggest that listeners find a piece of music more or less pleasant according to its complexity. Experimental results show that complexity and liking have different relationship according to the musical genre examined, and that these two variables are also affected by other factors such as familiarity to the music and expertise of the listener. Although previous experiments have examined several genres such as jazz, pop, rock and bluegrass, surprisingly, no study has focused on contemporary music. In this paper, we fill this gap by studying the relationships between complexity, liking, musical training and familiarity in the case of contemporary music. By analysing this genre – which is usually underrepresented in music cognition – it is possible to shed some light on the correlation between liking and complexity in the case of highly complex music. To obtain data, a multifactor experiment was designed in which both music experts and novices had to provide scores of subjective complexity and liking for four 30-second long excerpts of contemporary music with different degrees of complexity. Empirical results suggest that liking and complexity are negatively correlated in the case of contemporary music and that listeners’ expertise does not influence the perceived complexity of musical pieces, but it can significantly affect liking. This possibly indicates that experts have the musical knowledge needed to appreciate extremely complex music, while novices do not. I. INTRODUCTION Why does the majority of people listens to pop/rock music and not to contemporary music? At first glance, this question seems to be related to cultural habits only. Our society largely promotes pop music because of the huge profits it generates. Pop songs are everywhere on television and on the radio, and it is difficult for a person not to stumble upon the last hit. On the other hand, contemporary music is followed by a small niche of people and it is hardly aired even on classical music radio stations. However, what happens in our society might also be a reflection of basic cognitive processes related to the musical content of these two genres. Obviously, people listen to the music they like. Numerous studies (e.g, Vitz, 1966; Berlyne, 1971; Heyduk, 1975; Walker, 1980) suggest that listeners find a piece of music more or less pleasant based on its complexity. Listeners do not like music which is overly complicated, like most of contemporary music, because it is difficult to understand. On the other hand, listening to music which is too simple is boring. Therefore, people prefer music of intermediate levels of complexity like pop/rock music (Orr & Ohlsson, 2001). Several studies support this inverted-U hypothesis (Vitz, 1966; Crozier, 1974; North & Hargreaves, 1995; Orr & Ohlsson, 2001). Vitz (1966) found that the aesthetic assessment for a series of tones increases along with complexity until it starts declining when the melody becomes too complex. Likewise, North and Hargreaves (1995) discovered that pop songs of moderate complexity are preferred by listeners to songs that are perceived as too simple or too complex. However, there are some studies in the literature that contradict the inverted-U hypothesis. Russell (1982) and Smith and Melara (1990) found a negative correlation between liking and complexity. In both studies, simple musical excerpts were preferred to complex fragments. Although less frequent than the experiments which support the inverted-U hypothesis, these results might suggest that liking and complexity have different relationships depending on the musical style analysed and on other factors such as familiarity to the music proposed and level of expertise of the listener. Until now, a number of genres have been examined in controlled experiments. For instance, Orr and Ohlsson (2001) focused on jazz and bluegrass; North and Hargreaves used both pop songs (1995), and new-age music (1996). Both classical music and avant-garde jazz were tested by Hargreaves (1984). All studies which attempt to find a relationship between liking and complexity face a great challenge, i.e., defining musical complexity. Complexity is a fuzzy concept difficult to measure. To simplify this notion it is possible to divide complexity into two separated concepts: objective complexity and subjective complexity (Hargreaves, 1984). The former refers to the objective amount of complexity carried by a musical piece based on its properties. Previous studies measured objective complexity relying on the tools of information theory (Vitz, 1966), or by performing feature analysis (Steven & Latimer, 1991). Subjective complexity, on the other hand, is the amount of complexity experienced by people while they listen to a musical piece (Steven & Latimer, 1991). Subjective complexity is a function both of objective complexity and of the musical background of the listener. Regarding the musical background, a piece that is perceived as simple by a listener who grew up listening to Western music can be experienced as extremely complex by a listener who spent her life in an African musical environment. The opposite is obviously true as well. Although it is difficult to predict subjective complexity on a theoretical basis, there is a simple strategy to measure it: asking listeners to rate complexity in a scale while they listen to music (North & Hargreaves, 1995; Orr & Ohlsson, 2001). A major concern to address when studying the relationship between liking and complexity is the type of musical stimuli used during the experiments. For example, Vitz (1964; 1966) used computer-generated melodies and Smith and Malera