Musings of a Genre-Crossing Man
Well, not really my life story, or some version of it. This posting acts as more of a brief reflection on the importance of validation for one’s ideas.
I sometimes wonder if my interest in similarity of music from different genres, the focus of my doctoral research, which implicitly requires an “objective” perspective, is just based on a very personal agenda. A single-minded attempt to break down boundaries among musics that follow different conventions. After all, my interest in classical music reblossomed during adolescence, when “everyone else” was listening to popular music, and when musical tastes become especially important as a signifier of social identity. While I had thought about it off-and-on for a number of years, a paper on which I was a collaborator served as an epiphany. It made me start thinking about the marginalization of classical music (despite it being the stereotypical territory of wealthy white males) from the perspective of library and information science, with a focus on the direct and indirect privileging of genre in various systems that facilitate categorization and recommendation.
Fortunately, I’ve found that my topic elicits interest from people who best describe their musical interests as being “all over the map.” Just last week, I gave a guest talk to three sections of an online searching course. It focused broadly on “non-text” information initially (photos, still images, sounds, and the like), followed by a more in-depth discussion of my pet topic. As has been typical, a few students told me that they enjoyed my presentation, also talking about some of their own experiences with music. For the presentation, I start discussing my research by giving students a “random” listing of genres, asking students to name some of their traits. The integrity of those categories almost invariably falls apart within a few minutes. Aucoutruier and Pachet (2003) mention the inconsistency of taxonomies for genres:
Pachet and Cazaly (2000) compares 3 Internet genre taxonomies: allmusic.com (531 genres), amazon.com (719 genres) and mp3.com (430 genres). Results show that there is no consensus in the name used in these classifications: only 70 words are common to the three taxonomies. More importantly, there is no shared structure: among these common words, even largely used terms like “Rock” or “Pop” do not have common definitions, i.e., they do not denote the same set of songs. Finally, the location of these taxons (i.e., nodes) in the hierarchy differs from one taxonomy to the other. Semantic interoperability of “natural” genre taxonomies is clearly a dream.
So if the experts can’t agree, what’s that mean for genre-based categorization… or even for recommender systems, whose algorithms indirectly perpetuate notions of similarity based on genre? From the broader perspective of information seeking, we sometimes don’t know what we don’t know. In the case of music, we might “know” we prefer music from a specific genre. On the other hand, we might conflate that perception with other musical traits we actually like. They may include key, tempo, melodic contour (or “shape”), and so on… or at least what such traits might signify culturally (see Downie 2003 and Gabrielsson 2009 for perspectives from both music information retrieval and music psychology). To top it off, what we desire from music can vary over time, making recommender systems seem too superficial and market-driven to provide more adventurous and nuanced suggestions.
Discussion about cross-genre similarity also tends to appear piecemeal, most especially when discussing similarities between classical and popular music (with jazz occupying some kind of middle ground). So it’s always good to find some solid evidence of overlap beyond my own idiosyncratic and highly personal connections. Some resources are listed on a supplemental page of this blog. What prompted this posting, however, recently appeared on the blog Molly Makes Music, authored by an aspiring opera singer. In a posting entitled Are Classical and Pop Music Fundamentally the Same Thing?, Molly considers just that possibility. Somewhat related to the traits and facets listed by Downie and Gabrielsson, she states, “When it comes down to it, all any music is is just the careful arrangement of melodies, chords, and rhythms.” Furthermore, Molly’s most recent posting actually builds on one from a few days before, which tries to unpack the highly imperfect and inconsistent definition(s) of classical music. Join in the conversation over there!
As a bonus, I had a long chat just today with someone else in my Ph.D. program. As a part of our conversation, he gave me renewed encouragement in pursuing my topic, citing its relatively broad appeal… at least in relation to topics that some pursue for either self-aggrandizement or to just have “something to research” (even if they find it deadly dull).
What I’m researching can sometimes seem like navel gazing, at least based on my experiences as “the weird kid in my small rural hometown who listened to classical music.” With reminders that others share an interest in cross-genre similarities, along with the my friend’s pep talk, I don’t feel quite so alone.