view page | download magazine

Exchanging music—what could be a simpler or better way to get to know someone? As a kind of social experiment, I exchanged a mix CD of jazz, pop, and a little classical music with someone I had never met in person. In return, she sent me a CD of rap, hip-hop, and dance music. “Well,” she wrote me, “we certainly do have different tastes.” I listened to the CD and enjoyed the experience—both the musical exposure as well as the chance to get to know someone new. (I even liked one of the Kanye West songs—a little.) We both got new music, in the strong sense of “new”—something that was completely different from what we’d normally listen to.

Beyond person-to-person exchanges, new online music recommendation services have rapidly accrued popularity in the past several years. Even without discussing actual music, the relative merits of Last.fm, iLike, Pandora and others provide the basis for many a lively music conversation. iLike features integration with Facebook and tends to draw those interested in advertising their tastes; Last.fm draws on iTunes frequency statistics to develop an impression of a listener’s taste; Pandora cross-references the user’s musical preferences with a massive database of (theoretically) musicological song characteristics ranging from “Renaissance modalities” to “boastin’ lyrics.” Naturally, every service provides some means of purchasing the music that it recommends; none of them take any significant step toward hiding the essentially commercial nature of their enterprises.

In what was imaginably a competitive move against these existing music recommenders, Apple incorporated its own “Genius” recommendation service into iTunes 8. The popularity of iTunes let Genius bypass the usual slow buildup of popularity and data. Instead, like the plant from “Little Shop of Horrors,” Genius went straight to iTunes users’ music libraries and immediately began devouring collection and playback information. “Genius’ elegance is indisputable: to build the largest database of music, draw information from the most popular music player.” Genius has both a sidebar with recommendations, and the ability to create playlists of similar music from an existing library—especially useful for those with mammoth collections that stretch into the multiple dozens of gigabytes. It makes the chaos of one’s own consumption habits more approachable.

This is Genius’ main strength: to connect the appropriate dots and forge connections that an individual user might not be able to. The other day, when I asked it to generate a playlist based on Wilco’s “A Shot in the Arm,” it directed my attention to a Neil Young album that I’d forgotten to play after a friend had given it to me. Likewise, I would never have discovered the terrific acoustic group Nickel Creek if it hadn’t been for a quick listen to a Genius recommendation for a Beatles song. Genius is far less successful with jazz—I fail to see a connection on any level between Fats Waller’s stride piano and Ornette Coleman’s infamously chaotic free jazz. Its worst performance by far, though, is with classical music, where Genius recommends unrelated, overplayed music, and in some cases even makes dead wrong suggestions. (Just because Céline Dion uses violins does not mean that she should be categorized with Brahms’s second symphony.)

But despite its successes, the extent to which Genius promotes the discovery of new tastes for music listeners is questionable. Genius’ function is to recommend music based on the summed collections of people who listen to music like yours—it aims to give you music like what you already listen to. Like an echo chamber, Genius and similar aggregate-taste recommendation services connect you with people who enjoy the same type of music that you do, reinforcing your tastes while creating a pleasant, minor sense of discovery when you encounter a new artist who creates music fundamentally similar to what you already like. Even as it claims to expose people to new and different music, Genius reinforces the boundaries of genre and time period. Like bumper cars, Genius gives us the illusion of collision while actually keeping us safe and isolated within our preexisting aesthetic bubbles.

Encouragement of this musical insularity not only discourages exploration of the unfamiliar and the development of truly new tastes, it also robs music of some of the magic of human contact. The music to which I feel the strongest connection is often the music that’s pinned to some person, moment, or place in my memory: Beethoven sonatas on the first CD that my piano teacher gave me, the Radiohead album a friend introduced me to in my sophomore year of high school, the Ben Harper album I listened to the first time I went to Europe. That’s not to say that using Genius stops people from talking to one another about music, or that it keeps friends from sharing music with one another. Rather, the expectation of personalized, familiar-but-new recommendations tends to make listeners parochialize internally, to close themselves off to contrasting musical opinions before a conversation even takes place. It exacerbates the worst xenophobic tendencies in all of us, entrenching us in suspicion of anything truly different and formed from a disparate life perspective. If music is representative of a basic human need for self-expression and for culture in general, then closing off our tolerance for new music is a bad sign indeed.

I’m probably as guilty of this as anyone else. Why go to the trouble of listening to something exotic and challengingly new when I have a well-worn mp3 file of a favorite album at my fingertips? Why listen to LCD Soundsystem or Kanye West when I almost always dislike techno and rap? I’d rather have my tastes reaffirmed than assaulted. Why should I sacrifice my limited listening time to an artist or a genre that I don’t know or even actively dislike?

Despite these objections, I haven’t stopped sharing music with other people, and neither have the vast majority of music listeners. Listening to music offers a means to explore our own identities as well as others’; it is an inherently social activity. The point is not to like everything, but to be open to anything at least once. I have no idea whether my exchange partner ended up liking any of the music I sent her; I know that I didn’t like everything that she sent me, although I now think twice before I dismiss her type of music offhandedly. And I’m convinced that this type of sharing will not cease any time soon. In the end, the concern is not the extinction of sharing, but the excessive veneration of listening within a personal comfort zone.

Musical tastes give a certain immediate level of insight into personalities and backgrounds that very little else can do. Far more illustrative and personal than a favorite color or food, music preferences are also more immediately accessible than books or movies. Music subtly indicates identity. It follows, then, that understanding music is one attempt at understanding identity. So when I log into Pandora (don’t take me the wrong way; I do still use these services) and see their claim, “It’s a new kind of radio—stations that play only music you like,” I take it as a vague sort of threat. The day when everyone has a station that plays only music they like is ultimately as dystopian as the day when everyone listens to the same music. It’s disturbing that this claim is Pandora’s selling point.