A large sample of people was given forty-eight recorded songs to listen to by different artists. The sample was divided into two main groups. One was a “social” group in that the participants were able to see how many downloads the songs were getting. The other wasn't and thus its members were unaware of any emerging consensus. In the latter group (uninfluenced) tastes were evenly spread. In the former (influenced) the difference between best and worst songs was substantial.

To further their experiment the researchers divided the "social" group into eight sub-groups each of which had access to their own sub-group’s download numbers but not those of the other seven. Within each of these smaller groups a particular song emerged as a distinct favourite while another one bombed yet none of them returned the same best or worst song.

These results didn’t surprise me. I've thought for long enough that beyond a certain threshold of talent the popularity of music is determined by factors other than the music itself with peer influences being a crucial element. In broader terms value is determined by context as much as by content. Success in music is less due, as is commonly understood, to some innate quality in the work but more to the complex social, commercial and cultural factors that surround music-making.

Saying this out loud usually draws a blank. Perhaps that’s because it undermines the mythology of music. It's a more agreeable prospect to believe in the “essence” of a work; that the gifted creator is tapping into something transcendent; that a preternatural force is at work in the piece and your well-honed sensor is suitably tuned to discern it. The myth speaks better to the emotion felt from music than to the considerably more mundane aspects of appreciation such as liking something just because others do.

The experiment points to a consensus theory of quality. It implies that good music is simply music that gains sufficient numbers of followers. If it does so for a long period toward posterity then it can be said to be truly good. When you live in the woods like I do, holding fairly independent views on many things, it’s heartening to come across supportive findings like these.

Later:

The research was led by sociologist Duncan Watts titled Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market. Again, for me it demonstrated the importance of social factors in determining taste before anything else like intrinsic qualities in the music itself. The researchers were more concerned with how that social influence actually worked, rejecting the convention (such as in Malcolm Gladwell’s “Tipping Point”) that prominent individuals (influentials) were the crucial factor. Their argument was that cultural success tends to be arbitrary and subject to broader susceptibilities in society as a whole. I agree, not forgetting the role of the elusive serendipity which though difficult to analyse is easy to spot.

Further discussion:

Is the Tipping Point Toast? (Fast Company)

Can't Buy Me Luck (Scientific American)