I speculate that preferential attachment is relatively weak for traditional types, since such attachment depends on network-level knowledge that is not readily available, while the absence of mass media sharply limits attachments’ geocultural range. I hypothesize that knowledge about artists—primarily transmitted orally—drops off as the inverse square of distance, while transportation, language, and musical style/culture inhibit musical flows as the inverse of distance, roughly speaking. Without mass media, music that is extremely popular in one location is not likely to be popular far away, even if preferential attachment continues to play a role. Subject to the above-mentioned restrictions, selections depend primarily on fans’ perceptions of artists’ audio-visual quality: fame is essentially aesthetic.
By contrast, for popular types, distance plays a relatively small role in forming musical preferences, as global media (especially satellite TV and web) not only disseminate artists’ music in proportion to their fame, but also (by diffusing associated language and culture, e.g. English and hiphop) facilitate its broad receptivity. Further, driven by marketing considerations, global media diffuse not only musical information, but also tremendous quantities of meta-information about artists’ network degrees (and providing information, explicit or implicit, about popularity). As a consequence, preferential attachment predominates: fame is essentially social.
In the absence of empirical data, this speculative theory is here experimentally tested via a computer simulation developed using the open-source statistical programming language R. A spherical surface (deployed to avoid edge effects and emulate global reality) is randomly populated by fans and artists according to a uniform distribution. All distances are computed along great circles. In each simulated time slice, each fan selects an artist at random, up to a maximum of 4 such selections (“old” artists being cycled out on a FIFO queue), as modulated by two global parameters: a weighting for preferential attachment applied to artist degree, set at 1 for traditional and 10 for popular; and an exponent implementing an inverse distance weighting, set at 5 for traditional, and 0 for popular. 500 iterations appear sufficient to achieve steady-state results, and simulations were run for traditional and popular cases. The resulting degree distributions turn out to be quasi-normal for traditional music, and quasi-scale-free for the popular. These theories and simulation results await confirmation by empirical data.
Yuka Hayashi is a self-taught designer, from Japan. She graduated from Monterey Institute of International Studies, specializing in International Policy Studies with certification in Nonproliferation Studies (MA). She worked at a private think tank relevant to international policy as a research assistant. She also has an experience of working at a nonprofit organization promoting international cooperation. Currently, she is CEO of YR-Design together with her partner, Rob Oudendijk. She does design work, video shooting/editing, translation work, and IT assistant work together with their team member cat, Kitty-chan.
mcohen@u-aizu.ac.jp