Solveig Singleton at Tech Liberation Front comments on that Steve Levy column about mash-up artist Girl Talk I linked last week. While she’s certainly right that it seems like a bad idea to riddle copyright law with a tangle of ad-hoc exceptions (is this actually what’s being proposed?) her take seems unfair in a variety of ways. Referring to Girl Talk’s acclaimed Night Ripper album, which samples some 167 artists, she writes:

Fair use? Transformative use? Why bother with the technicalities? Levy and a legislator likes the fellow, so they weigh in on the side of legislating (yet another) exception. Maybe jam transformative and fair uses together into a whole new category, “rave” use, with a safe harbor for “hipster” use and for the older set “cool” uses? The principle behind it might be that if you offend only a little, you are liable, but if you offend multiple players a lot, you are home free.

Except that the legislator quoted in Levy’s column does, in fact, suggest that mash-ups are a “new transformative art” and that fair use safe harbors in copyright law should be clarified or amended so as to accommodate such works. So, pace Singleton, the point is not at all that we need a special loophole for a particular art form that somebody likes. It’s that this provides one sort of example of the kind of creative work that current rules tend to squelch, and constitutes an argument in favor of a general reform of those rules to make this kindThe last sentence quoted above also seems awfully glib if you think about it for a minute. Because it does, in fact, make perfect sense that a song constructed from pieces of a dozen other songs is more likely to be a significantly novel composition than one simply pasted together from two or three tracks. If I compose an essay or story by cutting and pasting whole paragraphs from two prior works, I’m a plagiarist. If I mix and match individual sentences from 20 different works, I’m probably doing something else entirely.

Probably Singleton’s right that the system for clearing rights to large numbers of work for use in mash-ups could be streamlined. But she doesn’t explain why that’s preferable. What you have here are two kinds of creation, and two sets of incentives for it. Under a licensing system, you have higher costs of mash-up production, but more potential revenue for creators, either in the form of licensing fees or because the system has deterred the creation of mash-ups that are consumed instead of the original works (though a popular mash-up seems as likely to spark interest in the source material). Under a broader safe harbor, you have the elimination of one major barrier to the creation of mash-ups, which are often homebrewed by talented amateurs, balanced by a reduction of the kind of revenue the licensing system produces. You could imagine the latter scenario reducing the amount of fodder for future mash-ups if the effect of the lost revenue were to severely reduce the production of original works, but this seems awfully unlikely.

It seems crystal clear to me that you get more net creation under the broader safe-harbor system. Under the most streamlined, reasonable sample-licensing system imaginable, you’re going to entirely deter large numbers of amateur DJs, whereas I have trouble believing even a single song is not going to get produced because labels or artists aren’t able to squeeze every last iota of value out of their creations. Is there any reason for thinking more liberal sampling rules are likely to depress music creation on net, rather than “screwing” one subset of creators out of one more source of potential revenue?