When Is a Phonorecord Not a Phonorecord?

Last time, we finally figured out how ReDigi operates and how it plans to get around the fact that it must make at least one (and often two) intermediate copies of a song file in order to complete the sale of the song file. ReDigi’s solution is to structure itself as an Amazon-style music locker and rely on space/format shifting for those intermediate copies.

But this doesn’t get around the other concern I raised (way back here), which we might call the “phonorecord problem.” Recall that the nub of the RIAA’s argument is that the First-Sale Doctrine is limited the distribution right. The RIAA’s point was that the intermediate copies exercised the reproduction right and, therefore, fell outside the scope of the First-Sale Doctrine. While I thought there might be a different way of looking at that issue, it turns out ReDigi is fine with the RIAA’s argument, since it thinks it has an alternate (and better) legal theory regarding those intermediate copies.

The “phonorecord problem” is more fundamental. Under a strict and plain reading of the Copyright Act, the distribution right is limited to the distribution of physical embodiments of the copyrighted works, e.g., a CD. Read literally, the distribution right is limited to the “distribut[ing of] copies or phonorecords of the copyrighted work.” The term phonorecords is defined as “material objects in which sounds … are fixed … and from which the sounds may be perceived, etc. ….” A computer file isn’t a material object.*

* Right? Can anyone think of a basis for saying a computer file is a physical object?

Normally, this doesn’t create a problem. Any digital distribution of a computer file will automatically implicate the reproduction right, and you don’t get extra points for having extra exclusive rights violated. But in ReDigi’s case, it presents a problem. At the time, I thought it was a problem for ReDigi, because if ReDigi’s right to make intermediate copies were tied to its right under the First-Sale Doctrine to distribute the audio files, the right to make those intermediate copies would disappear along with the right to distribute the underlying files. And if ReDigi wasn’t distributing material objects, but only computer files, it might not be able to take advantage of the First-Sale Doctrine. In the case of an audio file sitting on a user’s computer hard drive, the “phonorecord” would be the hard drive–and that’s a little hard to sell!

It turns out that this may be a problem for the music industry, not for ReDigi, now that ReDigi has an alternate (and stronger) argument for the right to make those intermediate copies. If ReDigi’s space/format-shifting argument succeeds, then all the music industry as left is a claim on the distribution right.*

* Well, that and a couple of very minor claims for violation of the public-performance right (update: for 30-second audio clips) and display right (in the cover art). I’ll comment on public-performance issue at the very end of this post.

I don’t know if ReDigi read this blog post, but they took that idea and ran with it. Or, more precisely, turned it on its head. For ReDigi, the “phonorecord problem” isn’t a problem; it’s a solution. ReDigi now argues that it does not violate the distribution right precisely because it doesn’t distribute physical embodiments of the song files; it only distributes digital audio files. It doesn’t need the First-Sale Doctrine at all!*

* The First-Sale Doctrine doesn’t go completely away. The restrictions that Amazon and others place on the right to transfer the license associated with the audio file (discussed here) still have some legal effect, even outside of Vernor. The buyer needs that license to play the audio file (because playing an audio files requires copying from storage to memory); otherwise, the buyer is infringing copyright, and ReDigi is possibly interfering with the contract between Amazon (or whoever) and the original buyer. Remember that Apple’s iTunes does not place such a restriction on transfer. This would explain why ReDigi appears now to be limiting its service to audio files obtained from iTunes.

Digital content has been distributed for quite some time now, and this is, honestly, the first time I had ever thought of it this way. Courts and almost everyone else* have been assuming that digital distributions infringe the distribution right. The music industry’s file-sharing cases were all mostly dependent on a claim of infringement of the distribution right, which led to the knotty question of whether merely making a file available on your shared folder, without more, constitutes a distribution.** Was there, perhaps, a more fundamental question?***

* Actually, some amici argued this in a file-sharing case in 2008, but the court ignored it. I’ve now also seen the issue raised in some law review notes (like this one).

** I think the majority rule is that it does not. You still need to prove that someone else accessed and downloaded the file.

*** It may not have made a difference. The music industry could have argued (persuasively in my opinion) that “making available” a file, plus evidence of downloading of file, constitutes contributory infringement. I suspect the reason the music industry didn’t make this argument was that it preferred to rely on the “making available” theory of distribution (which didn’t work out in the end), and that “distribution” is just conceptually neater.

Have we really been fooling ourselves for the last, I don’t know, 10-15 years*? The issue has come up obliquely a few times since 2001, but as far as I can tell, it’s never really been argued. In New York Times v. Tasini, 533 U.S. 483 (2001), the Supreme Court assumed without analysis that when LEXIS/NEXIS sold copies of articles through its electronic database, it was “distribut[ing] copies of the articles to the public by sale.” It is pretty clearly dicta, as no one appeared to argue otherwise.** In Perfect 10 v. Amazon, 508 F.3d 1146 (9th Cir. 2007), the court followed Tasini’s dicta, again without analysis, and again, no one appeared to have argued the point.

* Actually, we could go back much farther. As early as 1991, scholars were using the text-only Gopher protocol to digitally distribute articles and papers. In 1992, we called it “electronic research” and thought it was the coolest thing.

** Material in a judicial opinion unnecessary for the actual holding of the case is not binding authority and is called “dicta.” In particular, we do not consider authoritative conclusions reached by the court that were not the product of adversarial argument by the parties.

Adding somewhat to the confusion was the 1995 enactment of the Digital Performance Right in Sound Recording Act, which allowed others to make the “digital phonorecord deliveries” of music if they paid a statutory royalty. 17 U.S.C. § 115(c)(3)(A). This wouldn’t be necessary if delivering digital phonorecords weren’t illegal. Further, the same Act made clear that it didn’t affect the (apparently pre-existing) rights “to reproduce and distribute the song recording … including by means of digital phonorecord delivery.” Again, if Congress didn’t think this right didn’t exist already, it wouldn’t have felt the need to protect this right.

As I’ve explained previously, though, Congress’ opinion about the meaning of the Copyright Act in 1995 doesn’t affect the central inquiry, which is what Congress was intending back in 1976 when it drafted the distribution right. Consider the state of technology in 1976 (or even 1978 when the CONTU Report was issued). Congress was aware of computers and software, but the idea of distributing digital content was not well understood, if understood at all. DARPANet was just a baby. Pretty much all one could reasonably transfer over a network at the time was text, and even that took a long time. Could Congress really have had anything in mind other than the distribution of physical embodiments?*

* I understand there’s an article in the Journal of the Copyright Society of the USA, which argues that Congress actually intended the distribution right to be read very broadly, but I can’t seem to lay my hands on it right now. I’ll read it with interest when I do.

In addition, a literal reading of the distribution right would create problems elsewhere in the copyright system. For some reason, the definition of publication is identical to the distribution right, and several important dates are keyed off the first “publication” of a work.* If we read the distribution right literally, we’ll need to read the definition of publication literally, and this could cause all kinds of mischief. Are we really prepared to accept that a song that’s available only as a digital download, or a book or graphic novel serialized exclusively on the Web, isn’t a “publication”?

* Most notably the requirement that you must register your copyright within three months of first publication if you wish to avail yourself of statutory damages or attorney’s fees.

The funny thing is that ReDigi doesn’t actually care how this turns out. From its point of view, it wins either way. If we read the distribution right broadly to include purely digital distributions, then the First-Sale Doctrine applies. If we read the distribution right narrowly to be limited to physical objects, then the distribution right isn’t even implicated. I suppose the music industry could find some way to argue that, while the distribution right should be read broadly, the First-Sale Doctrine should be read narrowly, though that seems intuitively unfair (but not impossible, all the same). Alternatively, the music industry could just concentrate its fire on the reproduction right.*

* Final note: Apparently the public performance right has been raised by Capitol. If ReDigi really is following the Amazon model and maintaining separate copies of every file that is uploaded, i.e., not using the “single master” method of storage, then I don’t see how the public performance right is implicated (as I explained here.) The only person who can enjoy the particular song file is the owner of the file–and that’s not a performance to the public.

Update: Based on Capitol’s motion for preliminary injunction, it is simply relying on Tasini to get around the “phonorecord problem.”

Update: It also appears that Capitol isn’t basing its claim for infringment of the public-performance right on any type of “single master” theory, but on the fact that ReDigi allows potential buyers to listen to 30 seconds of an audio file. I guess I didn’t realize ReDigi did that. ReDigi would seem to have a decent, but not slam-dunk, fair-use argument here (though the “30 second rule” is something of legal urban legend), which I’m not going to get into here. Even if Capitol prevailed, it wouldn’t affect the core of ReDigi’s business–it’d just make it harder to promote the songs.

Thanks for reading!