Viacom v. YouTube: When Is It Storage? When Is It a Public Performance?

Today, a federal district court in New York granted YouTube’s motion for summary judgment in its long-running litigation with Viacom. Viacom and—separately,—the English Premier League sued YouTube and Google alleging that they were liable for infringing works that users posted on YouTube. Google is understandably trumpeting its victory though it of course is subject to appeal to the Second Circuit.

The issue in the case focuses on the scope of the Digital Millennium Copyright Act’s safe harbor for online service providers set forth in section 512 of the copyright statute. Section 512 is pretty chunky but most of the court’s analysis focuses on section 512(c). That subsection insulates service providers from liability “for infringement of copyright by reason of the storage at the direction of a user of material that resides on a system or network controlled or operated by or for the service provider” if the service provider can meet a three-part standard set forth in that subsection.

Continue reading

Scanning the Public Domain

I am working on a paper on the history of razors and blades (yes, I know that sounds obscure, even for an ivory tower sort; I’ll leave it to another day to try to persuade you that you should be fascinated, too). I have been reading turn of the century—that is the 19th century—catalogues. The 1895 Montgomery Ward & Co. catalogue was a wonder. Montgomery Ward was the Amazon of its day, a market leader in the mail-order business. Local stores, especially outside of big cities, might offer only a limited selection, but Montgomery Ward promised the world to the entire country. Catalogue No. 57—Spring and Summer, 1895—was a behemoth, running 624 pages and offering tens of thousands of items for sale.

You would think that it would be a trick to find the 1895 catalogue, and, in truth, putting together a run of catalogues from say 1895 to 1930 does involve an unpleasant mix of brittle paper and clumsy microfilm. But you can have the 1895 catalogue in your hands tomorrow: go to Amazon, spend $22.21, and jump into the Monkey Ward time machine to see what life looked like back then.

What you actually will receive is a Dover Publication described as an unabridged facsimile of the original 1895 catalogue. In a word, a copy. Dover, I assume, scanned an original version of the catalogue to produce the book that it is selling to us. Scanned, meaning, took pictures of the pages of the catalogue. We shouldn’t fear for Dover, of course, as whatever the original copyright status of the Ward catalogue, the U.S. cutoff date for the public domain is 1923. Anything published in the U.S. before that date has entered the public domain, so Dover is on safe ground in making its copy. And in selling the public domain, Dover is doing exactly what it is allowed to do. You can copy the public domain, remix it, turn it into movies and take in more than $330 million at the box office—the Alice in Wonderland remake, of course—and sell it in its original form as Dover is doing with the Ward catalogue.

But that isn’t what I care about. If you turn to the copyright page in the Dover facsimile, it says “Copyright © 1969 by Dover Publications, Inc.” This is where the plot thickens. What exactly—if anything—does Dover have a copyright in? There is an introduction to the catalogue and Dover can hold a copyright in that, as you can with any newly-created content, but what about the catalogue itself? And while I’d like to know where I stand as to the Dover book, of course, the real question is where do I stand as to Google? Google is the big enchilada when it comes to the scanned public domain and we are going to care a great deal about what rights we have as to the scanned images that Google has created. It isn’t clear yet what will happen to Google Book Search—I have views here, here and here—but Google is perfectly within its rights to scan works in the U.S. public domain and that will be a core part of its database going forward.

But what can I do with what Google—and Dover—have created? We should review quickly the basics of copyright in photographs. We start, of course, with the Supreme Court’s 1884 decision in Burrow-Giles and Sarony’s photograph of Oscar Wilde. Burrow-Giles helped itself to that photograph and then defended its use on the ground that the photograph was a “mere mechanical reproduction” and hence insufficiently original to qualify for copyright protection. The Court acknowledged that that might be the case for the “ordinary production of a photograph” though it declined to conclude that no copyright would attach even then. But Sarony had done much more in creating the photograph of Wilde: Sarony had composed the picture, arranged Wilde’s pose and the setting for it, adjusted the lighting and shading, and all of that together represented an original creation by Sarony.

But the line that Sarony seemed to suggest—we should look for originality in the creation of the subject of the photograph—went by the by as courts embraced copyright in uncreated subjects (outdoor settings such as the New York Public Library). As soon as we took that path, we were going to have huge numbers of original photographs. Go to a party, take a bunch of pictures, create original copyrighted works. Indeed, the more interesting question quickly becomes: what does an unoriginal photograph look like? Burrow-Giles could of course have copied the Sarony photograph of Wilde by getting Wilde to repose and making sure that every aspect of the new photograph matched the old one. That would be as much a copy of the Sarony photograph—and therefore unoriginal—as what Burrow-Giles did in the actual case.

Next have Burrow-Giles take a photograph of Sarony’s photograph of Wilde. That is just as much a copy of the original photograph as when Burrow-Giles recreated the picture per my hypo. That photograph of the photograph should be treated as a copy of the original photograph and unoriginal. You could of course make an original photograph that incorporated the Sarony photograph—perhaps with the photograph surrounded by other objects—but a pure photo of the photo should be treated as a copy. When you take such a photo of a copyrighted photo, you infringe. When the original photo is no longer copyrighted because it has entered the public domain, your photo doesn’t infringe but it doesn’t cease to be a copy and it still lacks originality.

Where does that put Google (and Dover) with its digital scanners? We are starting to see skirmishes over photographs and scans of public domain works. The British National Portrait Gallery got into a spat with Wikipedia when Wikipedia uploaded onto its website digital images created by the NPG of public domain works in its collection. Actual caselaw is scarce, with Bridgeman Art Library v. Corel Corp., a 1999 federal district court decision, as a leading case. The court characterized the dispute as one over “’slavish copies’ of public domain works of art” and concluded that such copies lacked the spark of originality and therefore could not be copyrighted.

That isn’t to say that an owner of such a digital file couldn’t try to control use of it through some means other than copyright. Go to Google Book Search and download Alice in Wonderland. The first page is from Google, not Lewis Carroll, and it offers a strong defense of the public domain: “Public domain books belong to the public and we are merely their custodians.” Then we get to the next word “nevertheless” and you can guess what follows: digitization is expensive—as indeed it is—so Google has imposed a series of limits on how the digital file can be used.

To be sure, there are oddities about this analysis. A photograph of a rose in your backyard is probably nothing more than a slavish copy of the rose and yet it is hard to imagine that a court would find insufficient originality in the photo. Yet a photo of a photo—or a digital scan of the 1895 Ward catalogue—will probably be found to be nothing more than a copy of the original work, infringing or not depending on whether the work has entered the public domain, but almost certainly lacking sufficient originality for the second photo or scan to be independently copyrightable.

DOJ Filing in Google Book Search

Let me start by saying thanks to Mike for the invitation to blog here. I will probably cross-post some posts here and at the Chicago Faculty blog.

This is a longish post and runs the risk of violating some filter Mike has undoubtedly set about the number of self-cites permitted in a post.



Late last night, the Antitrust Division filed its statement in the pending Google Book Search case. Once it became clear that DOJ would participate in the case, everyone understood that the filing would receive a great deal of attention. Filing at 10pm on a Friday—and a Jewish holiday to boot—sounds like DOJ was trying to hide the filing, but I suspect the timing reflects the reality of the real complexities of the case. DOJ’s bottom line is that the proposed settlement agreement should be rejected “in its current form” with the parties encouraged to modify it to comply with “Rule 23 and the copyright an antitrust laws.” (Early commentary on the DOJ filing is available by Danny Sullivan (here) and James Grimmelmann (here).)

The statement of interest is divided into three sections and leads with an analysis of whether the proposed settlement complies with Federal Rule of Civil Procedure 23 on class actions. In one way, that isn’t surprising at all. Take a quick scan at some of the filings—nicely available at the Public Index—and you will see much discussion of both deep conceptual issues and ordinary day-to-day class action mechanics. The Antitrust Division filing does both, but do remember what business they are in: antitrust. The Antitrust Division doesn’t have a general portfolio dealing with class actions, so the extent of the class-action discussion is surprising.

My guess is that it reflects real questions about whether a class action is an appropriate vehicle for the scope of changes proposed through the settlement agreement. DOJ characterizes these as “a matter of public, not merely private, concern” and “typically the kind of policy change implemented through legislation, and not through a private judicial settlement.” That said, DOJ still favors moving forward to improve the settlement: “Because a properly structured settlement agreement in this case offers the potential for important societal benefits, the United States does not want the opportunity or momentum to be lost.” I agree with that. Indeed, that was the starting point for my recent testimony on Google Book Search before the House Judiciary Committee (written here and video here (my direct starts around the 45th minute)).

The DOJ filing focuses on conflicts between active rights holders and inactive ones, especially orphan holders. Two conflicts jump out. First, orphan holders are locked into the settlement agreement. Active holders can opt out and indeed some publishers have and will negotiate separate deals with Google. The extreme version of this would be a settlement agreement that turns out to be binding only on orphan holders because the active holders opt out and do separate deals. It is hard to know what representativeness means in those circumstances. Second, active rights holders benefit from revenues that orphan works generate if those orphan holders never appear.

DOJ suggests a number of potential modifications, though on this core question of representation DOJ states “that it is difficult to see how any class representative could adequately represent the interests of all owners of out-of-print works (including orphan works).” At the House hearing, in questions, I noted that in other contexts—bankruptcy mass tort cases, for example—it was standard to appoint a separate representative for uncontactable—is that a word?—groups of interested parties.

The statement of interest then turns to antitrust law. For me, at least, this was like getting a test back, since my original paper focused on antitrust issues. My incoming Fall antitrust students will be comforted to know that I think that I passed. I focused on three issues in that paper (and for the quick version, read the four-page intro that I used for my House testimony). One was a technical but important issue involving Noerr-Pennington immunity. There is a risk that judicial approval of the settlement agreement will block a subsequent challenge by DOJ. That would be wrong but the caselaw doesn’t resolve this cleanly. The good news on that shows up in footnote 6, p. 16 of the statement of interest: “The parties have represented to the United States that, with regard to any potential challenge to the legality of the settlement, they will not raise a defense based on the Noerr-Pennington doctrine.”

My second issue was the collective pricing mechanism for consumer purchases. That is tricky—see my paper—but DOJ, while not concluding anything yet—they are very clear on that—suggests that “absent modification of these terms, there is a significant possibility that the Department will conclude that they violate the Sherman Act.”

That leaves us with my third issue, the licensing of the orphan works. The genius and curse of the settlement agreement is that it makes possible the use of the orphan works by Google, but, unfortunately, only by Google. In my paper and my testimony, I suggested a number of ways that that license might be expanded. The DOJ statement of interest frames that as a question of market foreclosure and concludes that their concerns could be minimized if “Google’s competitors could gain comparable access to orphan works.”

The third section is brief—less than a full page. DOJ favors what Google has promised to do for users with print disabilities. That is clear enough. Less clear to me is the second point: “data provided should be available in multiple, standard, open formats supported by a wide variety of different applications, devices, and screens. Once these books are digitized, the format in which they are made available should not be a bottleneck for innovation.” I am not sure what that means exactly. The settlement agreement will generate a great deal of data. Indeed, privacy advocates are very much concerned about what Google will be able to learn about users as they read books online. Other data will be data about rights ownership. For example, some post-1923 books may have entered the public domain because of formalities problems and the process for sorting through rights will generate tons of data on that and on which books aren’t actually orphaned. Finally, the digitized scans themselves are a kind of data as well. My guess is that DOJ is referring to rights-ownership data, but they will have to clarify that at some point.

To conclude, at the House hearing, I called the settlement agreement a beta and said that it would be “ungooglish” to not try to improve it. I understand DOJ to call for exactly those sort of improvements going forward. I hope this interesting process that we are running—hearings in the EU and Congress, a filing by DOJ—will result in an improved version of the settlement and one that will provide a framework for meaningful competition in online libraries.