Scrivener's Error: Warped Weft
originally posted SeptemberOctober 2005
1 Professor Solum has a useful introduction to one aspect of the procedural morass raised by this case… and points the way to the ultimate solution (which is probably about three posts down the road from here). So, for the next few posts, I'm going to rip into the procedural problems with the case as it was filed.2
Like Professor Solum, we should start with the class definition, which is:
The Class is initially defined as all persons or entities that hold the copyright to a literary work that is contained in the library of the University of Michigan.
This class definition was obviously crafted by someone who had never seen a commercial publishing contract, probably never seen an academic or professional publishing contract, had never heard of Rosetta Booksin short, knew nothing of publishing law and practice. The Class can't be that big for a very simple reason: the right to do what Google appears to be doing may have been licensed to a third party, which in turn may not object to Google's conduct. And that "may" is a nontrivial qualification, because it means that the three named plaintiffs are almost by definition not typical of the plaintiff class.3 Further, this isn't a simple matter of "does the contract explicitly transfer the rights?", either; that's what Rosetta Books concerned, and the only substantive opinion is a nonbinding district court opinion concerning one publisher's contracts from a specific time period.
At a more practical level, this class definition is unmanageable for another reasonthe problem of orphan works, which is another demonstration of a factor that will probably predominate in many claims. This alone will make any court that will ultimately be responsible for administering a remedy in any judgment against Google wary of too-broad a remedyand certainly a remedy involving unknowing parties.
The ultimate irony here is the usual mechanism of class remedies: Usuallyas in the horribly structured "database settlement" also "brokered" of late by the Author's Guildeveryone who falls into the class definition is bound by the settlement unless they opt out. Given that opt-in/opt-out is precisely what is at issue in Google's program, flowing in the opposite direction, this seems rather silly. It wouldn't have been that hard to both acknowledge reality and make the class definition logically consistent with the cause of action:
Class A is initially defined as all persons or entities that:
1. Hold the copyright to a literary work that is contained in the library of the University of Michigan; and
2. Have not transferred a license to a third party that would authorize the conduct alleged; and
3. Have not otherwise approved of the conduct alleged consistent with the signed writing requirement of 17 U.S.C. § 204.
Class B is initially defined as all persons or entities that:
1. By contract or operation of law hold a license that would authorize the conduct alleged to a literary work that is contained in the library of the University of Michigan; and
2. Have not otherwise approved of the conduct alleged consistent with the signed writing requirement of 17 U.S.C. § 204.
This should make one major procedural defect in the case as filed readily apparent: There is no adequate representative named for class B.4
Those who have been following my arguments with Amazon for the last couple of years will have spotted the unconsidered case that remains. Can you spot the unconsidered class that has inherent conflicts with both class definitions? Yes, Ms Sanchez? I said that "Not all books are single, complete manuscripts," and that I must be referring to collections and anthologies? Well, Ms Sanchez, you do get a brownie point for doing some research in my blawg, but I'll get back to you later in the session for what that really means. In any event, the key point is that the editor of such a volumeparticularly if there is more than one contributor, as in anthologieshas an inherent conflict of interest with the individual(s) who contributed works to it. We could try to define a Class C, perhaps like this:
Class C is initially defined as all persons or entities that:
1. Hold the copyright to a literary work (the "Component Work") that forms part or all of another literary work (the "Composite Work") that is in turn contained in the library of the University of Michigan; and
2. Have not transferred a license to a third party, including without limitation the publisher and/or editor of the Composite Work, that would authorize the conduct alleged; and
3. Have not otherwise approved of the conduct alleged consistent with the signed writing requirement of 17 U.S.C. § 204.
This is starting to look awfully ugly, isn't it? It is marginally manageable with all three classes defined… if, and only if, nobody looks under the surface at the substantive strengths of the claims. All three classes do have claims; however, the strengths are radically differentClass A is a strong copyright claim, Class B is a moderately strong copyright claim combined with a strong interference-with-contract claim, and Class C is a slam dunk under either theorythat each require different kinds of evidence to validate both liability and remedy.
It gets better. Class C should also be divided just like the original class definition was divided into Class A and Class B, for the same reasons. Where things really get interesting is in determining who might be an adequate representative. The Author's Guild cannot be an adequate representative for Class C (or Class D), because its membership consists almost entirely of those who publish at book length only. Further, there's a hidden landmine in the Tasini question for Class C (and Class D): that many, if not most, Composite Works include one or more Component Works that were previously published elsewhere.
If you're not convinced that the only thing certifiable about Author's Guild v Google is the individual who decided that this could be a unitary class action, there is more yet to comeand we haven't even gotten to the substance!
My remaining procedural grumble in this matter comes from some particularly inelegant drafting in the complaint.
7. This copyright infringement action arises under 17 U.S.C. § 101 et seq. This Court has jurisdiction over this action under 28 U.S.C. § 1331 federal question, and 28 U.S.C. § 1338 (acts of Congress related to copyright).
8. Venue is proper in this district pursuant to 28 U.S.C. §§ 1391(b) and 1400(a) because one of the named plaintiffs resides in this district and because defendant conducts business in this district.
To begin with, I suppose two out of three isn't bad; unfortunately, "isn't bad" isn't good enough. Note that there is no specific allegation of personal jurisdiction over the defendant! That's an easily fixable problem, but it shouldn't need to be fixed at allespecially as the theory of venue expressed in the complaint is that same as the theory that would give personal jurisdiction.5
The more-subtle problem, though, is the question of exactly what business is being conducted in the Southern District of New York that might give rise to jurisdiction. The complaint and class definition limit themselves to scanning taking place at the University of Michiganwhich is not, by any reasonable definition, in Manhattan. So, then, the only "business" that might be taking place in the Southern District of New York must be the dissemination of the final materialsan event that has not yet taken place. This is where we get into the interplay between the procedural posture of the matter and the substance, for two reasons.
First, a large portion of the case depends upon treating this as a declaratory judgment matter, requesting an injunction preventing the scanning from beginning. Leaving aside the logical problems with this issue, what we've really got is a court being asked to order a party not ordinarily found in the district (the reliance on "does business in") to not "do business" in another district. This is not the most obvious way to bring parties before the court, let alone make an enforceable remedy possible!
Second, this needlessly conflates the copyright problems. Google's program actually involves three distinct potential copyright violations (and I emphasize potential):
There is some case law on both sides of each of these issues, and the statute really isn't all that clear. To say the least, I think it was an incredibly silly decision to conflate the three violations into a single cause of action, because that encourages misapplication of analyses and privileges from one of the three into the others. For example, the argument that "Arriba Soft says this is fair use" can logically apply only to violation 2. The facts in Arriba Soft explicitly deny application to violation 3, and really don't apply to violation 1 at all.6
So, in the end, the Author's Guild's complaint forfeits some of its most powerful arguments for the sake of maintaining class uniformity (that, as I've already noted, is fatally flawed in any event). The public statements have focussed almost entirely on violation 2; that's where I'll start analyzing the substance, but I can guarantee that's not where I'll end up.
One of the justifications that many apologists for Google's actions have raised is "Arriba Soft says it's fair use." This is a woeful misunderstanding of Arriba Soft, unless one willfully misreads the opinion and wrenches it completely from context. The first two paragraphs of the most-recent Ninth Circuit opinion make this clear:
This case involves the application of copyright law to the vast world of the internet and internet search engines. The plaintiff, Leslie Kelly, is a professional photographer who has copyrighted many of his images of the American West. Some of these images are located on Kelly's web site or other web sites with which Kelly has a license agreement. The defendant, Arriba Soft Corp., operates an internet search engine that displays its results in the form of small pictures rather than the more usual form of text. Arriba obtained its database of pictures by copying images from other web sites. By clicking on one of these small pictures, called "thumbnails," the user can then view a large version of that same picture within the context of the Arriba web page.
When Kelly discovered that his photographs were part of Arriba's search engine database, he brought a claim against Arriba for copyright infringement. The district court found that Kelly had established a prima facie case of copyright infringement based on Arriba's unauthorized reproduction and display of Kelly's works, but that this reproduction and display constituted a non-infringing "fair use" under Section 107 of the Copyright Act. Kelly appeals that decision, and we affirm in part and reverse in part. The creation and use of the thumbnails in the search engine is a fair use. However, the district court should not have decided whether the display of the larger image is a violation of Kelly's exclusive right to publicly display his works. Thus, we remand for further proceedings consistent with this opinion.
Kelly v. Arriba Soft Corp., 336 F.3d 811 (9th Cir. 2003) (emphasis added). This is the first hint that the matter concerns more than indexing. In the context of reproduction of visual images already in electronic forma distinction that is quite significant, as we'll see later in looking at the merits of Author's Guild v. Google itself7the Court held that the District Court's judgment of fair use was inappropriate.
As mentioned above, the district court granted summary judgment to Arriba as to the full-size images as well. How-ever, because the court broadened the scope of both the parties' motions for partial summary judgment and Arriba's concession on the prima facie case, we must reverse this portion of the court's opinion.
With limited exceptions that do not apply here, a district court may not grant summary judgment on a claim when the party has not requested it. The parties did not move for summary judgment as to copyright infringement of the full-size images. Further, Arriba had no opportunity to contest the prima facie case for infringement as to the full-size images. Accordingly, we reverse this portion of the district court's opinion and remand for further proceedings.
(footnotes omitted) Sure, this is a procedural reason only; however, the third fair-use factor analysis indicates that a more-completely developed record would weigh against finding fair use as to the complete works.
This factor neither weighs for nor against either party because, although Arriba did copy each of Kelly's images as a whole, it was reasonable to do so in light of Arriba's use of the images. It was necessary for Arriba to copy the entire image to allow users to recognize the image and decide whether to pursue more information about the image or the originating web site. If Arriba only copied part of the image, it would be more difficult to identify it, thereby reducing the usefulness of the visual search engine.
Ah. Here we are. What this really reduces to is this question, which is both far more complex than the question raised in Arriba Soft and deceptively simple:
Is it necessary to copy the entirety of a textual work to "allow users to recognize the [work] and decide whether to pursue more information about the [work]"?
This is deceptively simple because it divides the inquiry into two aspects: "recognition" and "pursuit of more information". I think the conjunctive linking of these two concepts is probably an artifact of legal writing more than it is a logical necessity, which will bear analysis later. The problem, though, is that this specific statement has been shorthanded into the concept of an "index"a shorthand that professional index-preparers would deny, and a conclusion inconsistent with Second Circuit law.8
The culture clash of imposing "Internet values" on print publications results from some seriously flawed reasoning in the Copyright Act itselfflaws inherited from the 1870 revision and 1909 Actthat could have been avoided just by reference to the piano-roll case9 and its implications. The question this raises is also deceptively simple:
Must changes in technological capabilities and/or penetration always result in changes in both copyright owners' rights and copyright users'/exploiters' responsibilities?
Contrary to most analyses of Arriba Soft, which try to make the Internet somehow a "privileged" medium, the answer to this question is "no." Those changes under some circumstances may result in some changes; but reaching the result desired by Google requires that this question be a black-and-white imperative, not analysis sensitive to the context of not just the desired reuse, but the actual origin. That, in turn, distinguishes the "index" that Arriba Soft created from that Google would create; but this posting has already beaten that horse into a pulp.10
Keep in mind that fair use is not a privilege. It is instead an affirmative defense to an accusation of infringement. (Unfortunately, neither the statute nor the courts are very consistent in treating it as such.) The burden of proving an affirmative defense is on the party asserting the defense; this is one of the many "the tie goes to the runner" principles common in American jurisprudence.
In any event, let's take a fresh look at the way the fair use defense applies to each of the three potential violations that the Author's Guild should have alleged.
Admittedly, this doesn't look a whole lot like the analysis in Arriba Soft. That is primarily because, as I've tried to make clear, this case isn't Arriba Soft. It is not being heard in the Ninth Circuit, which (along with the Eleventh Circuit) has the least-stringent view of fair use; it is not based on materials merely gathered, but for which substantial and conscious copying must occur for any of the three "uses"; it is not based upon reuse of materials in exactly the same form, medium, and purpose/function as provided by the copyright holder; and does not concern a well-delineated final use and presentation.
Perhaps most importantly, though, that this case is being heard in the Second Circuit matters for a procedural reason: Second Circuit law on affirmative defenses in general is somewhat less favorable to defendants engaged in knowing and purposeful acts. That is, the nature of fair use as an affirmative defense matters more in the Second Circuit than it does in the Ninth Circuit (or would in the Sixth, Eighth, or Eleventh Circuits). Put another way, the umpires in the Second Circuit seem less appreciative of the well-turned, but close, double play than do those in the Ninth Circuit.
Ordinarily, this would be the first question one asks. If there would be no violation even without referring to the fair use defense, why bother with the four-factor weighing? Rhetorically, I think it was necessary to at least indicate that fair use is at best arguableif only because so many ill-informed commentators12 have immediately jumped to that defense because it's easier to sell the public on that rhetoric.
In any event, let's take a closer look at the three violations I outlined previously.
That leaves us with one clear violation and two potential violations with varying degrees (and causes) of mud. Mud they may be; but they cover the ground. Thus, we do need to consider the fair use doctrine; but can (and should) only do so in the context of the actual violations.
Most of my discussion of the merits thus far has assumed that the complaint asserted what it should havethree distinct, clear theories of copyright infringement, each sufficiently fact-bound to avoid a motion to dismiss (and probably summary judgment). Of course, that's not what will emerge from this fiasco; instead, we'll ultimately get an opinion founded on the complaint.16 So, then, what does the complaint actually allege?
39. Google has made and reproduced for its own commercial use a copy of some of the literary works contained in the University of Michigan Library, which contains the Works that are the subject of this action, and intends to copy most of the literary works in the collection of that library.
40. Google's conduct is in violation of the copyrights held by the Named Plaintiffs and other members of the Class.
41. Google's infringement of the copyrights of the works was willful.
Author's Guild, Inc. v. Google, Inc., No. 05cv8136 (S.D.N.Y.), complaint (Dkt. 1, Sep. 20, 2005).17
Admittedly, this isn't as bad as it could be. The real problems are twofold. First, it implies that only the copying itself (what I have labelled violation 1) is potentially at issue, although that conclusion is inconsistent with the factual allegations preceding it, particularly paragraphs 29, 31, and 33, each of which asserts that Google intends to display the copied works (violation 3). That this is only an implication, though, is not binding on anyone; under controlling law, the complaint is to be read broadly to include all theories and causes of action for which the defendant is fairly upon notice. Since the complaint alleges facts that would (or at least might) allow relief for violation 3, violation 3 has been pleadbut only by implication, and inelegantly at that.18
More damagingly, though, the complaint includes five unnecessary words that sharply limit the scope of any potential relief: "for its own commercial use". Remember, "commercial use" is just one possible conclusion concerning one of the four fair use factors. As a hypothetical, would it be a violation if the University of Michigan did exactly the same thing as Google is doing/is proposing to do? If so, what if the contract between Google and the University of Michigan et al. characterizes Google as a subcontractor, doing something that the universities therefore could do on their own? Would that transform Google's behavior into a commercial, as opposed to scholarly, use? Licenses, after all, can have transitive properties: Absent a specific prohibition, one can ordinarily sublicense an action one is authorized to do through a license from a third party. And, of course, this gets vastly more complex in light of the special privileges accorded libraries under 17 U.S.C. § 108.
So, through some poor drafting, the complaint leaves itself open to serious problems both in substance and in rhetoric. A case of this nature must be won in the court of law and in the court of public opinion; winning in only one of those courts isn't enough. The language of the complaint sabotages the legal prong and subtly undermines the public prong.
McGraw-Hill, on behalf of a number of publishers, has filed its own parallel action against Google regarding GooglePrint. Although it isn't exactly the complaint I would have written, it is vastly better than is the AG complaint.
To begin with, McGraw-Hill's19 counsel at Debevoise & Plimpton wisely evaded the question of class certification by not filing as a class. Not only does this avoid my initial criticisms of the particular problems with the AG class definitionand, sadly, these are problems that would also necessarily occur in any class of publishersbut the AAP suit explicitly requests "declaratory and preliminary and permanent injunctive relief" (Cmplt. ¶ 1), but not damages. By doing so, AAP would fashion a remedy that could be relied upon by any potential class member through offensive collateral estoppel20 without the sideshow of class certification.
The substance of the complaint is also much better conceived. Early on, it acknowledges the possible library privilege under 17 U.S.C. § 108, and properly distinguishes that privilege from Google's program (Cmplt. ¶¶ 67). Most importantly, the AAP complaint properly characterizes the opt-in/opt-out controversy concerning whole-work (what I have called "primary") copying.
33. Both the Google Library Project and Google's pronouncement that publishers must provide to Google detailed lists of books that they wish to be excluded are contrary to the black letter requirements of the Copyright Act. The exclusive rights enjoyed by copyright owners to reproduce, publicly distribute[,] and publicly display their works, and to authorize others to engage in such acts, 17 U.S.C. § 106, squarely put the burden on Google either to obtain the permission from the copyright owners to copy and make use of copyrighted books or exclude them from the Google Library Project. [remainder of paragraph redacted]
This is the real core of the controversy. The key point is that the Supreme Court has already spoken on this subject:
More to the point, even if the dissent is correct that some authors, in the long run, are helped, not hurt, by Database reproductions, the fact remains that the Authors who brought the case now before us have asserted their rights under § 201(c). We may not invoke our conception of their interests to diminish those rights.
The AAP complaint is also quite a bit shorter than is the AG complaint. Overall, it is a much cleaner statement of what is really at issue. That this "cleaner" statement comes from parties whose exclusive right is often far from clear (or clean)22 seems rather ironicbut all too typical in the murky world of copyright and the Internet.
The entire dispute arises from a multidimensional culture clash among authors,23 publishers, libraries, ordinary readers, and information entrepreneurs.24 In no particular order, those dimensions include:
The real problem here is inherent market failure at the boundaries of intellectual property, which itself has several dimensions. Most market theory is Coase-positive, in the sense that it assumes that most markets areor, at least, ideally should becompliant with the Coase theorem. Nowhere, however, is the endowment effect more apparent than in psychological, actual, and creative claims of authorship. Throw in grossly unbalanced initial power and economic resource allocations and information asymmetries that would be laughed off a law-school final exam as completely implausible, stir in the disjuncture between copyright (and, for that matter, patent and trademark) duration and both the perceived rate of Progress and lengthening lifespans,26 and we have the ingredients for a swirling brew of conflicts troubled by hotspots no amount of stirring seems able to disspell.
No, I don't have a universal solution. Neither, however, do I have a whole lot of respect for either side in Author's Guild v. Google (and only a little bit more for the AAP version). Both sides are so focussed on short-term enlightened self-interest that they've forgotten that overriding purpose: encouraging Progress.
and reality in publishing (seldom the same thing!) from the author's
side of the slush pile, with occasional forays into military affairs,
censorship and the First Amendment, legal theory, and anything else
that strikes me as interesting.
Ritual disclaimer: This blog contains legal commentary, but it is only general commentary. It does not constitute legal advice for your situation. It does not create an attorney-client relationship or any other expectation of confidentiality, nor is it an offer of representation.
All material © 200311 except where otherwise indicated. All rights reserved. N.B. This blawg does not use the Creative Commons License, although I'm usually pretty good-natured about permissions for attributed reuse.
Essays in Warped Weft are consolidated from Scrivener's Error with only formatting corrections (and the occasional typographical correction).