And the details...
The Copyright-y Stuff (wait, no, not quite yet)
Before getting to the meat of it all, it's worth noting that the copyright stuff is mostly in response to a Motion for Summary Judgment. MSJ's are filed early in the process of a case, and are basically an argument that it's not worth it to have a whole trial and the rest of the court process shebang, because the arguments for your side are so completely convincing that we can see right up front who's gonna win. Until pretty recently, it was unusual for copyright cases (especially fair use cases) to be decided at summary judgment or at other early points in court proceedings, because the copyright holders usually argued that they needed the whole court process to fully argue the fine points of fair use. (I worked on part of a brief, as a student intern, where we had to make the argument that it was even appropriate to consider a fair use defense at the summary judgment stage.)
Anyway, it has more recently become pretty common to make MSJ's in copyright cases (thanks to DMCA takedowns and Righthaven, basically.) So now, it's pretty common for both sides to file these motions (as they did in this case), which means they're both arguing that their own arguments are a slam-dunk. When both sides do have reasonably good arguments, the court usually denies both MSJs, and proceeds further with the case. But winning on summary judgment means the court agrees your arguments are a slam-dunk. It's a very good place to be; appeals of decisions on MSJ are often a significantly harder process than appeals from a regular judgment. As James Grimmelmann said, this opinion "makes the case seem so lopsided that it makes the appeal into an uphill battle."
The Copyright-y Stuff (for reals)
One of my least favorite bits of the plaintiff's filings argued that, because libraries have a list of specific kinds of copying they 100%, no-question, can do (in 17 U.S.C. § 108), anything outside that list is something libraries cannot do - that is, that the kinds of copying listed in section 108 are the only kinds of copying libraries can do. This is a pretty ridiculous argument (and I was not alone in my frustration with it), because section 108 says it doesn't affect section 107 (fair use) rights. The plaintiffs argued it should be read another way, and offered legislative history to support their arguments, but the judge didn't buy it. "nothing Plaintiffs submitted convinces me that fair use is unavailable as a defense, or that the manner of reproduction is prohibited simply because it does not fall within Section 108." (p. 13)
Because of what I mentioned above about summary judgment, in order to get the case dismissed, Hathi and their defendants had to prove that there was "no genuine issue of material fact" over whether their use was a fair use. Although Judge Baer looks at fair use quite holistically, he does examine each of the fair use factors individually - all the Authors Guild had to do was allege (not prove) that even one of the factors was worth arguing out at greater length, and they could have forced this to go to trial. But they didn't.
Factors 2 and 3: the quickies
Factor 2 inquires as to the nature of the copyrighted work. Here, the court skipped completely over the published/unpublished issue sometimes raised (probably because all the works in question were unquestionably published), and went right to the "factual/creative" distinction that is sometimes made. While acknowledging the standard that there is often more room for fair use of more factual materials, the court largely brushes this aside as not terribly relevant. Citing both the 2 Live Crew (Campbell v. Acuff-Rose) and the Grateful Dead posters (Bill Graham Archives v. Dorling Kindersley) cases, the court held that for transformative uses, this factor is "not dispositive" (p. 18), i.e., doesn't really affect the outcome of the case.
Factor 3 asks how much of the original work was used. In many fair use cases, the parties fight a lot about this: what fraction of the work was used, what the relevant frame of reference is (is an illustration part of a book, or all of that illustration), etc. Many providers of fair use advice and information will give hard number or percentage guidelines as to the upper limit "amount" for fair use. And in this case, HathiTrust unquestionably had made copies of entire books for indexing purposes, and was showing copies of entire books to users with print disabilities.
But the judge, again citing some of the foundational transformative use cases, including Campbell, affirmed that the relevant question is whether the amount taken was necessary to the use, and that "[s]ometimes it is necessary to copy entire works." (p. 18) Even if the whole copies of the books could have been discarded for searchability purposes once the index is constructed, retaining and showing whole copies to users with print disabilities (and preservation, maybe, too, N 26) is a sufficient reason to retain them long-term. (p. 19)
Factor 1: Purpose
There's a lot inside this one. First, a few more quick hits: Hathi's scholarly and research purposes weigh in their favor on fair use. Although the judge did say that preservation copying, on its own, may not be transformative, he also said that preservation copying for noncommercial purposes is likely to be fair use (which, hey, libraries waiting on preserving decaying VHS until you figure out whether section 108 applies? Might want to make note of this one.) (p. 15, N 19). The court also rejects the argument that partnering with Google and/or failing to buy extra copies of the works made the purpose commercial. (p. 17)
The court explores transformative use in some depth, affirming that transformative uses do not have to actually change the work, but can also simply serve "an entirely different purpose" (p. 16). Even though Hathi made copies of entire works, the court holds that since the purpose of the Hathi copies is "superior search capabilities rather than actual access to copyrighted material" (p. 16), this is a transformative purpose. The judge goes on to also affirm that providing access to people with print disabilities is transformative use because they're not a significant market for the publishers and therefore sales to them weren't part of the original purpose. That's pretty convoluted, and could be used to argue that if the publishers do start selling accessible copies, the use is no longer transformative, so it's nice that the court also cites a case suggesting that accessibility is probably a fair use regardless of purpose.
There is a great aside in one of the footnotes that both the plaintiffs' argument that fair use just shouldn't be available as a defense here, and that copying published books for search indexing is somehow different from copying online images for search indexing "relies heavily on the incorrect assumption that the scale of Defendants' copying automatically renders it unlawful." (p. 17 N23) I know I encounter a lot of people who do have some instinct that larger-scale copying is inherently less amenable to fair use justifications, so it's nice to see a court affirming that if it's fair use, it's fair use - the scale of copying is not really a relevant consideration!
Factor 4: Market Harm
The judge's insistence that "[a] copyright holder cannot preempt a transformative market" (p. 20) is a marked contrast to the deference courts often give to the supposedly delicate and highly imaginative "potential markets" copyright holders suggest they might possibly want to exploit down the road. The opinion's well-reasoned and detailed support for the lack of market harms from noncommercial and/or transformative use is almost enough to make this lawyerbrarian dance with glee!
There's a couple of quick hits here, too: because the use is noncommercial (and again, mere association with Google doesn't make it commercial), the plaintiffs actually had to prove that there was some market harm, or that a "meaningful likelihood of future harm exists." Their claim that each copy made by Hathi was a lost sale is dismissed outright, because "purchase of an additional copy would not have allowed either full-text searches or access for the print-disabled individuals, two transformative uses that are central to the MDP." (p. 19) Their claim that Hathi copies are "immense security risks" was unfortunately not supported by any factual evidence, since their expert admitted he didn't know what security procedures were in place, and Hathi detailed that they did take precautions (whatup, Cory), and are a trustworthy depository.
So the market harm eventually boils down to an argument that the Hathi copies are (or will) preclude the development of future licensing opportunities. The court relies on existing precedents to strongly state that where a use is transformative, copyright holders cannot undermine fair use just by saying that they intend to license that use in the future. "A copyright holder cannot preempt a transformative market." p. 20
The defendants had suggested that a licensing market for this kind of use was unlikely to develop because it would be prohibitively expensive, both in the licensing fees extracted, and in administrative overhead. Although plaintiffs characterized this argument as saying "it is permissible to steal the goods if it is too expensive to buy them," the court affirmed defendant's argument that "the high costs will prohibit the formation of a viable market in the first place" (p. 21).
Finally, the plaintiffs own arguments as to the small size of the population of users with print disabilities at the University of Michigan was used against it to affirm that the size of the market for accessible copies is negligible, and thus not a source of market harm. Again, the court emphasized that, in addition, the ADA and other laws intend for accessible copies to be made available.
"Balancing the Fair-Use Factors"
The decision wraps up the fair use discussion by looking at the use(s), as a whole. "The enhanced search capabilities that reveal no in-copyright material, the protection of Defendants' fragile books, and, perhaps most importantly, the unprecedented ability of print-disabled individuals to have an equal opportunity to compete with their sighted peers in the ways imagined by the ADA protect the copies made by Defendants as fair use" (p. 21) "Although I recognize that the facts here may on some levels be without precedent, I am convinced that they fall safely within the protection of fair use such that there is no genuine issue of material fact. I cannot imagine a definition of fair use that would not encompass the transformative uses made by Defendants' MDP and would require that I terminate this invaluable contribution to the progress of science and cultivation of the arts that at the same time effectuates the ideals espoused by the ADA." (p. 22)
The ADA and Copyright
The opinion also affirms that the Chafee Amendment to the Copyright Act (17 U.S.C. § 121) permits copying of "previously published, non-dramatic literary works", and that it is entirely appropriate for Hathi to have done so. Although plaintiffs argued again that, as with section 108, copying beyond the bounds of the Chafee Amendment was not allowed, the court also rejected that argument, affirming that fair use may cover some copying for accessibility that the Chafee Amendment does not. (p. 23, N 33). This is awesome, because access for users with print disabilities is such a no-brainer to me that I cannot believe publishers fight against it (but they do, see all of the last few years of WIPO negotiations!) I'm extremely happy to see an opinion affirming the ability of other entities to provide access, if the original publishers and/or copyright holders do not.
The Other Stuff
You may have noticed that my quotes above are all from pages 13 or higher. That's because the whole first part of the opinion is about other legal issues, such as whether the Author's Guild can sue on behalf of its members, and whether anyone can sue about the Orphan Works Project yet. More on that soon, hopefully.
The Other Other Stuff
I am intensely grateful to the folks at the HathiTrust who made the initial decisions to participate in these projects, knowing there was likely to be legal objection, and to fight the good fight (now, and possibly in the future) with this case. Too often, users with high, good, socially- and legally-valued purposes make the choice not to engage in a use because it is "too risky" to rely on fair use. Every time that happens, fair use shrinks and becomes more brittle. It does take resources to make a stand, and not every individual or institution can take these risks - but HathiTrust's decisions to take them on benefit all of us in libraries, and every user around the world. Copyright exists (at least in U.S. law) to "promote the progress of science and useful arts" - today's decision is a big win for fair use, and for progress!