Here is a quick look at what I perceive to be the major changes.

First, and probably most significant, the Settlement is now very clear that it applies only to books registered for copyright in the US or published in the UK, Canada or Australia.  This is an obvious attempt to avoid the objections and potential legal complications of including in the new Google products books subject to different copyright regimes and different author expectations; this comment on the Stanford Cyberlaw blog calls it “a policy-oriented maneuver intended to remove political pressure coming from abroad.”  It is interesting to consider what this means for the actual Google Book search results.  Will we continue to see snippets only of “foreign” books scanned by Google in the free database?  If so, does that indicate that Google really is continuing to assert fair use in some cases?  Or will these books disappear altogether?  In any case, we know that under the new terms, such works from other countries will not be part of the institutional subscription database, and that fact will substantially lower the value of the product.  Many of the works that will now be excluded are the very ones that research libraries have the most difficulty obtaining via interlibrary loan, so they are ones we were most counting on finding in the Google subscription database.  We must expect that the pricing of that database will be adjusted to reflect a significant decrease in value offered.

The next important thing to note about the amended settlement is what it does not do.  It does not address the concerns over reader privacy that have been expressed by many groups.  The Electronic Frontier Foundation expresses its disappointment about this lack of a privacy plan here.  In a conversation about this topic yesterday, a colleague of mine made the point that for Google to offer a privacy plan for this product would beg the question of privacy policies enterprise-wide, and that is a discussion Google, which depends on targeted advertising, does not want to have.

On the positive side, the amended settlement does clarify that rights holders can elect to allow free availability of their out-of-print but still in-copyright books, or to have them released subject to a Creative Commons license.  This will be a significant benefit to academics who have retained or reclaimed copyright in their own books; unfortunately, that is a much smaller class of authors than it should be.

In many ways the heart of the settlement, its huge compulsory license to Google to commercialize orphan works and other unclaimed titles, has not changed much.  There is a nice discussion of this aspect of the matter in this New York Times article.  The basic change being made is some restriction on how the money generated by sales of these works can be spent, and the appointment of an independent “fiduciary” to the Books Rights Registry board to protect the interests of this immense unrepresented class  The fundamental legal problem of an inappropriate use of the class action suit to create this license does not, and probably cannot, change.  Without that license, there is little if any incentive for the parties to bargain at all.

The great unknown in all of this is whether the Books Rights registry will be able to license other parties to exploit orphan works in ways similar to the opportunity the settlement creates for Google.  The provisions of the first settlement that had been known as the “most favored nation” clause, which said that the BRR could not give anyone else more favorable terms than Google got, has been removed — see the changes in sections 3.8(a) and 6.2(b) in the above-linked “redline” version of the settlement.  So it is now the case that if the BRR can license orphan works at all, it can give favorable terms to other parties.  But that is a very big “if.”  As James Grimmelmann explains in this post, it is not at all clear that the BRR will have this authority until and unless Congress acts to resolve the orphan works dilemma legislatively.  The NY Times article also seems to believe that this change depends on Congressional approval before the new independent fiduciary can assist with other orphan works projects.  So Google likely still has its exclusive position, which could only be threatened by an unlikely combination of Congressional action and very deep pockets.

Overall, it is hard to argue with the title of this article from Library Journal which suggests that the amended settlement agreement makes only minor changes, and, as noted above, the most significant change probably decreases the overall utility of the settlement to academic libraries.  The LJ article quotes a spokesman for the Open Book Alliance (which opposes the Google project) to the effect that the settlement is “sleight of hand.”  Even allowing for the bias source of this quote, I think it reflects a truth, that Google and its partners (which is what they are, as opposed to opponents, at this point) wanted to change the agreement as little as was necessary to slip by the complaints raised by the Justice Department.  As the NY Times article notes, that is really the only critic of the original deal that the settlement is designed to placate.  Whether it has gone far enough to satisfy the DoJ is still an open question.

On TechDirt, notice of this arrangement provoked a lot of angst.  Many of the comments expressed outrage at the “fact” that ordinary citizens who have to pay a copyright fee for photographs they take of public buildings, because of the copyright protection afforded to architecture.  That this is the state of the law is affirmed by several of those comments.  In contrast, the blogger who wrote the TechDirt post in the first place asked a differently focused question –”why should the painter have to pay a fee at all?”

All of the venting in the comments on this blog post reminded me of an article I have been reading by Professor Jessica Litman, about which I shall say more in another post, in which she discusses the “independent threat to the health of the copyright system” that arises from the “widespread perception of the current copyright system as illegitimate.”  The outraged comments point up this perception, even though they are largely misinformed.  The important question to address in this particular situation is really that other one — why pay a fee at all? — since the answer should allay some of the outrage.

The basic response to the concern for photographers and artists is that the copyright law provides an explicit exception, a limitation on the scope of the right in architectural works, that makes most drawing, paintings and photos of buildings non-infringing.  Section 120(a) of the Copyright Act (Title 17 of the US Code) says:

Pictorial Representations permitted — The copyright in an architectural work that has been constructed does not include the right to prevent the making, distributing or public display of pictures, paintings, photographs and other picotial representations of the work, if the building in which the work is embodied or is ordinarily visible from a public place.

So there is no way under copyright law for UT Austin or any other building owner to prevent, or extract fees, for paintings and photographs of such buildings, either because they were constructed before copyright protection extended to buildings (as opposed to just plans) or because of the exception quoted above.

Other types of protection that could allow the extraction of a fee from the artist would usually be trademark issues.  University emblems will almost certainly be subject to such protection.  The issue of the burnt orange color scheme, however, is much more doubtful, especially after the decision earlier this month involving a similar issue around an artist’s use of Alabama’s crimson colors in the U.S. District Court for Northern Alabama.

So to return to the question of why the artist should pay this fee, one possible answer, I think, is that he wants to.  For paintings of buildings on the public property of the University, such a fee is probably neither required or enforceable, unless there are trademark elements included.  But the artist could have entered into a voluntary licensing agreement with the University, perhaps out of a sense of loyalty or fairness.  Or, of course, he may just be badly misinformed.  Unfortunately, we frequently encounter situations in which someone asks about a license in a situation where none is required and ends up paying an unnecessary fee.  Copyright owners (or putative owners) have little incentive to correct these misapprehensions.

In the Autumn 2009 issue of The American Scholar, William J. Quirk writes an absolutely fascinating reflection on the finances of F. Scott Fitzgerald, whose tax returns and yearly financial ledgers were preserved and form the basis for Quirk’s essay called “Living on $500,000 a Year.”

The essay will be of interest to many people who are not obsessed with copyright issues, but one line struck my obsession very deeply.  According to Quirk, when Fitzgerald died in 1940, “his estate was solvent but modest – around $35,000, mostly from an insurance policy.  The tax appraisers considered the copyrights worthless.”  Fitzgerald’s copyright, of course, were not worthless over time.  As Quirk tells us elsewhere, the royalties from the sale of The Great Gatsby continue to generate about half a million dollars every year for the trust set up by Fitzgerald’s daughter Scottie to benefit her children.

This would seem to be one of the rare cases where the long term of copyright protection (Gatsby, published in 1925, will be protected until 2020) continues to benefit the descendants of the author.  Usually, of course, copyrighted works have ceased to generate any income at all after only a few years, and those who inherit the rights generally neither know that they hold them nor get any benefit.  Fitzgerald, however, arguably presents a strong, if unusual, case for long-lasting copyright protection.  But when we look deeper, we see that the incentive that copyright is supposed to provide probably was overblown even in this case.

First, it is not very likely that the knowledge that his writings would make his grandchildren wealthy really played any part in Fitzgerald’s decision to write his novels.  Indeed, Quirk’s essay tells the story of a man driven to work partly by the need to make money to keep a roof over his head from day to day (which is why he wrote short stories) and partly by the need to express himself (which is what he wrote his novels to do).  The copyright incentive worked in the short term – it made it possible for Fitzgerald to sell his stories and novels for relatively healthy sums – but all of that incentive was immediate.  The thought of riches two generations in the future was no part of what made Scott write.

Also, if the copyrights were considered worthless at his death, it is hard to see how Fitzgerald could have imagined profits for his grandchildren.  Even in this case, the incentive argument for an average copyright term of 95 years rests on the absurd assumption that authors can predict future success in the face of present failure (or perhaps that they are even more deluded than the rest of us).

Even if the copyright incentive can work in some cases, of course, it must remain in the hands of the authors.  If Fitzgerald had transferred his copyright to publishers, as is common practice today, there is little chance that his grandchildren would be enjoying all of that money, especially since there was a substantial period of time during which no one could anticipate the rebirth of interest in his books.  Once again we come to the inescapable conclusion that the justification for strong copyright protection, which is the incentive it provides to authors, is only valid if copyright remains in the hands of those authors.  To be effective at all, copyright must be, and remain, an author’s right.

First there are these stories about file-sharing of scholarly articles amongst medical researchers – one from the Chronicle of Higher Education here and the other from Techdirt here.  The Chronicle is particularly scolding in its tone, and it evokes the misnomer “piracy” in its title.  But aside from that commonplace rhetorical strategy, I want to emphasis two points raised by these reports.

First, the estimate of how much money journal publishers would lose by this practice, which the Chronicle sets at $1.4 million per year, should be taken with a ton of salt.  As has become very well known in recent times, these estimates are usually built on false assumptions.  The recording industry, for example, often assumes that each unauthorized song download costs the industry the full price of a CD.  But there is no reason to believe that that is true; the consumer might be unwilling to pay for an entire CD to get one song and prefer to forgo the music altogether or pay to download a single track if the free option were not available.  Likewise, an article shared over a peer-to-per network does not translate to a lost subscription or even, necessarily, to a lost per-article fee.  We just do not know how much access would be worth to a consumer, and the copyright monopoly has prevented us from ever getting reliable market data.

Second, we should remember that there is a big difference between music file-sharing and the swapping of academic journal articles.  In the latter case, academics are on both ends of the transaction; they are the authors as well as the consumers of the articles that are exchanged.   If it were not for the academic practice of giving away copyright to publishers, this would be no big deal at all.  Because of the lack of a financial incentive for academic authors, file-sharing of academic articles causes no economic lose to producers.  What makes it newsworthy, and, from the academic point of view, necessary, is that copyright is held by entities other than the authors.  By transferring copyright wholesale, instead of granting temporally-limited licenses to publish, academic authors have help create the access problem they are now trying to solve with file-sharing networks.  That doing so is potentially an infringement of copyright is evidence of how harmful this practice has become to the fundamental mission of colleges and university.

And this gets me back to the point about copyright as an author’s right.  For copyright to function, it must serve the needs of creators; if it does not, its fundamental purpose, which is to create incentives that encourage authorship and other forms of creative expression, is defeated.  When academic authors give away their copyrights, and then have to resort to “illegal” file-sharing to get access to fundamental research, the copyright system has broken down.  Only by reclaiming their copyright entitlement can academic solve this problem.

In a few days I will look at this problem of copyright as an author’s right from a very different perspective, based on an article about the finances of F. Scott Fitzgerald.

Earlier this month, a district court in Michigan handed down such a decision in yet another “course pack” case in that state.  There was a twist this time, however, which the court chose to ignore.  In this case, brought by Blackwell, Elsevier, Oxford, Sage and John Wiley publishers, the copy shop received photocopied course packs from professors, than handed them to individual students who made copies for their own use.  Amazingly, the court found that this practice constituted direct infringement by the copy shop of the copyright holder’s exclusive rights of reproduction and distribution.

The problem, of course, is that no employee of the copy shop took any action that actually implicated either of these rights.  The shop, owned by Excel Research Group, actually made no copies of any of the material over which the suit was brought.   Judge Avern Cohn contemptuously dismisses this point, asserting that “the fact that the students push a button on a copier in the manner described is of no significance.”  But such facts are exactly what are significant in legal reasoning, and the judge offers no principled reason for ignoring this fact other than his apparent desire to see the plaintiffs win their case.

Even more troubling was the treatment of the distribution right, which the court said was infringed by the shop handing out the course packs it was given.  For this to be true, those copies would have to be unlawfully made, so that the doctrine of first sale did not apply to them.  But Judge Cohn made no such finding; he simply noted that the packs were brought to the shop by professors who assigned the readings to their students.  Again, the judge seems disinterested in either the details of the law or the facts before him; he simply appears anxious to arrive at the conclusion he thinks is desirable.

This startlingly bad reasoning serves a purpose for both the court and the publishers, of course.  It is the only way, albeit one that requires ignoring both law and facts, to hold the copy shop liable without also saying that the copies made by faculty members and students were themselves infringing.  The shop’s liability, if any, is clearly contributory to direct infringement by students and their instructors.  But the court did not want to hold this way, and I am sure the publishers did not ask them to.  To find direct infringement would be so obviously to attack the basic necessities of education, and would so clearly contravene the intention of Congress when they included “multiple copies for classroom use” in the list of examples of fair use, that it was too politically dangerous.  And publishers would fear, no doubt, a decision that would suggest to their customers that their products truly are unusable. So rather than find direct infringement by teachers and students so as to hold the copy shop indirectly liable, the court rendered an incoherent decision in which Excel is held liable for directly infringing copyright without making a single copy.

There is some excellent analysis of this decision by Peter Hirtle here and by Shourin Sen here.

The result of this case is similar to what we have been decrying for some time now — a creeping expansion of the exclusive rights in copyright way beyond the boundaries Congress set for them.  Here that expansion has been abeted by a judge who apparently sees his role as a kind of knight-errant, righting every wrong he perceives, regardless of the legal foundation.

Jon is a Ph.D. candidate in Biomedical Engineering at Duke and a U.S. Marine Captain (Ret.). He is also an amputee, having lost his right arm in Iraq four years ago, and is thus a researcher with a personal interest in prosthetics.  He talked to us about why the money the government spends on R&D for prosthetic research does not produce the kinds of progress that it ought — the lack of coordination and such a small market that there is little incentive to move from workbench to marketplace once the research money is spent.  Jon offered potential solutions for this lack of progress that addressed both his very specific research and the broader problem of intellectual property restrictions.

In the very specific area of his own work on arm prosthetics, Jon envisions a remarkable collaboration, made possible by open hardware.  He would like to make the hardware being developed to improve neural control of prosthetic arms open and offer it to researchers in the video game industry.  His hope is that work undertaken to create new video game controllers (an area with a much larger market and much more money to spend) will also speed the development of better artificial arms, which has been largely stalled for quite a few years.

This is a remarkable vision, I think, of a win-win collaboration that would be founded on open sharing of technological development.  Openness, as some have been pointing out for quite a while, can breathe new vitality into innovation, in spite of claims from some industries that free access can only stifle and discourage it.  More information about the video controller project can be found at http://openprosthetics.wikispot.org/Open_Myoelectric_Signal_Processor

Jon Kuniholm does not stop with this vision of collaboration, however.  He has a concrete and well-informed notion of the mechanisms needed to bring it about.  I spoke with him briefly before the event about the intellectual property issues involved with this idea.  He pointed out that hardware can be shared openly from its inception because patent protection, unlike copyright, is not automatic and is, in fact, quite costly to obtain.  Where copyright does cover a work, regarding plans and specifications, for example, Jon advocates using the open source GPL, or General Public License.  The problem with open hardware, however, would come if another party saw profit in the hardware and filed its own patent application Since patent restricts the use of an idea, this would halt all other development based on the same hardware unless license fees were paid.  Since patents in the US law are granted to the first to invent (rather than the first to file a patent application), it would be possible, but very expensive, to fight such following-on patents.  Jon’s suggestion here is that the open hardware movement create mechanisms to publish what is called “prior art” — the science that leads up to new developments –in ways that will be very obvious to patent examiners.  The hope is that the ready availability of prior art will prevent patents from being issued that could shut down the kind of collaborative work based on open hardware that Jon and many others both need and foster.

Although it is often easy to spot the linguistic excess coming from the copyright industries, a recent letter to the Senate Judiciary Committee from the National Music Publishers Association took a more subtle, and even more dangerous, approach. There is a CNET news story about this letter here.  The theme of the letter, that copyright law should be technologically neutral, seems benign enough, but the work that the music publishing industry tries to get that rhetoric to do is very troubling.  The thrust of this “technological neutrality” appeal is a claim that music publishers should collect a fee for a public performance of a musical composition every time there is a digital download of a piece of music.

To call this grasp at a wholly new income stream “technological neutrality” shows amazing nerve; it is really the opposite of such neutrality.  Music publishers do not collect a public performance fee when a CD is sold because there is no way to prove or assume that a public performance (as opposed to a private one, over which rights holders have no control) will take place.  Why should a digital download be different?

Fred von Lohmann of the Electronic Frontier Foundation, who is quoted in the article, suggests that this is just a turf war between different rights societies over who will collect a fee and, hence, get a “cut.”  He is surely right about that, as he is when he points out that copyright law has never been technologically neutral.  Some exceptions (such as the section 108 library exceptions) apply only to certain technologies or treat different technologies differently.  There is a special rule, after all, for digital audio tape.  But pointing put the triviality of this use of “technological neutrality” may not be enough.  We should notice something really pernicious that is happening behind this smokescreen.

The language of copyright neutrality has quite a bit of appeal for copyright policy makers.  The fantasy of a law that adapts automatically to new innovation appeals to a legislative sense of economy.  That attraction is being used, in this letter, to attempt to vastly expand the scope of the exclusive rights protected by copyright.  And this is not the first time.  We should remember that copyright owners do not get absolute control over their works, only control within the scope of the enumerated rights.

A single line in the CNET story really encapsulates the problem here — “composers, songwriters and publishers are asking for a guarantee that they will get paid for a public performance even if there isn’t a public performance.”  In this letter, the apparently benign call for technological neutrality is being used to disguise an attempt to enlarge beyond all reason the scope of the public performance right.  This is not the first effort to use that right to expand the reach of the copyright monopoly.  As I wrote about here, the debacle regarding the Kindle text-to-voice feature was based on an attempt to expand “public” performance deeply into the private use of new technologies.  For another example, see this report on the unsuccessful attempt recently by music publishers to collect a fee for every ring-tone “performance” of copyrighted music.  So the desire to expand the reach of copyright control is well-established, what changes is the disingenuous rhetoric behind which these efforts are hidden.

One of the persistent criticisms — or perhaps reservations is the correct word — about the Creative Commons licensing scheme has been that one of the major terms used in CC licensing — non-commercial use — is too vague and subject to varying interpretations.  The core purpose of the Creative Commons, of course, is to allow copyright holders to license their works in a way that assures subsequent users that they can make use of the works within defined parameters.  Two of those parameters are attribution, which is protected by CC licenses even though not adequately ensured under U.S. copyright law alone, and, often, a restriction to non-commercial uses.  But if there is no agreement on what it means to call a use non-commercial, then there is a real problem with the licensing scheme; it would fail to provide that assurance, which reduces the need for transaction costs involved in seek permission, that is its basic purpose.

Now the Creative Commons has released a voluminous report it commissioned to study this potential problem.  Although “Defining Noncommercial” is a massive document that I have not read in its entirety, it is clear from the executive summary and a perusal of the survey data that the situation is not really as serious as some feared.  The report suggests that although a comprehensive definition of noncommercial remains elusive, there is not a major problem with its use in CC licensing.  Basically, most people seem to agree that “they know it when they see it.”

Two specific findings in the report struck me as particularly supportive of the continued use of “non-commercial” as a licensing term.  First, the marketing firm that did the research found that there was broad agreement on what non-commercial meant.  Most creators and users agreed that a use that made money for the user or involved advertising was commercial, while those that did not, were not.  This broad agreement helps explain why the millions of items licensed under CC licenses have generated so little litigation in the eight years since its founding.

Even more interesting was the finding that showed that when creators and users disagreed about whether or not a use was commercial, it was the users who were more likely to err on the side of seeing a use as commercial, and thus not covered by an “nc” license.  The reason this is such an encouraging finding is that it suggests that users will ask permission in doubtful cases, even when the creators (who hold the rights) do not think permission is needed in the particular situation.  Thus CC licenses can reduce the transaction costs involved in seeking permission, but they will not eliminate all need for permission and users are likely to ask when they are in doubt.  CC licensing, of course, facilitates asking permission as well, since works so licensed will have an identifiable rights holder.

This finding is consistent with our experience at the Duke University Libraries, where we placed most of our web pages under a CC license over two years ago.  We still do receive some requests for permission, even for pages that carry the CC license.  I try to inquire about why people are asking when the page carries a prior permission that almost always covers the proposed use.  Invariably I am told (mostly by other librarians) that they consider asking both the cautious and the courteous thing to do.  So while we believe that the license empowers many users and reduces transaction costs, we also see that users who are in doubt feel free to contact us for clarification.  This confirms that the non-commercial term is not the problem that some have feared.

Creative Commons has always striven to make its licenses effective and useful, and this study is one more tool for understanding how those licenses are and can be employed.  The CC itself suggests three lessons that we can take away from this study:

the findings suggest some reasons for the ongoing success of Creative Commons NC licenses, rules of thumb for licensors releasing works under NC licenses and licensees using works released under NC licenses, and serve as a reminder to would-be users of the NC licenses to consider carefully the potential societal costs of a decision to restrict commercial use.

Good advice, available for those who want to be sure that, in regard to non-commercial use, we “know it when we see it.”

Harkaway also has my agreement when he expands his discussion to the problem of orphan works, and suggests that the Google Books deal gives added incentive to a broader, more generalizable solution for the millions of works still protected by copyright yet for which no rights holder can be found.  Harkaway embraces a familiar solution to this problem when he endorses renewed recourse to a renewal system.  Under this plan, rights holders would have to renew their copyright claim periodically in order to prevent the work from dropping into the public domain.  Thus orphan works would become free for use once a renewal period passed without action by the rights holder.    There are other ways to approach the orphan works problem, but it clearly needs to be addressed, and the renewal suggestion would be one very effective approach.

Unfortunately, I stopped agreeing with Harkaway right at the end of his article, when he suggested that data-mining and other new uses for copyrighted works should be sources of new income for rights holders.  This is an old mistake based on thinking that whenever new technologies enable new uses, a new right is created.  But copyright does not work that way, and there has never been a “use right.”  Copyright holders do not get the right to control every use of their work, and thinking about how such a right might work should tell us why — it raises a huge problem of censorship; imagine, for example a book author or film producer who could use copyright to prevent negative reviews.  Instead, rights holders get the exclusive right to control copying, distribution, public performance and public display, as well as the creation of derivative works. This is a lot of control, but these rights do not impinge on using a lawfully obtained copy, at least for private purposes like research.  Everytime a new technology comes along, however, some rights holders are seduced into thinking that they should gain from it, even if it does not implicate any of these exclusive rights.

If digital copies of the world’s books are legally created, through a Google settlement or in some other way, use of those copies for data-mining and other research uses will be, and should remain, free for all users.  It may sound plausible when Harkaway complains that Google will be improving its search algorithm using his work and making money from that improvement.  But where does a use right stop?  Should the heirs on John Updike be reimbursed if digital copies of his work are used to create a Updike concordance?  Should an academic who wants to study a certain grammatical construction across a huge range of published literature, a use contemplated by the Google settlement, have to pay the copyright owner of every book in the corpus for that opportunity?  It quickly becomes clear why a separate use right within the copyright bundle would be a very bad idea.  I can follow Harkaway through most of his article, but when he gets to those last three paragraphs, it is clear he has gone astray.

The case is fairly complicated, and there is a nice summary of it here, on the IP Law blog.  The basic ruling, however, from the Ninth Circuit Court of Appeals, was that the doctrine of first sale, the rule that says that one who purchases a lawfully made copy of a copyrighted work may lend, resell, or otherwise dispose of that particular copy, does not apply to works that are manufactured and sold outside the US.  Basically, the court held, on reasonably good authority, that such works do not qualify as “lawfully made under this title (i.e. the Copyright Act),” which is a condition on the application of first sale.

Once I paid attention, it became very clear why this is a cause for concern in libraries.  Academic libraries especially buy lots of foreign materials, often from overseas distributors.  If first sale does not apply to those materials, can libraries lend them at all?  A negative answer could devastate our services in support of all kinds of language programs and area studies.  This possibility is raised in passing in this amicus brief urging the Supreme Court to review the case, filed by the Electronic Frontier Foundation.  Interestingly, however, the major library associations have not taken a position on the petition asking the Supremes to hear the case.  I was given two different reasons for this decision not to act, one which seems sound to me and one which leaves me with some concern.

One reason for not encouraging the Supreme Court to “take cert” (that is, agree to review the lower court’s opinion) is that there is real danger that the Supreme Court would affirm the decision.  That would make a problematic case from the West coast into binding law throughout the country.  Better, perhaps, that this remain an anomalous precedent only impacting libraries in the nine western states that comprise the Ninth Circuit.  Several authorities (Patry on Copyright and a concurring opinion in an earlier Supreme Court case) seem to support the position taken by the appeals court, and asking for cert might be asking for trouble.

More reassuring, but more problematic, is the other reason given for not taking action on this case — the exception for libraries that is built into the manufacturing clauses.  Section 602(a)(3) excludes certain copies purchased by libraries for lending or archival purposes from the general statement in 602 that importation of copies of copyrighted works purchased overseas into the US is an infringement of the distribution right.  That seems to let libraries off the hook.  But it is not entirely clear that this exception, specific as it is to section 602, actually solves the first sale problem created by the Ninth Circuit.  Even if it does, however, I am left with two concerns.

First, the section 602(a)(3) exception explicitly excludes audiovisual works from its scope.  For those works, only a single copy for archival purposes is allowed, and no mention of lending is made.  This suggests that even if print collections of foreign materials purchased overseas are OK, collections of film are not.  That would be a crippling lacunae for academic libraries.

The other problem is that, if the Ninth Circuit ruling stands, it might encourage textbook publishers to move their manufacturing and distributing operations overseas in order to be able to shut down secondary markets and thereby increase their profits.  The exception for libraries would not apply to resale of used textbooks, on which so many students depend to reduce their educational costs.  Closing off those used book markets would not directly harm academic libraries, but it would certainly hurt higher education.  Also, it hardly seems sensible to add to the incentives that are luring American manufacturing overseas.

I am thus left on the horns of a dilemma.  I want to see this decision overturned, but I agree that the review that would be necessary to reverse it would also carry a significant risk of an affirmation, which would be far worse.  It is an uncomfortable place to be, and one in which a good outcome is difficult to imagine.