The latest such lawsuit, filed on March 16 in the Southern District of New York, comes from three major textbook publishers – Pearson Education, Cengage Learning, and MacMillian Higher Education – and targets a new Internet-based business called Boundless Learning that purports, according to the complaint, to create textbooks, using open educational resources, that parallel and can replace the high-priced textbooks that each of these publishers offer.  The sum of the complaint is that Boundless copies the selection, arrangement and organization of their textbooks much too closely and so constitutes copyright infringement.

One should never form conclusions about a lawsuit based only on the complaint, which is inevitably just one side of the story and is always pitched to make the defendant’s behavior look as dastardly as possible.  Nevertheless, it seems safe to conclude that this lawsuit is not as farfetched as some of the ones we have seen recently.  The plaintiffs are really just asking the court to help draw a line that has never been very clear in copyright jurisprudence, the line between idea and expression.  Their description of the situation makes it sound like the placement of that line is obvious, but it probably is not.  What it certainly is, however, is a vital line for higher education to examine and understand.

I occasionally run across an interesting assumption about copyright, which I think is common both in academia and in the wider world.  It is the belief that if you want to reuse an image or a figure without permission, you can simply redraw it and avoid all copyright entanglements.  Like all simple rules in copyright, this one is not true; the standard for infringement is “substantial similarity,” and a redrawn figure or image that is substantially similar to the original could still be found to be infringing.  Far better, I tell those who raise this possibility with me, to use the original figure or picture in a way that supports the assertion of fair use, or to get permission.

From the complaint, it sounds like Boundless Learning is doing something similar to this “redrawing” in hopes of avoiding copyright problems.  The plaintiffs assert that Boundless creates “shadow versions” of their copyrighted textbooks, imitating the arrangement of topics and sub-topics, the depth of coverage for each area, and even the choice of illustrative examples.  I fail to see how a running and a fishing bear illustrate the First and Second Laws of Thermodynamics, but the complaint claims that the Boundless biology text imitates Pearson’s standard text on the subject down to such similar, but not identical, pictures.

So the question is directly posed about where the line between idea and expression is to be fixed.  Is the example of a bear merely an idea, which cannot be protected by copyright, or does it represent a creative and expressive decision by the Pearson authors which is infringed even by a different picture?  To put it another way, how much actual expression must be copied, if any, to infringe on another’s protected expression?  We have seen the “derivative works” right expand a great deal over the years, so that today even characters in a work of fiction are often protected even when the alleged infringing character copies no actual expression from the original (as in the recent Catcher in the Rye sequel case).  Is there a boundary on that right?

Since scholarship is an inevitably cumulative process, in which each new work builds, more or less explicitly, on what has been done before, the boundaries of the derivative works right and the line between protected expression and public domain ideas are very important to understand.  If this case brings us a little more clarity, that would be good.  But the judge will need to be very careful not to develop rules that would inhibit the basic processes of teaching and research.

Finally, I want to point out an oddity I noticed in the complaint that was filed to initiate this case and which will bear watching (no pun intended).  In addition to naming Boundless as a defendant, the complaint lists 10 “Doe” defendants; people who are alleged to be complicit in the claimed infringement but whose names are not known and can only be discovered through court orders.  Naming such John and Jane Does is a common technique in file sharing lawsuits and in some of the more notorious copyright “trolling” cases.  But in this complaint it is quite odd, because we have no real clue what role these people are alleged to have played or why they are being sued.  Are they students who have ordered “shadow” textbooks from Boundless, employees of Boundless, or faculty members who have recommended Boundless to impoverished students? As the case goes forward, we may learn something about how to draw the line between idea and expression, and we may also discover just who the publishers are willing to target in their campaigns to use litigation to protect their markets, if and when the identities of these Doe defendants are revealed.

One way to deal with this concern, which we have undertaken in regard to RePec, is to work with the disciplinary repository to feed article statistics from the institutional repository into the rankings produced by the disciplinary one.  That method provides a more comprehensive and accurate ranking of the articles.  And such rankings are, of course, a more useful measure of impact than impact factors, which apply to journals but not to individual articles, can ever be.

I do not know if it is possible to connect institutional statistics to SSRN or not, but Donovan and Watson describe a different approach to addressing this question.  They begin by pointing out an assumption behind the question, that article readership is a zero-sum proposition, that there is a defined number of readers for any given scholarly article, so that new means of access will simply divide up that readership, not generate new “eyeballs on the article.”  This is the same assumption made by publishers who insist that self-archiving, or even national funder policies, imperial their revenue, and by those who argue that libraries will never spend subscription dollars on works that will be made available freely.  Donovan and Watson begin the process of showing that this assumption is false.

In their article the authors report two different research methods they employed to study the question of whether one repository siphons readers away from another repository, or whether, instead, readership grows overall when an article is available from multiple OA sources.  Both methods lead them to the same conclusion: multiple outlets produce additional readers, so the sensible course for an author who wants her work to have maximum impact is, as they say, to “use both!”  Far from harming the ranking in one database, availability of an article in a second repository appears to increase substantially the overall number of downloads.

I like this article for two particular reasons.  The first is that it attempts to find solid data on which to base the discussion.  Instead of mere assertions of “obvious” truths in the open access debate, many of which are based on that zero-sum assumption, Donovan and Watson attempt to move the discussion to real evidence that actually places that assumption in some doubt.  As we continue to explore business models and look for dissemination options that more fully serve the needs of scholarly authors, basing our discussions on real data would be a refreshing trend.

The second reason I like this article is that it appears to offer empirical evidence, beyond the many anecdotes that we have collected over the years, of the role of “unexpected readers” in increasing the reach of scholarly research.  The zero-sum assumption gives rise to the presumption that the current system works in an acceptable way merely because the people I expect to see my work can see it.  But open access offers the possible of discovering a myriad of readers who are not expected, either by publisher or author.  If we take seriously the idea that academic research is undertaken, in the end, for the good of society, these are precisely the readers we would want to see find our scholarship.  And to rule them out on the basis of an unproven assumption would be to sell ourselves short as scholars.

When I got back to my office, one of my first tasks was to read the complaint in the lawsuit filed by the e-text platform company Kno against textbook publisher Cengage, alleging breach of contract.  The allegation is that Cengage has breached its contract to allow publication of content that it owns on the Kno platform, allegedly because certain features of the platform infringe Cengage’s copyrights.

Two qualifications are necessary.  First, a complaint is only one side of the story, so conclusions cannot be drawn from it.  Second, this may just be a contract dispute in which the two parties are bargaining for the best terms they can get, and litigation is simply a strategy in the negotiation.  That would be an unfortunate use of judicial resources, but it does happen.  Nevertheless, it is worth looking at the copyright issues that are raised, even if they never reach the stage of a decision on their merits.

According to Kno’s complaint, Cengage alleges that several features of the Kno platform create impermissible derivative works.  Specifically, Cengage allegedly objects to a “smart links” feature that inserts links to external educational resources into the text, to a “quiz me” feature that can create review quizzes from certain diagrams, and a “journal” feature that allows studies to record their own notes and pulls out from the text the excerpt to which the relevant notes refer.  The pedagogical value of each of these features is obvious, I think, but it is interesting to ask if they really do create derivative works.

Traditionally, derivative works are those that adapt the actual expression that is protected in a work, and usually they adapt the entire body of that expression.  Thus, a translation or an adaptation (novel to play, play to movie, etc.) are the paradigmatic examples.  Based on these criteria, it does not seem like inserting “smart links” into a text creates a derivative work, just a more useful one.  On the other hand, the “Quiz Me” feature does adapt some of the original expression in the text, but it adapts only a small portion.  Here, I think Kno could argue that this is a transformative fair use rather than a derivative work (they do raise fair use as one potential response to Cengage’s objections).

It is the “Journal” feature that seems to be most in dispute, based on how much the complaint has to say about it.   Students repeatedly tell us, of course, that one prerequisite to adopting e-texts is the ability to annotate the works, so this seems like a necessary part of any e-text platform.  It also seems like a classic fair use of the excerpts.  Insofar as the journal is just a layer over the top of the text, it hardly seems to implicate copyright at all.  And where excerpts are pulled out for the student to comment upon, that is exactly what fair use permits.  It is hardly different than if the student kept a separate notebook and copied out key phrases and passages, as I did throughout law school.  If that is fair use, and no one really disputes that it is, so, it seems to me, is the journal feature of the Kno platform (as described in the complaint).

My biggest concern about the dispute described in this complaint is the possibility that it shows us another publisher trying to disable they very possibilities that make e-books attractive to consumers because they do not understand how those features work and feel threatened by them.  E-texts specifically offer tremendous new potential for innovative learning, and ways to study a subject that work for a variety of different learning styles.  But these are possibilities only if the publishers get over their intense fear of the digital environment and their express desire to introduce “inconveniences” so that their digital products mirror the limitations of the print world.

After all this, let’s go back to the debate about derivatives versus transformative fair use.  My proposed criteria for what makes a derivative do not entirely solve the question.  Both derivatives and transformative fair uses adapt the original expression of the work in question.  In two examples above I suggest that the amount of the original work that is used may make a difference (it is, after all, one of the fair use factors).  This is helpful, I think, but probably not sufficient.  Perhaps the determinative question will be if there is market harm; courts that find transformative fair use usually remark that there is no direct market competition, and no “customary” licensing market, for the new, transformative use.   These reflections suggest, I think, a broad outline of how to make this slippery distinction, but they do not make it easy.  And they suggest that the dispute between Kno and Cengage really will turn on the terms of the license that is at the heart of the issue.

It seems the university system has been surprised by the opposition the policy has generated, but such surprise speaks poorly of their awareness of the realities on their own campuses.  I am reminded of the surprise that content industries have expressed at the opposition to the Stop Online Piracy Act or the European dissent over ACTA.  What, exactly, did they expect?  The days when no one pays attention to such policies are past, and that is a very good thing.

The works created by faculty raise a complicated situation in a couple of ways, and need to be treated differently than works created by employees in a corporate environment.  For one thing, there is the issue of academic freedom.  Although the copyright law could well support the claim that all faculty works, even traditional scholarship like journal articles and books, are work for hire, the case would be much more complicated than the university system seems to imagine.  As I say, academic freedom would pose a unique obstacle, since courts have recognized a First Amendment interest in academic freedom.  So there is a constitutional argument could be used to counter a work for hire claim at a public university.  Also, there would be an argument over the traditional “teacher exception” that courts recognized for many years.  Although there have been no decisions definitively invoking the teaching exception since the 1976 Copyright Act took effect, it would still provide a line of defense against work for hire claims that the universities would struggle to overcome (this article by one of the scholars quoted in the article explains this ambiguous situation).

Then, of course, there are the practical problems.  Having a policy like this is sure to make it harder for the universities in the Louisiana system to recruit top faculty.  Even if individual campuses modify the policy into something wholly different than the system-wide template, which they would be wise to do, the question will always hang over recruitment, and the balance of decision-making may be tilted in some cases involving highly productive  and savvy professors.  On the university side, the mountain of paperwork they will create for themselves if they really undertake to review every contract for publication seems not to have occurred to the drafters.  But any small additional profit they could hope to make by claiming a portion of royalties is sure to be devoured by increased administrative costs.

As I read the article, I was struck by the sense that the university system was trying to create a single policy that would treated patented works and those subject to copyright in the same way.  Any superficial sense that this might seems to make is easily dispelled, however; there a good reasons to deal differently with these two varieties of intellectual property.  Copyright, for one thing, is easy and cheap to get.  It is, in fact, automatic whenever original expression in fixed in tangible form.  There is no need for the university employer to intervene to help the employee creator protect her rights.  With patents, the situation is wholly different.  Patents are difficult to get and it usually requires an investment of tens of thousands of dollars to successfully “prosecute” a patent.  And patents, when they generate a profit (which few actually do), offer much larger gains.  So creators need more help from the university to get a patent, and universities have a greater incentive to provide that assistance, than is the case with copyright.

Also, the “significant use of university resources” is really quite different with copyrighted materials than it is with patents.  Often the resources employed to create something subject to copyright are resource the university would supply in any case — a library, computers in faculty offices, art supplies for artists who both create and teach, a video camera in a dance studio, etc.  Resources used in equipping labs to pursue patentable inventions, on the other hand, are often extremely specialized and sometimes costs millions of dollars.

This does not mean that the copyrighted productions of a faculty are less important than those subject to patents; their very ubiquity testifies to how vital such works are to the tasks of teaching and research.  But the conditions of creation are so different that policies that conflate the two are seriously misguided.

One point I try to make whenever I discuss copyright ownership policies on university campuses is that such policies should distinguish between ownership of the underlying rights and uses that other parts of a scholarly community can be licensed to make.  For example, a university need not claim ownership over the design of an online course by a faculty member as long as it assures, by policy or by separate agreement, that it has a perpetual license to use that courseware.  Thus the faculty member’s academic freedom and reasonable expectation to own her own work is upheld, but the university does not have to worry about having to redesign the course if the faculty member leaves or to pay her twice for it.  The designer, of course, is also able to reuse her work at a new institution, so everyone’s needs can be met.  A careful policy on copyright ownership should be a kind of matrix that identifies types of works and groups of users who might have an interest in those various kinds of works.  Then the decision about ownership can be made at an appropriately granular level, and the use rights of those with recognized interests in each category can also be assured.

Copyright ownership policies are important, and becoming more so everyday.  Campuses that do not have such policies need to remedy that situation sooner rather than later, and before conflicts develop that will leave the decision to courts applying the default rules of the copyright law.  The policy proposed by the University of Louisiana System is a poor model and an apparently ill-considered stab at such a policy.  But if the controversy it has generated causes faculty in Louisiana and elsewhere to think hard about these issues, perhaps some good will result.

This statement strikes me as deliberately misleading.  Publishers are not afforded any period of copyright protection by the copyright law, anymore than plumbers or ophthalmologist are.  This kind of misinformation is intended to create the illusion that publishers’ business models are somehow favored by federal law and thus inviolate, but that is not true.  Only one group is afforded copyright protection and the term for which that protection lasts — authors (under section 201(a) of the copyright law, Title 17 of the U.S. code).  If publishers hold any rights, they hold those rights only because they are transferred to them by the authors whose works they publish.  And if those authors choose, they can transfer less than the full copyright, and for less than the full term of protection.

Increasingly the transfer of copyright to publishers in exchange for using their distribution networks seems like a very bad bargain indeed.  As the ongoing boycott of Elsevier dramatically indicates, scholarly authors are becoming much more vocal and open as they demand a better solution for distributing their works.  It has always been problematic to give away the rights under copyright for free to publishers who then sold the works at a high profit, in which authors did not share.  Now there are many other options available to authors, many of which publishers are anxiously trying to undermine.  It is very important to some publishers that authors do not come to understand the power they have based on the fact that they hold all of the rights under copyright and can leverage those rights to do what is best for them.

Statements like the one from Wiley Blackwell reflect, I think, an increasing sense of panic in the publishing community.  Disinformation is seen as one way to fight the growing realization that they may become as irrelevant in the Internet age as blacksmith and buggy whip makers became in the age of the automobile.

We see this sense of panic manifest in several ways.  When Oxford University Press tries to claim that essays written for edited volumes are “work made for hire,” they are grasping at a legal straw that cannot hold up for them.  Likewise when publishers like Elsevier and the American Chemical Society write publication contracts that try to make authors’ retention of rights, or not, dependent on the kinds of internal policies that exist on the authors’ university campuses.  Such contracts are more cries of anger and fear than legal agreements.  In all of these cases, the publishers are looking for a legal lever they can push that will stave off irrelevance.  But the law does not work that way in general, and copyright is written to benefit authors and give them control over their works, not to prop up a particular business model.

Companies that survive are those that adapt to technological change, not those that desperately try to use legal coercion to prevent the change.  The movie industry learned this when their attempt to prevent home video recording failed; they were forced to adapt, and they found new ways to flourish.

Instead of resisting public access to taxpayer-funded research and writing byzantine contract language intend to punish authors who seek to exploit their legal rights, publishers need to look at how they can provide services to authors that will be necessary and desirable in the digital environment for scholarship.  Last month I had lunch with an official of a major publisher who talked about this approach to his business and was full of creative ideas.  Unfortunately, he is still a minority voice.  But misrepresenting the state of the copyright law is not the future for the publishing industry; services for authors is the future.  It is time for publishers to stop grasping at straws, for authors to stop giving away all of their rights under copyright, and for both groups to work together to figure out what the future of scholarship is going to look like.

There is something particularly galling for a scholar about having her article described as “work for hire.”  It implies a lack of academic freedom and even a “hired pen” approach to scholarship.  Most universities, which actually might have a strong case if they claimed faculty works as “works made for hire,” long since decided that the obvious ill-will and hassle that would attend such claims made them unpalatable.  OUP, on the other hand, does not seem to have learned the same lesson or, in fact, to even understand the law correctly.

I have to say first that I do not know how widespread this practice is, even within OUP.  This is the first time I have heard of this situation.  It may apply only to essays written for inclusion in collected, thematic volumes.  Or it may just be a test foray into a really bad idea.

By way of introduction, it is important to note that there are two ways for a work to be a work for hire.  First, it can be a work created by a regular employee within the scope of his or her employment.  That definition could likely fit faculty writings, but it has almost never been used to contest faculty ownership.  Alternatively, a work by an independent contractor — someone who is not a regular employee — can be a work for hire if two conditions are met.  First, the work must fall into one of nine categories enumerated in the law.  And second, there must be an express, written and signed agreement between the employer and the contractor “that the work shall be considered a work made for hire.”

OUP obviously hopes to take advantage of the second path to work for hire, since the first one would not apply.  A “contribution to a collective work” is one of the permissible categories for independently contracted  works made for hire.  But I think OUP has a big problem meeting the second requirement.

It is important to note also that the effect of work for hire is the same in either situation –  the employer is designated the “author” from the moment the work is created.  The person who actual puts pen to paper, as it were, has no rights at all in the work.  That fact probably explains some about why OUP would make this foolish move, and it is also part of the reason why their attempt to turn faculty writings into work for hire is likely to fail.

As to why OUP would do this, I think there are a couple of legal benefits for authors that OUP hopes to avoid having their contributors enjoy.  One would be the termination right, which allows an author or other creator to terminate a transfer of copyright after thirty-five years, regardless of the terms of the original contract.  This right, while it may seem obscure, has recently gotten attention as the legal window through which composers and performers of profitable music from the late 1970′s has just opened.  The one way to prevent an author from terminating a transfer of rights is to own the work as a work for hire, so that no transfer was ever required.  I suspect some legal beagle at OUP saw the controversies over music and thought this might be a good idea.  It is not.

The other thing that having these contributions classified as work for hire would prevent would be prior licenses.  As more faculties adopt open access policies, which usually take the form of a prior license to the institution for repository deposit, the possibility arises of an eventual contest between a prior license contained in such a policy and a later transfer of the copyright through a publication contract.  OUP may be test-driving an idea for avoiding that situation — if the faculty author is classified as a non-author by the work for hire doctrine, they would be unable to grant any prior licenses, since they never held any rights.

So why do I think this move is stupid, and doomed to fail?  Three reasons.

First, nothing is more surely designed to make faculty authors angry than to tell them they are not the authors of the scholarship they offer to publishers.  As a group, faculty authors have been pretty docile toward publishers for a long time, but foolish and tone-deaf moves by publishers have begun to stir faculty anger toward presses they once considered friends and colleagues.  If a claim like this, which denies the fundamental dignity of authorship to scholars, becomes widespread, that slow rebellion will speed up very quickly.

Second, in the work for hire battle, presses are likely to lose.  As I said above, universities could, if they choose, assert a convincing case that faculty are regular employees whose writings are created within the scope of their employment.  Were OUP really to assert its WFH claim to defeat a prior license, the institution could claim that, as the regular employer of the scholar, it was the author and so the agreement with OUP would be void as outside the ability of the faculty member to sign.

Finally, and most importantly, there are two cases in the U.S. courts that have held that, in an independent contractor situation, an agreement designating the work as a work made for hire must be signed, or at least formed (meaning that both parties understood), prior to the creation of the work.  There is an excellent discussion of those cases on the website of copyright attorney Ivan Hoffman.  By making the work for hire provision part of a submission agreement, OUP would be unable to show that such an agreement would even have been contemplated by the author, much less agreed to.  So this is a move which can only have negative consequences for OUP.  The cost in bad feeling is very high, and it cannot, I don’t think, succeed as a legal maneuver, even if OUP thinks it is worth that high cost.