Abuse of Trust?

U. of Michigan Library admits to flaws in process for identifying copyright "orphans" in HathiTrust Digital Library, in setback for project aimed at increasing access to digitized texts.
September 19, 2011

Less than a week after the University of Michigan brushed off a lawsuit by the Authors Guild over the university’s move to make copyrighted “orphan” works in its digital collection freely available to students and faculty, the Michigan Library suspended the practice Friday, admitting “serious” flaws in its process for identifying orphans.

Friday’s mea culpa followed a public flogging of the library and its nonprofit digital consortium, HathiTrust, at the hands of the Authors Guild, in which the guild quickly tracked down the owners of the copyrights on several works that HathiTrust had categorized as “orphans” -- books and articles that are in copyright but whose copyright owners cannot be located or identified.

“The close and welcome scrutiny of the list of potential orphan works has revealed a number of errors, some of them serious,” the Michigan library wrote in its statement. “This tells us that our pilot process is flawed.”

The librarians said they had “learned from [their] mistakes” and have “already begun an examination of our procedures to identify the gaps that allowed volumes that are evidently not orphan works to be added to the list.” The HathiTrust’s Orphan Works Project -- a Michigan-led effort to identify and increase access to the orphans from the consortium’s digital library -- has been suspended until the university can come up with “a more robust, transparent, and fully documented process” for making sure works are genuinely orphaned before categorizing them as such.

The Authors Guild, along with authors' associations in Australia and Quebec and a handful of individual authors, had filed suit last Monday against the HathiTrust, Michigan, and several other university libraries heavily involved in the Orphan Works Project. The plaintiffs claimed that by establishing its own set of procedures for clearing orphan works for wider accessibility, the libraries were taking copyright into their own hands. They argued that the orphans should stay under lock and key until Congress passes legislation governing how orphan works can be identified and displayed.

Michigan and other HathiTrust supporters argued that giving faculty members and students access to digital orphan works was protected by the "fair use" provisions of U.S. copyright law.

But the Authors Guild struck back on its blog, calling into question the integrity of Michigan’s process for attempting to find the copyright holders for its orphan candidates. In a series of "gotcha" blog posts, the guild documented its own efforts to find the copyright holders for HathiTrust orphans. It quickly tracked down several authors that HathiTrust had apparently been unable to reach.

One such author was J.R. Salamanca. Salamanca, who is an emeritus professor at the University of Maryland, wrote The Lost Country, one of the works listed as a "candidate" for HathiTrust's digital orphanage. According to the Authors Guild, Salamanca said he had not heard of HathiTrust and “wasn’t happy” to learn his book was marked to be made freely available online for Michigan’s students and faculty next month.

“If HathiTrust’s researchers can’t locate a bestselling author with a literary agent, an author who’s also a retired professor from a major East Coast university, how are they going to locate authors in other countries?” the guild wrote on its blog.

The guild later published the full list of candidates for HathiTrust’s digital orphanage, inviting readers to check the group’s work.

Prior to Michigan’s mea culpa, James Grimmelmann, an associate professor at New York Law School, lamented that the HathiTrust’s sloppy efforts to properly identify orphan works had cast a pall over good-faith efforts by libraries to properly identify orphans. “No one will ever be able to make the orphan works argument again without opponents bringing up the HathiTrust orphans that weren’t,” Grimmelmann wrote in a blog post entitled “HathiTrust Single-Handedly Sinks Orphan Works Reform.”

After Michigan admitted that its process was flawed, Grimmelmann told Inside Higher Ed the episode may have given the Authors Guild the political capital it needed to fight the larger battle over whether HathiTrust should be allowed to hold scanned copies of any copyrighted books at all.

The guild has said its main priority is to wrest back control over the millions of less ambiguously in-copyright works that make up the bulk of the HathiTrust Digital Library. The Orphan Works project was mainly a way to give the lawsuit “immediacy,” the guild has said. That has prompted some, including Brandon Butler, of the Association of Research Libraries, to argue that the Authors Guild does not have standing to try to sink the Orphan Works Project, since that project probably does not compromise the interests of any of its members.

However, “by finding authors whose works seem plausibly likely to have been declared orphans by mistake, I think the Authors Guild has done a fair amount to cure one of the procedural issues with its complaint: its ability, as a relatively small membership organization, to claim to speak on behalf of all authors,” Grimmelmann said on Sunday.

“It could make it easier for them to get their legal arguments ultimately considered on their merits by a judge,” he said, “instead of getting held up on procedural issues.”

But some critics say that the Authors Guild’s attempts to shame HathiTrust and its partners over its unreliable procedures by reporting on the surprise and outrage of authors such as J.R. Salamanca cast fog over the fact that the HathiTrust and its university partners are not trying to usurp authors, but instead to give their work new life.

“I checked with one of our members that has The Lost Country (the first work that the Guild ‘reunited’ with its owner) in its collection, and they were able to tell me that the book had not been checked out since 1993,” Butler, the ARL vice president of public policy, wrote in an e-mail. “The goals of the project are to make these kinds of under-used, out-of-print books easier to find by making them searchable digitally, and then making them easier to read by letting authenticated students and faculty access them online and download one page at a time as PDFs.”

Kevin Smith, a scholarly communications officer at Duke University, reinforced the point on his blog in an “Open Letter to J.R. Salamanca.”

“The sad fact is that The Lost Country has become a pretty obscure work,” Smith wrote, noting that the Duke library’s copy had not been checked out for decades. “…The current system, I am afraid, is failing to connect your book to new readers,” he wrote.

“It has to be said that the Authors Guild is not going to help you in this regard,” Smith continues. “Where you can find help for this problem is with the HathiTrust. Their goal, and the goal of the libraries that plan to participate in the orphan works project, is to make it easier for readers to find works like your novel, which might otherwise languish on shelves or in large warehouses of books.”

The Authors Guild did not respond to requests for comment. But it remarked in its legal filings that it thinks Congress, not individual libraries, should decide on a protocol for identifying and setting access rules on orphan works. Paul Courant, the dean of libraries at Michigan, told Inside Higher Ed he does not think Washington will take up the issue any time soon.

Grimmelmann, the New York Law School associate professor, says HathiTrust could have provided a framework for identifying orphans that federal regulators might have blessed, had they not been so sloppy.

“I think there was wasted opportunity here for the libraries to produce a process that could be demonstrated as trustworthy and provide a good model for Congress to draw on,” Grimmelmann said.

For the latest technology news from Inside Higher Ed, follow @IHEtech on Twitter.

Search for Jobs


  • Viewed
  • Commented
  • Past:
  • Day
  • Week
  • Month
  • Year
Back to Top