[NOTE: We are marking Fair Use Week with this post from Brandon Butler, Director of Information Policy at the University of Virginia Library. Earlier this week we published a new policy on "non-consumptive use" for services at the HathiTrust Research Center. Here Brandon explains the legal grounding of the policy and offers his take on the policy's significance.]
It is fitting to end fair use week with some words about a new policy that, although it may seem dry and technical, may in fact represent the most compelling invocation of fair use in the last decade or more.
Earlier this week the HathiTrust Research Center (HTRC) published a new policy on non-consumptive use, which outlines the allowable types of data access and exports for researchers conducting computational research on the HathiTrust collection. Based on the latest case law, it is designed to foster groundbreaking research that enriches the public without intruding on the market prerogatives of copyright holders. With this policy, HathiTrust is operationalizing fair use to empower researchers to mine its rich corpus of millions of in-copyright texts for new insights, understandings, and information of a kind that literally could not have existed only a few years ago.
The key legal ingredient in this effort is the concept of “non-consumptive use.” A clear articulation of this concept ensures the HTRC is only enabling genuinely new and productive forms of research grounded in computer analysis of the texts: core protected fair use. It also rules as clearly out-of-bounds the ordinary reading that will typically require permission or another kind of fair use rationale.
The policy is grounded in a deep, well-established principle in fair use law: the signature characteristic of a fair use is that it does not supersede the original work in its ordinary and likely markets. The emphasis on non-substitution traces all the way back to the seminal case of Folsom v. Marsh, in which Judge Story first articulated the modern four factor test for fair use. Story explained:
"[A] reviewer may fairly cite largely from the original work, if his design be really and truly to use the passages for the purposes of fair and reasonable criticism. On the other hand, it is as clear, that if he thus cites the most important parts of the work, with a view, not to criticize, but to supersede the use of the original work, and substitute the review for it, such a use will be deemed in law a piracy."
As Judge Leval explained at length in his own seminal opinion in the Google Books case, non-consumptive uses are non-superseding par excellence. They serve the purpose of copyright by separating facts from expression, and allowing the former to circulate freely. (I described this aspect of the opinion at some length right after it came out.) Access to these facts is an enormous boon to society, and at the same time poses no risk at all to copyright holders, whose interest is in access to the expressive content of their works.
The non-consumptive use policy is also a natural development of a modern line of fair use cases that blessed internet search engines. Starting with Kelly v. ArribaSoft, this line of cases found that copying entire webpages and images posted online and showing snippets and thumbnails from these pages as search results was fair use because none of these activities provided the public with a substitute for ordinary access to the underlying material. To the contrary, search engines create a vast new form of knowledge: facts about the content of the web, which can be queried in order to help users find what they're looking for online. As this line of cases grew, the question naturally arose: is this logic limited to internet search engines, or is there a broader category of fair use here?
Perhaps the most exciting thing about the new policy is that it represents the generalization of the principle favoring web search so that it favors all computer analysis of any text for an open-ended variety of research purposes. Judge Leval's opinion, together with the opinion in the HathiTrust case, laid out the broad principles favoring this generalization, but this policy may well be the first attempt to really move beyond search to cover the full range of research projects that might benefit from non-consumptive methods.
So, take a look at the policy. I am proud to have been among the many talented people who worked on it, and one of our goals was to make it as readable as possible while staying faithful to the complex concepts at the policy's core. I hope that other institutions interested in fostering this kind of research will find our policy a useful reference.