Promoting research and access to its products has always been a core purpose of copyright law, often expressed in limitations and exceptions for research uses. Recent legal scholarship has examined the need for copyright exceptions for text and data mining (TDM) methodologies, and the doctrines recently enacted to achieve this purpose. Empirical scholarship has highlighted the positive impact on scholarship of copyright exceptions for TDM and of more “open” exceptions for research uses. Until now, however, there has not been a collection and categorization of the world’s copyright laws according to the degree to which they provide exceptions for research. The Program on Information Justice and Intellectual Property’s recent report, Research Exceptions in Comparative Copyright Law, fills this gap.
Our study uses TDM research as a use case and asks whether each copyright law contained in the WIPO Lex database contains an exception that allows reproductions and sharing between researchers of whole works for an academic research project. Such exceptions normally require any such use to also pass a proportionality test such as the use be consistent with fair practice. We included analysis of exceptions that specifically mention “research” as a permitted purpose, those for libraries and for private uses even if they did not specifically reference “research,” and recent exceptions for TDM, computational use, and information analysis passed in a handful of countries.
For each exception, we scored the words of the statute on the degree to which the research exception is sufficiently open to permit the reproduction of whole works and the sharing of those works with other researchers in a TDM project. These uses were identified as critical to collaborative and cross border academic TDM research projects in interviews with TDM researchers. Thus, for each law, we analyzed the following dimensions:
- Works: Does the exception apply to all works, including whole works and including audio visual works?
- User: Does the exception apply to a research use by any user, including both individuals and institutions?
- Uses: Does the exception apply to all needed uses for research, including both the reproduction activities and for communications or “making available” to the extent needed to engage in limited sharing of works and corpuses with other researchers for purposes such as collaboration or verification?
Our work finds that every copyright law examined has at least one exception for research uses. But how open those exceptions are to TDM uses can be mapped along a spectrum from the most open to the most restrictive.
Fewer than a quarter of the countries in our study have research exceptions that are open to reproduction and sharing of any type of work by any user. These countries are labeled green in Figure 1. Open exceptions for research are ubiquitous in countries with copyright laws most influenced by British law. Such countries normally provide a general exception for any “fair” dealing or use for a number of purposes, including research. Open exceptions for research are not limited to common law countries, however. For example, Estonia provides a research exception that permits any research use of any work by any user, subject to the fairness criteria that the use be “to the extent justified by the purpose and on the condition that such use is not carried out for commercial purposes.” Thailand’s general exception using an enabling framing of the Berne Convention’s three-step test (art. 9(2)) similarly applies to all uses of all works by all users and specifically mentions research as a protected purpose.
Some, but not all, recently enacted exceptions specifically for TDM research are sufficiently open to be labeled green in our map. Some TDM exceptions, such as that of Switzerland and in arts 3 and 4 of the EU CDSM Directive, authorize only reproductions for TDM, with no express authorization to share works and corpuses with other researchers for purposes such as collaboration or validation.
At the opposite end of the spectrum, labeled red in Figure 1 below, are a small number of countries whose research exceptions extend only to the use of quotations or excerpts for a research purpose. Argentina, for example, limits research uses to “up to 1,000 words for literary or scientific works, or eight bars in musical works and, in all cases, only the parts of the text essential for that purpose.” Such exceptions for uses of mere excerpts do not authorize TDM research, which require uses of whole works.
Between the poles of open and highly restrictive research exceptions lies the majority of countries in the world – marked yellow in our map. In these countries, some kinds of TDM research projects may fall within the exceptions. But restrictions on some uses, by some users, of some kinds of works create complications for many projects.
Limitations and implications
Our research has limitations. Many of the laws in the WIPO Lex database are outdated. We did not review case law or interpretations of the law in each jurisdiction, and thus our examination is only of the law on the books, not as it applies in practice. And we do not analyze the entire copyright system in each country, including whether the scope of copyright protection would apply to many TDM uses which may be considered acts of reading and analysis not protected by copyright at all. Accordingly, one cannot conclude from our study whether a particular project is or is not legal. It does not give legal advice.
We hope the work will nonetheless be useful. Countries considering TDM exceptions may use our report to find models for the kind of exceptions they want to adopt. International policymakers can use the information in the exploration of new rules — such as on the cross border sharing of lawfully created TDM corpuses — that could help all countries produce and benefit from TDM projects. And empirical researchers may use our data as an independent variable to help assess the impacts of legal change.
 The gap in our analysis goes both ways. For example, we label Brazil a red country based on a plain text reading of its statute. We are aware, however, that Brazilian courts have recognized human rights-based exceptions to copyright that can operate like an open, general exception like fair use, See. S.T.J. Recurso Especial No. 964.404 ES (2007/0144450-5). Relator: Ministro Paulo de Tarso Sanseverino, 15.03.2011, Diario da Justica Eletronico [D.J.e.], 23.05.2011 (Braz.).