Compliance of National TDM Rules with International Copyright Law: An Overrated Nonissue?

Seeking to devise an adequate regulatory framework for text and data mining (TDM), countries around the globe have adopted different approaches. While considerable room for TDM can follow from the application of fair use provisions (US) and broad statutory exemptions (Japan), countries in the EU rely on a more restrictive regulation that is based on specific copyright exceptions. Surveying this spectrum of existing approaches, lawmakers in countries seeking to devise an appropriate TDM regime may wonder whether the adoption of a restrictive approach is necessary in the light of international copyright law. In particular, they may feel obliged to ensure compliance with the three-step test laid down in Art. 9(2) of the Berne Convention, Art. 13 of the TRIPS Agreement and Art. 10 of the WIPO Copyright Treaty. Against this background, the analysis raises the question whether international copyright law covers TDM activities at all. TDM does not concern a traditional category of use that could have been contemplated at the diplomatic conferences leading to the current texts of the Berne Convention, the TRIPS Agreement and the WIPO Copyright Treaty. It is an automated, analytical type of use that does not affect the expressive core of literary and artistic works. Arguably, TDM constitutes a new category of copying that falls outside the scope of international copyright harmonization altogether.


Introduction
Seeking to devise an adequate regulatory framework for text and data mining (TDM), countries around the globe have adopted different approaches. While considerable room for TDM can follow from the application of fair use provisions (US) 1 and broad statutory exemptions (Japan), 2 countries in the EU rely on a more restrictive regulation that is based on specific copyright exceptions. 3 The EU approach leads to considerable complexity and legal uncertainty. The invocation of available exceptions depends on individual criteria, such as use for scientific research purposes and the absence of a rights reservation by copyright owners. 4 Surveying this spectrum of existing approaches, lawmakers in countries seeking to devise an appropriate TDM regime may wonder whether the adoption of a restrictive approach is necessary in the light of international copyright law. In particular, they may be exposed to allegations of non-compliance with the three-step test laid down in Art. 9(2) of the Berne Convention for the Protection of Literary and Artistic Works (''Berne Convention'' or ''BC''), Art. 13 of the Agreement on Trade-Related Aspects of Intellectual Property Rights (''TRIPS Agreement'' or ''TRIPS'') and Art. 10 of the WIPO Copyright Treaty (''WCT''). 5 An imbalanced, restrictive interpretation of the three-step test, however, can lead to an overly cautious approach that stifles TDM and makes broad inroads into the fundamental right to research that supports the freedom of researchers to apply modern technologies for research purposes. 6 Instead of providing breathing space for TDM and establishing a proper balance between copyright protection and the right to research, 7 national legislation may place unnecessary constraints on the use of TDM in research contexts.  Senftleben and Margoni et al. 2022, pp. 72-73. 4 For an overview of complicating factors, see Senftleben (2022), pp. 36-47. For a critique of the approach chosen in the EU, see Griffiths et al. (2022), pp. 12-14; Ducato and Strowel (2021), pp. 322-329; Margoni and Kretschmer (2021), p. 26;Geiger (2021), pp. 383-394; Geiger and Schönherr (2021), §11. 109;Geiger et al. (2018), pp. 837-838;Hilty and Richter 2017, pp. 7-8;Handke et al. (2015), pp. 120-130. 5 For a more detailed discussion of the different criteria of the three-step test and the transformation of the three-step test into a general yardstick for assessing national limitations and exceptions, see Senftleben (2004). For a discussion of compliance questions leading to legal uncertainty in national policymaking and a ''ceilings'' approach as a potential countermeasure, see Kur et al. (2016), p. 133;Kur and Große Ruse-Khan (2011), pp. 359-407;Große Ruse-Khan (2009), pp. 56-102. 6 For a more detailed discussion of the right to research in the EU Charter of Fundamental Rights, see Van Daalen (2022), pp. 231-245; Geiger and Jütte (2022a, b). 7 The CJEU has recognized the need to strike a proper balance between copyright protection and competing fundamental rights, in particular freedom of expression and information, in several cases. See CJEU, 29 July 2019, case C-516/17, Spiegel Online, para. 54; CJEU, 29 July 2019, case C-469/17, Funke Medien NRW, para. 70; CJEU, 1 December 2011, case C-145/10, Painer, para. 134; CJEU, 3 September Against this background, the following analysis raises the question whether international copyright law, including the three-step test (Sect. 2), covers TDM activities at all. TDM does not concern a traditional category of use that could have been contemplated at the diplomatic conferences leading to the current texts of the Berne Convention, the TRIPS Agreement and the WCT. It is an automated, analytical type of use that does not affect the expressive core of literary and artistic works. 8 Arguably, TDM falls outside the scope of international copyright harmonization altogether. From this perspective, the discussion about compliance with international copyright norms is a nonissue (Sect. 3). International copyright law does not limit the freedom of national policymakers to devise appropriate domestic solutions to reconcile copyright protection with the right to research that underlies TDM activities (Sect. 4).

Three-Step Test Controversy
As already indicated, the three-step test in international copyright law -known from Art. 9(2) BC, Art. 13 TRIPS and Art. 10 WCT -is often invoked to cast doubt upon the compliance of national copyright exceptions and limitations with international law. 9 With its open-ended criteria, the three-step test adds considerable complexity to the assessment of use in the context of TDM research. At the international level, the decision of a WTO Panel in ''United States -Section 110(5) of the US Copyright Act'' yielded a detailed discussion and interpretation of the test criteria. 10 The WTO Panel discussed the three-step test laid down in Art. 13 TRIPS which stipulates that WTO Members: shall confine limitations or exceptions to exclusive rights to certain special cases which do not conflict with a normal exploitation of the work and do not unreasonably prejudice the legitimate interests of the right holder. 11 Evidently, the text of the provision offers several starting points for questioning the permissibility of a copyright limitation or exception that has been introduced in national law. Does a given form of TDM research constitute a ''special case'' in the sense of Art. 13 TRIPS? What is a ''normal'' exploitation of literary and artistic works? When does use for TDM purposes enter into ''conflict'' with a normal exploitation? Which interests of copyright holders must be deemed ''legitimate''? Footnote 7 continued 2014, case C-201/13, Deckmyn, para. 27;CJEU, 11 September 2014, case C-117/13, TU Darmstadt, para. 43. Cf. Geiger andIzyumenko (2020), pp. 292-298;Geiger (2006), p. 371. 8 Carroll (2019), p. 954; Ducato and Strowel (2021), p. 334;Ueno (2021), pp. 150-151. 9 For a more detailed analysis of this development, see Senftleben (2004), pp. 118-124. 10 World Trade Organization, 15 June 2000, United States -Section 110(5) of the US Copyright Act, Report of the Panel, WTO Document WT/DS160/R. For a discussion of this WTO Dispute Settlement decision, see Geiger, Gervais and Senftleben (2014), pp. 593-597;Senftleben (2006), p. 407;Senftleben (2004), pp. 134-230;Ficsor (2002), p. 111;Oliver (2002), p. 119;Brennan (2002), p. 213;Ginsburg (2011), p. 13. 11 Article 13 TRIPS. Explanation in brackets added by the author.
When does a prejudice to these legitimate interests reach an impermissible, ''unreasonable'' level? In the absence of sufficient legal guidance on the right interpretation of these elastic criteria, the three-step test can easily become a source of legal uncertainty that can impede, or even frustrate, the development of sufficiently flexible limitations and exceptions for TDM at the national level. Against this background, it is important to examine the guidance for the interpretation of the test criteria that can be derived from the aforementioned WTO Panel decision. 12 With regard to the first test (''certain special cases''), the WTO Panel distinguished between the word ''certain'' and the word ''special''. It understood the term ''certain'' to mean that a copyright limitation had to be clearly defined, while there was no need ''to identify explicitly each and every possible situation to which the exception could apply, provided that the scope of the exception was known and particularised.'' 13 From the term ''special,'' the Panel inferred the additional requirement that a limitation should be narrow in a quantitative as well as a qualitative sense. 14 It summarized this twofold requirement as narrowness in ''scope and reach.'' 15 The application to Sec. 110(5) of the US Copyright Act shows that, pursuant to the Panel's conception, it is, for example, the number of potential beneficiaries that must be sufficiently limited in order to comply with the quantitative aspect of speciality. 16 As to the qualitative aspect, the Panel eschewed an inquiry into the legitimacy of the public policy purpose underlying the adoption of a limitation. 17 In the Panel's view, the qualitative aspect of speciality does not have a normative connotation. A limitation need not serve a special purpose to be qualified as a special case in the sense of Art. 13 TRIPS. 18 Instead, the Panel raised conceptual issues, such as the categories of works affected by a copyright limitation and the circumstances under which it may be invoked. In this vein, the Panel lent weight to the fact that one of the use privileges flowing from Sec. 110(5) of the US Copyright Act was found limited to dramatic renditions of operas and homestyle receiving apparatuses. 19 Turning to the second test (''conflict with a normal exploitation''), the WTO Panel interpreted the term ''exploitation'' as a reference to ''the activity by which copyright owners employ the exclusive rights conferred on them to extract 12 For additional guidelines based on a broader analysis of relevant policy issues, see Geiger et al. (2014), pp. 581-626;Geiger et al. (2008), p. 707. 13 Report of the WTO Panel, id., para. 6.108.
14 Report of the WTO Panel, id., para. 6. 109. Cf. Lucas (2001), p. 430, who insists on the combination of both aspects of speciality to ensure a sufficiently rigid standard of control. 15 Report of the WTO Panel,id.,para. 6.112. 16 Report of the WTO Panel, id., paras. 6.127 and 6.143. 17 Report of the WTO Panel, id., para. 6. 111. 18 Report of the WTO Panel, id., para. 6.112. For literature statements supporting this view, see Ginsburg 2001, p. 13;Ricketson (2002), p. 31. For comments insisting on the importance of a policy analysis already at this first step of the test procedure, see Ficsor (2002), pp. 129-133, 223-229;Senftleben (2004), pp. 138-152, who propose to require that limitations be justified by some clear reason of public policy. 19 Report of the WTO Panel, id., paras. 6.145, 6.146 and 6.159. economic value from their rights to [musical] works.'' 20 In this context, the Panel distinguished between an empirical and a normative meaning of the word ''normal.'' While the first connotation of the term ''normal'' appeared to be ''of an empirical nature, i.e. what is regular, usual, typical or ordinary'', the second connotation reflected ''a somewhat more normative, if not dynamic, approach, i.e., conforming to a type or standard.'' 21 With regard to the empirical aspect, the WTO Panel accepted the US approach asking ''whether there are areas of the market in which the copyright owner would ordinarily expect to exploit the work, but which are not available for exploitation because of [the exemption at issue].'' 22 Accordingly, uses from which an owner would not ordinarily expect to receive compensation were not regarded as parts of a normal exploitation. Seeking to give meaning also to the normative aspect of the word ''normal'', the Panel had recourse to the work of a study group preparing the 1967 Stockholm Conference for the revision of the Berne Convention. 23 In particular, it attached importance to the conclusion that ''all forms of exploiting a work, which have, or are likely to acquire, considerable economic or practical importance, must be reserved to the authors.'' 24 The Panel inferred from this formula that it was appropriate to consider in addition to those forms of exploitation that currently generate significant or tangible revenue, those forms of exploitation which, with a certain degree of likelihood and plausibility, could acquire considerable economic or practical importance. 25 On its merits, the normative aspect, therefore, served as a vehicle to widen the perspective. It allowed the Panel to factor into the equation both currently existing and potential future markets when determining a conflict with ''a normal exploitation of the work.'' Finally, the WTO Panel's analysis of the test of ''no unreasonable prejudice to legitimate interests'' remained limited to the interest in the economic value of the exclusive rights conferred by copyright. In the absence of any objections raised by the parties, the Panel could readily qualify this interest as legitimate. 26 With regard to the expression ''not unreasonably prejudice,'' the Panel noted that any copyright limitation, by definition, caused some detriment to right holders because it reduced the scope of exclusive rights. This led to the insight that in order not to erode copyright limitations altogether, ''a certain amount of prejudice has to be presumed justified as 'not unreasonable'.'' 27 The Panel concluded that ''prejudice to the 20 Report of the WTO Panel, id., para. 6.165. 21 Report of the WTO Panel, id., para. 6.166. 22 Report of the WTO Panel, id., paras. 6.177-6.178. 23 For a description of the work of the study group, see Ficsor (2002), pp. 115-121. 24 Report of the WTO Panel, id., para. 6.179. The study group was composed of representatives of the Swedish government and the United International Bureaux for the Protection of Intellectual Property (BIRPI). 25 Report of the WTO Panel, id., para. 6.180. 26 Report of the WTO Panel, id., para. 6.226. 27 Report of the WTO Panel, id., para. 6.229. legitimate interests of right holders reaches an unreasonable level if an exception or limitation causes or has the potential to cause an unreasonable loss of income to the copyright owner. '' 28 As this overview of assessment factors shows, the open-ended criteria flowing from the three-step test offer various starting points for right holders to challenge the compliance of a given national TDM exemption with Art. 13 TRIPS. For instance, it may be argued that TDM research -requiring large-scale use of copyrighted data resources and affecting various categories of literary and artistic creations 29cannot be qualified as a special case in the sense of the first test (''certain special cases'') because the use is not sufficiently confined from the quantitative perspective which the WTO Panel introduced.
At the same time, the openness of the test criteria offers room for counterarguments. While, in TDM cases, the number of works may be high and the spectrum of affected work categories may be broad, the circle of beneficiaries -researchers using TDM tools -is narrowly drawn and highly specific. From a normative perspective, it can be added that use for scientific TDM research, by definition, constitutes a special case in the sense of the three-step test because breathing space for use in this category is indispensable to arrive at a proper balance between copyright protection and the freedom of expression and information, and freedom of sciences, which researchers can invoke. 30 Another recurring theme in the three-step test debate is the argument that exemptions of research use cause a conflict with the normal exploitation of works that are specifically made for research use, such as academic publications. Again, counterarguments are conceivable. The use of academic publications for TDM research does not aim at providing alternative access to scientific publications and undermining the market for the original information products. Instead, protected publications are only used in the framework of an automated, computational analysis. 31 If TDM datasets containing protected works are shared within a group of researchers belonging to a research consortium, this is merely done to jointly carry out the research project. To enable replicability studies confirming the validity of TDM results, selected members of the broader academic community may obtain access -again without eroding the primary market for publications used as source materials. Insofar as final research publications contain traces of protected works used for the TDM analysis, these traces are also unlikely to substitute the original publications and erode the primary market. Because of the specific research context, 28 Report of the WTO Panel, id., para. 6.229. 29 Cf. Senftleben et al. (2022), pp. 70-71. 30 As to the recognition of the right to research as a fundamental right in the EU, see TDM researchers will thus be able to dispel concerns about an erosion of the normal exploitation.
In the case of academic publications, it must also be considered that publishers of academic books and journals benefit greatly from the work of researchers in scientific projects, including TDM projects. 32 Without paying for the research work, they can reap the benefits of research results by publishing books and articles that share the new knowledge -once again without paying for the time invested in the drafting and reviewing of publications. Against this background, the use of existing academic publications as data sources for TDM projects -leading to new knowledge for new publications -appears as a good investment in the continuous evolution of publishable material. The cyclic process in which new knowledge arises from the analysis of pre-existing academic sources 33 begs the question whether academic publishers have any legitimate interests at all in the sense of the three-step test (see the final step prohibiting use that ''unreasonably prejudice[s] the legitimate interests of the right holder'') 34 that could be asserted against national TDM exemptions in the discussion about compliance with the three-step test. 35 Following this line of reasoning, arguments based on an erosion of the market for academic publications appear unfounded from the outset.
These examples of potential lines of reasoning show that the three-step test is an effervescent source of both compliance and non-compliance arguments. While strong arguments against allegations of incompatibility are available, the criteria of the test can be used strategically to cast doubt upon exemptions of TDM use in domestic copyright law and destabilize national legislative efforts to create breathing space for TDM. Article 13 TRIPS creates legal uncertainty that can have a corrosive effect on national legislative initiatives seeking to offer room for TDM research. Even if non-compliance arguments can finally be rebutted, the three-step test gives right holders munition to attack TDM exemptions and the legal position of researchers.

Scope of the International Right of Reproduction
Considering the problems that can arise from an obligation to ensure compliance with the three-step test, it is important to clarify whether copies made in the context of TDM research have copyright relevance at all. If TDM copying does not amount to a relevant act of reproduction in the first instance, there is no need to introduce a 32 For a more detailed analysis of this point, see Peukert (2012), pp. 146-148. 33 For a discussion of the role of copyright law in enabling cyclic innovation in the literary and artistic sector, see Senftleben (2020), pp. 26-35. 34 Article 13 TRIPS. 35 As to the validity of this argument, see WTO Panel, 17 March 2000, WTO Document WT/DS114/R, Canada -Patent Protection of Pharmaceutical Products, para. 7.69, stating that both society and the scientist have a legitimate interest in using the patent disclosure to support the advance of science and technology. Hence, it has already been recognized in WTO jurisprudence that exclusive intellectual property rights must not thwart the achievement of overarching societal goals in the field of research use. Cf. Senftleben (2011), pp. 162-163. limitation or exception to justify the use, and there is no need to defend this limitation or exception against allegations of incompliance with the three-step test. At first glance, however, international copyright law seems to leave little room for assuming that copies made in the context of TDM fall outside the scope of the right of reproduction that has been recognized in Art. 9(1) BC. The provision is often understood to grant a broad, flexible right of reproduction -an exclusive right capable of absorbing whatever new form of copying arises as a result of technological developments. 36 Article 9(1) BC reads as follows: ''Authors of literary and artistic works protected by this Convention shall have the exclusive right of authorizing the reproduction of these works, in any manner or form.'' A closer look at the international discussion surrounding Art. 9(1) BC, however, reveals that in the light of new settings for the making of copies, Berne Union Members did not agree on a broad, ''general'' scope. In fact, the right of reproduction has been understood as a much less comprehensive exclusive right than the words ''in any manner or form'' may suggest. The deliberations at the 1996 WIPO Diplomatic Conference on Certain Copyright and Neighboring Rights Questions, which led to the adoption of the WIPO Copyright Treaty and the WIPO Performances and Phonograms Treaty (''WPPT''), offer important insights in this respect (Sect. 3.1). These insights can be put to good use in the context of the current debate on TDM (Sect. 3.2).

Lessons from the 1996 WIPO Diplomatic Conference
Article 7(1) of the Basic Proposal for the international treaty that finally became the WCT contained a provision that sought to confirm the comprehensive nature of Art. 9(1) BC and fortify the position of copyright holders in respect of forms of copying that were deemed new at the time of the 1996 Diplomatic Conference. 37 These new forms included in particular temporary acts of copying that are inherent in technical processes, such as streaming, caching and browsing: The exclusive right accorded to authors of literary and artistic works in Article 9(1) of the Berne Convention of authorizing the reproduction of their works shall include direct and indirect reproduction of their works, whether permanent or temporary, in any manner or form. 38 Articles 7(1) and 14(1) of the Basic Proposal for the later WPPT set forth corresponding provisions for performers and phonogram producers. The notes accompanying Art. 7(1) of the WCT Basic Proposal pointed out that: [t]he scope of the right of reproduction is already broad. The expression ''in any manner or form'' could not be more expansive in scope. It clearly includes the storage of a work in any electronic medium; it likewise includes such acts as uploading and downloading a work to or from the memory of a computer. Digitization, i.e. the transfer of a work embodied in an analog medium to a digital one constitutes always an act of reproduction. 39 The notes also reflected the intention to clarify the ''widely held'' 40 understanding that both permanent and temporary reproduction constituted reproduction within the meaning of Art. 9(1) BC. 41 The confirmation that temporary copying falls within the scope of the right of reproduction was seen as the ''only way'' 42 to harmonize effectively the interpretation of the scope of the right. 43 More specifically, the notes accompanying Art. 7(1) of the WCT Basic Proposal pointed out that, on the one hand, the result of reproduction may be ''a tangible, permanent copy like a book, a recording or a CD-ROM.'' 44 On the other hand, it may as well be ''a copy of the work on the hard disk of a PC, or in the working memory of a computer.'' 45 A work stored for ''a very short time may be reproduced or communicated further, or it may be made perceptible by an appropriate device.'' 46 As a counterbalance to the confirmation and recognition of this broad right of reproduction, Art. 7(2) of the WCT Basic Proposal allowed for the introduction of limitations within the confines of the international three-step test. 47 The notes explained in this respect that it would be a matter for the legislation of Contracting Parties to limit the right of reproduction in the case of temporary reproduction of a work, in whole or in part, in certain specific cases, namely where the purpose of the temporary reproduction is solely to make the work perceptible or where the reproduction is of a transient or incidental nature. Moreover, the temporary reproduction must always take place in the course of use of the work that is authorized by the author or permitted by law. 48 39 WIPO (1999), Vol. 1, Document CRNR/DC/4, 188, para. 7.01. 40 WIPO (1999), Vol. 1, 188, para. 7.05. As to the preparatory work in expert committees, see Reinbothe and von , paras. 7.1.31-7.1.36. 41 Cf. Ricketson and Ginsburg (2006), para. 11.72, who point out that ''implicit in the call for effective harmonization may be a recognition that the scope of the application of the reproduction right to digital media was not completely 'obvious'.'' As to the critique, see Litman (1994), pp. 39-42;Sheinblatt (1998), p. 541. 42 WIPO (1999 Cf. Samuelson (1997) WIPO (1999), Vol. 1, 188-190, para. 7.07. As to the policy rationale underlying the last element of this provision, requiring the authorization of copyright holders, see Samuelson (1997), p. 385: ''However, it was sufficiently narrowly drawn that it would not, for example, have relieved telephone companies or online service providers from potential liability for temporary copies of infringing material made in company equipment as the material passed through their systems en route from sender to recipient. Such copies would, of course, meet the transient or incidental standard of draft article 7(2), but they would not have 'take [n] place in the course of use of the work that [wa]s authorized by the author or permitted by law.''' As to concerns of the telecommunication industry and internet services providers, see Goldstein The negotiations at the 1996 WIPO Diplomatic Conference, however, showed that the position described in the notes accompanying Art. 7(1) of the WCT Basic Proposal was less widely shared than expected. 49 As a guideline for the deliberations in Main Committee I -the Committee dealing with the substantive provisions of the proposed WIPO ''Internet'' treaties -the President, Esther Mshai Tolle, had highlighted in her opening speech the urgent need to adjust copyright protection which had evolved in response to earlier technologies, ''to the new means of exploitation that resulted from digital technology and global information networks.'' 50 Member state delegations, however, did not unanimously share the opinion expressed in the WCT Basic Proposal that the adaptation of copyright law to the new digital environment necessitated the recognition of an all-embracing, comprehensive right of reproduction. 51 Singapore stated instead that the inclusion of temporary reproduction in the reproduction right ''had far reaching consequences, in that fair use rights in the areas of researchers, libraries and private use were in danger of being curtailed.'' 52 Departing from the Basic Proposal, Singapore expressed the view that the digital environment did not require such a broad extension of the right of reproduction. On the contrary, the delegation felt that it should be made clear that: temporary reproduction of works should be permissible when the reproduction merely made perceptible works which would otherwise be imperceptible, when it was of a transient or incidental nature, or when it facilitated transmission of a work and had no economic value independent from facilitating transmission. 53 In the work of Main Committee I, this position received support from Scandinavian countries. Interestingly, the discussion also became more fundamental in the sense of addressing the concept of ''reproduction'' as such. While Singapore had expressed concerns about the loss of ''fair use rights,'' 54 Denmark explained in Main Committee I that temporary reproduction made for the sole purpose of making a work perceptible, or reproduction of a purely transient or incidental character, as part of a technical process, ''did not constitute reproduction within the meaning of Art. 9(1) of the Berne Convention.'' 55 Sweden confirmed that these temporary acts of reproduction ''were not really reproductions.'' 56 In the same vein, Norway stated that, instead of merely allowing for the adoption of copyright limitations to keep Footnote 48 continued and Hugenholtz (2019), p. 367; Ricketson and Ginsburg (2006), para. 11.73; von Lewinski (2008), para. 17.53;Sheinblatt (1998), p. 542;von Lewinski and Gaster (1997), pp. 624-625; Vinje (1996), p. 588. 49 Reinbothe and von , para. 7.1.29; Samuelson (1997), p. 388. 50 WIPO (1999, Vol. 2, 597, para. 311. Cf. Ricketson and Ginsburg (2006), para. 11.68. 51 Reinbothe and von  acts of temporary copying free, it should be made clear in the text of the future treaty that temporary reproduction made for the sole purpose of making a work perceptible, or of a purely transient or incidental character as a part of a technical process, ''did not as such constitute a reproduction within the meaning of Article 9(1) of the Berne Convention.'' 57 To better illustrate its position, Norway tabled a concrete proposal for amending Art. 7(1) of the WCT Basic Proposal which consisted of adding the following sentence: Temporary reproduction made for the sole purpose of making a work perceptible, or which are of a purely transient or incidental character as part of a technical process, does not as such constitute a reproduction within the meaning of Article 9(1) of the Berne Convention. 58 The negotiations in Main Committee I, thus, did not confirm the broad, unlimited scope of the right of reproduction laid down in Art. 9(1) BC. As explained, the Basic Proposal for the later WCT sought to establish the comprehensive, general scope of the right of reproduction and create breathing space for temporary acts of copying only in the form of narrowly-defined exceptions satisfying all requirements of the three-step test. 59 In the discussion in Main Committee I, however, this proposal gave rise to the question whether the right of reproduction -and the underlying concept of ''reproduction'' -could be understood to cover temporary copying at all. 60 Instead of confirming the broad scope of the right, several delegations cast doubt upon the comprehensive nature of Art. 9(1) BC.
The position taken by the Scandinavian countries did not remain an outlier. Switzerland declared that it shared the concerns expressed by the delegations of Denmark, Sweden and Norway. 61 Australia proposed an amendment to Art. 7 of the WCT Basic Proposal that pointed in the same direction as the proposal made by Norway: instead of providing limited room for the introduction of copyright exceptions, it was advisable to make a statement clarifying that temporary copying fell outside the scope of the right of reproduction altogether. 62 The African Group synthesized this approach by proposing a new Art. 7(2) for the later WCT that would replace the mere possibility for contracting parties to adopt copyright exceptions in the original Art. 7(2) of the WCT Basis Proposal: Temporary reproduction does not as such constitute a reproduction within the meaning of Article 9 (1)  (ii) such temporary reproduction is part of a technical process incidental to the transmission or utilization of the work concerned; or (iii) such reproduction is incidental to the use of the work as authorized by the right holder concerned or permitted by law. 63 In an explanatory note, the Group of Latin American and Caribbean Countries described the need for this remarkable change of course as follows: Article 9(1) of the Berne Convention establishes the right of reproduction ''… in any manner or form.'' In view of the fact that new technology has broadened traditional ''forms'' and ''manners,'' it has become necessary to redefine the scope of the right of reproduction and exclude from it those reproductions that are technically necessary for making the work perceptible or form part of a technical process and do not in themselves constitute the production of originals or copies of the work. If the foregoing is not expressly stated, it could give rise to one-sided, excessively broad interpretations of the scope of Article 9(1) of the Berne Convention. 64 Therefore, the deliberations at the 1996 Diplomatic Conference culminated in statements and proposals that, instead of confirming the comprehensive scope of the right of reproduction at the international level, did not hesitate to articulate the need for a redefinition of the conceptual contours and reach of the exclusive right granted in Art. 9(1) BC in the light of technological developments. 65 With initiatives seeking to carve out temporary copying from the scope of the right of reproduction altogether, a consensus on Art. 7 of the WCT Basic Proposal was beyond reach. Against this background, Main Committee I decided to delete the proposed Art. 7 of the WCT Basic Proposal altogether. 66 With regard to the right of reproduction, the final treaty text only contains the following Agreed Statement accompanying Art. 1(4) WCT: The reproduction right, as set out in Article 9 of the Berne Convention, and the exceptions permitted thereunder, fully apply in the digital environment, in particular to the use of works in digital form. It is understood that the storage of a protected work in digital form in an electronic medium constitutes a reproduction within the meaning of Article 9 of the Berne Convention.  Ricketson and Ginsburg (2006), paras. 11.68 and 11.74;Samuelson (1997), pp. 389-390.
The impact and interpretative authority 67 of this Agreed Statement is limited because it has not been adopted by consensus. 68 The Statement was established with 51 votes in favour, 5 against, and 30 abstentions. 69 Explaining the reasons for its vote against the Agreed Statement, Brazil pointed out that the Diplomatic Conference had not managed to reach consensus on the definition of the term ''storage.'' 70 It was the understanding of Brazil that neither the access to make a work perceptible by browsing nor the transmission of a work through a computer network in the occurrence of a temporary or non-temporary storage resulting from a technical procedure infringed the exclusive right of reproduction within the meaning of the Berne Convention. 71 Korea explained that it had opposed the Agreed Statement because reproductions that were not relevant in economic terms should not be considered as reproductions all the time. The acts of browsing or providing telecommunication facilities had an economic value in a number of cases. It was, however, difficult to distinguish acts which had an economic significance from those which had not. Against this background, Korea believed that ''acts of browsing should not be covered by the exclusive right of reproduction without exceptions.'' 72 3.2 Application of the Conference Experiences to TDM Quite clearly, the discussion about temporary copying that emerged with regard to browsing, caching and streaming at the 1996 Diplomatic Conference is not directly applicable to the current TDM discussion. TDM research does not always remain limited to temporary copying. TDM may require the establishment of a stable ''corpus'' of source material for the computational analysis -a corpus that is not of a mere temporary or transient nature. 73 If the TDM corpus includes (parts of) works that enjoy copyright protection, it cannot be said that the negotiations at the 1996 Diplomatic Conference offer clear guidance on how to deal with the question of potential infringement in the light of the international right of reproduction laid down in Art. 9(1) BC. As explained, Art. 9(1) BC covers reproduction ''in any manner or form.'' In line with Art. 31(1) of the Vienna Convention on the Law of Treaties (''Vienna Convention'' or ''VCLT''), this text of the provision plays a central role: ''A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.'' 67 Cf. Reinbothe and von Lewinski (2015), para. 7.1.29, who refer to ''interpretative value'' in the absence of the ''status of an Article of the WCT.'' 68 For a more detailed description of the circumstances leading to the adoption of the Agreed Statement at the Conference, see Samuelson (1997), pp. 390-391. 69 WIPO (1999), Vol. 1, 628, para. 497. Cf. Reinbothe and von , para. 7.1.50; Samuelson (1997), pp. 390-391. 70 For a more detailed discussion of use of the word ''storage'' in the Agreed Statement, see Ricketson and Ginsburg (2006) With this formula, Art. 31(1) of the Vienna Convention establishes the primacy of the text as the basis for its interpretation. The starting point of treaty interpretation is the elucidation of the meaning of the text. The parties are to be presumed to have the intention that appears from the ordinary meaning of the terms which they used. 74 Admittedly, these terms are not to be determined in the abstract. Article 31(1) VCLT does not demand a purely grammatical or linguistic analysis. The interpreter is obliged to consider the ordinary meaning of the term at issue systematically, in the context of the whole treaty, and in the light of its object and purpose. Together with this context, several further sources are to be considered in accordance with Art. 31(3) VCLT. These additional sources include ''any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions.'' 75 The crucial point resulting from the debate at the 1996 Diplomatic Conference, however, is the absence of agreement. 76 While Art. 7(1) of the WCT Basic Proposal sought to confirm the comprehensive scope of the right of reproduction with regard to the digital environment and digital modes of reproduction, the Conference did not yield a clear result. An agreement in the sense of Art. 31(3) VCLT is missing.
Nonetheless, it is possible to lend weight to the fact that the scope of the right of reproduction in Art. 9(1) BC was deemed less ''expansive'' at the Conference than the Basic Proposal for the later WCT assumed. 77 The Vienna Convention does not inhibit interpreters from consulting extrinsic material besides the mere terms of a treaty. Separated from the described general rules set out in Art. 31 (1) and (3) VCLT, the rules for going beyond the terms of the treaty are laid down in Art. 32 VCLT: Recourse may be had to supplementary means of interpretation, including the preparatory work of the treaty and the circumstances of its conclusion, in order to confirm the meaning resulting from the application of article 31, or to determine the meaning when the interpretation according to article 31: (a) leaves the meaning ambiguous or obscure; or (b) leads to a result which is manifestly absurd or unreasonable.
Access to material besides the mere treaty text is thus allowed. The formulation ''including the preparatory work of the treaty …'' (emphasis added) leaves no doubt that relevant external sources are not limited to the travaux pre´paratoires of the relevant international provision itself. As a supplementary means of interpretation, the discussion at the 1996 Diplomatic Conference can play a role as well. As Ian Sinclair has pointed out: it is clear that no would-be interpreter of a treaty, whatever his doctrinal point of departure, will deliberately ignore any material which can usefully serve as 74 Cf. Rauschning (1978), pp. 252-253; Sinclair (1984), p. 115;Bernhardt (1967) As encouraging as this statement may appear, it is important to point out that a cautious approach is necessary. The clear distinction between the general rules of interpretation, set out in Art. 31 VCLT, and the rules concerning supplementary means, laid down in Art. 32 VCLT, underlines that the supplementary sources should not be misused to establish an alternative, autonomous method of interpretation. Article 32 is no loophole for undermining the primacy of the treaty text by switching to a subjective approach that aligns the treaty interpretation with intentions of the parties that have not made their way into the treaty text itself. Recourse to supplementary sources may only serve as a means to aid an interpretation governed by the principles set forth in Art. 31. 79 More concretely, this implies that it is indispensable to identify a gap in the treaty text: an ''ambiguous or obscure'' 80 meaning that requires the consultation of supplementary sources to obtain a proper understanding of the international provision under examination.
The rules of the Vienna Convention, thus, give rise to the question whether the text of Art. 9(1) BC contains an ambiguous or obscure element that requires recourse to supplementary means of interpretation, such as the records of the 1996 Diplomatic Conference reflecting the described controversy concerning the scope of the right of reproduction. The expression ''in any manner or form'' features prominently in Art. 9(1) BC. These words, however, seem to offer hardly any starting point for establishing an ambiguity. Once it is clear what ''reproduction'' is, Art. 9(1) BC seeks to give authors control over all manners and all forms of reproduction.
Therefore, the ambiguous and obscure element is not the expression ''in any manner or form.'' It is the very concept of ''reproduction'' that lacks precision. 81 At the 1996 Diplomatic Conference, it was conceded in the notes accompanying the Basic Proposal for the later WPPT that ''no definition of 'reproduction' has been proposed. This is because the meaning of the term is developed fully in Articles 7 and 14.'' 82 Articles 7 and 14 of the WPPT Basic Proposal, however, did not contain a definition of ''reproduction.'' 83 Instead, these provisions were fully aligned with the clarification of the right of reproduction that had been proposed in Art. 7(1) of the WCT Basic Proposal. 84 Article 7(1) WPPT Basic Proposal read as follows: Performers shall enjoy the exclusive right of authorizing the direct or indirect reproduction, whether permanent or temporary, of their 78 Sinclair (1984), p. 116. 79 Cf. Rauschning (1978), p. 255. 80 Article 32(a) VCLT. 81 Cf. Ricketson and Ginsburg (2006) Alternative A: musical performances fixed in phonograms, Alternative B: performances fixed in any medium, in any manner or form. 85 Article 14 WPPT Basic Proposal comprised the same central elements: ''direct or indirect,'' ''permanent or temporary'' and ''in any manner or form.'' 86 As indicated, these building blocks could also be found in the corresponding copyright provision in Art. 7(1) WCT Basic Proposal. 87 A comparison with Art. 9(1) BC shows that this family of draft provisions in which the concept of reproduction had been ''developed fully'' 88 -to such an extent that a separate definition was deemed unnecessary -contains the additional element ''permanent or temporary.'' Hence, the Basic Proposal for the later WCT and WPPT, indeed, sought to remove an ambiguity in the sense of Art. 32(a) VCLT: an ambiguity in the text of Art. 9(1) BC that had been adopted at the 1967 Stockholm Conference for the revision of the Berne Convention. 89 At the time, acts of temporary copying in the digital environment, such as browsing, caching and streaming, did not exist. Accordingly, the 1967 Stockholm Conference did not address this phenomenon. Therefore, it was an open question whether Art. 9(1) BC could be understood to cover not only permanent but also temporary copying. As a supplementary means of interpretation in the sense of Art. 32(a) VCLT, the records of the 1996 Diplomatic Conference demonstrate that no agreement on this question could be established. The ambiguity which the text of Art. 9(1) BC has in this respect remained unresolved. 90 To add a further example of Member State positions, it is of particular interest that China proposed an amendment of Art. 7(1) WCT Basic Proposal that avoided any reference to temporary copying: The exclusive right accorded to authors of literary and artistic works in Article 9(1) of the Berne Convention of authorizing the reproduction of their works in any manner or form shall include direct and indirect permanent reproduction of their works. 91 In sum, the concept of ''reproduction'' underlying Art. 9(1) BC is ambiguous and obscure in the sense of Art. 32(a) VCLT. While the text of Art. 9(1) BC gives evidence of the intention to include reproduction ''in any manner or form'' known at the time of the 1967 Stockholm Conference, the status of new categories of reproduction that emerged after the adoption of the current text of Art. 9(1) BC, such as the category of temporary copying in electronic networks, remains unclear.  WIPO (1971), Document S/1, 113. Cf. Senftleben 2004, p. 47. 90 Ricketson and Ginsburg (2006), para. 11.71 and 11.75;Samuelson (1997), p. 392;Sheinblatt (1998), p. 550. 91 WIPO (1999 of Berne Union Members also shows that countries avail themselves of the flexibility that arises from this unclear status of temporary copying. Art. 13a of the Copyright Act of the Netherlands (Auteurswet), for instance, contains the following ''carve out'' of temporary acts of reproduction: The reproduction of a literary, scientific or artistic work does not include the temporary reproduction that is transient or incidental and an integral and essential part of a technological process, the sole purpose of which is to enable: (a) a transmission in a network between third parties by an intermediary, or (b) a lawful use of a work to be made, and which has no independent economic significance. 92 With regard to TDM, these insights raise the question whether copying for the purposes of an automated, computational TDM analysis can also be qualified as a reproduction category that falls outside the scope of the general right of reproduction that was adopted and enshrined in Art. 9(1) BC in 1967. Obviously, it is necessary to walk a fine line in this regard. The mere fact that modern technology for carrying out a certain form of reproduction was not known in 1967 can hardly be sufficient. Otherwise, Art. 9(1) BC would be rendered incapable of keeping pace with technological developments and become outdated and meaningless. 93 At the 1967 Stockholm Conference, for instance, photocopying played an important role in the discussion about the right of reproduction. 94 Therefore, this form of copying clearly falls within the scope of the concept of ''reproduction'' underlying Art. 9(1) BC. The fact that photocopying technology has evolved since 1967 and may differ quite substantially from the photocopying machines known at the time of the Stockholm Conference does not change the equation.
For TDM to fall outside the province of Art. 9(1) BC, it is thus necessary to demonstrate that this is a new category of copying -a category that, like temporary copying, has not been considered and included in the reproduction concept in 1967. Arguably, this threshold for assuming an outright ''carve out'' does not pose major difficulties. The parallels between the fundamental discussion of the conceptual contours of reproduction at the 1996 Diplomatic Conference and the current TDM debate are striking. As the President of the 1996 Conference for the WIPO ''Internet'' Treaties recalled, the negotiations aimed at developing an appropriate regulatory response to challenges that had arisen from new technologies. 95 The same can be said about the questions raised by TDM technology: questions about an unprecedented and formerly unknown type of using copyrighted source material in the context of an automated analytical process.
Even more importantly, there is also a parallel to be drawn with regard to the crucial conceptual question relating to the scope of the right of reproduction laid 92 Van Eechoud (2012), pp. 511-512; Spoor (2012), p. 204. 93 Cf. Spoor (2012), p. 205. 94 WIPO (1971, Document S/1, 112, footnote 1, and Report on the Work of Main Committee I, pp. 1145-1146. 95 WIPO (1999), Vol. 2, 597, para. 311. down in Art. 9(1) BC. At the 1996 Diplomatic Conference leading to the WIPO ''Internet'' Treaties, acts of temporary copying, forming part of browsing, caching and streaming activities, triggered fundamental questions about the Berne Convention concept of reproduction and the corresponding scope of the international right of reproduction. As discussed, the proposal to bring temporary copying under the umbrella of Art. 9(1) BC had no success. Many delegations rejected the proposal and, instead, insisted on a clarification that this new category of making copies fell outside the scope of the right of reproduction altogether. 96 In the TDM debate, similar points -underlining the fundamentally different nature of TDM copies and the lack of a relevant act of reproduction in the traditional sense of making copies for the purpose of consulting and enjoying a work -have been made around the globe. From a US perspective, Michael Carroll has pointed out that in the context of TDM: copies are made only for computational research and the durable outputs of any text and data mining analysis would be factual data and would not contain enough of the original expression in the analysed articles to be copies that count. 97 Explaining the outright exemption of TDM activities in Art. 30-4(ii) of the Japanese Copyright Act, Tatsuhiro Ueno has pointed out that: if an exploitation of a work is aimed at neither enjoying it nor causing another person to enjoy it (e.g. text-and-data mining, reverse engineering), there is no need to guarantee the opportunity of an author or copyright holder to receive compensation and thus copyright does not need to cover such exploitation. In other words, exploitation of this kind does not prejudice the copyright holder's interests protected by a copyright law. 98 Criticizing the restrictive regulation of TDM in the EU, Rosanna Ducato and Alain Strowel described the following alternative approach: when acts of reproduction are carried out for the purpose of search and TDM, the work, although it might be reproduced in part, is not used as a work: the work only serves as a tool or data for deriving other relevant information. The expressive features of the work are not used, and there is no public to enjoy the work, as the work is only an input in a process for searching a corpus and identifying occurrences and possible trends or patterns. 99 In fact, the distinction between use of ''works as works'' and use of ''works as data'' is not entirely new in the European copyright debate. In 2011, Mauricio Borghi and Stavroula Karapapa already developed the concept of ''de-intellectualized use'' 100 96 WIPO (1999), Vol. 2, 764, paras. 877-881. Cf. Goldstein and Hugenholtz (2019), p. 287; Ricketson and Ginsburg (2006), paras. 11.68, 11.71 and 11.74;Samuelson (1997), p. 392;Sheinblatt (1998), p. 550. 97 Carroll (2019, p. 954. 98 Ueno (2021), pp. 150-151. 99 Ducato and Strowel (2021), p. 334. 100 Borghi and Karapapa (2011), p. 45. against the background of mass digitization projects, such as the Google Book Search. As Borghi and Karapapa point out, mass digitization turns protected content into mere data -with the result that ''the expression of the idea embodied in the work is not primarily used to communicate the 'speech' of the author to the public but rather to form the basis of machine-workable algorithms. '' 101 Considering these comments and observations, it becomes apparent that TDM raises issues that concern the conceptual core of the right of reproduction: the question whether a relevant category of reproduction can be assumed in TDM cases at all. Hence, it is important to recall that the negotiators at the 1996 Diplomatic Conference leading to the WCT and the WPPT did not jump to the conclusion that, because of the reference to reproduction ''in any manner or form'' in Art. 9(1) BC, the grant of the right of reproduction at the international level, by definition, must be understood to cover each and every form of reproduction, regardless of the individual circumstances surrounding the use. By contrast, many delegations preferred a nuanced approach and sought to leave room for ''carve outs'' that place new categories of copying, such as temporary copying, outside the scope of the international right of reproduction. 102 Additional support for this position can be derived from the preparatory work for the 1996 Diplomatic Conference. Seeking to clarify the scope of the international right of reproduction prior to the Conference, the 1992 WIPO Committee of Experts on a Possible Protocol to the Berne Convention proposed, as one option for tracing the conceptual contours of the right of reproduction in Art. 9(1) BC, to determine whether a protected work ''with the aid of a machine or other device can be so perceived [by seeing or hearing it] and, if so desired, further reproduced or communicated.'' 103 This criterion 104 already played a role in the 1982 deliberations of the Second WIPO/UNESCO Committee of Governmental Experts on Copyright Problems Arising from the Use of Computers for Access to or the Creation of Works. With regard to input-related questions, this latter Committee had developed the following formula for identifying acts falling within the scope of the right of reproduction: As for the act of input of protected subject matter for storage purposes, it was generally agreed that it included at least reproduction of works on a machinereadable material support and their fixation in the memory of a computer system. The Committee agreed that whatever this act may be it involves fixation of works in a form sufficiently stable to permit their communication to an individual, and therefore it should be considered as governed by the international conventions and national legislation on copyright and therefore was subjected to the author's exclusive rights. Thus, prior authorization by the right owner is required for the said act. Some delegations precised that sufficient stability of a form in which a work is fixed should be considered from the functional side, in the sense that the work can be perceived, reproduced or otherwise communicated to the public with the aid of a computer system. 105 Considering the current TDM debate and the question of relevant acts of reproduction, this statement is of particular interest. Its first sentence seems to support the conclusion that ''at least reproduction of works on a machine-readable material support and their fixation in the memory of a computer system'' 106 should be qualified as a form of copying that requires the prior authorization of copyright holders. Following this approach, reproductions leading to a stable TDM corpus in a computer memory would amount to copyright infringement in the absence of prior consent. The second sentence, however, adds an important nuance by articulating the overarching criterion of ''fixation of works in a form sufficiently stable to permit their communication to an individual.'' 107 Evidently, the Committee also attached importance to the underlying aim: ''communication to an individual.'' 108 This configuration of the test offers room for the exclusion of TDM copies from the scope of the right of reproduction. As copies for TDM purposes merely concern a computational analysis 109 and do not constitute use of a work ''as a work,'' 110 they are not intended to permit ''communication to an individual'' in the sense of the criteria developed by the WIPO/UNESCO Committee. 111 The final part of the statement confirms this understanding: Some delegations precised that sufficient stability of a form in which a work is fixed should be considered from the functional side, in the sense that the work can be perceived, reproduced or otherwise communicated to the public with the aid of a computer system. 112 In the 1982 WIPO/UNESCO Committee, there was thus support for a ''functional'' approach asking whether ''the work can be perceived, reproduced or otherwise communicated to the public.'' 113 As indicated above, this element reappeared ten years later in the work of the 1992 WIPO Committee of Experts on a Possible Protocol to the Berne Convention. According to this 1992 Expert Committee, relevant acts of reproduction can be identified by asking whether a protected work ''can be so perceived and, if so desired, further reproduced or communicated.'' 114 As Jörg Reinbothe and Silke von Lewinski explain, this criterion has been developed against the background of temporary, transient copies. They argue that, as long as temporary copies can be perceived, reproduced, or otherwise communicated, they are ''by definition as much acts of reproduction under copyright as permanent ones.'' 115 However, this position did not receive sufficient support during the 1996 Diplomatic Conference leading to the WCT. A more accurate description of the outcome of the Conference is offered by Pam Samuelson: The most honest thing that can be said about the temporary copying of works in computer memory is that there is no international consensus on this subject. Still, it is significant that the copyright treaty signed in Geneva does not include a provision on temporary copying given how intent the U.S. and E.U. delegations had been about getting such a treaty provision. 116 As explained above, the delegates at the Diplomatic Conference were not convinced that temporary copying should be qualified as a relevant act of reproduction. 117 This outcome of the Conference was reached even though, for instance in cases of browsing, temporary copies enable an individual user to perceive and enjoy the work. 118 Against this background, there is even more reason to assume that the international right of reproduction laid down in Art. 9(1) BC does not apply when, 113 WIPO/UNESCO (1982), id., para. 33. 114 WIPO (1992, 68, para. 75. Cf. Reinbothe and von Lewinski (2015), para. 7.1.31. 115 Reinbothe and von , paras. 7.1.53-7.1.56. For a similar line of argument invoking Art. 20 BC, see von Lewinski (2008), para. 17.57. Article 20 BC offers room for special agreements, such as the WCT, ''in so far as such agreements grant to authors more extensive rights than those granted by the [Berne] Convention, or contain other provisions not contrary to this Convention.'' Quite clearly, however, it cannot be inferred from Art. 20 BC that the 1996 Diplomatic Conference adopting the WCT led to the recognition of a right of reproduction covering temporary copying. In the absence of agreement on this point at the 1996 Diplomatic Conference itself, an international obligation to grant a right of reproduction with this scope cannot be inferred from Art. 20 BC. As Ricketson and Ginsburg (2006), paras. 11.69-11.71 and 11.75, explain, the status of temporary copying under Art. 9(1) BC is controversial and unclear. It cannot be said that Art. 9(1) BC covers temporary copying. Cf. Samuelson (1997), p. 392;Sheinblatt (1998), p. 550. 116 Samuelson (1997 from the outset, copies are not made for the purpose of allowing an individual to perceive, reproduce or otherwise communicate the work. 119 In the case of TDM copies, works are only included in a harmonized dataset to allow the analysis with the help of an automated, computational process. The requirement of a reproduction allowing an individual to consult and enjoy the work is not fulfilled. More generally, it seems safe to assume in the light of the experiences with the 1996 Diplomatic Conference that the international copyright community has reserved for itself the freedom of taking individual decisions on each new category of copying that has not been considered previously. 120 As discussed above, the debate at the 1996 Diplomatic Conference -and the lack of consensus on a comprehensive, all-embracing right of reproduction -has the status of a supplementary means of interpretation in the sense of Art. 32(a) VCLT for determining the scope of the reproduction concept underlying Art. 9(1) BC. Instead of confirming the comprehensive scope of the provision and allowing the internationally recognized right of reproduction to automatically absorb new types of copying, the 1996 Diplomatic Conference demonstrated that the negotiators considered themselves free to decide otherwise and keep new categories, such as temporary acts of copying, outside the scope of the right of reproduction. From this perspective, Art. 9(1) BC cannot be understood to cover TDM until such time as an international consensus is reached on this very question. Given the considerable conceptual differences between TDM copying for the purpose of an automated, computational analysis, and traditional forms of reproduction that have been considered at the 1967 Stockholm Conference, such as photocopying for the purposes of reading and enjoying a work, TDM copies can be qualified as a distinct category that requires international agreement before it can be brought within the scope of the 1967 reproduction concept underlying Art. 9(1) BC.
The present analysis, however, cannot be concluded without revisiting the Agreed Statement that was established at the 1996 Diplomatic Conference to address reproductions in the digital environment. As doubtful as its authority may be because of the missing consensus at the Conference, 121 the Agreed Statement concerning Art. 1(4) WCT adds further nuances to the international concept of reproduction that must not be disregarded. As already explained, the Agreed Statement reads as follows: The reproduction right, as set out in Article 9 of the Berne Convention, and the exceptions permitted thereunder, fully apply in the digital environment, in particular to the use of works in digital form. It is understood that the storage of a protected work in digital form in an electronic medium constitutes a reproduction within the meaning of Article 9 of the Berne Convention.
For a proper assessment of the impact of this Statement, it is important to distinguish between the first and the second sentence. Above, it has already been pointed out that the Agreed Statement was not adopted by consensus in the final plenary Conference session. 122 However, the earlier discussion in Main Committee I -where the substantive provisions of the later WCT have been prepared -reveals that the second sentence constituted the controversial element. The first sentence was adopted unanimously in Main Committee I. 123 In the light of the foregoing analysis, however, the first sentence does not militate against the assumption that Art. 9(1) BC fails to cover TDM: a new category of copying outside the frame of the 1967 Stockholm Conference that was still unknown at the 1996 Diplomatic Conference. The first sentence confirms that the international right of reproduction fully applies in the digital environment and covers the use of works in digital form. Quite clearly, however, this clarification aims at allowing the right of reproduction to keep pace with technological developments that affect reproduction categories that have been considered previously in the context of international policymaking. The Agreed Statement concerns ''[t]he reproduction right, as set out in Art. 9 of the Berne Convention'' and, accordingly, focuses on the continuation of the reproduction concept underlying Art. 9 BC. Copying for the purpose of private study and enjoyment, for instance, was a controversial issue at the 1967 Stockholm Conference. 124 In this respect, the first sentence of the Agreed Statement is of utmost importance. It ensures that Art. 9(1) BC remains fully applicable. The switch from analogue private copying to digital private copying does not inhibit right holders from invoking the protection following from Art. 9(1) BC. In this way, the first sentence of the Agreed Statement shields right holders from an erosion of the right in the digital environment. If the use of digital technology for a traditional form of copying, such as private copying, were sufficient to escape the verdict of infringement, the international right of reproduction would already have been deprived of most of its practical importance.
However, the first sentence -only forming part of an Agreed Statement, not of a substantive and formally adopted treaty provision -can hardly be understood as being sufficient to take more radical steps and include a completely new (and at the time of the 1996 Diplomatic Conference unknown) category of copying, such as copying for TDM purposes, in the concept of reproduction. Otherwise, the same could be said about temporary copying. This, however, would be highly inconsistent in view of the described critical remarks and the strong opposition of Member States to the inclusion of temporary copying at the 1996 Diplomatic Conference itself. 125 Rightly understood, the first sentence of the Agreed Statement concerning Art. 1(4) WCT does not offer support for the extension of the reproduction concept to new categories, such as temporary copying and copying for TDM purposes. Commentary literature discussing the first sentence stresses that the confirmation of full applicability in the digital environment is ''possibly meaningless'' 126 and that the central point lies in the first sentence's ''emphasis on exceptions to the reproduction right'' 127 -a point that does not support the assumption that the first sentence provides a basis for extending the traditional reproduction right in Art. 9(1) BC to new categories, such as temporary and TDM copying.
With regard to attempts to interpret the words ''fully apply'' in the first sentence in a way that allows the extension of the right of reproduction to temporary copying, Sam Ricketson and Jane Ginsburg add that ''the second sentence of the agreed statement muddies the waters.'' 128 As they explain, the text of the second sentence scrupulously avoids detailing the meaning of ''storage''. If ''storage'' means retention for some more than transitory period of time, then it might add nothing to ''any manner or form'' [in Article 9(1) BC], because ''form'' could imply a manifestation more concrete than a transient communication. If, however, the term ''storage'' was meant also to cover fleeting passage through a digital network, then one may wonder at the selection of a term whose ordinary meaning implies more lingering than just passing through. 129 With its reference to ''storage of a protected work in digital form in an electronic medium,'' the second sentence of the Agreed Statement may also fuel discussions about the extension of the reproduction concept to TDM. In Main Committee I, Brazil expressed concerns relating to temporary copying. It was the understanding of Brazil that: the access to make a work perceptible by browsing, and the transmission of a work through a computer network in the course of a temporary or nonpermanent storage resulting from a technical procedure, did not infringe the exclusive rights of reproduction within the meaning of Article 9 of the Berne Convention. 130 The discussion, thus, focused on the issue of browsing, caching and streaming. The temporary nature of these forms of copying occupied centre stage. As already indicated above, TDM will often require the creation of a stable TDM corpus. TDM practices go beyond mere temporary acts of reproduction. To bring TDM under the umbrella of Art. 9(1) BC on the basis of the second sentence, however, it would be necessary to infer from the focus on temporary, non-permanent storage at the 1996 Diplomatic Conference that the extension of the international right of reproduction to more permanent storage, such as storage of protected fragments of a work in a digital TDM corpus, would be uncontroversial. 126 Ricketson and Ginsburg (2006), para. 11.74. 127 Ricketson and Ginsburg (2006), para. 11.74. 128 Ricketson and Ginsburg (2006), para. 11.75. 129 Ricketson and Ginsburg (2006), para. 11.75. 130 WIPO (1999), Vol. 2, 785, para. 1054 For several reasons, this line of argument is indefensible. At the 1996 Diplomatic Conference, the negotiators had no opportunity to discuss TDM because this category of copying was unknown at the time. The absence of stronger resistance against the inclusion of the word ''storage'' in the second sentence, thus, does not indicate that the delegations generally agreed on a broad understanding of the second sentence -an agreement solid enough to provide a basis for the extension of the international right of reproduction to TDM copies that are more stable than mere temporary copies. As consensus on temporary, non-permanent copying was beyond reach, it is much more convincing to assume that the delegations simply sought to express, as a minimum result, their agreement on more permanent forms of storage that are the result of traditional reproduction categories, such as private copying leading to storage in an electronic medium.
Finally, it must be recalled that the second sentence of the Agreed Statement has very limited authority anyway. The second sentence constitutes the controversial element that led to a situation where the Agreed Statement could only be adopted with 51 votes in favour, 5 against, and 30 abstentions in the final plenary session. 131 In total, 130 countries and 83 organizations participated in the 1996 Diplomatic Conference. 132 Therefore, more than 40 countries chose not to attend the session adopting the Agreed Statement. 133 This fact further diminishes the impact of the Statement and, in particular, the impact of the second sentence. Even if the second sentence could be understood to cover storage in a TDM corpus, the international consensus is missing that would be required for the far-reaching policy decision to subordinate TDM research to the right of reproduction. 134 Hence, neither the first nor the second sentence of the Agreed Statement offer support for the assumption that, as a result of the 1996 Diplomatic Conference, the reproduction concept underlying Art. 9(1) BC must be understood to encompass copies made for the purposes of TDM.

Conclusion
In the light of the experiences at the 1996 Diplomatic Conference that led to the adoption of the WCT and the WPPT, it can hardly be said that international copyright law covers copying that takes place in the context of TDM research. Admittedly, Art. 9(1) BC sets forth an exclusive right of reproduction covering reproduction ''in any manner or form.'' The Agreed Statement concerning Art. 1(4) WCT may give rise to the question whether TDM activities leading to the ''storage 131 WIPO (1999), Vol. 1, 628, para. 497. Cf. Reinbothe and von , para. 7.1.50; Samuelson (1997), pp. 390-391. 132 WIPO (1999), Vol. 2, 819-882. Cf. WIPO (1997, 5, para. 19. of a protected work in digital form in an electronic medium'' 135 amount to relevant acts of reproduction in the sense of Art. 9(1) BC.
Considering the deliberations at the 1996 Diplomatic Conference, however, the true story is this: despite the seemingly broad scope of Art. 9(1) BC, the international copyright community has reserved for itself the freedom of taking individual decisions on each new category of copying that has not been considered previously. Instead of confirming the comprehensive scope of Art. 9(1) BC and allowing the internationally recognized right of reproduction to automatically absorb any new category of copying, the 1996 Diplomatic Conference demonstrated that the negotiators considered themselves free to decide otherwise and keep new categories of reproduction, such as temporary acts of copying, outside the scope of the international right of reproduction.
In other words: Art. 9(1) BC cannot be understood to cover TDM until such time as an international consensus is reached on this very question. As the topic of TDM has not yet been discussed at an international conference clarifying the scope of the right of reproduction, current TDM regulations that presume copyright relevance and seek to prevent inroads into the right of reproduction, such as the restrictive approach taken in the EU, 136 constitute Berne/TRIPS/WCT-plus regimes that have been adopted voluntarily in domestic legislation to offer more copyright protection than necessary under international law.
This voluntary policy decision, however, does not preclude more flexible approaches in other countries. In the absence of a binding international consensus on the question of TDM regulation, lawmakers outside the EU are free to weigh the pros and cons of flexibility in the area of TDM without any obligation to factor the international matrix of protection, including the requirements of the three-step test, into the equation. In this way, they can develop the solution that appears most promising in the light of their domestic social, cultural and economic needs and safeguard the right to research.
Rightly understood, the regulation of TDM, thus, does not constitute an issue of copyright exceptions or limitations at the international level. Instead, international policymakers must first establish that the making of copies in the context of TDM has copyright relevance and falls within the right of reproduction in Art. 9(1) BC at all. The discussion will have to commence at the level of tracing the conceptual contours of the term ''reproduction'' and determining the corresponding scope of the right granted in Art. 9(1) BC. In this discussion, policymakers are free to opt for a general ''carve out.'' They can place TDM copies outside the scope of the right of reproduction from the outset. Adopting this approach, the international copyright community would follow the precedent that has been established at the 1996 Diplomatic Conference with regard to temporary copying. 135 Agreed Statement concerning Art. 1(4) WCT. 136 With regard to the critique of the approach taken in the EU, see the literature references supra note 4.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creativecommons.org/licenses/by/4.0/.