Case 12-4547, Document 159, 06/04/2013, 955881, Page1 of 53
`
`12-4547-cv
`
`United States Court of Appeals
`for the
`Second Circuit
`
`
`
`
`
`AUTHORS GUILD, INC., AUSTRALIAN SOCIETY OF AUTHORS
`LIMITED, UNION DES ECRIVAINES ET DES ECRIVAINS QUEBECOIS,
`ANGELO LOUKAKIS, ROXANA ROBINSON, ANDRE ROY, JAMES
`SHAPIRO, DANIELE SIMPSON, T.J. STILES, FAY WELDON,
`
`(For Continuation of Caption See Inside Cover)
`
`_______________________________
`ON APPEAL FROM THE UNITED STATES DISTRICT COURT
`FOR THE SOUTHERN DISTRICT OF NEW YORK
`
`BRIEF OF DIGITAL HUMANITIES AND LAW SCHOLARS
`AS AMICI CURIAE IN SUPPORT OF DEFENDANTS-
`APPELLEES AND AFFIRMANCE
`
`
`
`
`
`
`
`
`
`
`
`
`
`On the Brief:
`MATTHEW SAG*
`ASSOCIATE PROFESSOR
`LOYOLA UNIVERSITY OF
`CHICAGO SCHOOL OF LAW
`
` *
`
`JASON SCHULTZ*
`ASSISTANT CLINICAL PROFESSOR OF LAW
`UC BERKELEY SCHOOL OF LAW
`396 Simon Hall
`Berkeley, California 94720
`(510) 642-1957
`jschultz@law.berkeley.edu
`
`Attorneys for Amici Curiae
`
` Filed in their individual capacity and not on behalf of their institutions.
`
`
`
`
`
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page2 of 53
`
`
`AUTHORS LEAGUE FUND, INC., AUTHORS’ LICENSING AND
`COLLECTING SOCIETY, SVERIGES FORFATTARFORBUND, NORSK
`FAGLITTERAER FORFATTERO OG OVERSETTERFORENING,
`WRITERS’ UNION OF CANADA, PAT CUMMINGS, ERIK GRUNDSTROM,
`HELGE RONNING, JACK R. SALAMANCA,
`Plaintiffs-Appellants,
`
`v.
`
`HATHITRUST, CORNELL UNIVERSITY, MARY SUE COLEMAN, President,
`University of Michigan, MARK G. YUDOF, President, University of California,
`KEVIN REILLY, President, University of Wisconsin System,
`MICHAEL MCROBBIE, President, Indiana University,
`
`NATIONAL FEDERATION OF THE BLIND, GEORGINA KLEEGE,
`BLAIR SEIDLITZ, COURTNEY WHEELER,
`
`Defendants-Appellees,
`
`Intervenor Defendants-Appellees.
`
`
`
`
`
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page3 of 53
`
`TABLE OF CONTENTS
`
`
`
`TABLE OF AUTHORITIES ......................................................................................................... iv
`STATEMENT OF INTEREST OF AMICI ..................................................................................... 1
`SUMMARY OF ARGUMENT ...................................................................................................... 2
`ARGUMENT .................................................................................................................................. 4
`I. The Freedom to Make Non-expressive Use of Copyrighted Works is Vital to the “Progress of
`Science” in the Digital Humanities ................................................................................................. 4
`II. Copyright Law Does Not Protect Non-expressive Aspects of Works .................................. 14
`A. The Idea/Expression Distinction ....................................................................................... 14
`B. Section 102(b) ................................................................................................................... 15
`C. Merger and Scènes à Faire ................................................................................................ 16
`D. Fact/Expression Distinction .............................................................................................. 17
`E. Non-expressive Metadata Does Not Implicate the Statutory Rights of the Copyright
`Holder ....................................................................................................................................... 18
`F. Non-expressive Metadata Is Also Noninfringing Because It Does Not Allow the Public to
`Perceive the Expressive Content of a Work ............................................................................. 22
`III. Text Mining Creates Value by Facilitating the Advancement of Our Collective Knowledge;
`and Other Non-expressive Purposes Should Be Considered "Fair Use" ...................................... 24
`A. Non-expressive Copying to Expand Our Knowledge in the Digital Humanities Is An
`Activity of the Sort that Copyright Law Should Favor, Through Fair Use .............................. 24
`B. The Nature of the Works in Question Is Favorable to the Fair Use Analysis of Mass
`Digitization for the Advancement of Digital Humanities Research and Scholarship .............. 27
`C. To the Extent Relevant, Mass Digitization Uses a Reasonable “Amount and
`Facilitating Data Mining for the Advancement of the Digital Humanities .............................. 28
`D. Allowing Intermediate Copying in Order to Enable Non-expressive Uses Does Not Harm
`Implicate the Works' Expressive Aspects in Any Way ............................................................ 29
`
`To Protect That Value, Mass Digitization and Similar Intermediate Copying for Data Mining
`
`Substantiality” of the Works in Question, in Light of the Socially Beneficial Purpose of
`
`the Market for the Original Works in a Legally Cognizable Manner, As The Practice Does Not
`
`ii
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page4 of 53
`
`CERTIFICATE OF COMPLIANCE WITH FRAP 32(a) ............................................................ 32
`
`
`
`
`
`
`
`iii
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page5 of 53
`
`
`TABLE OF AUTHORITIES
`
`Cases
`
`A.V. ex rel. Vanderhye v. iParadigms, LLC,
`562 F.3d 630 (4th Cir. 2009) ......................................................... 4, 29, 30
`
`
`Basic Books, Inc. v. Kinko's Graphics Corp.,
`758 F. Supp. 1522 (S.D.N.Y. 1991) ........................................................ 27
`
`
`Bill Graham Archives v. Dorling Kindersley Ltd.,
`448 F.3d 605 (2d Cir. 2006) .............................................................. 25, 27
`
`
`Bond v. Blum,
`317 F.3d 385 (4th Cir. 2003) ....................................................... 26, 27, 29
`
`
`Campbell v. Acuff-Rose Music, Inc.,
`510 U.S. 569 (1994) ..................................................................... 25, 28, 30
`
`
`Cariou v. Prince, No. 11-1197-cv,___ F.3d __, slip op. at 13
`(2d Cir., April 25, 2013) ......................................................................... 25, 29
`
`Castle Rock Entm’t v. Carol Publishing Grp.,
`150 F.3d 132 (2d Cir. 1998) ........................................................ 20, 21, 22
`
`
`Davis v. United Artists, Inc.,
`547 F. Supp. 722 (S.D.N.Y. 1982) .......................................................... 23
`
`
`Eldred v. Ashcroft, 537 U.S. 186 (2003). ..................................................... 13
`
`Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., Inc.,
`499 U.S. 340 (1991) ........................................................................... 17, 20
`
`
`Fisher v. Dees,
`794 F.2d 432 (9th Cir. 1986) ................................................................... 30
`
`
`Fuld v. Nat’l Broad. Co., Inc.,
`390 F. Supp. 877 (S.D.N.Y. 1975) .......................................................... 23
`
`
`
`iv
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page6 of 53
`
`Golan v. Holder,
`132 S. Ct. 873 (2012) ............................................................................... 15
`
`
`Harper & Row Publishers, Inc. v. Nation Enters.,
`471 U.S. 539 (1985) ........................................................................... 14–15
`
`
`Hasbro Bradley, Inc. v. Sparkle Toys, Inc.,
`780 F.2d 189 (2d Cir. 1985) .................................................................... 21
`
`
`Hoehling v. Universal City Studios, Inc.,
`618 F.2d 972 (2d Cir. 1980) .............................................................. 16, 18
`
`
`Kelly v. Arriba Soft Corp.,
`336 F.3d 811 (9th Cir. 2002) ............................................................. 25, 29
`
`
`Kregos v. Associated Press,
`937 F.2d 700 (2d Cir. 1991) .................................................................... 16
`
`
`Madrid v. Chronicle Books,
`209 F. Supp. 2d 1227 (D. Wyo. 2002) ..................................................... 23
`
`
`MyWebGrocer, LLC v. Hometown Info, Inc.,
`375 F.3d 190 (2d Cir. 2004) .................................................................... 17
`
`
`Nat’l Basketball Ass’n v. Motorola, Inc.,
`105 F.3d 841 (2nd Cir. 1997) ............................................................ 17, 18
`
`
`New Era Publ’ns Int’l, ApS v. Carol Pub. Grp.,
`904 F.2d 152 (2d Cir. 1990) ............................................................... 27-28
`
`
`N.Y. Mercantile Exch., Inc. v. IntercontinentalExchange, Inc.,
`497 F.3d 109 (2d Cir. 2007) .................................................................... 16
`
`
`N.Y. Times Co. v. Tasini,
`533 U.S. 483 (2001) ........................................................................... 22, 23
`
`
`NXIVM Corp. v. Ross Inst.,
`364 F.3d 471 (2d Cir. 2004) .................................................................... 26
`
`
`
`
`v
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page7 of 53
`
`Perfect 10, Inc. v. Amazon.com, Inc.,
`508 F.3d 1146 (9th Cir. 2007) ....................................................... 4, 25, 29
`
`
`Peter F. Gaito Architecture, LLC v. Simone Dev. Corp.,
`602 F.3d 57 (2d Cir. 2010) ...................................................................... 15
`
`
`Religious Tech. Ctr. v. Lerma,
`908 F. Supp. 1362 (E.D. Va. 1995) ................................................... 26, 27
`
`
`Reyher v. Children’s Television Workshop,
`533 F.2d 87 (2d Cir. 1976) ...................................................................... 15
`
`
`Sega Enters. Ltd. v. Accolade, Inc.,
`977 F.2d 1510 (9th Cir. 1992) ....................................................... 4, 28, 30
`
`
`Sony Computer Entm’t, Inc. v. Connectix Corp.,
`203 F.3d 596 (9th Cir. 2000) ............................................................... 4, 28
`
`
`Sony Corp. of Am. v. Universal City Studios, Inc.,
`464 U.S. 417 (1984) ............................................................................ 14-15
`
`
`Stromback v. New Line Cinema,
`384 F.3d 283 (6th Cir. 2004) ................................................................... 23
`
`
`Tufenkian Imp./Exp. Ventures, Inc. v. Einstein Moomjy, Inc.,
`338 F.3d 127 (2d Cir. 2003) .................................................................... 15
`
`
`Ty, Inc. v. Publ’ns Int’l Ltd.,
`292 F.3d 512 (7th Cir. 2002) ................................................................... 21
`
`
`Warner Bros. Entm’t Inc. v. RDR Books,
`575 F. Supp. 2d 513 (S.D.N.Y. 2008) ......................................... 19, 20, 21
`
`
`Walker v. Time Life Films, Inc.,
`615 F. Supp. 430 (S.D.N.Y. 1985) .......................................................... 23
`
`Statutes
`17 U.S.C. § 102(a) (2006) ............................................................................. 20
`
`17 U.S.C. § 102(b) (2006) ............................................................ 4, 14, 15, 16
`vi
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page8 of 53
`
`17 U.S.C. § 106(2) (2006) ............................................................................ 21
`
`17 U.S.C. § 107 (2006) ................................................................................. 25
`
`17 U.S.C. § 201(c) (2006) ............................................................................. 22
`
`Constitutional Provisions
`U.S. Const. Art I., Sec. 8 ............................................................................... 13
`
`Other Authorities
`Sophia Ananiadou et al., Text Mining and its Potential
`Applications in Systems Biology,
`24 TRENDS IN BIOTECHNOLOGY 571 (2006) .................................................... 5
`
`Leif Isaksen, Elton Barker, Eric C. Kansa, Kate Byrne, GAP: A NeoGeo Approach
`to Classical Resources, 45 LEONARDO 82-83 (2012) ..................................... 7
`Christian Blaschke et al. Information Extraction in Molecular Biology, 3
`BRIEFINGS IN BIOINFORMATICS 154 (2002) ..................................................... 5
`
`Patricia Cohen, Digital Keys for Unlocking the Humanities’ Riches,
` N.Y. TIMES, Nov. 17, 2010, at C1 .................................................................. 7
`
`James M. Hughes, et al., Quantitative Patterns of Stylistic Influence in the
`Evolution of Literature, 109 PROC. OF THE NAT’L ACAD. OF SCI. OF
`THE U.S. 7682 (2012) ............................................................................... 10-11
`
`Matthew Jockers, MACROANALYSIS: DIGITAL METHODS FOR LITERARY
`HISTORY (2013) ..................................................................................... 6, 7, 10
`
`Brian Lavoie & Lorcan Dempsey, Beyond 1923: Characteristics of
`Potentially In Copyright Print Books in Library Collections, 15 D-Lib Mag.,
`http://www.dlib.org/dlib/november09/lavoie/11lavoie.html ........................ 28
`
`Pierre N. Leval, Toward A Fair Use Standard,
`103 HARV. L. REV. 1105 (1990) ................................................................... 26
`
`Steve Lohr, Dickens, Austen and Twain, Through a Digital Lens,
`N.Y. TIMES, Jan. 26, 2013, at BU3, available at
`http://www.nytimes.com/2013/01/27/technology/literary-history-seen-through
`-big-datas-lens.html?pagewanted=all&_r=2&. ............................................ 11
`
`vii
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page9 of 53
`
`MALLET: MAchine Learning for LanguagE Toolkit, http://mallet.cs.umass.edu/
`(last visited May 31, 2012) ........................................................................... 10
`
`Jean-Baptiste Michel, Yuan Kui Shen, Aviva Presser Aiden, Adrian Veres,
`Matthew K. Gray, The Google Books Team, Joseph P. Pickett, Dale Hoiberg,
`Dan Clancy, Peter Norvig, Jon Orwant, Steven Pinker, Martin A. Nowak, and
`Erez Lieberman Aiden; Quantitative Analysis of Culture Using Millions of
`Digitized Books. 331 SCIENCE 176 (2011) .................................................... 10
`
`MONK: Metadata Offer New Knowledge, http://www.monkproject.org/
`(last visited May 31, 2013) ........................................................................... 10
`
`Franco Moretti, Graphs, Maps, Trees: Abstract Models for
`Literary History (2005) ................................................................................... 6
`
`Toshihide Ono et al., Automated Extraction of Information on Protein–Protein
`Interactions from the Biological Literature,
`17 BIOINFORMATICS 155 (2001) ...................................................................... 5
`
`Matthew Sag, Copyright and Copy-Reliant Technology,
`103 NW. U.L. REV. 1607 (2009) ..................................................................... 3
`
`Matthew Sag, Orphan Works as Grist for the Data Mill,
`27 BERKELEY TECH. L. J. 1503 (2012) ........................................................... 3
`
`Software Environment for the Advancement of Scholarly Research (“SEASR”)
`http://seasr.org (last visited May 31, 2013) .................................................. 10
`
`Text Analysis Portal for Research (“TAPoR”), http://www.tapor.ca/portal/
`portal (last visited July 2, 2012) .................................................................... 10
`
`Tracking 18th-century “social network” through letters, STANFORD
`UNIVERSITY (Dec. 14, 2009) (video), http://www.youtube.com/watch?v=
`nw0oS-AOIPE ................................................................................................ 7
`
`
`
`
`
`viii
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page10 of 53
`
`STATEMENT OF INTEREST OF AMICI1
`Amici are over 100 professors and scholars who teach, write, and research in
`
`computer science, the digital humanities, linguistics or law, and two associations
`
`that represent Digital Humanities scholars generally.2 Amici have an interest in this
`
`case because of its potential impact on their ability to discover and understand,
`
`through automated means, the data in and relationships among textual works.
`
`Legal Scholar Amici also have an interest in the sound development of intellectual
`
`property law. Resolution of the legal issue of copying for non-expressive uses has
`
`far-reaching implications for the scope of copyright protection, a subject germane
`
`to Amici’s professional interests and one about which they have great expertise.
`
`Amici speak only to the issue of copying for non-expressive uses. A complete list
`
`of individual amici is attached as Appendix A.
`
`
`
`
`1 Pursuant to Fed. R. App. P. 29(a), (c)(4), (c)(5) and Rule 29.1 of the Local Rules
`of the United States Court of Appeals for the Second Circuit, Amici hereby state
`that none of the parties to this case nor their counsel authored this brief in whole or
`in part; no party or any party’s counsel contributed money intended to fund
`preparing or submitting the brief; and no one else other than Amici and their
`counsel contributed money that was intended to fund preparing or submitting this
`brief. Amici also hereby state that all parties have consented to the filing of this
`brief, and we rely on that consent as our source of authority to file.
`2 See Association for Computers and the Humanities, http://www.ach.org/;
`Canadian Society for Digital Humanities, http://csdh-schn.org.
`.
`
`1
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page11 of 53
`
`
`
`
`
`
`
`SUMMARY OF ARGUMENT
`
`Mass digitization, especially by libraries, is a key enabler of socially
`
`valuable computational and statistical research (often called “data mining” or “text
`
`mining”). While the practice of data mining has been used for several decades in
`
`traditional scientific disciplines such as astrophysics and in social sciences like
`
`economics, it has only recently become technologically and economically feasible
`
`within the humanities. This has led to a revolution, dubbed “Digital Humanities,”
`
`ranging across subjects like literature and linguistics to history and philosophy.
`
`New scholarly endeavors enabled by Digital Humanities advancements are still in
`
`their infancy but have enormous potential to contribute to our collective
`
`understanding of the cultural, political, and economic relationships among various
`
`collections (or corpora) of works—including copyrighted works—and with society.
`
`The Court’s ruling in this case on the legality of mass digitization could
`
`dramatically affect the future of work in the Digital Humanities.
`
`This Court should affirm the decision of the district court below that library
`
`digitization for the purpose of text mining and similar non-expressive uses present
`
`2
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page12 of 53
`
`no legally cognizable conflict with the statutory rights or interests of the copyright
`
`holders. Where, as here, the output of a database—i.e., the data it produces and
`
`displays—is noninfringing, this Court should find that the creation and operation
`
`of the database itself is likewise noninfringing. The copying required to convert
`
`paper library books into a searchable digital database is properly considered a
`
`“non-expressive use” because the works are copied for reasons unrelated to their
`
`protectable expressive qualities — the copies are intermediate and, as far as is
`
`relevant here, unread.
`
`The type of non-expressive use at issue here – statistical analysis of text – is
`
`common among copy-reliant technologies: for example, Internet search engines
`
`and plagiarism detection software do not read, understand, or enjoy copyrighted
`
`works, nor do they deliver these works directly to the public. Such platforms copy
`
`the works only incidentally, in order to process them as “grist for the mill”—raw
`
`materials that feed various algorithms and indices. See Matthew Sag, Copyright
`
`and Copy-Reliant Technology, 103 NW. U.L. REV. 1607 (2009); Matthew Sag,
`
`Orphan Works as Grist for the Data Mill, 27 BERKELEY TECH. L. J. 1503 (2012).
`
`Further, generating data about a copyrighted work (often called “metadata”)
`
`does not infringe the original work because, as has been recognized for over a
`
`century, copyright law protects only an author’s original expression, not facts. That
`
`a “fact” might pertain to or describe an expressive work does not change its factual
`
`3
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page13 of 53
`
`character—or render it an author’s exclusive intellectual property under the law.
`
`Indeed, making such factual information freely available to all is crucial to the
`
`harmony between copyright law and the First Amendment—hence the existence of
`
`rules like the “idea/expression” distinction (see 17 U.S.C. § 102(b)), the doctrine of
`
`scenes à faire, and the “merger” principle.
`
`The act of copying works into a database in order to enable the generation of
`
`metadata about those works should thus be deemed noninfringing. As numerous
`
`courts have found, making intermediate copies that enable socially beneficial
`
`noninfringing uses and/or outputs constitutes a protected “fair use” under Section
`
`107 of the Copyright Act. See, e.g., A.V. ex rel. Vanderhye v. iParadigms, LLC,
`
`562 F.3d 630, 645 (4th Cir. 2009); Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d
`
`1146, 1168 (9th Cir. 2007); Sony Computer Entm’t, Inc. v. Connectix Corp., 203
`
`F.3d 596, 609 (9th Cir. 2000); Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510,
`
`1527-28 (9th Cir. 1992). Similarly, the mass digitization of books for text-mining
`
`purposes is a form of incidental or “intermediate” copying that enables ultimately
`
`non-expressive, non-infringing, and socially beneficial uses without unduly
`
`treading on any expressive—i.e., legally cognizable—uses of the works. The Court
`
`should find such copying to be fair use.
`
`I.
`
`ARGUMENT
`
`The Freedom to Make Non-expressive Use of Copyrighted Works is
`Vital to the “Progress of Science” in the Digital Humanities
`4
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page14 of 53
`
`
`
`Where large-scale electronic text collections are available, advances in
`
`computational power and a proliferation of new text-mining and visualization tools
`
`offer scholars of the humanities the chance to do what biologists, physicists, and
`
`economists have been doing for decades—analyze massive amounts of data.
`
`
`
` “Digital Humanities” scholars fervently believe that text mining and the
`
`computational analysis of text are vital to the progress of human knowledge in the
`
`current Information Age. The potential of these non-expressive uses of text has
`
`already been revealed in the life sciences, where researchers routinely use a variety
`
`of text-mining tools to facilitate the search for relevant research across disparate
`
`fields and to uncover previously unnoticed “correlations or associations such as
`
`protein-protein interactions and gene-disease associations.” See Sophia Ananiadou
`
`et al., Text Mining and its Potential Applications in Systems Biology, 24 TRENDS IN
`
`BIOTECHNOLOGY 571, 571 (2006) (citing Toshihide Ono et al., Automated
`
`Extraction of Information on Protein–Protein Interactions from the Biological
`
`Literature, 17 BIOINFORMATICS 155 (2001) and Christian Blaschke et al.
`
`Information Extraction in Molecular Biology, 3 BRIEFINGS IN BIOINFORMATICS 154
`
`(2002)).
`
`Similar breakthroughs are on the horizon in the humanities. Traditionally,
`
`literary scholars have relied upon the close and often anecdotal study of select
`
`works. Modern computing power, advances in computational linguistics and
`5
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page15 of 53
`
`natural language processing, and the mass digitization of texts now permit
`
`investigation of the larger literary record.
`
`Digitization enhances our ability to process, mine, and ultimately better
`
`understand individual texts, the connections between texts, and the evolution of
`
`literature and language. As University of Nebraska Professor Matthew Jockers
`
`explains, by exploring the literary record writ large, researchers can better
`
`understand the context in which individual texts exist, and thereby better
`
`understand the texts themselves. See Matthew Jockers, MACROANALYSIS: DIGITAL
`
`METHODS FOR LITERARY HISTORY (2013). Along similar lines, Stanford University
`
`Professor Franco Moretti has noted that “a field this large cannot be understood by
`
`stitching together separate bits of knowledge about individual cases, because it
`
`isn’t a sum of individual cases: it’s a collective system, that should be grasped as
`
`such, as a whole . . . .” Franco Moretti, GRAPHS, MAPS, TREES: ABSTRACT MODELS
`
`FOR LITERARY HISTORY 4 (2005) (emphasis in original).
`
`Researchers working in the field of information retrieval frequently use text
`
`mining and computer-aided classification to identify and retrieve relevant
`
`documents. Using similar techniques, researchers in the Digital Humanities are
`
`able to identify and retrieve relevant texts, often from unlikely places. Humanities
`
`researchers can thereby expand their traditional study of a few canonical works to a
`
`study of several million in the larger archive of literary history—an archive that
`
`6
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page16 of 53
`
`has hitherto remained hidden because of the limitations of humans’ reading
`
`capacity. As part of this process, such non-expressive uses often leads to additional
`
`expressive uses, expanding the audience (and the potential market) for enjoyment
`
`of individual works.3
`
`Mass digitization also results in the creation of data that enables scholars to
`
`reimagine relationships between texts—for example, by linking texts with maps.
`
`Thus, Google’s “Ancient Places Project” links the text of public domain books like
`
`Gibbon’s Decline and Fall of the Roman Empire to a map of the ancient world.4
`
`The interface allows the user to browse the books, including the full text, at the
`
`same time as she browses a map. The places mentioned are marked on the map and
`
`hyperlinked.5 Similar maps could be made with reference to works still under
`
`
`3 For example, Matthew Jockers used text mining and computer aided
`classification to identify an overlooked tradition of whaling fiction predating (and
`arguably informing) Melville’s writing of Moby Dick. See Jockers, supra.
`4 See Leif Isaksen, Elton Barker, Eric C. Kansa, Kate Byrne, GAP: A NeoGeo
`Approach to Classical Resources, 45 LEONARDO 82-83 (2012).
`5 In a similar vein, researchers at Stanford University have mapped thousands of
`letters exchanged during the Enlightenment and thereby devised a theory of how
`these individual networks fit into a coherent whole, which the scholars refer to as
`the “Republic of Letters.” See Tracking 18th-century “social network” through
`letters, STANFORD UNIVERSITY (Dec. 14, 2009) (video),
`http://www.youtube.com/watch?v=nw0oS-AOIPE. Such aggregation yields
`surprising insights: for example, “the common narrative is that the Enlightenment
`started in England and spread to the rest of Europe,” but the relatively low volume
`of correspondence between London and Paris suggests otherwise. See Patricia
`7
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page17 of 53
`
`copyright—importantly, without ever making the text of the book available for free
`
`viewing. Extracting such data from texts to create these maps is a quintessential
`
`non-expressive use of the underlying texts that does not implicate any copyright-
`
`protected use—let alone infringe the copyrights of—the works in question.
`
`Google’s “Ngram” tool provides another example of a non-expressive use
`
`enabled by mass digitization—this time easily visualized. Figure 1, below, is an
`
`Ngram-generated chart that compares the frequency with which authors of texts in
`
`the Google Book Search database refer to the United States as a single entity (“is”)
`
`as opposed to a collection of individual states (“are”). As the chart illustrates, it
`
`was only in the latter half of the Nineteenth Century that the conception of the
`
`United States as a single, indivisible entity was reflected in the way a majority of
`
`writers referred to the nation. This is a trend with obvious political and historical
`
`significance, of interest to a wide range of scholars and even to the public at large.
`
`But this type of comparison is meaningful only to the extent that it uses as raw data
`
`a digitized archive of significant size and scope.6
`
`
`Cohen, Digital Keys for Unlocking the Humanities’ Riches, N.Y. TIMES, Nov. 17,
`2010, at C1.
`6 Google Ngram is available at http://books.google.com/ngrams.
`8
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page18 of 53
`
`Figure 1: Google Ngram Visualization Comparing Frequency of
`“The United States is” to “The United States are”7
`
`
`
`To be absolutely clear, 1) the data used to produce this visualization can only
`
`be collected by digitizing the entire contents of the relevant books, and 2) not a
`
`single sentence of the underlying books has been reproduced in the finished
`
`product. In other words, this type of non-expressive use only adds to our collective
`
`
`7 Figure 1 is a reconstruction of data generated using Google Ngram, sampled at
`five-year intervals. The y-axis is scaled to 1/100,000 of a percent, such that 1 =
`0.00001%.
`
`9
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page19 of 53
`
`knowledge and understanding, without in any way replacing, damaging the value
`
`of, or interfering with the market for, the original works.8
`
`Google Ngram is just the tip of the iceberg.9 In Macroanalysis: Digital
`
`Methods and Literary History, Professor Jockers draws on a corpus of Nineteenth
`
`Century novels to demonstrate how literary style changes over time. See generally
`
`Jockers, supra. Examining word frequencies, syntactic patterns, and thematic
`
`markers in the metadata-enriched context of author nationality, author gender, and
`
`time period, opens up literary study to an entirely new perspective.10 Trendsetters
`
`
`8 For additional examples of Ngram’s uses, see, e.g., Jean-Baptiste Michel, Yuan
`Kui Shen, Aviva Presser Aiden, Adrian Veres, Matthew K. Gray, The Google
`Books Team, Joseph P. Pickett, Dale Hoiberg, Dan Clancy, Peter Norvig, Jon
`Orwant, Steven Pinker, Martin A. Nowak, and Erez Lieberman Aiden;
`Quantitative Analysis of Culture Using Millions of Digitized Books. 331 SCIENCE
`176 (2011) (a study of linguistic and cultural changes in over five million digitized
`books).
`9 The toolkit available to Digital Humanities researchers is becoming increasingly
`sophisticated. See, e.g., Text Analysis Portal for Research (“TAPoR”),
`http://portal.tapor.ca/portal/portal (last visited May 21, 2013) (tools to map word
`usage over time, including peaks, density, collocations, and types); MALLET:
`MAchine Learning for LanguagE Toolkit, http://mallet.cs.umass.edu/ (last visited
`May 31, 2013) (a Java-based package for statistical natural language processing,
`document classification, clustering, topic modeling, information extraction, and
`other machine learning applications to text); MONK: Metadata Offer New
`Knowledge, http://www.monkproject.org/ (last visited May 31, 2013) (a digital
`environment designed to help humanities scholars discover and analyze patterns in
`the texts); Software Environment for the Advancement of Scholarly Research
`(“SEASR”), http://seasr.org (last visited May 31, 2013).
`10 A recently published study, led by mathematicians at Dartmouth, makes a
`similar point. See James M. Hughes et al., Quantitative Patterns of Stylistic
`10
`
`
`

`

`Case 12-4547, Document 159, 06/04/2013, 955881, Page20 of 53
`
`and outliers are revealed, as when Jockers’ text mining and computational analysis
`
`demonstrated that Harriet Beecher Stowe’s fiction is far more similar to the work
`
`of male authors of her generation than to the female-authored works of
`
`“sentimental fiction” among which her work has traditionally been categorized.11
`
`Figure 2 provides another fascinating example of Professor Jockers’ research.
`
`The chart shows the extent to which British, American, and Irish authors focused
`
`on the theme of American slavery during the Nineteenth Century, based on a
`
`corpus of 3,450 novels from that time period. Although it comes as no surprise that
`
`slavery was most often addressed by American authors, the strong Irish reaction to
`
`the American Civil War (note the spike in the light gray line beginning in 1860)
`
`compared with the decidedly muted response by British aut