Benson Snippets:
Digitized Copies of
Books from Latin
American Collection
Appear Online
by
María Elena González Marinas
I
In January 2007, the University of Texas
Google’s Goals: Online Discovery of Book Contents
Libraries signed a cooperative agreement with
Google to digitize no fewer than one million books
from its collections over the next six years. By summer 2008, Google had digitized more than 300,000
volumes from the unique holdings of the Benson
Latin American Collection and made available online digital copies of
many of the photographs, illustrations, maps, graphs, and charts as
well as the text of two to three hundred of these books. Titles from the
Benson that appeared online include such forgotten classics as Paul
Groussac’s El viaje intelectual and a rare account by Argentine Navy
officer Jose M. Sobral of his exploration of Antarctica, Dos años entre
los hielos, 1901–1903.
Through the parallel efforts of Google Library Partners in the United
States and Europe, Latin Americanists everywhere in the world will
soon have online keyword, author, title, publisher, or publication date
access to the full text of tens of thousands of rare and out-of-print
books from the finest research collections.
One of four large-scale digitization projects under way in the United
States, the Google Library Partners now includes twenty-eight research
libraries with total holdings in the hundreds of millions of books.
The combined holdings from the libraries of the original five Google
partners—Harvard, Michigan, New York Public Library, Oxford, and
Stanford—have been estimated at 58 million volumes. Of these, 15
million are planned to be digitized and online within ten years. Other
mass-digitization initiatives include the Carnegie Mellon University’s
Million Book Project, Microsoft’s Live Search Books, and the well publicized Open Content Alliance.
Google’s main goals are to make the contents of books discoverable
by as wide an audience as possible and ultimately help authors and
publishers sell more books. Google has the resources, the indexing
capacity, the search engine, and so far the incentive to digitize, instantaneously sort through, and disseminate the millions of bytes that the
imaged books are taking up on the Google servers. Since many of the
books take up an average of 20 megabytes each, the first lot of 15 million will absorb a minimum of 300 terabytes of server space. This is
about four times the bytes the Library of Congress estimates the U.S.
National Digital Information Infrastructure and Preservation Program
has processed to date.
Like all other initiatives to digitize and make gargantuan masses of
materials available online, Google is restricted by copyright law. A web
of international copyright laws protects recently published books as
well as many obscure, hard-to-find, and out-of-print books, precluding their dissemination online. To accommodate legal restrictions,
Google developed a three-tiered system for delivering desired texts
to users as Snippets, in Limited View, or in Full View. That is, when
searched in Google Book Search, texts protected by copyright appear
as phrases on a fortune cookie, separated from essential bibliographic
information including pagination; these are called Snippets. At most,
the Snippet returns a few lines of text before and after the keyword
or phrase entered by the user.
When permission is granted by the copyright owner—often the
publisher of a book—Google more liberally displays as much as
50–75 percent of the book. If a book is clearly in the public domain,
then the book appears in full text ready for downloading by the user
llilas
portal
40
as a document file. In some cases, there is
no preview of the book at all but basic bibliographic information is provided including
author, title, publisher, date of publication,
number of pages. When the holder of a text
is one of the library partners, the name of the
library is identified. Additional links in the
“About This Book” section of the Google Book
Search results guide the reader to the libraries
nearest them that also have the book in their
collections. Google directs readers interested in
purchasing books to links to publishers, online
vendors, and used book dealers.
Google earns revenues from publishers
willing to pay for each reader clicking on the
publishers’ Web sites and by selling links to
nonpublisher sites related to the book’s contents; so far, these last are few and appear
unobtrusively as footers in the Google Book
Search results page. Although revenue figures related to publisher site visits have been
impossible to tally from public records, Google
claims to have agreements with more than
10,000 publishers. As long as these agreements
are profitable, Google Book Search revenues
remain relatively secure and unprofitable clicks
to library Web sites will likely continue to be
subsidized.
Inevitable Controversy
Controversy about the Google initiative broke
out shortly after the first Library Partner pilot
projects were announced in 2004. Complaints
revolved around two key points: Google
intended not only to scan and add to their
databases the entire contents of all partner
library books—whether or not protected by
copyright—but also to index the material. Commercial publishers and university presses sued
on grounds that the scanning of copyrighted
texts as well as the proprietary indexing by
Google went beyond fair use and thus violated
the rights of copyright holders. In turn, Google
proposed an opt-out program for publishers
and authors who did not want their works digitized or publicized by Google. It should be very
clear, however, that Google never intended to
make public the full text or even parts of text
of materials protected by copyright. Although
the case has not been resolved, substantive
arguments as to whether the use of text snippets or image thumbnails falls under the fair
use provisions outlined by U.S. copyright law
have taken on new dimensions as lawsuits
unrelated to research and teaching purposes
Like all other initiatives to
digitize and make gargantuan
masses of materials available
online, Google is restricted
by copyright law.
To accommodate legal
restrictions, Google developed a three-tiered system for
delivering desired texts to
users as Snippets, in Limited
View, or in Full View.
make their way through the courts.
Fearing Google as a colossal advertising
mechanism as well as a popular search engine,
many suspect that it will likely be the enterprise to implement a yet-to-be devised business
model to exploit the scholarly texts commercially. If that turns out to be the case, no one
can determine the long-term consequences of
trusting Google to control such a vast collection of research material. Competitors are also
concerned about Google’s incomparable capital
resources and power to squelch innovation in
the mining of large digital collections. For a
specialized bibliography that traces speculation about the impactof the Google initiative
on publishing and research patterns, one may
refer to Google Book Search by former University of Houston librarian Charles W. Bailey,
Jr. See http://www.digital-scholarship.org/
gbsb/gbsb.htm
llilas
portal
41
On the other hand, anxiety about the likelihood that Google will abandon the Google
Library Partnerships before reaching its stated
goals peaked in May 2008 when Microsoft
announced that it would end similar massdigitization projects, Live Search Books and
Live Search Academic Services. By that date,
Microsoft and its partners had digitized threequarter million books and tens of millions of
scholarly journal articles. Microsoft turned
over the digital files to Ingram Digital, which
already had functioned as the host for the
Microsoft Live Search Books files. Presumably,
publishers will now have to renegotiate terms
with Ingram or opt out of the project.
Less concerned about business practices,
some scholars fear that popular mass digitization programs will reduce the motivation
of research libraries to preserve the original copies. Others question the rationale for
imaging millions of printed pages at a time that
scholars demand massive data sets they can
manipulate. To date, mapping, notation, and
data aggregation tools provided by Google have
proven inadequate for scholarly use.
individual works, with the intent of uncovering
the maximum number of titles in the public
domain. Books in the public domain may be
presented online in full view and prepared
for downloading by anyone with adequate
technology.
Library Partner Goals
Many research libraries, including the Library
of Congress, have undertaken ambitious book
digitization programs but not on this scale.
Given the limited resources and lack of access
to the necessary technology, individual research
libraries have not been able to digitize more
than a few thousand of their books, confirming
librarians’ estimates that digitization of their
holdings would take generations to complete.
Google’s proposal to digitize and service the
digital archive of their collections at no cost
to the research libraries not only accelerated
the rate at which the digitization processes can
take place, but it promises to quickly multiply
the number of printed books to be digitized
from the thousands to the millions.
Alliance with Google assures research
libraries that their books will be more easily
discovered by users worldwide and their collections made known beyond the walls of an
institution. These outcomes advance the mission
of universities to disseminate knowledge and
fulfill the mandate of libraries to expand and
diversify access services. Librarians are beginning to accept digital copies as an important
means of preserving their collections much as
they once did microfilm reproductions of books
and manuscripts so that partnering with Google
has been repeatedly justified as an affordable
long-term preservation strategy. The creation
of digital copies for preservation is defensible
as a fair use and is sensible as a solution to
replacement for lost and fragile items.
Several research libraries, such as those
at the University of Michigan and Stanford
University, also have leveraged these largescale projects to develop user-oriented search
applications and improve the integration of
disparate digital collections. In connection with
the Google Library Partnership, the University
of Texas Libraries staff is testing the means to
link the corresponding digitized texts to the
main online catalog author entries.
In addition, research staff began an investigation of copyright laws in various Latin
American countries and joined colleagues at
the University of Michigan in refining procedures for determining the copyright status of
Copyright Laws in Latin America
The global distribution of potential readers
and authors complicates the determination
of the copyright status of individual works
disseminated online. Although the language
of current copyright laws flows from that of
the Berne Convention and is harmonized
through multilateral treaties among countries, definitions and duration of copyright
protection vary widely from country to country.
The laws of individual countries distinguish
authored works, anonymous and pseudonymous works, edited works, and translations
as well as works-for-hire, among others, while
copyright protection terms vary from fifty to
one hundred years, sometimes calculated from
the date of publication for compiled works or
from the year of death for authored works. In
the case of multiple authors, copyright protection is calculated from the year of death of the
last living author.
The classification of creative works, the
identification of authors, the accurate determination of author death, and the correct
calculation of the duration of copyright are
complex, time-consuming, and painstaking
tasks. They are also risky, as errors could prove
embarrassing, if not expensive. Consequently,
the rules for determining the copyright status
of large numbers of books must be extremely
conservative. That is the case with Google’s
algorithms.
According to the U.S. law, books published
anywhere before 1923 are no longer protected
in the United States and are categorically in
the public domain. After that date, a series of
conditions related to copyright formalities,
trade agreements, treaties, and court decisions
must be conjugated in order to arrive at the
copyright status of a work. Ironically, at different times U.S. law has provided copyright
protection beyond the term the works enjoyed
in the country of origin. Many of the books
in the Benson Latin American Collection are
works that had been in the public domain in
the country of origin but were republished
with a new foreword, annotations, or similar
editorial additions, thus making it difficult to
llilas
portal
42
separate the original text for dissemination
throughout the digital public domain. Edited
and republished works make up a substantial
portion of research libraries like the Benson
Latin American Collection, which increased
book purchases and exchanges during the
1960s. All of the republished work will be
protected by copyright law for many more
years and will continue to appear to the Google
Book Searcher as Snippets.
Creative Uses of Google Book Search
Although small-scale and mass digitization
projects have been recognized as a boon to
classicists, Medieval, Renaissance, and Early
Modern scholars in general, Latin Americanists
have not had enough time to use the newly
available materials and to publicize their experiences accessing and utilizing mass digital
libraries. Recent postings on discussion lists
and formal publications by historians, print
history, and book culture scholars in general
have been mostly negative, creating additional
doubt about the value of the endeavor. The
more positive comments admit to the usefulness of mechanisms that allow user-chosen
searches over millions of books and recognize
Google Book Search as an effective finding
aid for pinpointing the location of arcane
and regional materials currently scattered
worldwide without having to consult foreign
language union catalogs.
Because the books that are being digitized
come off the shelves as raw inventory from
dozens of research libraries, few attempts have
been made to select or curate subcollections
in relation to the interest or research needs of
faculty and students. Indeed, the individual
researcher must carefully select from a huge
mass of digitized titles for their personalized collections. It will be up to researchers
to prompt their home institutions to assist
in those searches and to provide the tools to
restructure the contents within the digitized
books so that they can be analyzed and manipulated as a mass—of snippets.
Explore Google Books Search, http://books.
google.com/advanced_book_search and have
a say on what will be done next!
María Elena González Marinas is Assistant
Professor in the Library and Information Science
Program at Wayne State University. She received
her Ph.D. from the School of Information at the
University of Texas at Austin. ✹
