Working Paper Series
Villanova University Charles Widger School of Law Year 2005
Creative Commons and the New
Michael W. Carroll
Villanova University School of Law, firstname.lastname@example.org
This paper is posted at Villanova University Charles Widger School of Law Digital Repository.
CREATIVE COMMONS AND THE NEW INTERMEDIARIES
MICHAEL W. CARROLL*
In many early conversations about the Internet, the story that
dominated was one of liberation from the intermediaries. Record companies,
retailers of all stripes, and the mainstream media all were dinosaurs whose
days were numbered. The Internet’s end-to-end architecture enabled end-to-
end commerce, end-to-end culture, and end-to-end news. Even the new
intermediaries, like Internet service providers (ISPs), merely supplied
infrastructure because end-to-end architecture greatly limited the kinds of
control ISPs might try to assert.
After the revolutionary euphoria died down, however, many
acknowledged that intermediaries are necessary to all kinds of transactions in
commerce, culture, and news. Reintermediation soon follows from
disintermediation, and the real question the Internet posed was not whether
intermediaries are necessary but what kinds of intermediaries are necessary.
When contemplating this question now, fifteen years after the invention of
the World Wide Web, I want to highlight the disintermediating and
reintermediating roles that Creative Commons licenses currently play on the
Web and to also suggest that these licenses deserve lawyers’ attention as a
species of machine-readable law.1
Creative Commons licenses respond to the explosion of “copyright
events” that digital technologies has let loose.2 Explosions usually have
violent consequences. The copyright explosion certainly has disrupted a
* Associate Professor of Law, Villanova University School of Law. Thanks to Barry
Brewer for excellent research assistance. [Disclosure: I serve on the Board of Directors of
Creative Commons, Inc. The views expressed herein are mine alone and do not necessarily
reflect the views of Creative Commons or those associated with it.] Information current as
of July 1, 2005. This Article is licensed under a Creative Commons Attribution-Non
Commercial 2.5 License, at http://creativecommons.org/licenses/by-nc/2.5/legalcode.
Attribution should be to Michael W. Carroll as author and to the Michigan State Law Review
as original publisher.
1 See generally Creative Commons available at www.creativecommons.org.
2 I use the term “copyright events” to explain why the scope of copyright law’s domain has
expanded so dramatically with the growth of digital technology. A “copyright event” is any
action in the world that entails the exercise of one or more of a copyright owner’s exclusive
rights to copy, distribute, perform, display or adapt information. Some copyright events are
infringing and others are not. All implicate copyright law. The courts’ response to digital
technologies which require copying to function has been to permit copyright law to infiltrate
almost every digital interaction. See, e.g., MAI v. Peak, 991 F.2d 511 (9th Cir. 1993) (holding
that every copy written to the Random Access Memory of a computer is a copy for the
purposes of the Copyright Act).
number of industries and relationships that rely on copyright law. What is
perhaps more interesting is how this radical expansion of copyright law’s
domain has not led to chaos. Instead, a number of implicit understandings
have grown up around digital technologies, and these understandings have
led to norms and implied licenses that serve important coordinating
functions. As robust as these informal mechanisms are, however, greater
clarity and coordination can often be had when copyright owners explicitly
designate which copyright events they consider to be permissible. Enter
Creative Commons licenses.
Moreover, the proliferation of Creative Commons licenses on the
Web points up a new relevance dimension – the copyright status of
information found on digital networks.3 Imagine that you are an independent
filmmaker in need of some music to accompany a montage in your film. You
have no time or budget to clear the rights to the music. If you search for
“Chopin,” what is relevant is not simply whether there is information – such
as a music file – that is accurately associated with your term, but also whether
that information is available to you on terms that permit your desired use.4
Recently, the creation of a Creative Commons search engine, followed by
Yahoo!’s offering of a specialized Creative Commons search, enables
searching along both the topical and copyright dimensions.
Creative Commons licenses act as a disintermediating force because
they enable end-to-end transactions in copyrighted works. The licenses have
reintermediating force by enabling new services and new online communities
to form around content licensed under a Creative Commons license.
Intermediaries focused on the copyright dimension have begun to appear
online as search engines, archives, libraries, publishers, community
organizers, and educators. Moreover, the growth of machine-readable
copyright licenses and the new intermediaries that they enable is part of a
larger movement toward a Semantic Web. As that effort progresses, we
should expect new kinds of intermediaries that rely on machine-readable law
II. CREATIVE COMMONS LICENSES AS INTERMEDIARIES
A Creative Commons license is a form copyright license that can be
linked to via the Web. In addition to the legal code, the license is described
by a “human-readable” Commons Deed, which identifies the key terms of
3 The idea of “relevance dimensions” is familiar to many who undertake quantitative study.
See, e.g. INEX Relevance Assessment Guide available at
4 Of course, there also are many others searching the Net for music files for personal use
who consider the copyright status of the works to be irrelevant.
the license and machine-readable metadata that associates the Internet
location of the licensed resource with the Internet location of the license
document. As of this writing, there are nearly 16,000,000 digital objects
accessible over the Internet linked to a Creative Commons license.5 These
resources include scientific journal articles, music files, picture files, and
Creative Commons licenses permit certain royalty-free uses of the
licensed copyrighted work. The most permissive license permits all uses so
long as the copyright owner’s directions concerning attribution are followed.
Other conditions include a requirement that derivative works be licensed
under the same terms, a limitation to non-commercial uses, and a prohibition
on the creation of derivative works. These can be combined to create six
permutations.6 There are also some tailored licenses that respond to requests
from particular communities. Musicians asked for a “sampling” license that
permits commercial uses involving creation of derivative works through
digital sampling. The sampling license comes in three flavors.7 The
Developing Nations license differentiates permission by geography, granting
an Attribution license for uses in developing nations while reserving default
copyright protection for uses in developed nations.8 Creative Commons has
also coupled its metadata with the pre-existing legal code from the Free
Software Foundation for the use of creators of software who wish to license
their creations under the GNU General Public License (GPL) or Lesser
General Public License (LGPL).9 In addition to these licenses, Creative
5 See Mike Linksvayer, CC in Yahoo! Advanced Search, Creative Commons:About: Weblog, at
http://creativecommons.org/weblog/entry/5456 (May 27, 2005).
6 See Creative Commons: About: Choosing a license, at
http://creativecommons.org/about/licenses (listing the various CC licenses) (last visited
June 9, 2005). They are: "Attribution," "Attribution-NoDerivs"(No Derivative Works),
"Attribution-NonCommercial-NoDerivs," "Attribution-NonCommercial," "Attribution-
NonCommercial-ShareAlike," and "Attribution-ShareAlike." See Creative Commons:
Creative Commons Licenses, at http://creativecommons.org/licenses (last visited June 9,
2005) (replacing the original four licenses, which could be combined into 11 permutations).
7 See Creative Commons: Publish: Choose Your Sampling License Options, at
http://creativecommons.org/license/sampling (last visited June 9, 2005) (including
"Sampling,"which allows sampling for non-advertising purposes, "Sampling Plus," which is
the same, but allows non-commercial copying of the entire work as well, and "Non-
Commercial Sampling Plus," which allows only non-commercial sampling and copying).
8 See Creative Commons: Commons Deed: Developing Nations 2.0, at
http://creativecommons.org/licenses/devnations/2.0 (last visited June 9, 2005).
9 See Creative Commons: Publish: Creative Commons GNU GPL, at
http://creativecommons.org/license/cc-gpl?lang=en (last visited June 9, 2005), Creative
Commons: Publish: Creative Commons GNU LGPL, at
http://creativecommons.org/license/cc-lgpl?lang=en (last visited June 9, 2005).
Commons offers a service through which copyright owners can dedicate
their works to the public domain.10
Creative Commons licenses facilitate cheap speech. For example, a
teacher who wishes to find materials to copy for a course pack can see
immediately that she can use content licensed under an Attribution license
without asking for permission. In addition, by using Creative Commons
licenses, millions of bloggers ensure that “news reader” programs may copy
their respective RSS feeds and compile them into a derivative works. These
speech transactions are made faster and cheaper by simple, machine-readable
Moreover, all of these licensed objects will function as a common
pool. There will be new functions to be performed, similar to traditional
functions related to traditional creative works, but within the context of the
freedoms associated with digital objects licensed under Creative Commons
licenses. In addition, Creative Commons licenses can be complemented by
new licensing intermediaries who can facilitate transactions with respect to
the rights reserved to the copyright owner under a Creative Commons
license. For example, if a user finds a work licensed under a
NonCommercial license, he or she can negotiate with the copyright owner
for permission to use the content for commercial purposes. Existing
intermediaries, such as the American Society of Composers, Artists and
Publishers (ASCAP),11 Broadcast Music, Incorporated (BMI)12 and the Harry
Fox Agency13 in the music industry, for example, or new intermediaries may
emerge to broker such negotiations.
III. NEW INTERMEDIARIES ENABLED BY CC LICENSES
The intermediaries enabled by Creative Commons licenses include
search engines with added relevance dimensions, archives and libraries that
include content tagged with CC licenses, new producers and publishers who
facilitate uses of “some rights reserved” material made possible by Creative
Commons, communities of Creative Commons creators, and even
educational institutions. This section highlights primarily U.S.-based new
intermediaries, but it is important to note that a whole range of such
intermediaries also are emerging internationally.
10 See Creative Commons: Public Domain Dedication, at
http://creativecommons.org/licenses/publicdomain (last visited June 9, 2005).
11 See ASCAP, at http://www.ascap.com (last visited June 9, 2005) (licensing music for
12 See BMI, at http://www.bmi.com (June 9, 2005) (licensing music for public
13 See HFA, at http://www.harryfox.com (last visited June 9, 2005)(licensing copyrighted
music for mechanical purposes such as Compact Discs).
One of the earliest “reintermediaries” were the search engines. As
the amount of information on the Net and on the Web continued to grow,
connecting people with information they desired became increasingly
difficult. The race was on to produce results that were most relevant to the
terms used in a searcher’s query. For the time being, Google’s PageRank
algorithm dominates along this dimension.
Most people measure relevance along more than one dimension,
however, and the next stage in the search race will be to deliver
multidimensional results. Creative Commons uses RDF for its metadata.14
Potentially that metadata could be read by search engines to yield results that
respond to results with both topical and copyright relevance. Recognizing the
importance of finding licensed content, Creative Commons developed its
own search engine.15 The Firefox web browser now provides a toolbar link to
Searching along the copyright dimension took a giant forward stride
on May 23, 2005, when Yahoo! released the beta version of the Yahoo!
Search for Creative Commons.17 Searching Yahoo!’s far more
comprehensive database, the search engine finds sites that have a Creative
Commons license.18 The site allows a searcher to choose among four
14 See Creative Commons: Frequently Asked Questions, at
http://creativecommons.org/faq#faq_entry_3329 (last visited June 9, 2005).
15 Creative Commons Search, at http://search.creativecommons.org/index.jsp (last visited
June 9, 2005). Creative Commons Executive Director Neeru Paharia and [who helped?]
deserve credit for this advance.
16 For those who use Firefox, the upper right corner defaults to a Google toolbar, but it is a
pull-down menu that permits use of other search engines, including those provided by
Yahoo!, Amazon, and Creative Commons.
17 Yahoo! Search Blog, Larry Lessig on Searching Creative Commons (Mar 23, 2005), at
18 Yahoo! Search: Creative Commons Search Beta, at http://search.yahoo.com/cc (last
visited June 9, 2005). The Yahoo! search greatly increases the number of sites found with
Creative Commons licenses. To make the copyright dimension visible the following test was
run. Using the keywords “Eiffel Tower,” a standard Yahoo! search yielded 1,970,000 results.
See Yahoo! Search, at http://search.yahoo.com (last visited June 9, 2005). A test comparison
between the Creative Commons engine and the Yahoo! Creative Commons Search yielded
the following results: (a) with no restrictions other than a search for Creative Commons
licensed content, Yahoo! produced 3,430 results and CC produced 32; (b) with the "Find
content I can use for commercial purposes" option selected, Yahoo! produced 510 results
and CC produced 1; (c) with the "Find content I can modify, adapt, or build upon" option
selected, Yahoo! produced 2,250 results and CC produced 4; (d) with both options selected,
Yahoo! produced 375 results and CC produced 1.
As with all new technologies, there is room for improvement. Web site owners
criteria.19 The searcher can type in keywords to find any topically relevant
Creative Commons licensed content, or the searcher can specify, “Find
content I can use for commercial purposes,” or “Find content I can modify,
adapt, or build upon,” or both.20 This search works by adding a parameter
for the Creative Commons license to the standard Yahoo! search.21 The
copyright relevance dimension has gone mainstream.22
B. Archives and Libraries
Traditionally, libraries have performed at least four basic functions.
They collect and preserve information, disseminate information, index that
information by creating and maintaining metadata about their collections in
their card catalogs, and they enable searching of the index of metadata.23 In
the United States, copyright law traditionally facilitated libraries’ performance
of these functions.24 Copyright in digital works is less hospitable to these
traditional practices.25 Creative Commons licenses facilitate a rebalancing
that frees libraries to better perform their traditional roles as well as new ones
called for by the digital environment.
Among online librarians seeking to perform these roles, Brewster
Kahle stands out as a visionary.26 Recognizing early on that the malleability
sometimes tag a web page with a Creative Commons license but do not also license audio or
video files available through the site with a Creative Commons license. In such a case, the
search engines will identify the site as relevant even though the content the searcher wants is
not available under a Creative Commons license.
19 See Id.
21 See Yahoo!: Developer Network Beta, at
http://developer.yahoo.net/web/V1/webSearch.html (last visited June 9, 2005).
22 Yahoo! Search Blog, Larry Lessig on Searching Creative Commons (Mar 23, 2005), at
23 Libraries also perform a latent authentication function. We do not think about
authentication in a physical library. Generally, we assume that when a library has a book on
the shelves called “Oliver Twist” by Charles Dickens, it really is that book. Manipulating a
physical book is not easily done, and to the extent that there are multiple editions of this
book, the differences among them are readily discernible. With digital objects, however,
there are usually many versions and digital objects are easily manipulated. Authentication
now emerges as a potential function for an online library. Online libraries will have to
decide what will be archived, and which, if any, of the many manipulations or versions is
24 See, e.g., 17 U.S.C. § 109 (limiting exclusive right of distribution to not include lending of
legally-acquired copy of a copyrighted work).
25 See, e.g., LAWRENCE LESSIG, FREE CULTURE: HOW BIG MEDIA USES TECHNOLOGY AND
THE LAW TO LOCK DOWN CULTURE AND CONTROL CREATIVITY 226-28 (2004)
26 See Lessig, supra note XX, at 110-15 (describing Kahle’s vision for comprehensive online
of content on the Web presented an immediate challenge for preservation, he
created the Internet Archive,27 a non-profit organization that built and
maintains an Internet Library.28 The site provides access to historical
material in digital format.29 The Internet Archive stores texts, audio, moving
images, and software as well as archived web pages.30
The Internet Archive has agreed to host content tagged with a Creative
Commons license. Responding to this generosity, Creative Commons has
written an easy-to-use piece of software, CC Publisher, which uses a drag-
and-drop method for tagging content with a Creative Commons license and
publishing the content to the Internet Archive.31 This combination of a
software tool for tagging and uploading content with a central repository for
that content serves the intermediary function of enabling creators and users
to more easily share. Creative Commons licensed content on the Internet
Archive also appears in the results of a Yahoo! Creative Commons search.
C. Producers & Publishers
One large and important role for Creative Commons licenses is to
facilitate amateur-to-amateur communication.32 However, Creative
Commons licenses also enable new intermediaries to create new business
models for the distribution of creative works created by professional authors.
For example, Magnatune, an on-line record label, was created to distribute
music over the Internet and eliminate the problems inherent with traditional
recording contracts.33 Its business model is to target Internet radio listeners
and “fans of music that gets little radio airplay or major record distribution,
but has a fairly large audience.”34
Magnatune is a new intermediary that incorporates Creative
27 Internet Archive: Universal access to human knowledge, at http://www.archive.org (last
visited June 9, 2005).
28 See About the Internet Archive, at http://www.archive.org/about/about.php (last visited
June 9, 2005).
29 Id.; see also Kahle v. Ashcroft, 2004 WL 2663157 *2 (N.D. Cal. Nov. 19, 2004) (describing
Internet Archive’s collection).
31 See CC Publisher at http://creativecommons.org/tools/ccpublisher (last visited June 9,
32 See generally Dan Hunter & Greg Lastowka, Amateur-to-Amateur, 46 WM. & MARY L. REV.
__ (2004) (describing growth of amateur-to-amateur communication online and the
obstacles that copyright law imposes on such communication).
33See John Buckman Why I created Magnatune, at http://www.magnatune.com/info/why (last
visited June 9, 2005).
34See The Business Model: how we (and our artists) pay the rent., at
http://www.magnatune.com/info/model (last visited June 9, 2005).
Commons licenses into a profit-driven business model. To market is music,
Magnatune provides free radio stations that allow listeners to preview music
from many different genres.35 Royalty-free downloads are available under a
Creative Commons Attribution Non Commercial ShareAlike license.36 If
listeners like what they hear, they can pay for downloadable albums or
physical CDs.37 The listener chooses what to pay, between $5 and $18 per
album.38 Purchasers can make non-commercial derivative works based on
the works they purchase.39 Magnatune also licenses music for commercial
purposes.40 The contract and the price are set by the type of use,41 and the
process is completely automated.42 There is also no review of the use of the
The benefits of Magnatune for musicians are that the label splits
revenue between itself and the artist on a 50/50 basis, which is much higher
for the artist than a traditional record label.44 Magnatune, unlike other online
music sources, does not accept everyone.45 It evaluates the artists like a
traditional record label to maintain quality control.46 According to
Magnatune, “top artists make several thousand dollars per year.”47 With non-
major artists on traditional labels, often no money is made by the artist.48
Creative Commons licensing is also being used in the publication of
35 See Id.
36 See What is "Open Music"?, at http://magnatune.com/info/openmusic (last visited June
37 See Id.
38 See Id.
39 See Id.
40 See The Business Model: how we (and our artists) pay the rent., at
http://www.magnatune.com/info/model (last visited June 9, 2005).
41 See Id. (stating that the standard practice is for wealthier companies to be charged more for
42 See Id.
43 See Id.
44 See Magnatune: Information: What's in it for musicians, at
http://www.magnatune.com/info/musicians (last visited June 9, 2005). See also Magnatune:
Information: Distribution contract terms, at http://www.magnatune.com/info/terms (last
visited June 9, 2005) (Magnatune gives artists 50% of the gross on music downloads and
licensing, but due to production costs, artists get 50% of the profits on physical items like T-
shirts, posters, etc.)
45See Magnatune: Information: The Plan: problems with the music industry and how
Magnatune is trying to fix them., at http://www.magnatune.com/info/plan (last visited June
46 See Id.
47 See Id.
48 See Magnatune: Information: What's in it for musicians, at
http://www.magnatune.com/info/musicians (last visited June 9, 2005).
scientific research.49 Two publishers, the Public Library of Science
(hereinafter “PLoS”) and Biomedcentral, use Creative Commons licenses to
facilitate their respective missions to make the world’s scientific and medical
literature a public resource.50 The vision is to give unlimited access to the
latest scientific research, make it possible to search the text of every article to
locate specific ideas, methods, experimental results, and observations,51 and
to facilitate innovative ways to explore and use the world's treasury of
scientific ideas and discoveries.52
The journals published by both groups are peer-reviewed and feature
established, well-regarded editorial boards. Rather than assign copyright to
the publisher, authors grant the public a Creative Commons Attribution
license, which enables these publishers to post articles on the public Web
immediately upon publication. The immediate availability of this research
has had noticeable effects. For example, PLoS began publishing two
Journals, PLoS Biology and PLoS Medicine, and added PLoS Computational
Biology in June 2005.53 ISI Thomson, which assigns “impact factors” to
scholarly journals based on the quantity and quality of citations received,
assigned PLoS Biology an impact factor of 13.9, after only one year of
publication.54 BioMed Central’s journals also have received impact factors
that compare favorably with competing subscription-based journals,
particularly in light of how young these journals are.55 These “open access”
publishers can use Creative Commons licenses in this way because they rely
primarily on supply-side funding rather than the traditional demand-side
funding through paid subscriptions.56
These two new business models show how Creative Commons can
facilitate changes in the way we obtain both entertainment and important
information. With Magnatune, the Creative Commons license helps listeners
49 See Public Library of Science, at http://www.plos.org/index.html (last visited June 9,
50 See About PLoS, at http://www.plos.org/about/index.html (last visited June 9, 2005);
What is BioMed Central? at http://www.biomedcentral.com/info/ (last visited Jun. 9, 2005).
51 See id.
52 See id.
53 See About the PLoS Journals, at http://www.plos.org/journals/index.html (last visited
June 9, 2005).
54 See The First Impact Factor for PLoS Biology—13.9, at
http://www.plos.org/news/announce_pbioif.html#note. PLoS is launching two more
journals in the near future. See PLoS Genetics, at http://www.plosgenetics.org (last visited
June 9, 2005) (launching July 2005), PLoS Pathogens, at http://www.plospathogens.org (last
visited June 9, 2005) (launching September 2005).
55 See Frequently Asked Questions, at
56 See PLoS Publishing Model, at http://www.plos.org/journals/model.html (last visited
June 9, 2005).
and licensors find high-quality music that may not have mass appeal while
creating revenue streams for artists who would have difficulty earning
revenues under a traditional recording contract and would not be likely to
reach as broad an audience. Open access publishers embrace the public
goods nature of valuable information and use Creative Commons licenses in
conjunction with a new financing model to make use of the Internet’s
D. Creative Commons Communities
In some cases new intermediaries have adopted Creative Commons
licenses as community norms. In other words, sharing is not just allowed, it
is the point. A sampling of these includes the following:
Music. Opsound is an Internet record label,57 but unlike Magnatune,
this site contains an “open pool” into which all artists are invited to
contribute music.58 Opsound describes itself as “a kind of laboratory for
looking at how artists can release music in a manner synergistic with the
internet's capacity to encourage communication and sharing.”59 The site also
describes itself as creating a “gift economy” among musicians.60 The only
requirement for adding music to the open pool is that the artist use the
Creative Commons Attribution ShareAlike license, or place the music in the
Creative Commons also launched its own musical sharing site, CC
Mixter, with the help of a number of volunteers.62 The site invites users to
“sample, mash, & share music ... legally.”63 Everything on the site is licensed
with a Creative Commons license.64 Creators can sample and alter the music
they find on the site to create their own works.65 The only requirement is
that the artist abides by the CC license used by the source artist.66 CC Mixter
also hosts contests in which artists can obtain material from the site and
submit their creations.67 In fact, CC Mixter recently held a contest in
conjunction with Magnatune, the winner to receive a Magnatune contract.68
57 See Opsound, at http://www.opsound.org/opsound/about.html (last visited June 9, 2005).
58 See Id.
59 See Id.
60 See Id.
61 See Id.
62 See CC Mixter, at http://ccmixter.org (last visited June 9, 2005).
63 See Id.
64 See Id.
65 See Frequently Asked Questions, at http://ccmixter.org/isitlegal (last visited June 9, 2005).
66 See Id.
67 See CC Mixter, at http://ccmixter.org (last visited June 9, 2005).
68 See Id.
Visual Art. A similar community is the Open Clip Art Library,
which contains more than 3,400 clips contributed by more than 200 artists.69
The Open Clip Art Library “aims to create an archive of clip art that can be
used for free for any use.”70 It requires that all pieces of clip art submitted be
placed into the public domain using the Creative Commons statement.71 The
clip art is then available to anyone to use for any purpose. The site’s clip
collection has been accumulated in little more than one year.72 As the archive
grows, this site will be a particularly useful resource for non-professionals
with small budgets who create things like newsletters or promotional flyers.
It also allows those who create small pieces of graphic artwork to disseminate
their creations to the public.
Photographs. One of the fastest-growing communities that use
Creative Commons licenses is flickr.73 Flickr is a website that allows
members to show photos either to everyone only to select friends and
family.74 Flickr enables but does not require users to post photographs under
a Creative Commons license. Nonetheless, as of June 1, 2005, there were
more than two million photographs hosted by flickr under a Creative
Commons license.75 Flickr serves as an intermediary both for those who
only wish to view photos and for those who wish to use photos for their
own creative works. The relevance dimensions added by the various
searchable and browsable Creative Commons-licensed photos makes flickr a
significant resource for creators who seek to share the works of others.
Blogs. Creative Commons is also an important part of the
Blogosphere. Technorati, an online weblog (“blog”) search engine, describes
the blogosphere as a conversation in which millions of people express their
ideas and millions respond to them.76 About 32 million Americans are
regular blog readers.77 There are 38,000 new blogs a day, and roughly
500,000 posts daily.78 A Yahoo! Creative Commons search for the term
69 See Open Clip Art Library, at http://www.openclipart.org (last visited June 9, 2005).
70 See Id.
71 See Id.
72 See News, at http://www.openclipart.org, (announcing their one year aniversary as of
April 1, 2005).
73 flickr, at www.flickr.com (last visited June 9, 2005); see also supra notes XX and
74 See What is flickr?, at http://www.flickr.com/learn_more.gne (last visited June 9, 2005).
75 See Id.(containing171,172 Attribution licensed photos, 58,567 Attribution-NoDerivs
licensed photos, 582,019 Attribution-NonCommercial-NoDerivs licensed photos, 302,800
Attribution-NonCommercial licensed photos, 751,548 Attribution-NonCommercial-
ShareAlike licensed photos, and 159,224 Attribution-ShareAlike licensed photos).
76 See About Technorati, at http://www.technorati.com/about (last visited June 9, 2005).
77 See Id. (citing the Pew Internet Study: The state of blogging, available at
"weblog" returns 3,230,000 hits, and a search for "blog" returns 5,960,000.79
Creative Commons licenses facilitate the conversation. Since the
blogosphere includes not just separate blogs, but blogs that respond to, cite
and quote other blogs, the Creative Commons licenses allow bloggers to
build the community conversation with the legal convenience provided by
those licenses. Considering the large number of blogs searchable through the
Yahoo! Creative Commons search, this facilitation may be one of the most
powerful uses of the Creative Commons licenses.
Last but not least, an important intermediary function facilitated by
Creative Commons licenses is in the field of education. Colleges and
universities serve the societal intermediary function of disseminating
knowledge to members of society, preparing them for a productive life in
their chosen field and enriching their lives. Creative Commons licenses
enable institutions to disseminate information to an audience beyond the
university community while retaining some control over copyrighted works.
At a time when numerous institutions of higher education looked at
teaching materials produced on campus as a potential revenue source
through distance education, the Massachusetts Institute of Technology
(“MIT”) launched OpenCourseWare, “a free and open educational resource
for faculty, students, and self-learners around the world.”80 The program
contains 900 courses81 from thirty-four departments.82 MIT will evaluate this
experiment over the next five years, measuring its access, use and impact.83
MIT uses a Creative Commons license for nearly all of its content.84 The
license has enabled people from all over the world, who have Internet access,
to obtain, informally, many of the benefits of an MIT education. MIT not
only hopes to spread its educational material but also to promote the concept
79 See http://search.yahoo.com/cc, (all hits will not be separate, distinct weblogs, but it is
indicative of the importance and popularity of the form).
80 MITOPENCOURSEWARE, at http://ocw.mit.edu/index.html (last visited June 9, 2005).
81 See MITOPENCOURSEWARE: About OCW, at
http://ocw.mit.edu/OcwWeb/Global/AboutOCW/about-ocw.htm (last visited June 9,
2005) (stating that there were 900 courses as of September, 2004).
82 See MITOPENCOURSEWARE: Master Course List, at
http://ocw.mit.edu/OcwWeb/Global/all-courses.htm (last visited June 9, 2005).
83 See MITOPENCOURSEWARE: About OCW, at
http://ocw.mit.edu/OcwWeb/Global/AboutOCW/about-ocw.htm (last visited June 9,
84 See MITOPENCOURSEWARE: Legal Notices, at
http://ocw.mit.edu/OcwWeb/Global/terms-of-use.htm (last visited June 9, 2005) (noting
exceptions for some licensed 3d party material included in some courses).
of opencourseware in general.85 At least ten other universities from the
United States, Japan, and Viet Nam have launched opencourseware
programs, indicating the concept’s attraction.86
Connexions from Rice University represents another educational use
of Creative Commons licenses.87 Connexions disaggregates learning
materials by using small “chunks” known as “modules” as the basic unit of
course material.88 These modules can be organized and linked into courses.89
Learning need not be linear, and the use of modules can show “relationships
both within and between topics,” and show that “knowledge is naturally
interconnected.”90 The goal of Connexions is to create a commons of high-
quality diverse content through grassroots collaboration,91 facilitated by use
of a Creative Commons Attribution license.92 According to the site, “[m]ore
than one million people from 157 countries are tapping into over 2,400
modules and 90 courses developed by a worldwide community of authors in
fields ranging from computer science to music and from mathematics to
biodiversity.”93 Because of the open nature of Connexions, quality control is
handled by allowing third parties to review the content, presented in the
form of “lenses” that include ratings based on popularity, feedback by
universities and other reliable sources, and peer assessments.94 The modules
are also being translated into several languages.95 This shows how Creative
Commons facilitates not only dissemination, but also collaboration and
community building in the educational context.
Finally, Berklee Shares is a collection of music lessons prepared by
the faculty of the Berklee College of Music licensed under Creative
85 See MITOPENCOURSEWARE: About OCW, at
http://ocw.mit.edu/OcwWeb/Global/AboutOCW/about-ocw.htm (last visited June 9,
86 See MITOPENCOURSEWARE: Other Opencourseware Programs, at
http://ocw.mit.edu/OcwWeb/Global/AboutOCW/otherocws.htm (last visited June 9,
2005) (including Johns Hopkins University School of Public Health, the Fulbright
Economics Teaching Program in Viet Nam, and the Tokyo Institute of Technology).
87 See Connexions, at http://cnx.rice.edu (last visited June 9, 2005).
88 See Id.
89 See Connecxions: Philosophy, at http://cnx.rice.edu/aboutus/philosophy (last visited June
90 See Id.
91 See Connexions Web Tour: Vision, at http://cnx.rice.edu/aboutus/tour/8.html (last
visited June 9, 2005).
92 See Connections Content Commons, at http://cnx.rice.edu/content/browse_content (last
visited June 9, 2005) (linking to Creative Commons Deed Attribution 2.0, at
93 Connexions at http://cnx.rice.edu/.
94 See Connexions: Tour: Quality, at http://cnx.rice.edu/aboutus/tour/10.html (last visited
June 9, 2005).
95 See Id.
Commons licenses.96 The goal here is to provide free music lessons for the
musical community around the world and to promote the Berklee College of
Music.97 While not as broad in scope as MIT OpenCourseWare or
Connexions, the Berklee use stems from the same philosophy that learning
should be more widely available. It also demonstrates the potential to use
content offered under a Creative Commons license for promotional
purposes. Berklee Shares specifically states that one of its reasons for
making its content available is “to reach interested students and make them
aware of the possibility and potential of a Berklee education.”98
* * * * *
The rapid adoption of Creative Commons licenses by individual
copyright owners and by a variety of new intermediaries demonstrates the
utility of standardized understandings that enable some sharing of
copyrighted works while reserving other rights to the copyright owner. To
date, this utility has been derived primarily from the simplicity of the human-
readable Commons Deed and associated icons, which quickly communicate
the essential permissions and restrictions for each Creative Commons license.
Soon, however, chances are that the machine-readable description of these
licenses is likely to become paramount as efforts to build a Semantic Web
IV. THE SEMANTIC WEB
Frustrated by technological inabilities to share documents across
computing platforms, Tim Berners-Lee invented the hypertext mark-up
language (HTML) and other protocols that are the foundation for the World
Wide Web.99 Having achieved document interoperability, Berners-Lee and
his colleagues at the World Wide Web consortium (W3C) share a vision of a
next-generation Web that takes interoperability to a higher level, a Web in
which machines mine mountains of metadata in order to automate a wide
variety of transactions. They call this the Semantic Web.
The idea is to add logic to the Web, meaning to use “rules to make
inferences, choose courses of action and answer questions.”100 Two
96 See Berklee Shares.com: Welcome, at http://www.berkleeshares.com (last visited June 9,
97 See Id.
98 Berklee Shares.com: Frequently Asked Questions: Why is Berklee doing this?, at
http://www.berkleeshares.com/faq (last visited June 9, 2005).
99 See Lawrence Lessig, The Future of Ideas:
100 Tim Berners-Lee et al., The Semantic Web: A New Form of Web Content That Is Meaningful To
technologies exist for developing the Semantic Web, “eXtensible Markup
Language (XML) and the Resource Description Framework (RDF).101 XML
allows creators to extend the standardized tags used in HTML to tag their
content however they like, and RDF gives meaning to that content.102 The
goal of RDF is to enable machines to identify relationships among data at a
conceptual level by using XML tags to create “triples,” much like subject,
verb, object in a normal sentence.103 Each part of the triple is identified by a
Universal Resource Identifier (URI), rather than a normal phrase.104 This
allows similar but different concepts, universally defined, to be distinguished
RDF uses “ontologies” to describe relations of terms.106 Ontologies
enable machines not only to distinguish between similar but different
concepts, but also, through the use of “equivalence relations,” to understand
that some things are the same though they are described using different
terms.107 In the field of Internet search, for example, ontologies can improve
the accuracy of Web searches along the familiar topical dimension by looking
for only those pages that refer to a precise concept, ignoring those that use
Computers Will Unleash a Revolution of New Possibilities, SCIENTIFIC AMERICAN 3 (May 2001),
available at http://www.scientificamerican.com/article.cfm?articleID=00048144-10D2-1C70-
84A9809EC588EF21&catID=2 (last visited June 9, 2005) (contemplating a future web
where web connected devices will use "agents" to communicate and perform tasks). The site
that claims to be the first site on the Semantic Web, see Mindswap at
102 Id, (tagging means to attach a hidden label to content that can be used by programs).
103 Id. (noting that this allows web pages to assert that "things . . . have properties . . . with
104 Id. (explaining that the most common example of a URI is a Uniform Resource Locator
or URL, which is the format for the location of all pages on the current Web).
105 See Id. (using as an example the difference between an address that is a post office box, an
address that is a street address, and a speech).
106 See Id. (ontology includes taxonomy, meaning the definition of "classes of objects and the
relations among them" and inference rules, which allow the computer to manipulate the
terms ... in ways that are useful and meaningful to the human user"); see also Scientific
American.com: Sidebar: May 18, 2001: Glossary: Ontologies, at
B4A8809EC588EEDF (defining ontologies as " Collections of statements written in a
language such as RDF that define the relations between concepts and specify logical rules for
reasoning about them. Computers will "understand" the meaning of semantic data on a Web
page by following links to specified ontologies") (last visited June 9, 2005).
107 See Tim Berners-Lee et al., The Semantic Web: A new form of Web content that is meaningful to
computers will unleash a revolution of new possibilities, SCIENTIFIC AMERICAN (May 2001), available
84A9809EC588EF21&catID=2, at 6 (last visited June 9, 2005) (using the example that zip
code and postal code are different phrases that describe the same thing).
ambiguous keywords.108 Moreover, ontologies theoretically could facilitate
The dream of the Semantic Web has been elusive, but progress is
being made. Last year, the W3C approved RDF and the Web Ontology
Language (OWL), as standards.110 Berners-Lee has encouraged developers to
create applications to “justify the Semantic Web in the short term.”111 An
application called Haystack, developed at MIT, reportedly “knocks down the
partitions that separate e-mail clients, file systems, calendars, address books,
the Web and other repositories so that information can be worked with
regardless of its origin.”112
If realized, the Semantic Web vision has profound consequences for
law – deeper than the now-familiar concerns about electronic agents113 and
machine-enforceable rules.114 Creating machine-interpretable and machine-
108 Id. at 4.
109 If successfully deployed, the Semantic Web also would greatly increase the role of
electronic agents. Id. at 5. Berners-Lee and his colleagues offer a hypothetical in which a pair
of siblings make a doctor's appointment for their mother using their respective web agents.
See Id. at 1 (theorizing that the agent could find a doctor in their mother's insurance plan, the
best office location with considerations for traffic and the best time to schedule the
appointment to avoid major conflicts with existing obligations, all automated). These agents
communicate, verify the identity of other agents, ask for “proofs” of the data they receive to
ensure accuracy, and locate agents across the web that provide desired services through a
directory. See id. Agents will understand each other through the exchange of ontologies, and
indeed agents will be able to acquire new “reasoning capabilities” as they find new
ontologies. Id. at 6. Eventually, this will extend from the web to the physical world when
other items become web enabled. See id. (theorizing that in the future, devices will be able to
communicate and control each other, like a phone call triggering a reduction in the volume
of a stereo or television). According to the authors, even microwaves may be able to contact
the manufacturer of a frozen meal to learn the perfect way to cook that meal. See id.
110 See Darryl K. Taft, W3C Approves Pair of Semantic Web Specs, eWEEK: Web Services, at
http://www.eweek.com/article2/0,1759,1524304,00.asp (Feb 11, 2004) (citing support from
twenty-four organizations involved in technology including IBM, Adobe, and the U.S.
Department of Defense).
111 See Anne Chen, Semantic Web Is 2 Steps Closer, DevSource: Add Ons, at
http://www.devsource.com/article2/0,1759,1621521,00.asp (July 6, 2004) (quoting Tim
112 Lisa Vaas, Berners-Lee Maps Vision of a Web Without Walls, eWEEK.com: Infrastructure, at
http://www.eweek.com/article2/0,1759,1734926,00.asp (Dec. 2, 2004) (describing Berners-
Lee's vision and describing Haystack as a prototype that utilizes the ideas of the Semantic
Web); Haystack at http://simile.mit.edu/hayloft/index.html
113 See, e.g., Anthony J. Bellia, Jr., Contracting With Electronic Agents, 50 EMORY L.J. 1047
(2001); Stephen T. Middlebrook & John Muller, Thoughts On Bots: The Emerging Law Of
Electronic Agents, 56 BUS. LAW. 341 (2000); see also Margaret Jane Radin, Online Standardization
and the Integration of Text and Machine, 70 FORDHAM L. REV. 1125 (2002); Margaret Jane Radin,
Humans, Computers, and Binding Commitment, 75 IND. L.J. 1125 (1999).
114 See generally Symposium Issue, The Law & Technology of Digital Rights Management, 18
Berkeley Tech. L. J. 487 (2003); Dan L. Burk & Julie E. Cohen, Fair Use Infrastructure For
actionable concept maps of the law will enable more radical departures from
the default rules the law supplies than we have previously experienced.115
Further, the process of building machine-interpretable concept maps is likely
to alter our understandings of the concepts being mapped. Creative
Commons licenses, which use RDF at the machine-readable layer, are just
the tip of this particular iceberg. Efforts to create a “policy aware” Web,
appear to be a next step that lawyers should keep an eye on.116 Although
developers imagine the policies of which the Web should be aware to be
private policies adopted by those who provide Web resources, the
technologies also could be adapted to reflect public policies as well.
Some see the Semantic Web project as fundamentally flawed.117
These critics charge that the vision requires too much complexity and
demands that users adapt to the needs of machines instead of adapting
machines to the needs of users.118 Machines use rules to process information,
and rules require classification of information to be useful. People may use
rules to classify information, but we often use different rules depending on
context, and we may not agree about which rule to apply in any given
situation. In a well-argued essay Clay Shirky writes that semantics are in the
users not the system and that ontological classifications, such as those
required for the Semantic Web, work in certain limited domains but will not
Rights Management Systems, 15 HARV. J.L. & TECH. 41 (2001); Julie E. Cohen, Lochner In
Cyberspace: The New Economic Orthodoxy Of "Rights Management", 97 MICH. L. REV. 462 (1998).
115 This point should not be confused with the argument that “code is law.” See generally
LAWRENCE LESSIG, CODE AND OTHER LAWS OF CYBERSPACE (1999); Update at
argument is that software regulates behavior separate and apart from the way that law does
and that on the Internet these regulatory modalities are interchangeable, with code being the
more effective in many cases. For further discussion, see, e.g., Tim Wu, When Code Isn’t Law,
89 VA. L. REV. 679 (2003); R. Polk Wagner, On Software Regulation, 78 S. CAL. L. REV. 457
(2005); James Grimmelmann, Note, Regulation By Software, 114 YALE L. J. 1719 (2005).
116 See, e.g., Policy Aware Web at http://www.policyawareweb.org/; Daniel J. Weitzner, Jim
Hendler, Tim Berners-Lee & Dan Connolly, Creating a Policy-Aware Web: Discretionary, Rule-
Based Access for the World Wide Web, in WEB AND INFORMATION SECURITY ___ (E. Ferrari &
B. Thuraisingham, eds. forthcoming), also at http://www.w3.org/2004/09/Policy-Aware-
Web-acl.pdf. I thank Hal Abelson for this point.
117 Berners-Lee reminds that many also saw the vision for the World Wide Web as
fundamentally flawed. See An [sic] parenthetical discussion to the Web Architecture at 50,000 feet. [sic]
and the Semantic Web Roadmap, at http://www.w3.org/DesignIssues/RDFnot.html (last
visited June 9, 2005).
118 See, e.g., Eric Nee Web Future is Not Semantic, Or Overly Orderly, CIO|INSIGHT, at
http://www.cioinsight.com/article2/0,1397,1817758,00.asp (May 5, 2005) (quoting Google
cofounder Sergey Brin as saying “I'd rather make progress by having computers understand
[sic] what humans write, than by forcing –humans [sic] to write in ways that computers can
work for the Web at large.119
Shirky and other critics, influenced by recent thinking about complex
systems, argue that simple technologies like Really Simple Syndication (RSS)
and “social bookmarking” better enable user-defined complex organization
and classification.120 RSS enables users to automatically check to see if web
pages marked with the appropriate XML tags and to aggregate results. Used
by millions of bloggers and now most mainstream news sites, RSS has been
one of the most rapidly-adopted Internet technologies in recent years.121
Social bookmarking is a development destined to warm every
postmodernist’s heart. Social bookmarking and tagging enable quick
publication and aggregation of metadata about resources, such as web pages,
available on the Internet. Tagging theoretically enables us to forgo
hierarchical classifications – such as “organizing your favorites” into folders
– and the habits of mind associated with such classification.122 Moreover, the
technology enables probabilistic classifications that democratize and make
explicit the social construction of meaning.123 By publishing the list of web
pages that you have bookmarked in your Web browser, you implicitly make a
statement that of all the resources available on the Web, these are relevant to
you in some way. Social bookmarking sites, such as www.deli.icio.us, offer to
host a user’s bookmark file – thereby making the file available to the user on
any Net-connected computer – and to publish the file, or parts of it, to all, or
selected, Web users.124
These sites also permit users to also to associate “tags,” i.e.,
keywords, with these Web addresses and make more explicit the ways in
which these sites are relevant. Flickr’s photo hosting site, discussed above, is
one of the fastest-growing uses of social tagging, enabling searches for
119 See, e.g., Clay Shirky, Ontology is Overrated: Categories, Links, and Tags at [LINK]
120 See id.; Nee, supra note XX, (promoting the technology of Google and Really Simple
Syndication (RSS) as pragmatic alternatives to the Semantic Web theory).
121 CITES Made popular by bloggers, nearly all major news sites provide RSS feeds now.
122 I say “theoretically” because even though tagging does away with the need for visual
representations of conceptual hierarchies – such as a file folder organization scheme – most
people use conceptual hierarchies to make sense of the world and we should expect to see
those hierarchies reflected in their tags.
123 A probabilistic classification asserts that Z% of users think that X is relevant to Y rather
than asserting that X is relevant to Y. See Shirky. It is much easier to give a computer the
authority to make the former statement than the latter. Id.
124 As scholarly research continues to migrate to the Web, some scholarly publishers see the
value of social bookmarking for communities of researchers as well. The Nature Publishing
Group’s Connotea site targets scientific researchers to signal to each other which Web
resources, such as online articles, they deem to be most important or relenvant. See generally
photographs along the topical dimension by popular tags,125 or along the
copyright dimension for photographs available under a Creative Commons
Much of the opposition to the Semantic Web is misdirected. The
spread of these simple technologies is not antithetical to the Semantic Web.
Indeed, the Semantic Web vision requires that there be rich metadata
associated with information available on the Web. The creation of metadata
is costly. It may well be that simple technologies that supply incentives for
the creation of such metadata are prerequisites to realization of a Semantic
Web. RSS tags give you the news of the day, social bookmarks can influence
what you read, and Creative Commons metadata tells you about the
copyright status of the information you encounter.
Moreover, RDF’s first mission is to enable interoperability. As
various social bookmarking and tagging systems emerge, RDF can serve as a
bridge between these systems. Similarly, as machine-readable licensing
becomes more common, RDF can be used to identify equivalence relations
between licenses and/or license terms. When applied to public law, RDF
could also be used to identify equivalence relations between the legal codes
of various jurisdictions – taking international legal harmonization in a new
direction. In many ways, Creative Commons licenses are a test case for the
possibilities of machine-readable law, and this development is worth
The number of copyright events occurring in our daily lives
continues to grow as our collective use of digital media continues to expand.
Creative Commons licenses facilitate coordination and regulation of these
events by enabling end-to-end copyright transactions and by fueling the
growth of new intermediaries that rely on the common pool of Creative
Commons-licensed content. As this pool expands, it is possible to imagine
the growth of a cultural counter-canon along the copyright dimension, but
Creative Commons-licensed content need not compete with content
available on more traditional copyright. Indeed, Creative Commons licenses
are being integrated into traditional commercial licensing practices, although
this remains an underexploited growth opportunity for new and old
intermediaries. Finally, intermediaries increasingly will begin to use and rely
upon the machine-readable descriptions of Creative Commons licenses,
expressed in RDF, as the importance of the copyright relevance dimension
See Hot Tags, at http://www.flickr.com/photos/tags/
126See Creative Commons, at http://www.flickr.com/creativecommons/ (grouping
photographs by license terms); see also infra Section III.F. (discussing Flickr further).
increases and as the idea of machine-readable law becomes better
Michael W. Carroll