March 13, 2007  ·  Lessig

Nathan Weinberg has a piece on WebProNews responding to my response to Tom Rubin’s criticism of Google’s “respect” for copyright. The piece is (well written and) useful, because it packs in its couple hundred words all the fallacies that haunt this anti-Google debate.

Fallacy one: “When you make a copy of a copyrighted work, you are in essence stealing it, and even when I download music and movies, I never kid myself that what I am doing is legal.”

First, Nathan, don’t download music and movies without the permission of the copyright owner. Bad, bad, bad.

Second, even if making “a copy of a copyrighted work” is illegal, it is not “in essence stealing it.” See the Supreme Court’s rejection of this false analogy in footnote 33 of Sony v. Universal.

But third, and most important, not every “copy” violates copyright law. In particular, if a copy is “fair use,” then copyright law has not been violated. The question in this case is thus, as always, is the copying for purposes of making snippet access available “fair use.” As much as you know that it is wrong to download music without the permission of the copyright owner, I hope you also know that it is right to make copies — even without the permission of the copyright owner — when such copies are fair use.

Fallacy two: “While there are many authors who care more about getting their books out there than making money, the vast majority is trying to earn a living. Those authors whose books are out of print, but still in copyright, would love an opportunity to make some money off their older books, but Google’s plan involves copying them without permission.”

True, most authors would prefer to make more money from their works than they did before. True, Google plans on copying them (for purposes of making snippet access available) without permission. But the implication — that Google’s copying will reduce author’s opportunity to make money — is false. Out of print books are — by definition — books the authors are not making money on. Google’s Book Search will refer people who discover the out of print book to book sellers. If demand for a currently out of print book grows, then it is more likely than before that the book will go back into print, meaning again, the author can make money. So contrary to the “if Google copies, authors lose” fallacy, if Google enables access, at least some authors will get something they don’t have right now — their out of print book back in print.

Fallacy three: “Even if all the authors want their books in Google, I’ve always felt that to respect me, you have to show respect, and that means asking me if what you are doing is okay. Don’t tell these authors what’s best for them, that shows no respect at all.”

First, both sides in this debate are effectively “tell[ing] these authors what’s best for them.” Microsoft and the Publishers are telling the authors that it is best for them that the law ban companies like Google from securing access to snippets of their works. They want a rule that says “ask first” not for permission to distribute copies of their books, but for permission to enable access through a 21st century card catalog. When in the history of man did the law require permission from an author (or publisher) for a work to be included in a card catalog?

Second, the point ignores the central point in this debate: Given the insanely inefficient system of copyright that the government has created, there’s no way to identify the current owners of these copyrights (for works out of print). So how, again, are these people to be “asked”?

Fallacy four: “Mr. Lessig, even if you want Google to scan and index your book, even if you want the knowledge in your book spread throughout the earth, wouldn’t you want Google, a company that will make money off your book you will never see, to at least ask permission first?”

There are lots of people who make money off of my work without asking me, and it’s a good thing too. Look at the bump in Stephen Manes Google rank from his criticism of me. How much money do you think Forbes got in ads from the click-throughs that this article — calling me an “idiot” and “moron” and other such stuff? Do you think Manes asked me before he tried to profit as he did. (No.) Do you think he should have to ask me? I certainly don’t. If the use is “fair use” — which even Manes article is — then we should encourage people to make money on it. The more money people can make, the more the economy around spreading ideas can grow.

  • Mark Kraft

    Another thing to point out, Larry, is the “fair use” comparison between what Google is doing with books, vs. what is commonly done with audio recordings.

    It is perfectly legal for people to use “fair use” sound samples of people’s recordings of no more than thirty seconds, without violating any kind of law. We see this done all the time in the case of sites that do music sales or reviews, for example. For an average 4:30 song, this amounts to “giving away” 1/9th of their work.

    In contrast, Google’s snippets are perhaps giving away 1/500th of someone’s work. Is this really outside the bounds of fair use?

  • Kevin

    Great post.

    To the second point, that “the vast majority [of authors] are trying to earn a living,” I would point out that only a small minority of authors make their living of books alone.

    Most hold day jobs because very few books ever sell enough to make it economically viable to be just a writer.

  • mmmbeer

    I appreciate your attempt Mr. Lessig, but I think that there are some serious holes in your arguments.

    First, I think that the problem is that Google’s fair use argument is virtually non-existent under the current law. Which of the factors tilts in Google’s favor so heavily to outweigh the others? In addition, any defense would have to apply to all of the exclusive rights. Even if the there is a strong fair use argument that around displaying only a snippet, well, that’s only part of the exclusive copyright rights (distribution or public display). Among other things, there’s also the slavish copying of books for its own database.

    Second, I also think that the analogy of the card catalog is inaccurate. Listing the titles of the books with some other descriptive, identifying information (even snippets) is far from what is being done here. How exactly is the book project like a card catalog? In addition, a library, presumably, has an authorized copy of the work to lend on a one-reader-per-copy time.

    Finally, shifting the burden to the authors to defend their copyrights is exactly opposite of how the law is currently structured. This was reaffirmed recently by the Ninth circuit, I believe. If an author wanted anyone to be able to distribute or use their works, they (or their publishers) have the ability to do that: grant the license. It doesn’t work the other way. By analogy to real property (this is one of the times I actually think such an analogy works), someone cannot just use my vacation property without my permission for their own purpose simply because I’m not using it. The law doesn’t require that I ask them to stop first, but rather requires them to seek permission first and empowers me with a legal remedy. In this case, why should any intellectual property work any differently?

    In addition, the current law doesn’t require authors to state why they don’t want their works redistributed. They have the right to authorize the use. Further, there could be a lot of reasons authors don’t want their works online (a funny reason: they can’t believe that they ever wrote that and have reserved the remainder of copyrights for themselves).

    I think Google is squarely in novel legal territory. I think that they have persuasive arguments, but they are not the ones you’ve cited above.

  • Pat

    mmmbeer: In response to your first point, I think that the factors that most obviously make this a fair use is the amount (they only show a snip it) and the effect on the potential market (I still have yet to hear any logical argument as to how someone reading a snip it can effect sales). I think that there is also a strong argument about the purpose and character of the use; this project is enabling an easer search of literature. Also note that the four factors are not exclusive; the courts can consider other factors as well. I think that when you look at Google Book Search in relation to the stated objective of copyright (promote useful arts and sciences), I think that it is clear that the ability to search such a huge number of books (especially rare and out of print books) for background and previous research is a very valuable tool.

    Also, I hardly find the copying of books into a database “slavish.” In fact, if google didn’t make the database available to anyone then I am hard pressed to find any sort of a copyright violation (they are copying for private, non-commercial use). If this is the case then the copying of the books has no bearing on the issue of infringement, only the availability of the database through search.

    In response to part 2, while a card catalog does only have the things you listed, the purpose of the card catalog is still to help you find information on a certain subject or book. All google is doing is creating a better card catalog by showing you not just a small description of the book, but a small snip it of the book that shows you exactly how this book might relate to the information you are looking for.

    In response to part 3, see about a billion earlier posts on this blog as to why intellectual property =/= real property. Also, this is not shifting the burden of the authors to defend their copyright because this is a fair use of the works in question, so the authors don’t have a claim of infringement to bring against Google.

    While I agree that google book search is in novel legal territory, I do think that fair use is very relevent in these debates.

  • Seth Finkelstein

    The anti-Google argument is that Google is, itself, making commercial use of a large portion of the book in aggregate. See this skeptical analysis (not by me, just passing it along) for some arguments against the application being fair use. It’s about “microsales” (really, micro-commercial use). What’s new, in an evolutionary sense, is that Google has found a way to make large amounts of money off accumulated small sales.

    The key issue issue is, do you look at the usage per-transaction, or overall?
    A court could go either way.

    I have an old blog post on this:
    Google Print: Copyright vs. Innovation vs. commercial value

    “… the technology company can’t be right every time, almost by definition. Because copyright as a limited monopoly fundamentally restricts innovation in some ways. That’s the trade-off.”

  • ACS

    I think the real legal argument lies in “fallacy number 2″, that is, what legal ramifications lie in Google Print reproducing a work (book) for the purpose of providing a snippet from a search engine?

    Well, that really relates to two questions:-

    1. Is google conduct PRIMA FACIE infringement of copyright??
    2. Is that prima facie infringement voided by the Fair Use provisions??

    Well the answer to number one is certain – yes – reproduction of a work is prima facie infringement (Don’t even get started on the whole “Fair Use is not infringement” line – I said prima facie) . The fact is that Google must reproduce the whole work onto its databases to operate the search engine.

    So, does fair use apply, –> What are the definitions of Fair Use under the US law –> Does Google conduct have any commercial purpose?? —> I think we are now familiar with the potential advertising revenue (What’s google share price $490 something) available to popular search engines. Do you think Google is doing this out of altruism??

    Finally have a think about the remedies for copyright infringement (in Australia) damages or account of profits. The legislation is there (and copyright theory is reliant upon) the author having the exclusive economic rights in his/her work. Could googleprint

    (a) Provide economic advantages to persons other than the copyright owner?? or

    (b) Should those exclusive economic rights be swayed to the public interest??

    I only ask that if you follow (b) that you put it under that guise rather than use the argument put forward by the professor in his response to “fallacy number 2″, that is, argue the public interest – dont argue it (potentially) increases the authors economic interests.

    I think this case, if legal action is ever heard, would highlight the technology v rights holder debate. This case comes down to point of view.

  • BombaySam


    is perfectly legal for people to use “fair use” sound samples of people’s recordings of no more than thirty seconds, without violating any kind of law. We see this done all the time in the case of sites that do music sales or reviews, for example. For an average 4:30 song, this amounts to “giving away” 1/9th of their work.

    What is this pile of junk. Where does it say that 30 second soundbites are always legal. Id like to see that.

  • Brian Utterback

    The weakest part of Google’s argument and Mr. Weinberg’s main premise is that for Google to provide the indexing and the snippet to be displayed, Google must make a copy of the whole book and by and large, they do not own that book. I think we all get so hung up over the thousands of copies made of each new mp3, that we don’t realize that people might still argue over the legality of a single copy. While I hope that Google prevails, I have to admit I find it somewhat troubling. For me, the key issue is whether or not a complete copying (which would normally be infringing) be allowed if it is done in order to provide a fair use? Even if Google does not provide access to the whole copy, doesn’t it retain such a copy internally?

    Then there is the opt-in versus opt-out issue. If what Google is doing truly is fair use, then it shouldn’t have to provide an opt-out option. I hope that the fact that they do and that the service is itself fair use persuades the courts that opt-out is sufficient and not require an opt-in.

  • Jim Carlile

    I support Google’s scanning efforts, but has anyone else noticed something a little peculiar about their providing access to works that are clearly in the public domain?

    Excuse me if this has been noted before, but whenever I have tried to download 19th century books that have been reprinted within the last 50 years or so by somebody, Google won’t provide the “full view” or download capability. Only if a book has NOT been re-released will Google provide complete access.

    This is the case for scans of the actual 19th Century volumes, not just the reprints.

    Maybe this has changed lately, but Dickens is a great example. Last Christmas I couldn’t find any full view access on any of the Christmas Stories, or his novels as well. But the Microsoft site provides complete downloads of everything they have in the public domain, even if subsequent commercial reprintings have been done.

    Is this a Google policy, or only an anomaly? If it’s the first, it indicates to me that Google’s intent in all of this is not quite so lofty as they claim– it’s more about selling books than it is about providing access.

  • Anonymous

    “The question in this case is thus, as always, is the copying for purposes of making snippet access available “fair use.””

    Ah, but, what if we shift the focus from the snippets, to the entirely commercial purpose of copying the entire books in order to make the books searchable in an electronic index (G. Book Search). Surely there is great commercial value in reproducing/copying an entire work to make an electronically searchable index; I can envision copyright holders demanding payment from database vendors for the privilege of copying their entire book to electronically index.

    I.E. I might write a book and have LexisNexis pay me for the privilege of electronically indexing it. LexisNexis’ service would be all the more valuable to users for the ability to search through my book electronically, let alone a whole library-full. LexisNexis could indicate to a user that my search was relevant even without displaying snippets; and if snippets are displayed, the utility and functionality still depended entirely upon the creation of an entire digital reproduction through scanning of my work. So, I think it is slightly disingenuous to focus solely on the snippets. It’s not just the snippet display that needs to be measured under fair use, but also Google’s copying of entire works to index them, and making the entire work searchable to users through an electronic index; the search requires the entire work to be indexed, and this is for a commercial purpose.

    Displaying snippets implicates the public display right of copyright, but reproducing/digitizing an entire work to electronic index the work necessarily implicates the reproduction right of copyright. Is one right necessarily more worthy of protection? Then I wonder why the arguments I’ve seen mostly focus on the snippets and not the reproduction of the entire work? Well, perhaps it is because this argument could be used to attack indexing/caching of web content; and perhaps also because of the Arriba Soft precedent.

    To distinguish Google’s scanning of physical books from web indexing and to distinguish Arriba Soft, let’s first examine who is the least-cost avoider in terms of negotiating reproduction rights. People on the internet can easily put in robot.txt code to and other metadata to prevent web indexing. I would say there is a much higher transaction cost for copyright holders to call up/e-mail write Google to opt out of the Book Search project than for web publishers to put robots.txt in their html code.

    Also, I would argue that there’s a large difference in degree between Google’s indexing of digital information already publicly available on the web, and Google’s physically entering libraries and transforming paper-based information through scanning into digital form, and creating an electronic index (G. Book Search) from those scanned works. Arriba Soft did not go to museums and galleries and scan entire works for their electronic indexing purposes; the material was already digital, and voluntarily on the web for all to see, and robots.txt and other code perhaps could have been used to prevent electronic reproduction of the works by webcrawlers. Furthermore, arguably, use of the entire works is more pronounced in Book Search than with Arriba Soft. With Arriba Soft you could see a whole thumbnail (but you could also see the entire larger work on the web publicly). With Book Search, Book Search depends on the entirety of the works whenever you enter search terms; in fact Book Search must, for every search, rely upon having digitally reproduced and indexed every book in the index. The words of each of these millions of books are being relied upon with every Book Search, in my understanding of how electronic indexes work; sure, they may no longer be arranged in human-readable book order, but their entirety is in there logically coded (minus stopwords probably). Using the index coding, one could probably go back and reassemble a human readable form of each entire work (once again, minus stopwords). So an argument is that the operation of the Google Book Search search functionality actually requires usage of the entirety of all the works in the database to a much larger degree than Arriba Soft actually depended upon the thumbnails (which were just displayed for reference).

    Finding Google Book Search to not constitute fair use does not necessarily mean that a judge’s ruling would also threaten web indexing; nuances could be preserved, such as by examining which party is the least-cost avoider (in the web context the least-cost avoider I argue is web publishers, by putting robots.txt or other metadata into their html code to prevent indexing whereas in the print context, Google is the least-cost avoider, by simply not scanning books it knows it does not have permission to scan in the case of Google Book Search (publishers are not the least-cost avoider by any stretch of the imagination: avoidance of scanning would consist of publishers having to do an inventory of their books in and out of print, checking their contracts to see what each works’ copyright status is (did we enter any exclusive licensing/distribution deals with anyone that might be violated for this particular work? Did we actually transfer the copyright to another party entirely?), and then requiring the publisher to phone/e-mail/write to Google to opt out, and then have the publisher have to police Book Search to verify that their opt-out choice is being honored). For Google Book Search, I would say the least-cost avoider seems to most definitely be Google. What transaction costs would accrue if Google were to simply not scan books they are not aware of the copyright status of, versus the transaction costs required for publishers to inventory and assess the copyright ownership and licensing status etc. of works they’ve published, alert Google, and police Book Search? Once again, the least-cost avoider is definitely Google, in my opinion. Google seems to need to work on licensing, and not unilaterally copying. As may become more evident with the Viacom suit.

    A fair use analysis of the Google Book Search project:
    1) nature of work: runs the gamut from non-fiction to poetry, all of which can be considered highly expressive, unless they’re also scanning a whole bunch of alphabetically organized phone books.
    2) purpose of work: oh, so strictly commercial. Google Book Search draws users to Google’s network of web search services and other services, enhances the breadth and depth of search services, raises the profile of the brand name, adds more web real estate on which to post contextual ads.
    3) Amount of work: oh, entire works are scanned. Then every search a user does depends upon the entire work having been scanned and index. The amount displayed, whether snippets or not, is only one aspect: let’s not fool ourselves that the entire work is not being copied and in essence the entire bodies of works of millions of books are being relied upon each time the index is searched by users.
    4) Harm to market: as IP licensing grows in importance (see the Google v. Viacom suit and expect many more), obviously a copyright holder’s bundle of property rights regarding their IP will be of paramount importance. I foresee a market for electronic indexing of works, whether it they be CDs, DVDs, or books. Pay me for the privilege to even electronically index my works. That’s part of my bundle of copyrights, to exclude you from reproducing my work, unless it’s fair use, and if your index is for profit, and my work is expressive, and you need to copy my entire work for your index to be effective, and I can make money from allowing certain companies to index my work, and you unilaterally entering a library and digitizing and indexing my entire work might interfere with my bargaining ability to demand payment from others to reproduce/digitize and index my work, then, I assert that your use most certainly isn’t fair; the harm to my ability to market licenses allowing electronic indexing of my work, your profit motive, the expressiveness of my work, and the amount of my work you index (all of it) may outweigh your arguments that your use is transformative even though it is a commercial use. With the amazing ability with modern technology to effortlessly and cheaply distribute digital information goods,* the contours of copyright law must definitely be examined carefully, and the ability of copyright holders to exclude entities from reproducing their entire works must be carefully studied to see if, in fact, projects like Google’s unilateral reproduction of/digitizing of entire copyrighted works are not worthy of a “fair use” designation.


    *(I acknowledge that the term “goods” in the context of information may be controversial to those who do not want to view intellectual creations as “goods” or “property” but it seems fairly well established that “property” is a bundle of rights, and creators of intellectual works certainly are recognized to “exclusive rights,” with respect to their creations, hence, the terms “information goods” and “intellectual property” are appropriate. (Of course, those “exclusive rights” are subject to exceptions such as fair use, which is the balance we are all searching for…) Article I Sec. 8, US Constitution, “To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries”)

  • Kevin Carson

    As Woody Guthrie said, “Anyone found copying this without our permission will be mighty good friends of ours.”

  • Anonymous

    I must say, I do wish fair use were actually more liberally construed. I’m all for more liberal orphan works legislation too. It’s just interesting in that I don’t think Google’s case is very clear cut; I hope they have good lobbyists! Maybe there could be some sort of statutory fees for reproducing others’ work, for indexing, like there are statutory fees for making cover songs. That would be great if there were statutory fees for music sampling too.

  • b. eckenheimer

    I detest the term “intellectual property.”

    The purpose of restrictions on usage and duplication of creative or innovative works is to encourage continued creativity and innovation by allowing the creator/inventor the exclusive right to profit from the work for a (no longer very) limited period of time. Such works then revert to the public domain for the benefit everyone.

    In other words, creators and inventors are permitted to enrich themselves for a time so that we may all be enriched by their efforts in the long run. The fact that the law has been perverted to grant obscenely long protections to corporations (think Mickey Mouse), does not mean it is either ethical, moral, or right.

    The issue is most cogently enunciated in two speeches made over 160 years ago by Thomas Babington Macaulay to the British Parliament, which can be found here:
    Here’s an excerpt: “The principle of copyright is this. It is a tax on readers for the purpose of giving a bounty to writers. The tax is an exceedingly bad one; … In order to give such a bounty, I willingly submit even to this severe and burdensome tax. Nay, I am ready to increase the tax, if it can be shown that by so doing I should proportionally increase the bounty.”

  • bowerbird

    it would seem to me that the _consideration_ an author receives
    when making a contract with a publisher is that the work will be
    _made_available_to_the_public_. if that’s not happening — and
    it’s not, once a book is out of print — then the publisher cannot
    reasonably expect for the contract to remain in effect, thereby
    relinquishing any rights they receive under it. so orphan works
    could be argued to be free of encumbrances _if_ the publisher
    does not use available means (e.g., the web) to publicize them.
    if publishers don’t _use_ the right to make copies, they _lose_ it.