June 15, 2024

The Digital Death of Copyright's First Sale Doctrine

The legal media’s attention has been focused this past week on Supreme Court oral arguments in Golan v. Holder, an important copyright case involving the power of Congress to “restore” private rights in creative works that are already in the public domain. In this post, I’d like to focus on an important copyright case that won’t be argued in the Supreme Court. On October 3, the Supreme Court declined to review Vernor v. Autodesk, a Ninth Circuit Court of Appeals decision involving the applicability of copyright’s first sale doctrine to transactions involving software and other digital information goods.

The first sale doctrine is the provision in copyright law that gives the purchaser of a copy of a copyrighted work the right to sell or otherwise dispose of that copy without the permission of the copyright owner. If there were no first sale doctrine, there would be no free market for used books, CDs, or DVDs, because the copyright owner’s right of distribution would reach beyond the first sale, all the way down the stream of commerce. Without the first sale doctrine, movie rental services like Netflix and Redbox wouldn’t be able to lend DVDs without authorization from studios, and you wouldn’t be able to lend the bestseller you just finished to a friend without authorization from the book’s author or publisher. Along with fair use, the first sale doctrine promotes public access to culture and information by functioning as a crucial limit on the right of a copyright owner to control the disposition of a copyrighted work. A world without the first sale doctrine in it is a world I wouldn’t want to live in, but it’s one that’s quickly taking shape, thanks in part to legal decisions like the one in Vernor.

The Ninth Circuit’s decision in Vernor significantly erodes the first sale doctrine with respect to software and other mass-licensed digital goods. The plaintiff in the case, Timothy Vernor, bought several copies of an AutoCad software package from a direct customer of the software publisher. Vernor then resold the copies on eBay, only to be accused of having infringed the software publisher’s copyrights. He sued, seeking a declaration from the court that his sale of the software on eBay was protected by the first sale doctrine because the software publisher’s distribution right was exhausted by its “first sale” of the copies to its direct customer. By Vernor’s logic, software should be able to be purchased and resold in the same way that a book can be purchased and resold. Why should the two be treated any differently under copyright law, after all? The buyer of a book owns her copy of the book as personal property, and her ownership of that individual copy does not at all interfere with the author’s ownership of the copyright in the work. Once the author places a particular copy of the work into the stream of commerce, the author loses the right to control the fate of that copy; the author retains, however, all of the rights the statute gives her in the copyrighted work itself, including the exclusive right to make and sell new copies of it.

The Copyright Act makes an explicit distinction between ownership of a copy of a copyrighted work–whether that copy is printed on paper or burned onto an optical disk or stored in a computer’s memory–and ownership of the copyright in the work. The copy is tangible property owned by the purchaser; the copyrighted work embodied in the copy is intangible property owned by the author. A copy is a copy, the work is the work, and never the twain shall meet. In Timothy Vernor’s case, however, the publisher of the AutoCad software argued that it never actually sold the copies Vernor bought, so there was no “first sale” for copyright purposes. Under the software publisher’s logic, which the Ninth Circuit adopted in the case, both the copy and the intellectual property embodied in the copy were only licensed, and quite restrictively so, pursuant to the terms of a mass end user license agreement (EULA); nothing was ever sold, despite the retail transaction that put copies of the software into the hands of the initial purchaser, and despite the downstream transaction that put those copies into Timothy Vernor’s hands.

Existing copyright case law makes it clear that digital copies of works, even those stored only ephemerally in RAM, are “copies” within the meaning of the Copyright Act. Moreover, the the Copyright Act is clear on its face that there is a difference between ownership of a copy of a work and ownership of the work embodied in the copy. Decisions like the Ninth Circuit’s in Vernor v. Autodesk, however, permit copyright owners to conflate the copy and the work to the detriment of consumers in cases involving digital goods. Under Vernor, software copyright owners not only own the work embodied in every copy of a program they sell, they own every copy, too. Consumers are left with both empty pockets and empty hands.

As the transition from physical to streaming or cloud-based digital distribution continues, further divorcing copyrighted works from their traditional tangible embodiments, it will increasingly be the case that consumers do not own the information goods they buy (or, rather, think they’ve bought). Under the court’s decision in Vernor, all a copyright owner has to do to effectively repeal the statutory first sale doctrine is draft a EULA that (1) specifies that the user is granted a license; (2) significantly restricts the user’s ability to transfer the software; and (3) imposes notable use restrictions. Sad to say, it’s about as easy as falling off a log.


  1. The obvious distinction is that when you sell a book, you can no longer read the book.

    A buyer can copy software easily and then resell the original media.

    After the sale, the bit pattern on their computer won’t be recognized as a forgery, and it’s it’s a legal copy while the original media is owned, whereas an illegal copy of a book is usually easily identified as a forgery.

    Copying a copyrighted book is illegal. Copying copyrighted software to the hard disk of your machine is generally legal.

    I suppose it should be legal to remove software from your computer and then sell the original media, but how can this be verified that you no longer have a copy? Should the buyer be given the benefit of the doubt in this case, similar to the way the original copyright lawyers who wrote the law apparently presume that the seller didn’t photocopy an entire book? Perhaps. The presumption of innocence is the basis of our legal system. However, the courts probably do recognize the prevalence of software piracy, there are lots of people openly writing about it on the Internet and lots of software is available for download, some of it even cracked so no license is necessary. (I do not download this, I just Google and read this is true! I am a software developer and the idea of stealing other people’s software is greatly distasteful to me).

    The courts want to protect sellers as well as buyers. It’s much easier to do that with books.

    So, I suppose everyone loses out because some people can’t be trusted. I can’t say I blame the courts in this case – if the theft wasn’t so rampant, I expect the decision would be different. (By the way, I don’t buy the argument that the people who steal would never purchase the software only because there is no way to verify that – it might be true, but it seems likely that it often wouldn’t be true).

    • Annemarie Bridy says

      It’s a good point that the copy of the book is a rivalrous good, whereas the software is not. Software publishers can, though, use DRM to police the number of installations active for particular key or serial number. In an effort to accommodate the expectations of consumers based on their experience with physical books, Barnes & Noble has built lending functionality into its Nook e-reader that works in this way. It’s true that DRM can be, and often is, cracked. But I think it’s effective for the majority of users, who aren’t looking to get something for nothing. There are ways for corporate copyright owners to be more solicitous of consumer expectations without altogether throwing in the towel when it comes to copyright enforcement.

      • It’s funny, then, that the ones that do “throw in the towel” on “copyright enforcement” often find they can make more money that way.

  2. Early on in the history of recording, in about 1900, there was a whole legal mismash over patents, which resulted at one point in the manufacturer of Victor Records (with the Dog on the label!) being forbidden to sell records. He responded by putting this text on the label of his records:

    “This record is leased for the purpose of producing sound directly from the record, and for no other purpose; any attempt at copying or counterfeiting will be construed as a violation of this condition and a basis for legal proceedings.”

    It seems to have been sufficient to continue “selling” records, since Victor became the largest manufacturer of disks in the early part of the last century.

    It is also interesting to note that Victor was trying to invent some sort of copyright out of this “contract”, since copyright did not apply to sound recordings at the time.

    BTW, the history of patents for phonographs and records is quite interesting, and shows that patents were always pretty useless and misguided, though it is much worse of course now.

  3. Steve Roosa says

    Two thoughts. Because copies are becoming cheaper to make (electrons on the wire being less expensive than chewing up trees, etc.), then perhaps the new regime, which involves the erosion of the first sale doctrine, ultimately benefits the consumer by ushering in lower prices. The bargain is this: we the sellers are going to restrict rights associated with subsequent sale or transfer, but you the buyer will pay less. Another thought: If the public-at-large found restrictions associated with the first sale doctrine objectionable, wouldn’t there be a market for alternative arrangements? Isn’t it fair to infer from the absence of such a market that oppressive EULAs have low or lower consumer costs associated with them? Compare, for example, the market for privacy, and how the major browser companies have made efforts to embody ideas around “do not track.” There appears to be a competitive market for pro-privacy browser options even though the end-user does not have to pay for them. Why? The answer presumably has something to do with the relative power of demand. Having said all that, I am fond of the first sale doctrine and dislike oppressive EULAs, but apparently these sentiments are not shared by the public-at-large or any sufficiently powerful (in a market sense) minority.

    • Annemarie Bridy says

      If the bargain Steve describes actually materializes, then consumers may actually have something to gain from the demise of first sale. I’m skeptical. I’m also skeptical concerning whether the market will adequately supply the growing demand for online privacy. If browser manufacturers actually had it within their power to deliver on “do not track,” I’d be more optimistic, but as I understand the functionality, the browsers rely on web site operators to opt in to the consumer’s browser-based opt out. It’s not at all clear that web sites will get on board. What seems more likely to me is that consumers will be lulled into a false sense of privacy by flipping a toggle within their browser settings. (In the same way that they mistakenly think they own copies of software they’ve bought.) By doing this, however, all they’ll really be doing is expressing an automated preference–sort of a hope–not to be tracked. No one is bound to respect it. I suppose consumers could refuse to visit sites that don’t respect the opt out, thereby voting with their virtual feet, but I’m guessing it will be pretty hard to figure out which sites those are, and that the sites themselves won’t be forthcoming about it. There’s a reason that software manufacturers bury their licensing provisions in the smallest of small print. If I felt more confident that consumers really understood the arrangements into which they’re entering, I’d be more inclined to accept that they’re affirmatively choosing the new reality.

  4. Annemarie Bridy says

    Though I agree with his conclusion about courts muddling through, I disagree somewhat with Another Kevin’s characterization of the Vernor case. There were two separate questions in the case, both turning on the issue of whether the party from whom Vernor bought the software “owned” the copies he bought within the meaning of the Copyright Act. The first question was whether Vernor’s distribution/resale was protected by the first sale doctrine. The second question was whether Vernor could invoke the “essential step” defense in section 117 of the Copyright Act to insulate him from indirect infringement claims based on the downstream installation and use (i.e. reproduction) of the software by the people to whom he sold the copies. Vernor wanted a declaration from the court that he wasn’t directly infringing the publisher’s right of distribution or indirectly infringing the publisher’s right of reproduction. These are two separate rights, subject to two separate legal analyses. Both defenses require proof of ownership of the copies of the software, and Vernor failed on both counts.

    Vernor argued that he had the right to resell the copies because the publisher’s right of distribution was exhausted by its “first sale” of the software to its customer, from whom Vernor bought the copies. The publisher argued successfully that Vernor was not entitled to invoke the first sale doctrine because the publisher never actually sold the copies to its customer; it only licensed them. Because the software publisher’s customer was never an “owner,” only a licensee, there was never a first sale to exhaust the publisher’s right of distribution. Because there was never a first sale, Vernor was liable for infringing the publisher’s distribution right. In other words, Vernor was not protected from infringing the right of distribution by the first sale doctrine.

    The court also held that Vernor could not invoke the “essential step” defense from section 117 to defend against claims for indirect/contributory infringement because that defense is available only to an owner of a copy of software. Vernor’s customers weren’t owners, because Vernor wasn’t an owner, because the party from whom Vernor bought the software wasn’t an owner. Lack of ownership foreclosed that defense for Vernor, too.

    Bobbs-Merrill, which announced the first sale doctrine, which was later codified at section 109 of the Copyright Act, is distinguishable from Vernor’s case, because there was no question in Bobbs-Merrill that the copy of the book in question had been sold. To quote from the case, “[t]he facts disclose a sale of a book at wholesale by the owners of the copyright.” Vernor relied on Bobbs-Merrill in his defense, but the court said that Bobbs-Merrill couldn’t help him, because he didn’t acquire the copies from someone who actually owned them.

    • Once you’ve said that the difference was that in one case title was transferred but in the other it wasn’t, that still doesn’t answer the real question. If Bobbs-Merrill had simply used different language when attempting to impose its resale conditions (“you are licensed to possess this book, or to transfer your license it to another entitiy for compensation provided the compensation is no less than blah blah blah…”), would the court back then have upheld the restrictions? If not, then the distinction is clearly unsatisfying.

      I think there’s a good case to be made that courts not uncommonly make distinctions without a difference, just because overturning precedent directly would cause too much disruption. And that this is one of those times.

    • Isn’t this circular reasoning by the software companies? They claim you need to “license” the software rather than just own a copy because you have to make more copies to use it — but the “essential step” defense in section 117 says you don’t have to — except you don’t have that defense because you only “licensed” the software!

      On top of that, the typical transaction in a Best Buy sure looks like a sale, not a licensing. Can some fine print in some text the software itself displays when first run magically retroactively change “a purchase of goods from Best Buy” into “a licensing of some other thing from some other company entirely”?

      Why do the courts even allow such an interpretation? Why did they ever? It clearly benefits no-one. Software publishers could get by just fine if software sales were treated the same, copyright-wise, as book sales, with the section 117 defense making use of that software not infringing in particular. Indeed, the section 117 defense was obviously written with exactly that outcome in mind, so the intent of Congress was clearly that software should be sold, like books and records, to the general public.

      The requirement for ownership to be covered by the section 117 defense is the one serious “bug” in Congress’s program — not only does it allow the nasty loophole discussed here to trash first sale, but it also seems to mean that sitting down at someone else’s computer and, say, booting it up or launching Word is copyright infringement, even if you have the computer owner’s permission. The same is obviously not true for sitting at someone’s couch and flipping through their magazines. Both activities clearly should be equally legal (“borrowing” someone else’s copies of copyrighted stuff, with their permission) but apparently are not — and the lack in the latter case has a real, and massive, consumer cost, though not directly. Software publishers use it to force other businesses to shell out per-user rents, instead of per-copy rents, for software they use, and these added costs will cascade down the supply chain and eventually adhere to consumer goods as price increases. And it’s all deadweight loss, rather than paying for actual parts and labor for anything. All it does is line software publishers’ executives’ pockets, rather than actually pay for scarce materials, pay labor and employ people, or even pay for R&D.

  5. Another Kevin says

    The First Sale Doctrine was judge-made law in Bobbs-Merrill Co. v. Straus, 210 U.S. 339 (1908) – in which a book publisher attached an EULA stating, “The price of this book at retail is $1 net. No dealer is licensed to sell it at a lower price, and a sale at a lower price will be treated as an infringement of the copyright.” The Supreme Court’s decision in that case was actually inconsistent with the copyright law in the rest of the world: in Civil Law countries, a copyright holder is entitled to a fee when a work is resold.

    This case rather hinges upon the fact that the software cannot be used without making a copy. Unlike a book, which can be read and exchanged without copying it, software on a DVD must be copied into at least the computer’s main memory before it can be executed. The courts have held that the operation of doing so is making a copy under the copyright law, and therefore requires a license from the rightsholder – separate from the ownership of the DVD itself.

    What this case does is to transfer liability onto the reseller in advance. Presumably, if the seller were transferring the DVD as scrap plastic or an objet d’art, he might escape liability, but by presenting the DVD as containing the licensed software, he is, the court holds, guilty of contributory infringment by inducing the buyer to copy it.

    This is all a lengthy legal workaround to avoid having the decision affect established caselaw regarding the markets for used books and analog phonograph records. The real way to clarify the law would be to have the legislative branch draft coherent legislation. But we know that in the current political climate, if they were to do so, they’d sign over all our creative rights to the RIAA and MPAA – perhaps to the extent of holding copyright enforceable only by a collecting society. So the courts muddle through trying to interpret law that makes no sense according to precedents that are founded upon misunderstanding of how computers actually work.

    A paper that highlights the fundamental confusion in this area of the law is found at http://www.bc.edu/bc_org/avp/law/st_org/iptf/articles/content/1997041501.html.

    • What about a shrink-wrap license around a book? If the publisher attached a EULA that the buyer accepted by unwrapping the book?

  6. David Karger says

    It would appear that this decision could be “backported” to books, if a publisher chose to sell a book with a EULA stating that they were only selling the purchaser a license to read the book. This of course would destroy the used book industry (and possibly libraries).

    Surely the supreme court would be motivated to step in should such book EULAs ever come into use. Would it make sense to proactively publish a book with such a EULA, then find a reseller to sue, in order to get the issue settled quickly, and also applied to software?

    • People have put EULA’s in physical books as political statements going back to the 80s, but I don’t believe anyone has ever spent the money to try and enforce one. On the other hand, what you’re talking about is already the status quo with ebooks. Licenses and DRM forbid sale or lending, except within narrow parameters (and libraries are knuckling under to ebook contracts that forbid more than two dozen or so patrons from reading the electronic copy before additional costs kick in). In addition, a lot of ebooks are now sold as “apps” — essentially a barely-customized ebook reader wrapped around the text, so that the software rules apply directly.

      And as with the Vernor case, even physical transfer of the media containing a copy of an ebook doesn’t authorize someone to read it. If I give or sell you my ipod/pad/kindle/nook/whatever the books will vanish as soon as I “deauthorize” the device from my ebook-purchase account(s).

      (Oh, and somewhat offtopic here, but interesting from a tragedy-of-the-commons point of view, even out-of-copyright books get caught up in this mess, as ebook sellers choose to distribute low-quality OCR texts of classics instead of the proofread versions available from Gutenberg, and make it difficult to load Gutenberg texts onto their readers.)

    • Annemarie Bridy says

      The strategy you mention–of slapping a “not for resale” sticker on a physical book and treating it like a shrinkwrap license–is actually currently being used by law textbook publishers. Such stickers started to appear a couple of years ago on the complimentary instructor review copies I receive, and they are becoming the norm. On the IP professor listserv to which I belong, people disagree about whether such “contractual” restrictions would be legally effective to negate a first sale defense. I think they might be, given that the cases focus on whether or not title to the copy has been transferred, as opposed to whether or not mere possession of the copy has been transferred. If the sticker is somehow read to reserve title to the copy in the hands of the publisher, the backporting would be successful, despite the transfer of possession of the copy. I suspect, however, that people would be very reluctant to actually shell out money for physical books subject to such a restriction, given existing norms and expectations around the purchase of physical books. Buying a physical book implicitly and historically entailed both a transfer of title and a transfer of possession. Software publishers have been successfully shifting this norm–through all of those licenses that no one ever reads before clicking “I agree.” Ebook publishers are very savvy about this shift and are hoping for a world in which all content is only ever rented, and consumers don’t make any fuss about it. I’m not so sure that the Supreme Court would have any issue with the backporting you describe.

      • That was common practice with records way back when I was reviewing music for a newspaper. Review copies came stamped “not for resale.” However, the basis for it was not an EULA, but that the record was loaned to the reviewer. The reviewer of course hadn’t bought it and so couldn’t really complain (not that anyone ever paid any attention to it. the used record stores were full of old review copies.) I am sure that the basis for the restriction on resale of review copies of law books is the same.

        • It was my understanding that the relevant precedent, at least in the 9th Circuit, was UMG v. Augusto.


          The court rejected the theory that the “not for resale” stamps prevented a transfer of title. On the other hand, it distinguished this from the fact pattern in Vernor, which it had recently decided.

          The fact that the CDs were sent for free without solicitation by the recipient actually cuts against restrictions on redistribution… in at least two ways.
          1.) Under the First Sale Doctrine analysis, there doesn’t seem to be a preservation of title by the sender. Such preservation would require a license agreement that does not seem to exist because, among other things, “promotional CDs are dispatched to the recipients without any prior arrangement as to those particular copies.”
          2.) The court also held that the CDs were “unordered merchandise” under the Unordered Merchandise Statute, which states that the recipient “shall have the right to retain, use, discard, or dispose of it in any manner he sees fit without any obligation what- soever to the sender.”

          Whether or not this analysis would hold for the particular means and terms attached to the law books that Annemarie mentions seems to be somewhat fact-specific.

          • Annemarie Bridy says

            As Steve points out, the Augusto case was nicely clear cut for the court because the Unordered Merchandise Statute made it plain that the CDs at issue were owned outright by their recipients. In other words, the non-copyright statute permitted the court to definitively answer the copyright/first sale question. (“Because the statute grants to the recipients the right to treat the CDs as their own, shipping the unordered CDs to the recipients rendered the recipients owners, not licensees, of the CDs for purposes of the first sale defense.”) The gnarly “Is it a sale, or is it a license?” debate was thereby avoidable.

            In cases involving materials that are not unsolicited, the clarity diminishes–even, I think, with respect to hard goods. The law books I receive are not entirely unsolicited, or at least it’s plausible to argue that they aren’t, in the sense that I have asked publishers to send review copies of new books as they are published in the areas in which I teach. That can be understood as a sort of standing request. I don’t have any agreement with these publishers, express or otherwise, that I am not the owner of the copies I receive, but I can foresee being asked to acknowledge at some point in the not very distant future that the copies are only being licensed to me. I will resist (in the perhaps vain hope that resistance is not futile). But I don’t see any legal reason why such “backporting” could not occur in the non-digital world. Sure, the norms are different, but the norms are changing, particularly for the “born digital” generation.

          • John Millington says

            ‘..the Unordered Merchandise Statute made it plain that the CDs at issue were owned outright by their recipients … The gnarly “Is it a sale, or is it a license?” debate was thereby avoidable.’

            Has anyone ever tried using the Unordered Merchandise Statute with the license-vs-copy bait’n’switch that sometimes occurs with mail-order software? If you look at the Amazon page for some proprietary software, say Mac OS X, everything about that page suggests Amazon is selling a copy of the software, with nary a hint that they’re sneakily substituting licenses at shipping time.

            Let’s say that, unknown to you, it turns out they’re really sending you a license. That is without question, not what you ordered, even if obscure laws cause it to be what is actually inside of the box. Could one successfully argue that the license which was sent despite the order for the copy, was unsolicited? And if unsolicited, it could be discarded.. err.. which leaves the copy in your possession (though I’m not sure who has title to it; you ordered it and paid for it, so it’s not unsolicited, so the statue doesn’t cause the copy to become yours).