Post by Christopher Newman, George Mason University.
Thanks to Jorge for his thought-provoking post. It’s a thrill to have smart people find your work worth picking up and using for something. Here are a few responsive thoughts of my own.
Because the term “license” encompasses a variety of interests comprising differing sets of jural relations, I think it’s important to specify what exactly is at stake in any given context when we ask “does a CNS amount to a license”? As Jorge notes, the actual posture of Meso makes it a bad vehicle for this (nevertheless important and interesting) question, because the argument there really has nothing to do with identifying the substantive content of the property interest (if any) that accrues to the obligee of a CNS. Meso is trying to obtain standing to enforce use restrictions that are purely contractual; bizarrely, they are in the position of arguing the equivalence of contract and property not to draw implications that sound in property, but simply to make a round trip back into contract.
Before addressing the question myself, let me just try to clarify my approach: A license at root is merely a (presumptively revocable) jural status that results from consent to the use of property. Consent can be given in many ways (including by making a contract), so I agree with Jorge that whether a license is granted can be inferred from the overall transaction. I term all licenses “property interests” in the broad sense that their effects sound in property law, however I don’t claim that all licenses necessarily have the attributes we often attribute to property interests such as immunity from revocation, transferability, etc. I call interests having these latter attributes “ownership interests.” Whether a license amounts to an ownership interest usually depends on whether applicable norms of property transfer have been followed. There are specific doctrinal exceptions, such as the irrevocable implied copyright licenses at issue in cases like Effects v. Cohen (discussed at length here), but usually I would want to say that a licensee’s interest amounts to an irrevocable ownership interest only if there was a written document making this clear.
So as to the relationship between a CNS and a license, I see several distinct issues: 1) Does entering into a CNS signify consent to use? 2) If so, does a CNS purport to grant that consent in such a way as to make the use-privilege irrevocable as a matter of property law? 3) Does a CNS trigger exhaustion when the obligee makes and sells items within the patent scope?
On the first question, one can certainly imagine a conceptually coherent regime in which a CNS is held not to give rise to a license. In this regime, we would say that by choosing to use the language of CNS rather than license, the patentee is signaling not consent, but forbearance—much as a ceasefire doesn’t imply that one surrenders or accepts the rightness of the other side’s casus belli. The obligee’s actions remain acts of infringement, but the patentee is contractually bound not to sue over them. This would imply that in the event of termination for breach, the patentee would be free to sue for infringement on all the obligee’s pre-termination uses still within the SOL. There may be transactional contexts in which affording this sort of extra leverage to the patentee would be useful to the parties; I don’t know whether anyone has ever attempted it.
FWIW, I don’t think the Supreme Court’s opinion in De Forest Radio Telephone Co. v. United States, 273 U.S. 236 (1927) need be read as ruling out the possibility of a CNS that does not give rise to a license. The case is oft-described as holding categorically that the one amounts to the other, but I think the Court’s reasoning is actually a bit more nuanced. They don’t announce any such categorical rule; rather, they hold that in the circumstances of that case, where the patentee not only promised not to sue, but actively assisted in the practice of the patent, there was no infringement. To me, De Forest is really about implied license and the principle that permission to use can be granted by conduct. The key point it makes is that “license” at root just means “permission.” I think its facts could (I do not say should) be distinguished from a case in which there were a bare CNS without any conduct on the patentee’s part to actively facilitate or further the obligee’s actions in practicing the patent. Note too that in De Forest, the only thing at issue was whether the government had acted under a license that prevented its use of the patent from infringing; not whether it had any ownership interests in any such license.
In TransCore, on the other hand, the Federal Circuit did expressly repudiate in fairly broad terms the idea that a CNS might not entail a license. It did so, however, in a context where the only practical matter riding on the use of the term “license” was whether the obligee’s actions triggered exhaustion. One could limit the holding to that context, treating it as a ruling about the scope of exhaustion doctrine rather than about the purported lexical meaning under federal law of the term “license,” which is how Meso wants to use it. In other words, there are two (nonexclusive) ways in which the holding in TransCore could be construed:
- Exhaustion is triggered whenever the patentee consents to distribution, and a CNS necessarily signifies such consent.
- Whether or not a CNS constitutes consent to distribute, the purpose of exhaustion doctrine requires that it be triggered by any transaction in which a patentee purports to enable a party to sell free of patent claims.
Note again that neither of these propositions says anything about whether a CNS results in an irrevocable ownership interest; they address only the effects of a CNS while it is in force. I think the opinion in TransCore reads much more like version 1), and while the discussion is broader than I’d like, I’m not at all sure that its thrust is wrong. If I promise not to sue you for X, with full knowledge that you intend to do X, it does feel like a stretch to construe that as something less than consent to your doing X. In other words, while a license is not itself a contract, a CNS may well be an example of one of those acts of contracting (I have other examples in my article) that have the ineluctable effect of giving rise to a license.
Alright, but what if the written contract is very explicit about the fact that the patentee is not consenting, that it still regards the obligee’s actions as infringing, and is merely binding itself contractually to refrain from suit under certain conditions? Even assuming we are willing to effectuate this attempt by the parties to define their own relations, once we face the exhaustion question we are now considering effects on third parties. My version 2) of TransCore above could be viewed as a sort of “numerus clausus” or “protection strategy” ala Merrill & Smith, in which we place structural limits on the ability of parties to choose what downstream legal implications they wish their transaction to have. I don’t have any position right now on whether I think the principles underlying exhaustion doctrine really do require that parties be prevented from using a CNS to avoid triggering it; I merely note that analytically this is a question separable from the one of whether a CNS constitutes consent that prevents any infringement from taking place (i.e., whether it results in a license). Of course, just because things are analytically separable doesn’t mean that it is functionally desirable to separate them. It may be that the Federal Circuit’s gut reaction is the most sensible one on both infringement and exhaustion, and that the class of beneficial transactions we might enable by recognizing a non-license CNS aren’t worth the candle.
There remains the question whether a CNS amounts to an ownership interest—i.e., is it irrevocable (so as to survive the patentee’s transfer of the patent or the obligee’s bankruptcy), is it transferable, and so on. Again, this question is not answered simply by concluding (assuming we do) that a CNS necessarily gives rise to a license. Not all licenses are ownership interests. Once we conclude that a CNS gives rise to a license, we will mostly likely construe the covenant not to sue as including a covenant not to revoke the license either. But if that’s all the transaction contains as an assurance against revocation of the license, I’d say there is no ownership interest, just a right under a contract that’s only good in personam against the obligor, subject to the same doctrine concerning assignability that we apply to other contracts. For there to be a conveyance of an ownership interest, I’d want to see the written instrument use language evincing such an intent on the part of the grantor, which would include at minimum language unequivocally consenting to use and identifying a temporal duration within which this permission is to be immune from revocation at will. It seems to me that on this dimension, the distinction between contractual obligation and ownership interest does likely serve useful purposes and ought to be preserved.
Thanks again to Jorge for spurring me to write about this issue, which has been sort of percolating in the back of my mind for a while now. I’d be very interested to hear from people who actually have practical experience negotiating and drafting patent transactions, to understand what people in the trenches actually wish to accomplish (or think they are accomplishing) when they choose to use the language of CNS rather than license.
Cogent analysis, Chris. The analytical-functional distinction is an important one that is often missed by the courts and practitioners, and there is much confusion in the case law on these issues.
There is a further analytical wrinkle here recognized in the literature–namely, unlike copyrights, patents provide no rights to use (i.e., Hohfeldian privileges) the patented invention to the patentholder. Rather, patents are merely exclusionary devices, providing negative rights to their owners to prevent others from making, using, selling, offering to sell, and in some cases, importing and exporting, the patented invention. For instance, if I patent an improvement to a pioneering invention that itself is patented, I may be precluded from using my own invention by the pioneering patent.
In this sense, a patentee does not have the Hohfeldian power to grant use rights to others over an invention, at least under the guise of the patent. One might immediately conclude from this observation that the only effect of a patent “license” is for the patentholder to credibly commit not to sue for infringement. In other words, a patent license by definition is a commitment not to sue, nothing else.
Yet, from a Hohfeldian perspective, the patentee has the power to change a duty not to infringe (i.e., make, use, sell, etc.) to a Hohfeldian privilege. This power, of course, is merely vis-a-vis the patentee (and does not guarantee any in rem use privileges), but it is a power nonetheless. So from an analytical perspective, even for patents, a license is different from a covenant not to sue.
The license is an exercise of a power (one sounding in property on your view) to change duties to privileges. A bald covenant not to sue places a contractual duty on the patentee without lifting any of the non-infringement duties binding the obligee.
Of course, shifting duties to privileges erases the ability to sue, but that alone does not convert a license into a covenant not to sue. So we are left at the same place analytically.
Whether one should treat a license and covenant as equivalent in a given situation would, as Hohfeld would say, sound in matters of justice and policy, not analysis.