Frischmann Predicts Prometheus
Thanks so much to Frank, Danielle, and Deven for inviting me to participate in this symposium. It’s a great pleasure to discuss my colleague Brett Frischmann’s timely, engaging, and important book about infrastructure.
I’m going to focus my comments on Frischmann’s theory of intellectual infrastructure and how it relates to the structure of intellectual property law. Just a few days after the release of Infrastructure, the Supreme Court handed down its decision in Mayo Collaborative Services v. Prometheus Laboratories. That case presented the question whether certain diagnostic claims were within the scope of patentable subject matter under section 101 of the Patent Act. The Court held that they were not, in a manner that is strikingly consistent with Frischmann’s theory. Hence the title of my post. But Frischmann’s theory may also go a long way toward bringing some order to an area of patent law that has long been confused.
Let’s start with the concept of intellectual infrastructure. Frischmann explains that intellectual or cultural resources can be infrastructural in the same manner as physical goods. So long as the resource is a “nonrival input into a wide variety of outputs” (275), it satisfies the characteristics of infrastructure that Frischmann so richly describes. In turn, that suggests that the case for managing the resource as a commons is strong. Frischmann then explains how this concept applies to ideas. Ideas, he writes, often are infrastructure (subject to a number of complications that I’ll put to the side). So in his view, intellectual property should protect implementations of ideas but not the ideas themselves (286). To sort one from the other, Frischmann turns to the concept of abstraction in copyright law and argues that patent law should follow a similar path.
Now consider Prometheus. The inventors in that case discovered a correlation between the effectiveness of a drug and the amount of certain metabolites of that drug in a patient’s blood. Their patent claimed a method of optimizing the dosage of the drug based on that correlation. The method was simple: (1) administer the drug; (2) determine the amount of metabolites in the patient’s blood; (3) make an inference about drug dosage based on the correlation. Doctrinally, the question before the Court was whether this amounted to a claim on a “natural law” – the correlation between drug dosage and metabolism that happens in the human body – which would be unpatentable under a long-standing exception to the scope of patentable subject matter, or a patentable application of that law.
Substitute “idea” for “natural law” and the analogy to Frischmann’s analysis becomes clear. Indeed, there is no reason why a “natural law” cannot be an “idea” as Frischmann describes it. A natural law is the articulation of a principle; whether that principle is an infrastructural idea turns on its characteristics, not its origins. In Prometheus, the Court, with Justice Breyer authoring the unanimous decision, held that the patent claimed the natural law itself and not an application of that law. The Court held that the the patent claims did not “add enough to their statements of the correlations . . . to qualify as patent-eligible processes that apply” those correlations (slip op. 8). The “administering” and “determining” steps, in the Court’s view, comprised “well-understood, routine, conventional activity already engaged in by the scientific community” and so were “not sufficient to transform unpatentable natural correlations into patentable application of those regularities” (slip op. 11). Or, in Frischmann’s terms, the patent claims did not amount to an implementation of an idea; they claimed the idea itself.
The rationale underlying the Court’s decision also resonates with Frischmann’s argument. Justice Breyer cast the decision in expressly functional terms. The danger, he wrote, is that because natural laws are “the basic tools of scientific and technological work,” patents on natural laws “foreclose more future invention than the underlying discovery could reasonably justify” (slip op. 17). So too, the rationale for managing infrastructural resources as commons turns on the demand-side benefits: the spillovers and externalities that non-rival consumption of infrastructure resources enables. In this way, Prometheus is quite consistent with Frischmann’s injunction against propertizing ideas.
Frischmann’s theory also suggests some provocative solutions to problems that have long plagued the doctrine of patentable subject matter. Section 101 of the Patent Act defines as patentable any “new and useful process, machine, manufacture, or composition of matter.” But the Supreme Court has long carved out from those categories several broad exceptions. They are usually described by the terms “laws of nature, natural phenomena, and abstract ideas,” but at times appear also to encompass “products of nature,” “mental processes,” “mathematical formulae,” “algorithms,” and the like. Infrastructure may offer a unifying principle for choosing to exclude these things from the scope of patentable subject matter. Frischmann writes that “[t]he Supreme Court should stop referring to abstract ideas” in its patentable subject matter jurisprudence and should instead “make clear that ideas are not patentable” (300). Viewed through the lens of infrastructure, Prometheus is a step in this direction. The correlations at issue in that case were actually quite narrow. At the very least, they were not “abstract” in the sense that the “concept of hedging risk” found to be unpatentable in Bilski v. Kappos was abstract. Yet the correlations in Prometheus and the concept of hedging in Bilski are both ideas. And they are ideas that have the characteristics of infrastructure — nonrival inputs into a wide range of outputs. The Court’s functional analysis suggests that the problem in both cases was that granting exclusive rights to ideas that have the characteristics of infrastructure might foreclose future innovation.
I would go even further. The next big patentable subject matter case is likely going to be the Public Patent Foundation’s ongoing challenge to gene patents held by Myriad Genetics. Doctrinally, the question raised in that case — whether the patents claim “products of nature” – is different than the question whether a patent claims an idea. But to the extent that genes can be characterized as infrastructure — and I think there is a reasonable case that they can be — the functional analysis described above should apply. The exclusion of products of nature from the scope of patentable subject matter may therefore be consistent with the exclusion of ideas from the scope of patentable subject matter.
It has also been a matter of debate among academics and practitioners whether and to what extent the inquiry into patentable subject matter serves a function different from other requirements of patentability like novelty or nonobviousness. To the extent that infrastructural resources ought to lie beyond the scope of patent protection for functional reasons, the patentable subject matter screen will capture this insight in a way that other patentability doctrines will not. But this is a topic for further exploration.
Frischmann’s book is an important contribution across a wide range of fields. It is a terrific achievement and I think its influence will continue to grow as we grapple with the implications of its analysis.