Framing the Debate: The Question Presented
The fractured views of the world begin with the question presented, and reflect how different parties frame the debate in very different terms. Alice’s merits brief presents the question before the Court as “whether claims to computer implemented inventions…are patent-eligible.” Putting the question this way allows Alice to place its inventions and claims in the larger context of all computer-implemented inventions, the subtext being that if the Supreme Court holds that computer-implemented inventions are patent eligible—which is a fair bet—then Alice’s patents should be valid. Further, phrasing it this way allows Alice to distance itself from pure business method claims from the invalid claims in Bilski v. Kappos.
CLS does not see the question this way, and instead argues the question as “whether the courts below correctly concluded that all of the asserted claims are not patent eligible.” This phrasing transparently asserts that the Federal Circuit was correct all along, and allows CLS to argue that “to hold that the claims asserted by Alice are patent-eligible, the Court would have to disavow if not overrule both Bilski and Mayo,” obviously something the Court is not going to do. The Solicitor General, filing in support of CLS, also offered an alternative take on the question: whether Alice’s “computer-implemented claims for mitigating financial risk by using a third party intermediary” are patent eligible. This approach also subtly begs the question by reducing Alice’s patent claims to simply the abstract idea of risk mitigation itself.
This division is central to the themes and legal arguments of the parties and the amici. Alice and many of its supporters consider its inventions and claims as exemplary of application-specific software technology, no different in kind from other application-specific domains. By contrast, CLS, and many of its supporter including the Solicitor General, seek to lump Alice in with Bilski-like disfavored business method patents, and either directly or indirectly with the plague of the patent trolls. If, as Judge Rich told us, the “name of the game is the claim,” then the framing of the question before the Court is a way for the parties to subtly adjust the rules.
What’s an Abstract Idea?
The underlying problem in Section 101 generally and in this case particularly is what counts as an abstract idea, and what tests can be used to determine the eligibility of a patent claim. Many of the briefs, particularly on CLS’s side, argue along the lines that for “more than 150 years, this Court has held that abstract ideas” cannot be patented. Despite these exhortations, the Supreme Court has never provided an express articulation of what is and is not an abstract idea—if it had, there would not have been seven different opinions by the en banc Federal Circuit, Alice wouldn’t be before the Supreme Court, and we wouldn’t have two dozen briefs offering contradictory theories of the law. Thus, we expected to see various attempts by the parties and amici to define what is an abstract idea and offer tests for determining whether a claim is “merely” for an abstract idea, and thus ineligible by definition, or for something else.
Alice argues that the problem in Section 101 arises because there are two distinctly different meanings to the term “abstract idea.” First, in terms of the judicial exception to § 101, “this Court has used the term [“abstract idea”] as a synonym for a fundamental truth—a preexisting concept analogous to a law of nature, such as, for example, a mathematical formula or relationship.” Abstract ideas in this sense are not patent eligible because being “fundamental truths” they are not “new” in the language of § 101. Accordingly, “[i]f a claimed invention uses a computer simply to execute a mathematical formula [or other “fundamental truth”], the claim is ineligible, because limiting the claim to use of the formula on a computer, like limiting it to use in catalytic conversion or some other field, does not change the fact that the claim would preempt use of the formula as such.” In its reply brief, Alice emphasizes that the abstract idea or law of nature must be “actually recited on the face of the claim.”
According to Alice, the other meaning of abstract idea is as “a disembodied concept or pure mental conception,” and this form of abstract idea is not patent eligible for an entirely different reason: “A disembodied concept, standing alone, does not fall within any of the four statutory categories.” Alice argues that “the problems plaguing the lower courts in applying the abstract ideas exception can be traced directly to their failure to distinguish disembodied concepts from fundamental truths—and their application of the same analysis and rationale to both.” Alice argues that for this interpretation of abstract idea, “a computer-implemented invention that is either a system (in § 101 terms, a “machine”) or a process cannot be an abstract idea in the disembodied concept sense,” because “every computer operation involves using software to manipulate microscopic switches within computer hardware to bring about a particular result.”
Alice is generally correct about how the Supreme Court has understood the scope of the exceptions to § 101, and that the two distinct interpretations of abstract ideas have been incorrectly lumped into one bucket. The Court in Benson, Flook and Diehr believed that “mathematical algorithms” were ineligible because they are “like” a law of nature—which seems to be the quintessential type of so-called fundamental truths: “Reasoning that an algorithm, or mathematical formula, is like a law of nature, Benson applied the established rule that a law of nature cannot be the subject of a patent.” Flook, 437 U.S. 584, 589. Nowhere in Benson, Flook or Diehr did the court discuss the issue of “abstract ideas” directly; instead the focus was specifically on mathematical formulas, on the false assumption that they are, in all cases, scientific truths. It was only in Bilski that the Court extended Benson to the category of “abstract ideas”—without defining what an abstract idea is—thereby linking the “fundamental truth” rationale (itself based on an incorrect assumption that mathematical algorithms are always scientific truths) to the disembodied concept rationale, a maneuver that was neither contemplated nor required by Benson, Flook or Diehr.
The brief of Trading Technology (in conjunction with more than 40 other technology companies) supports Alice and takes a similar view, distinguishing “between inventions claiming a scientific truth or an abstract idea itself and inventions claiming the practical application of such an idea.” They argue that “[a]s stated in Benson, the proper focus of the exception is on ‘abstract intellectual concepts’—that is, purely mental impressions or processes,” and thus, “abstract idea in this general sense is any human thought that has not been reduced to some specific, concrete practice or application.” Their test is essentially one of whether the claim recites “the use of such tangible technology,” in which case it “cannot be considered impermissibly ‘abstract,’ at least without stripping that term of all meaning.”
The approaches of Alice and Trading Technologies offer the courts a way to distinguish between two different questions: whether a patent claim preempts a scientific or other “fundamental” truth (the real concern motivating Benson, Flook, Diehr and Mayo), or whether a claim is utilizing the standard and necessary tools of abstraction to define process or machine steps (the mistaken concern of certain members of the Federal Circuit). The courts are sensitive to the first issue but generally do not understand the second, and typically conflate the two. As we explained in our brief for Advanced Biological Laboratories (ABL), abstract ideas must be distinguished from “abstractions,” which are generalizations—a term or definition that identifies the principal aspects or features of the concept that are relevant to a given context, while removing features that are not important. This is because patent claims necessarily use abstractions to define inventions—whether mechanical such as “container” or “support member” or software based, such as “database” or “user profile” or “shadow record”—but that does not make them “abstract ideas.” “[T]he Patent Act requires that the claims themselves set forth the limits of the patent grant, but also because persons of ordinary skill in the art rarely would confine their definitions of terms to the exact representations depicted in the embodiments.” Phillips v. AWH Corp., 415 F.3d 1303, 1323 (Fed. Cir. 2005) (en banc). Thus, the only way to draft a claim so that it is not precisely a “picture claim” of the exact representations is to use more general nouns and verbs to describe the invention. This is no more “clever drafting” than choosing the right words in a contract, license or a will. Recognizing that the use of abstractions in claims is entirely proper and conventional allows a court to focus on the real issue: Does the claim actually cover human thinking and judgment (the focus of the original mental steps doctrine, as explained in the amicus brief for Ronald M. Benrey) or a scientific truth (the core concern of Benson, Flook and Diehr)?
CLS argues that “an abstract idea is a building block of technology (Benson), of innovation (Flook,), and of the economy (Bilski),” and that the judicial exceptions to Section 101 all ensure “that the patent system cannot withdraw from public use the building blocks of innovation and advancement in the useful arts.” The “building block” argument is repeated in various forms by the Solicitor General, American Antitrust Institute (“More fundamentally, patents that preempt abstract intellectual ideas harm all competition by interfering with access to the building blocks of competitive innovation”), Public Knowledge, Prof. Robin Feldman, and the American Civil Liberties Union. CLS asserts that “Alice’s claims recite the basic principle of intermediated settlement, a building block of economics and risk management just like the concept of hedging in Bilski.” CLS argues that the Court “has never suggested that mathematical formulas possess talismanic significance in the abstract idea inquiry,” and that three sitting justices of the Court have stated that “a patent may be ineligible for claiming an ‘abstract idea’ even if it does not claim a ‘fundamental truth,’” citing J. Stevens’ concurring opinion in Bilski.
It’s a bit much that CLS chides Alice’s argument for a specific definition of abstract idea as being merely “talismanic,” since CLS repeats the “building block” phrase ten times (far more than any of its supporters) but never explains what makes something a building block. The “building block” metaphor does not tell us anything useful about what is or is not an abstract idea. How would a court (or a patent examiner) distinguish an ineligible “building block” of “innovation” from a patent-eligible specific application? The early patents on integrated circuits—DRAM, microprocessors, GPS—were all for technologies that have become building blocks, but that was not by any means understood at the time they were filed, and even it were recognized, that would not have made the technologies “abstract ideas.” Whether an invention is a “building block” can only be determined years after its creation and patenting, when the industry and market have adopted the technology. In 1968, an engineer at IBM’s Advanced Computing Systems Division is reputed to have said about the microchip, “But what…is it good for?” And the “innovation” building block absolutely fails to distinguish Flook from Diehr. In Flook, the court assumed that the equation was novel (i.e., innovative), and in Diehr it was admitted all of the elements were conventional, yet Flook was ineligible and Diehr was eligible. With no explanation of how to know whether a given claim would cover a building block or not, CLS offers yet another “I know it when I see it” test, precisely the kind of subjective test that has led to the current problems.
The overall thrust of the Solicitor’s brief is the patent law should be confined to its “traditional domain of technological, scientific, and industrial progress.” Accordingly, abstract ideas are “non-concrete ideas” and “claims to non-technological methods of organizing human activity,” which not surprisingly excludes business methods. However, these two ideas conflict with each other. One can define a claim for a business method that requires very concrete operations in the real world. For example, consider a claim for operating a restaurant that includes receiving orders from telephone boxes positioned in numbered parking stalls, and the delivering of the food by waiters on rollerskates. This would cover only physical operations in the real world, not thinking, speaking or merely “calculating.” Furthermore, the “traditional domain” analysis is meaningless. If patents were limited to the traditional types of “technology” known in 1952 (or worse, 1792), we would by definition not extend patent protection to entirely new and different fields, such as biotechnology. Who decides what is “traditional”? As Alice’s reply brief points out, the Solicitor’s argument is essentially a technological arts test, which the Supreme Court rejected in Bilski and should likewise reject here.
Our brief for ABL argues that under Benson, Diehr and Mayo, the core concern of § 101 is preemption: whether the claims preempt “all practical applications” of a law of nature, abstract idea or natural phenomenon. Therefore, since the emphasis is on what are “practical applications” and what happens in the “real world,” the only objective way to address the question is through the use of POSITA, the person of ordinary skill in the art. Reliance on POSITA is necessary to avoid the subjectivity that now dominates the abstract idea exception. This conforms the eligibility analysis to how obviousness is determined, a question of law based on the Graham factors, as well as claim construction, enablement, written description and doctrine of equivalents infringement, all of which use POSITA. The perspective of POSITA would be used to understand the scope of the claim itself (not a plain English gist), and “whether the claim covers all and every practical application of the abstract idea in the real world, not merely some applications.” Real evidence, not mere attorney argument or judicial speculation, would be required of a patent defendant to show that all practical applications are blocked in order to hold a patent claim ineligible. On the other hand, if there is at least one practical application that does not come within the scope of the claim, then the claim is eligible because it does not “preempt” all applications of the idea in the “real world.” The primary objection to this approach is that it makes the determination of Section 101 an expensive question of fact, since expert testimony and perhaps even claim construction would be necessary. That may be true, but then it was the Supreme Court that set forth preemption as the concern, not Congress, and preemption can only be addressed in terms of facts, not speculations.
Claiming How, Not What
A number of amici argue that patent claims are for abstract ideas if they merely claim the desired result using a computer—essentially “apply a computer” to abstract idea—and instead must claim “how” something is implemented. For example, Google argues that Alice’s claims fail because they “do not recite a specific way of implementing financial intermediation,” “they impose no limit whatsoever on how a computer program or specialized hardware accomplishes the financial-intermediation function” and they “set forth nothing inventive about how to implement the abstract idea of financial intermediation on a computer.” Similarly, RichRelevance argues that “for a computer based limitation to make an otherwise abstract idea patentable, the limitation must be addressed to the how and not just the what.” The BSA tells us that “taking an abstract idea, saying ‘apply it on a computer,’ and adding nothing more is not enough to overcome the implicit exceptions to Section 101.” Microsoft takes this approach as well, arguing that the claim in Flook “did not recite a sufficiently specific method for how to [use the equation]…. The claim did not, for example, explain how to select the appropriate margin of safety, the weighting factor, or any of the variables.”
This suggestion seems rational on its face, but the problem is there is no agreement or guidance on what constitutes a sufficient limitation of “how” versus “what,” since the determination is very context dependent. For example, consider a claim limitation that recites using a computer to perform a “Fast Fourier Transform” on a specifically defined type of audio signal. On the one hand, this would be a “how” limitation if the invention could use any type of Fourier analysis, since there are a number of different ways to compute a Fourier transform and the FFT is only one of them, and its implementation is well known. On the other hand, this could be considered a “what” limitation since there are many different ways to actually implement the FFT itself (and indeed, there are many patents on different implementations of the FFT alone). Whether this is a how or a what limitation all depends on the specific details of the specification and the nature of the invention. Thus, a rigorous application of the “what”/“how” approach would require expert testimony, if not claim construction to determine whether a limitation is a what or a how—precisely what Google and many of the amici supporting CLS say is unnecessary and undesirable. Even in this very case, one man’s how is another man’s what: The BSA argues that “all [Alice’s patent claim] steps do is recite how one can implement intermediated settlement.” Apparently, Alice’s patent claims were not “how” enough for the BSA—and so we’re back to square one.
Use of General Purpose Computer and the Hunt for the Inventive Concept
Another key issue is whether the recitation of a general purpose computer, or its components such as a processor, memory, storage device, etc., contributes anything to patent eligibility. Alice’s treatment is straightforward: software is the “instructions that make computers work” and ultimately, therefore, every “operation performed by a modern computer, however complex, is the result of switching transistors on or off in different combinations” under software control. “A claim that recites specific use of a computer therefore cannot be an abstract idea in the disembodied concept sense of the term,” because “systems that contain a computer, and methods that are implemented with a computer, necessarily exist in the physical, rather than purely conceptual, realm.” Similarly, Microsoft, even though it supports CLS, explains that “simply put, it is software that transforms the single set of circuits in a general-purpose computer into a word processor, camera, telephone, video player, or gaming device.” The IEEE-USA correctly explains that a “general purpose computer” does not mean a personal computer with a browser, a word processor or other software applications—a common assumption by both the courts and litigants. Instead, the term is about the hardware architecture: “used in its broadest sense, as referring to the core internals of how a device operates, and not to the externals, as a user or consumer may experience them.”
Alice argues that because a computer is essential to all its claims, and not merely “applying” a computer to the concept of third party intermediation, it does not matter whether its invention is implemented on a general purpose or special purpose computer, because either way a physical computer is a real machine, not an abstract idea. Alice argues that the so-called “inventive concept” notion set forth in Flook and Mayo only requires “steps that go beyond such essential ancillary steps” that “anyone using the law or truth necessarily must employ.” Alice gives specific examples of how its claims require very particular computer-based operations that are beyond the necessary elements of third party risk intermediation. This section of Alice’s brief, grounded in the details of specific claims, provides the best argument that it is not merely “applying” a computer to the concept of risk mitigation.
CLS and many of its supporters argue the opposite. CLS argues that “the recitation of a computer (or computing functions) neither guarantees nor precludes eligibility,” because “[a] computer, like any other utensil of the relevant art (be it a hammer to the carpenter, a sextant to the navigator, or a syringe to the doctor), neither adds to nor subtracts from eligibility when used conventionally,” under Mayo. CLS argues that “[c]alculation is another of the most general, basic, and routine uses of a computer.” CLS’s core argument is that using a computer does not provide an “inventive concept” to an underlying abstract idea. Similarly, Red Hat argues that “merely adding a general purpose computer to an otherwise unpatentable algorithm is far from an inventive concept.” RichRelevance, PublicKnowledge and Software Freedom Law Center make similar arguments.
Again, this is one of those arguments that seems sensible but falls apart upon close scrutiny. In doctrinal terms, this is simply Flook’s “point of novelty” approach to patent eligibility, which identifies that novel contribution of the inventor, apart from any so-called conventional elements, and decides eligibility on that particular feature or combination of features. The dissent in Flook vigorously disagreed with this approach, stating: “it strikes what seems to me an equally damaging blow at basic principles of patent law by importing into its inquiry under 35 U.S.C. § 101 the criteria of novelty and inventiveness.” Flook, 437 U.S. at 600 (Stewart, J., dissenting).
The Flook dissent became the majority in Diehr, and plainly dispatched this theory. “The Court in Diehr rejected the ‘point of novelty’ analysis saying ‘the “novelty” of any element or steps in a process…is of no relevance in determining whether the subject matter of a claim falls within the § 101 categories of possibly patentable subject matter.’” In re Taner, 681 F.2d. 787, 791 (C.C.P.A. 1982). Judge Michel’s amicus brief goes even further, arguing that the Diehr “overruled…the aberrational Parker v. Flook, 437 U.S. 584 (1978).” But good things never last, and in Mayo, the Court specifically considered the “conventionality,” that is novelty of individual steps, directly contradicting Diehr: “this step tells doctors to engage in well understood, routine, conventional activity previously engaged in by scientists who work in the field.” While Alice argues that Mayo did not limit Diehr, and that Diehr’s approach to the claim as a whole is still intact, the Court’s continued heavy emphasis on Flook, and recasting of Diehr as allegedly having some unconventional elements (it did not) which contributed to its patent eligibility, suggests otherwise.
Second, a good test of any rule of patent eligibility is whether it applies equally well in other fields outside of computers. Here, CLS’s approach fails. Consider a machine made up of novel and complex arrangement of gears, levers, motors, springs, cams and other off-the-shelf mechanical parts, all providing their “normal” functions for the sole purpose of cracking nuts; there is no “innovative” hardware other than the entire machine itself. No one would argue that this machine is not patent eligible because it was made up of entirely conventional elements all performing the “well-understood, routine, conventional activity” of nut cracking, or that the use of motors, gears, etc., added nothing to the eligibility of the machine because “providing motive power” is the “the most general, basic and routine use” of a motor. The same analysis can be applied to electronics (e.g., consider the enormous number of patents with conventional and routine uses of transistors, logic gates, amplifiers or circuits) or chemistry. These rules are obviously wrong when applied to the mechanical, electrical and other arts, and they do not become valid when applied to software and computers.
While CLS and most of its supporters simply deny any contribution of a general purpose computer, the Solicitor General offers up a set of factors that “should guide consideration of the computer’s role within the claim,” including whether the “computer constitutes insignificant extra-solution activity”; “whether specific, unconventional computer equipment” is required; whether the claim “manipulates particular data in specific and useful ways”; “whether the computer is recited solely for its generic functions, including automating and synchronizing” and “whether the invention involves an improvement in the functioning of the computer…by making it more efficient.” The Solicitor General says that the “factors will tend to exclude business methods that do no more than recite a general purpose computer in order to facilitate transactions or other endeavors.” However, most of these factors are the same generalizations that historically have not provided any clear results, but instead devolve into a subjective analysis by the court. Further, to the extent these factors are to be evaluated seriously, with more than superficial arguments, they would require actual evidence such as expert testimony to establish. For example, apart from attorney argument, how is a court supposed to know whether an invention improves the functioning of a computer? Again, we’re back to expert testimony and thus the curse of expensive litigation. No easy out here.
In Part II, I will address additional issues being considered in the Alice v. CLS oral arguments.