Bilski and (Thankfully) Continued Uncertainty in Patentable Subject Matter

Thanks again to the folks at Prawfs for allowing me to loiter long enough to post about today's opinion in Bilski v. Kappos. The title of this post is nearly identical to the post I wrote when the Federal Circuit ruled (see that post if you have no idea what Bilski is about) - with the addition of thankfulness. Why should we be thankful? I'll discuss three reasons below:

1. My general prediction about the outcome came true

2. My general view -- that all attempts to define patentable subject matter outside the statutory categories fail -- is vindicated

3. The result could have been worse for invention and innovation

I discuss each in turn below the fold. 

First, the short version of the Court's ruling:

A. The claims at issue here are abstract ideas, and thus not patentable (9 Justices). This is true even if "token" activities are undertaken outside the mathematical equation. (9? Justices)

B. Limiting patentable processes to only those that use a machine or transform matter is a useful test but is not the only test. (5 Justices - Kennedy, Roberts, Scalia, Thomas, Alito)

C. Section 100(b) defines process, and nothing in that definition excludes business methods per se or makes machine or transformation the only test. (5 Justices) Section 273 (which provides a defense for potential infringers of business methods) is evidence that business methods cannot be read out of the statute.

D. Just because business methods patents were not issued early in history, that does not mean that they should not issue now with changing times and technology (4 Justices - Kennedy, Roberts, Thomas, Alito)

E. In addition to being unpatentable as an abstract idea, the term "process" should not be read in the ordinary sense, but instead should be interpreted in light of history, context, and patent policy goals to exclude business methods, despite Section 273. (4 Justices - Stevens, Breyer, Ginsburg, Sotomayor) Interestingly, the concurrence implies that this is a statutory, not constitutional question.

F. The majority explicitly gives no further guidance about identifying an unpatentable abstract process.

Now, on to some analysis...

My Prediction

Just about everybody predicted that the patent would be invalidated 9-0. However, few focused on the statutory outcome. In my discussion about oral argument, I wrote the following:

Fifth, to that end, I was extremely disappointed that -- in an hour of argument about the meaning of process -- there was not one single mention of 35 USC 100(b), which states the statutory definition of process. I would think that a court filled with textualists would want to know what the text of the statute says. Indeed, the definition, which defines a process as a "new use of a machine" would have cleared up much of the confusion about why a method might be unpatentable in the abstract, but patentable in a machine - when you program software to do something different, you are programming a new use of a machine. Even if you ignore the computer as "insignificant post-solution activity," the new use falls under the statutory definition.

In the end, ignoring 100(b) may be the largest problem with the machine or transformation test, which requires that all processes be tied to a machine or transform matter (and I do mean all - the Federal Circuit recently applied it to a method of treating patients with drugs). There are many, many processes from the inception of process patenting until the present, that are new uses of machines, matter, etc., but that are not tied to machines or "transform" matter. Iron working, glass blowing, medical diagnostics, new uses for old drugs, new uses for old machines, etc., are all historically patentable despite not satisfying the machine or transformation test. The fact that no one even mentioned it one way or the other is pretty surprising.

And, in fact, that's how the majority ruled. The court said:

Section 101 similarly precludes the broad contention that the term “process” categorically excludes business methods. The term “method,” which is within §100(b)’s definition of “process,” at least as a textual matter and before consulting other limitations in the Patent Act and this Court’s precedents, may include at least some methods of doing business. See, e.g., Webster’s New International Dictionary 1548 (2d ed. 1954) (defining “method” as“[a]n orderly procedure or process . . . regular way or manner of doing anything; hence, a set form of procedure adopted in investigation or instruction”). The Court is unaware of any argument that the “‘ordinary, contemporary, common meaning,’” Diehr, supra, at 182, of “method” excludes business methods.

Note the irony that Justice Scalia the originalist voted in favor of broad subject matter (though he did not join the opinion's discussion of evolving times, while "living constitutionalists" fell back on history to limit subject matter in Justice Stevens's concurrence. (Side note, I am working on a paper that analyzes historical patents to test the varying assertions about early method patenting).

The Undefinable Standard

In my article Everything is Patentable, I make clear that patentable subject matter jurisprudence is a mess that can never be cleaned up.  While this opinion doesn't make the mess any worse (as discussed below), it hardly cleans it up. It seems like the Court could barely muster a majority, and it may well be that the majority shifted at some point. A decision unanimous in outcome but yielding a 5-4 (with 1 partial abstention) split is pretty good evidence that defining a clear standard is hopeless.

Indeed, though the concurrence doesn't provide a great solution, Justice Stevens rightly points out that the majority opinion does nothing to help identify abstract ideas. Indeed, as the concurrence points out, the Bilski claims are far more complex and concrete than the simple and "abstract" mathematical formulas rejected in earlier Supreme Court cases. Nonetheless, the concurrence doesn't exactly provide a clear solution by outlawing business methods patents without defining what a business method might be. Finally, the Court does not identify what postsolution activity is sufficient to make a claim non-abstract.

Instead, the better approach - the one I advocated in this case and advocate for the future, is to read process exactly as it is written - as the Court does - but to do away with the supposed historical limitation on abstractness. Instead, other requirements are more than sufficient if rigorously implied (and indeed, were considered by the Court here, even if implicitly.)

First, the most basic claims here lacked practical utility - they didn't do anything. Instead it is a simple concept that does not necessarily lead to any practical result. (The Court calls this the "concept" of hedging, and in fact the claim is for no more than that.) Second, the claims were obvious. (The Court discusses the patent's "use of well-known random analysis techniques.") Third, the broader claims were not enabled or described. The patent specification describes a particular application with a particular formula tied to a very particularize set of data, but did not describe how that formula might apply to fields other than those listed.

Perhaps in light of the continued uncertainty in patentable subject matter, the PTO and Federal Circuit will adopt my proposal to reject unworthy patents based on rigorous application of patent standards rather than trying to make new rules that simply cannot be defined.

Impacts on Business

But enough about me. What about the rest of the world? What impact will this case have? It's too soon to tell, but I suspect not too much outside of the increased cost and uncertainty associated with prosecuting patents relating to intangible subject matter.

The good news (though I am certain there are people - including four Justices, apparently - who would disagree with me):

    a. Software patents seem to be clearly patentable unless they are too simple. Software patents were certainly not rejected - this was a big fear.

    b. Business methods (less software and more human activity) might be patentable, but they also might not. I suppose this might appease all sides. Aggressive inventors can seek patents, and opponents can continue to argue that such patents are too abstract. This might be more costly to inventors than a clear rule that allowed all patents and more costly to potential defendants, but certainly less costly than the alternative for either side.

I suspect that how you come out on the result on innovation and development depends on how you view the role of patents as it relates to these types of patents. If you think that patents are critical to innovation (I tend to think so), then this ruling is a bad thing only to the extent that it increases patenting costs. If you think patents are unnecessary or even harmful for innovation, then you are probably not thrilled with this decision, but it could have been worse - the Bilski claims might have been allowed. Opponents are no worse off than they were pre-Bilski.

In the end, this decision seems like a draw - it doesn't expand eligibility, but it also puts a halt to the trend of limiting eligibility. It appears that this was exactly the intent. The Court did not want to foreclose the debate, and - in my view - did the best it could short of adopting my point of view. Justice Kennedy wrote:

Rather than adopting categorical rules that might have wide-ranging and unforeseen impacts, the Court resolves this case narrowly...

The Court's discussion seems to imply that it simply does not know what effect a broad limitation might have, and thus leaves the determination up to a case by case review. The result is not surprising given recent Court rulings in other patent cases that also advocate a case by case analysis.