Professor Robert P. Merges
University of California at Berkeley School of Law
January 4, 2012

1.  Introduction

Now that In re Bilski has been decided, there is an enormous amount of speculation about the impact on patent applicants, litigants, and other participants in the patent system.  Most of the commentary is concerned with the holding in Bilski, and how this holding will be applied by courts and the Patent Office, and ultimately, the effect on inventors and those who hold and seek patents.

I am interested in another approach: how to minimize the cost and confusion that accompany a review of patents for Sec. 101 subject matter eligibility.  To this end I argue that the Sec. 101 issue of Bilski be considered only when doing so is absolutely necessary to determine the validity of a claim or claims in a patent.  Under this approach any claim that can be invalidated under one of the less controversial and less complex requirements for patentability – Secs. 102, 103, and 112, for instance – ought to be disposed of without considering subject matter patentability.  The Bilski issue, in other words, should be avoided wherever it is not strictly necessary.

To begin, it is inaccurate to visualize patentability as a stepwise series of tests applied to a single “invention.”  It is not true for example that “invention X” passes Sec. 101 and should thus proceed in logical sequence to be tested under Sec. 102.  One claim growing out of inventive insight X might present no Sec. 101 problems at all, yet another claim in the same patent application might raise a difficult issue under this provision.  Each claim, being a unique slice of the overall inventive insight, ought to be considered on its own terms, and in whatever order makes the most sense.  Put another way, the mental model of a stepwise sequence of patentability determinations overlooks the highly granular nature in which different slices of the inventive concept are presented for validity testing.

When a claim fails to pass muster under any single test of validity, that claim should be invalidated.  No further tests should be applied.  Beyond that point, any expenditure of resources on validity questions is inefficient.  Pragmatic considerations enter at this point.  Issues of cost and complexity are perfectly appropriate in determining the actual sequence in which validity tests are applied with respect to any particular patent claim.  The non-linearity of patent validity tests, together with the principle of efficient administration, yields a simple rule: Start with chain links that are, in general, easiest and cheapest to test, and when the chain fails, stop the process.  That way, the costliest and most complex doctrines – the trickiest “links in the chain” – are often avoided, and in any event are put off until later.  This means in practice that Sec. 101 will often be avoided, both at the Patent Office and in the courts.  This can be justified not only on efficiency grounds, but also by analogy to the Supreme Court rule of avoidance, which holds that the Court should avoid putting its legitimacy at stake by resolving cases on non-constitutional grounds whenever that is possible.  Legitimacy can be seen here as a special form of institutional currency – which the Court chooses to conserve whenever possible.

This leads to a very simple approach.  First, the Patent Office ought to start with Sec. 102 and then move on to Sec. 112 and Sec. 103.  Next, the courts should consider patent validity in a similar order, making exceptions or changing the order of doctrines as dictated by pragmatic considerations.  In all cases, the complex and costly process of deciding whether a claim presents patentable subject matter under Sec. 101 should be deferred until very late in the process.  Courts should in effect hold off on the difficult task of evaluating clams under Sec. 101 – ideally deploying the full Sec. 101 analysis only when that is essential, i.e., when a claim passes muster under the other validity doctrines.

2.  Patentable Subject Matter: Bilski and Beyond

The issue of patentable subject matter was given fresh prominence when the Supreme Court handed down its decision in Bilski v. Kappos in June 2010.2  In Bilski, the Supreme Court sustained the invalidity of claims to a method of hedging the risk price changes in commodity markets.  The rationale behind the ruling was that Bilski’s “claims are not patentable processes because they are attempts to patent abstract ideas.”3  The Court rooted the exclusion of “abstract ideas” in a long line of precedent stretching back to the 19th century, which, the opinion said, confers legitimacy despite the absence of any definition or even mention of the term “abstract” in the text of the Patent Act.4  The primary source of the Court’s holding was three earlier opinions, all touching on the patentability of software-related claims.  In explaining its ruling, the Court explicitly rejected several comprehensive standards for patentability that had been proposed by academics and practitioners.  These more comprehensive standards were attempts to create order in the complex area of patentable subject matter.  Instead of relying on these, however, the Court chose to rest its holding primarily on its case law from the decade 1970-1980:

Rather than adopting categorical rules that might have wide-ranging and unforeseen impacts, the Court resolves this case narrowly on the basis of this Court’s decisions in Benson, Flook, and Diehr, which show that petitioners’ claims are not patentable processes because they are attempts to patent abstract ideas.5

While it is understandable that the Court would choose to reject “atextual” tests, the organizing impulse behind these proposed standards was that they would transcend the Court’s software cases, which were widely thought to be too fact-specific and outdated to serve as an effective source for a workable test.

One problem is the lack of guidance in the words of the statute.  The literal terms of Sec. 101 – “process, machine, manufacture and composition of matter” – are so open-ended that they provide very little traction for a court that feels the need to rein in the scope of patentable subject matter.  This is where the historical non-textual exclusions from patentable subject matter come in.  Long ago a sort of gloss was placed on the Patent Act: Despite the statute’s open-ended terms,6 the Court decreed that “laws of nature, physical phenomena, and abstract ideas” were implicitly excluded from the realm of what is patentable.7  Because none of these exclusions has ever been defined in legislation, their contours have taken shape over the years in the traditional way of common law principles.  The chief advantage of this approach is well understood: flexibility and adaptability.  By the same token, the lack of a comprehensive definition can also create uncertainty – especially where the volume of case law is relatively low.  When cases are few and far between, those who must rely on a common law rule are placed in the difficult position of reading the specific facts of the few decided cases for whatever clues and signals they might throw off.

This is precisely the situation that now confronts patent lawyers, inventors, and everyone else who is interested in the scope of patentable subject matter under Sec. 101.  The hard kernel of legal authority after Bilski is this: The risk-hedging claims in that case were too abstract.  Therefore, claims that are similar enough to those at issue in Bilski will also be unpatentable.  It is easy enough to state this general principle, but – because of the great variety and complexity of patent claims that will be subject to the post-Bilski standard – very difficult in practice to apply it to a specific case.  The virtue of the Bilski opinion is that it does not tie the patent system down to a restrictive test for patentability.  But there is a matching vice: a bedeviling lack of guidance over what patent applicants and patentees can expect when Sec. 101 is applied to a specific patent claim.

If the volume of Supreme Court case law were higher, this guidance might emerge relatively quickly.  But typically, the Court takes few cases in this area.  Consider also that the Court was deeply split (5-4) in the Bilski case itself.  Because of this, the Court might well want to avoid the issue for some period of time.  This probably leaves the Federal Circuit to sort through this area on its own for a period of time.  What is the best course of action for them to take?  In the next section, we argue that where possible the Federal Circuit, as well as the district courts and the PTO, ought to simply avoid the Sec. 101 issue whenever possible.

3.  Lexical Priority and Patentability Doctrines

It is often implicitly assumed that when a Sec. 101 issue arises in a patent case, the relevant decisionmaker should deal with it first, before other requirements for patentability are even considered.8  In this section, I explain the appeal of the broader idea that the various sections of the Patent Act are meant to be considered seriatim.  I show that although there are a few sections in which lexical priority may imply logical priority, in general this is not true of the Patent Act.

The idea that the patent validity provisions are meant to be applied in the order in which they appear in the Patent Act has some appeal.  If we take the linguistic details of our statutes seriously, why then not take the lexical ordering of the statutory provisions seriously as well?  Sec. 101 appears before Secs. 102 and 103, after all – and surely there is some good reason for this.  Who would advocate skipping willy-nilly around the statute when a complex patent case is being decided?  Surely the result would be something like chaos.

To begin, it is important to see that the various sections of the Patent Act certainly do proceed in a logical order.  It is certainly not illogical to start the substantive requirements of patent law with patentable subject matter (after the administrative details of the very first sections of the Patent Act).  It makes sense in a way to begin with a general definition of the types of things that are patentable, and then proceed to the question whether a particular invention is novel; then nonobvious; then enabled; and so on.  But we would argue that although the statute unfolds in a logical order, this is not the only order that might make sense.  And, most importantly, it is not essential to apply the statutory requirements in the precise order they are set out in the Patent Act.

It is black letter law, for example, that novelty under Sec. 102 is determined strictly on the basis of all the elements recited in a given claim.  But if one or more of those elements cannot be pinned down, it makes sense to start with a consideration of the claim itself.  Thus, for example, the definiteness provision of Sec. 112, para. 2 might logically be considered first in some cases.9  Other claim-related requirements must also be considered first sometimes, before non-claim-related validity doctrines can be applied.  So, for example, a means-plus-function claim that recites only a single means – invalid under Sec. 112, para. 1 – need not first pass through the analysis required under Sec. 102 and Sec. 10310

4.  Conclusion: A Pragmatic Approach

Put simply, what I have been saying is this: There is no compelling logical reason why Sec. 101 must be applied first when considering the validity of a given patent claim.  The lexical priority in the statute is certainly not the result of deep thought about the most effective order in which to consider patentability issues.  And because claim sets present complex, overlapping subject matter, while patent doctrines build on interrelated policy concerns, there is good reason to take a more open-ended approach to questions of patentability.  In other words, it is best to proceed pragmatically.  And when looked at from this point of view, the primacy of Sec. 101 makes no sense at all.

Instead, it makes much better sense to move Sec. 101 to the back of the line.  What is needed is an approach to patentability that permits courts to knock out patent claims using less contentious and hard-to-apply doctrines.  Why labor to apply the complex and difficult Sec. 101 if an easier ground of invalidity is available, such as novelty or enablement?  This approach conserves on effort and time, and preserves as well the legitimacy of patent tribunals, whose pronouncements under Sec. 101 are often controversial and confusing.  The result of such an approach would be to take Sec. 101 off the hot seat and move it to the back of the line, where it will be applied only when all other validity doctrines have been fully satisfied.  Which will make Sec. 101 far less prominent, and thus, it is hoped, far less controversial as well.