“. . . the intolerable wrestle with words and meanings . . .” East Coker, by T.S. Eliot
Congress enacts laws. The courts interpret and apply them in cases. Often, there is disagreement over what the words mean, and judges debate the meaning in published decisions. Judges on the same court may agree, disagree, dissent, concur, and form shifting majorities and minorities. Occasionally, congress will take notice and attempt to clarify a law by amendment. Sometimes, this only adds to the confusion.
The eleven active judges on the Court of Appeals for the Federal Circuit, the patent appeals court, exemplify this dynamic in their August 31, 2012 en banc decision in two cases consolidated on appeal, Akamai v. Limelight and McKesson v. Epic Systems (link). The decision, totaling over 100 pages, is comprised of three opinions, each with dramatically different views of a fundamental issue in patent law.
The core issue the judges on the CAFC were unable to agree upon on is this: does patent infringement occur when separate entities perform the steps of a patented method?
To take a simple example, assume that a patent claim involves just two steps: first, delivering a web page to a customer’s server, second, the some form of manipulation of the page by the customer (e.g., tagging or data hashing).*
In fact, this is, in highly simplified form, analagous to some of the patent claims at issue in this appeal. (U.S. Patent No. 6,108,703).
Until August 31st, the patent owner could not successfully claim patent infringement, because this was “divided” infringement – multiple parties (in this example two parties) performed different parts of the single claimed patent method. As the CFAC stated in BMC Resources, Inc. v. Paymentech, L.P.), “infringement requires, as it always has, a showing that a defendant has practiced each and every element of the claimed invention.”
In Akamai/McKesson the CAFC agreed to reconsider the divided infringment doctrine en banc. However, the eleven judges were barely able to eek out a 6-5 majority opinion.
Six of the CAFC judges — the thin majority — formulated a new doctrine of “induced infringement”: a party can be liable for inducing infringement if it either (1) induces several parties to jointly carry out the steps necessary for infringement, or (2) performs some of the steps of the claimed method itself and induces a third party to perform the remaining steps claimed. Under this new interpretation of the law all the steps of a claimed method must be performed in order to find induced infringement, but all the steps need not have been performed by a single entity. However, importantly, it appears that the CAFC intends only for the “inducer” to be liable for patent infringement, not the parties that directly infringe the patent claim.
Four of the eleven judges dissented, arguing that there should be no patent infringement where the steps are “divided.” In other words, this group argued for adherence to the “single-entity” rule that had been in effect.
One judge, Pauline Newman, dissented separately, arguing that under the patent statute when more than one entity performs all of the steps, the claim is directly infringed. As she states, “The court should simply acknowledge that a broad, all-purpose single-entity requirement is flawed, and restore infringement to its status as occurring when all of the claimed steps are performed, whether by a single entity or more than one entity . . ..”
What is the statutory language the judges cannot agree upon? It is this:
whoever without authority makes, uses, offers to sell, or sells any patented invention . . . during the term of the patent therefor, infringes the patent (35 U.S.C. 271(a))
While the dissenters cannot agree whether the word “whoever” requires a single actor or permits several actors to combine to infringe a patent, the majority ducks this question altogether, focusing instead on section 271(b) of the statute: “Whoever actively induces infringement of a patent shall be liable as an infringer,” and concludes that this provision permits liability based on inducement. Of course, as the dissenters point out, this approach begs the question whether there has been an “infringement” in the first place if multiple actors perform the elements of a patent claim.
The decision leaves unanswered questions and poses contradictions, not the least of which is that it appears the patent owner cannot sue the direct infringers when more than one entity participates in the infringement, an odd result indeed. It leaves unanswered whether the patent holder can obtain injunctive relief against the direct infringers, or whether the should even be named as defendants. It muddies the lines between direct, contributory and vicarious patent liability, which lines were already far from clear before this decision. And, most fundamentally, it fails to answer the question whether direct patent infringement can exist where the steps are “divided” – if it does not, there can be no “induced” infringement, since it is beyond question that direct infringement must exist before induced infringement can occur.
The fact that the en banc court was unable to reach consensus on such a fundamental issue of patent law is disturbing. Speculation that the Supreme Court will attempt to straighten out this mess began almost immediately after release of the decision. Perhaps the Supreme Court will answer the central question dodged by the majority in Akamai/McKesson, whether multiple-actor infringement constitutes patent infringement, and thereby serve as the basis for induced infringement, and vindicate one of the dissenting positions. However, the Supreme Court has been known to make matters worse, not better, when it comes to “the intolerable wrestle with words and meanings,” particularly in cases involving patent law.
Akamai Techs., et al. v. Limelight Networks and McKesson Techs., Inc. v. Epic Systems Corp.
[Update: Reversed by U.S. Supreme Court on June 2, 2014. Link]