The Process of Claim Construction
The peripheral-claiming system seeks to define the outer boundaries
of the invention. In theory, the process works as follows: The applicant
and the PTO examiner negotiate over the scope of the invention,
limiting it in view of the prior art and the range of examples that
the applicant has enabled. The resulting claims define the scope of
the patent. Competitors can then read the patent claims and know
whether their actions will infringe the patent, and therefore whether
they need to avoid the patent or take a license. If they decide to infringe,
the patentee sues, and courts determine validity and infringement
by comparing those claims to the prior art or the defendant’s
product, respectively.
Every stage of this theoretical story depends critically on the ability
of participants in the process—patentees, examiners, competitors,
judges, and juries—to understand what the patent claims cover. Further,
at several critical points, the story depends on different parties
(applicant and examiner, patentee and competitor) having a shared
understanding of the meaning of the patent claims. If the applicant
and the examiner agree that a patent should issue, but disagree about
what the patent actually means, something is wrong. Similarly, if a
competitor thinks that a patent means one thing and the patentee
thinks it means something different, they are unlikely to be able to
conclude a licensing transaction efficiently.
In fact, disagreements over the meanings of patent claims are
pandemic. Lawyers are paid to interpret language—whether in statutes,
contracts, or patent claims—in ways that serve their clients’ interests.
And they are, as a general matter, quite good at it. Before
1995, disputes over what the patent covered were folded into general
disputes about patent validity and infringement. Those disputes ultimately
went to the jury, which presumably made an implicit determination
of claim scope in finding the patent valid or not and infringed
or not. Judges periodically determined the meaning of patent claims,
usually in the context of bench trials or summary judgment motions
requiring resolution of a claim-construction dispute. But claim construction
before 1995 was not a separate inquiry in a patent case. It
was a part of the overall determination of either validity or infringement.
That changed in 1995 with the Federal Circuit’s Markman decision,
affirmed by the Supreme Court the following year. In Markman,
the Court held that the construction of patent documents, like
the construction of other legal documents, was to be done as a matter
of law by judges, not juries. The result was the so-called “Markman
hearing,” a pretrial proceeding in which judges hear argument on the
meaning of disputed patent claims, sometimes hear evidence, and issue
a written ruling defining the words of the claims in other, theoretically
more precise, words. Courts are to construe patent claims
based on some combination of their “plain meaning,” the description
of the invention in the specification, the prosecution history of the
patent, technical treatises and dictionaries, expert testimony, and a series
of legal canons of claim interpretation. Virtually every judge in
the country conducts Markman hearings sooner or later in the trial
process, though a few judges defer claim construction until shortly before
trial when jury instructions are written..The creation of the Markman hearing has made the process of peripheral
claiming more transparent. Because judges now hold hearings
on claim-construction disputes and have to issue written rulings
interpreting disputed claims, the nature and extent of those disputes
is now clearer than it was before 1995. And the results are quite
alarming. First, there is essentially always a dispute over the meaning
of the patent claims. Patent suits normally include Markman hearings,
and every district that has local patent rules provides for them. Indeed,
there is rarely just one dispute in a patent claim; patent lawsuits
frequently involve fights over ten or more claim terms. The Federal
Circuit has held that courts must resolve every dispute over the scope
of the patent claims as a matter of claim construction, issuing a written
ruling that “interprets” even simple patent claim terms that jurors can
understand. And district court judges regularly issue orders limiting
Markman hearings to only ten or so claim terms. Without demanding
agreement from the parties, judges might well face arguments over
the meaning of every word in a patent claim.
These disputes turn out to be surprisingly hard to resolve. Even
after a district court issues its claim-construction order, the meaning
of the claims is uncertain. The Federal Circuit reverses more than
one-third of the claim-construction cases presented to it on appeal, a
far larger percentage than its general reversal rate. These reversals
aren’t just a problem of lack of education; David Schwartz has found
that the most experienced district court judges and ITC administrative
law judges who specialize in patents have their claim-construction de
cisions reversed just as often as district judges without patent experience.
And despite agreement by the Federal Circuit in Phillips on
the criteria for interpreting patent claims, Federal Circuit judges cannot
agree on how to apply those criteria: the rate of dissents in claim construction
appeals tripled after Phillips was decided, and the overall
reversal rate remained high. In short, claim-construction disputes
aren’t just make-work; they seem harder for courts to resolve consistently
than other types of patent disputes.
Why are there so many disputes over the meaning of patent-claim
terms? One possibility is that patents cover new scientific terminology
that doesn’t have a fixed meaning in the art, so that scientists in the
field can reasonably disagree over the meaning of those terms. No
doubt this does happen from time to time. But it is by no means a full
or even primary explanation for claim-construction disputes. At least
four more structural problems make certainty in patent boundary drawing
unlikely.
First, lawyers are paid to create, identify, and exploit ambiguities
in language. And lawyers are good at their jobs. The problem begins
with the patent applicant. Many claim-construction disputes are
over the meaning of terms that are well understood in the art, but
which are insufficiently specified in the claim itself.
Determining the meaning of patent claims necessarily requires a judge
to break the text of a claim into discrete “elements” or units of text corresponding
to the elements or units that comprise the claimed invention,
essentially organizing the language of the claims into “chunks” or
“quanta” of text. Define an element narrowly—limit it to a single word,
say—and you will tend to narrow the resulting patent. By contrast, defining
an element broadly tends to broaden the patent.
For each discrete packet identified, the courts must determine the
meaning of the constituent words. They can assign those words definitions
that range from narrow, specific meanings to broad, general meanings.
In determining the meaning of terms within a particular element,
judges practicing patent claim interpretation are engaged in an exercise
that to some degree resembles the famous “levels of abstraction test” articulated
by Judge Learned Hand for analysis of infringement under
copyright law’s “idea/expression” doctrine.
There are no hard and fast standards in the law by which to make the
“right” decision as to either the size of the textual element or the level of
abstraction at which it will be evaluated. Indeed, the indeterminacy is so
acute that courts generally don’t even acknowledge that they are engaging
in either inquiry. They define an element almost arbitrarily, and
even when judges disagree as to the proper definition they offer no
principled basis for doing so.
At the same time, not all claim construction occurs in the context
of infringement. Claims are also construed in the context of invalidity,
against the prior art. Some prior art references do constitute actual
devices or practices, such as anticipatory prior art known or used
in the United States before the date of invention But most references
will be textual, and even those that are not textual must typically
be reduced to a text before they can be evaluated by a court. For
that matter, even infringement analysis can require evaluation of textual
prior art—for example, to determine whether a range of equivalents
is precluded by the prior art or by foreseeable technologies.Thus, patent law frequently requires the mapping of text onto
text, construing claims against the prior art references. Courts ask
whether every element in a patent claim is found in an anticipatory
reference but have no explicit method for determining how to parse
the prior-art text to determine the presence of the invention’s elements.
Indeed, under the statutory-bar provisions of section 102(b),
the standard is one of substantial similarity—a classic statement of
central-claiming analysis. Where obviousness is concerned, courts
measure the claimed invention against the knowledge of the person
having ordinary skill in the art—the PHOSITA. The PHOSITA is in
some sense a composite of relevant prior-art references as of the date
of invention. But again, there are no explicit interpretive rules to determine
how such prior art is to be read, either individually or in the
aggregate. The result is that while we construe patent claims, theoretically
providing certainty in meaning, we proceed to undermine
that certainty in the subset of cases that involve comparing those
claims not to devices but to other, unconstrued documents.
Finally, and perhaps most fundamentally, is the problem of what
we might call metaconstruction. Peripheral claiming presumes that
the claims set out boundaries. The process of claim construction itself
presumes that the words of the claims are insufficiently precise to delineate
those boundaries. The solution that claim construction offers
is to substitute theoretically clearer words for the unclear words of the
patent claim. But what happens when—as seems inevitable—the parties
dispute the meaning of those new words? We should be skeptical
that the substitution of litigation-driven claim constructions for
equally well-understood terms that were not written with particular
litigation in mind will advance the understanding or clarity of patent
scope. It may end up resolving particular cases—indeed, it normally
will, at least if the lawyers advocating those claim constructions are
good at their jobs—but there is no reason to think these other words
will be less ambiguous or more likely to help a jury.
How great these problems are depends, to a large extent, on the
industry in which the patent exists. A patent claim that covers a DNA
sequence—a list of As, Ts, Cs, and Gs—is perfectly clear to a biotechnologist.
Similarly, a chemical formula is perfectly clear to a chemist.By contrast, we have a much harder time defining machines in words,
and a still harder time writing words that clearly delineate the scope of
software inventions. Not surprisingly, therefore, disputed claim constructions
are more likely in the latter sorts of industries. Similarly,
David Schwartz has found significant differences in the claim construction
reversal rates in different industries, suggesting that
resolution of those disputes is also harder in some industries than it is
in others. Our point is not that no one can ever understand patent
claims. Rather, it is that, in the industries that account for the overwhelming
majority of patents, figuring out the boundaries of a peripheral
claim is difficult, if not impossible. And it is certainly not
something on which the relevant audiences—patentees, examiners,
competitors, and judges—can often agree.
University of Pennsylvania Law Review(Excerpts)
AI
