In a comment on the More on Supposedly "Clear" Texts thread, commenter von writes (in response to my text, which he italicized):
More on Supposedly "Clear" Texts: In the post below, I made a simple claim: You can't tell whether a text is "clear" or "unambiguous" by simply consulting the text — you also need to see whether there is external evidence that the meaning of some term is something other than what you assume it to "clearly" or "unambiguously" be (whether we refer to the meaning intended by the author or the meaning likely understood by most of the author's intended readers). This shouldn't be politically or ideologically controversial. It's just a description of how language and communication works.
A bit off point, but it's worth noting that the process used to construe a patent's claims (the part of the patent that tells you the scope of the patentee's zone of exclusion) *generally* proceeds in the opposite direction: A court looks first to instrinsic evidence -- roughly, the four corners of the patent document + the back and forth with the patent office during the application process -- to construe a claim or claim term. Only if the claim or claim term remains ambiguous does the Court typically consider so-called "extrinsic evidence," e.g., expert testimony, dictionary definitions, and the like.
I say "generally" because different panels of the Federal Circuit have, at times, expressed different views on when and how extrinsic evidence should be considered. The most recent in banc decision on the subject suggests a slightly preferred view of some extrinsic evidence (expert testimony), but doesn't quite elevate it to the level of intrinsic evidence in the claim construction process.
I'm no patent maven, but I wonder whether this can really be so. The commenter lists "dictionary definitions" as "extrinsic evidence" that can only be checked when the "intrinsic evidence" yields ambiguity. But surely the judges construing the patent claims are using their own mental dictionary to determine what the intrinsic evidence means in the first instance. Yet all of our mental dictionaries are incomplete, especially as to technical areas in which we aren't specialists. How could it make sense for a judge to rely entirely on his likely flawed mental dictionary, coupled with what's in the claim and in the back-and-forth with the patent office, and to simply refuse to consult a written dictionary if the possibly flawed mental dictionary gives a supposedly "unambiguous" answer?
Say a judge, using his own mental dictionary, thinks some term is unambiguous, because he's aware of only one definition -- perhaps partly because he's not a specialist in the relevant technical field, and thus doesn't know the field's technical definitions. A litigant comes and says "No, wait, look at this dictionary [either a general dictionary or a scientific dictionary], which lists this other definition, which is the definition that I and the patent examiners used, though we didn't mention it in the patent because we took it for granted."
Would a judge really say, "nope, I may not be a specialist in computers, but I think the term 'dump' unambiguously means 'throw out' [which is to say, it unambigously means 'throw out' in my own mental dictionary], so I'm going to refuse to consider a computer science dictionary, even though you claim it gives an alternate definition"? If so, how is this remotely sensible?
Related Posts (on one page):
- Consulting Dictionaries Only After "Intrinsic Evidence" Shows Ambiguity:
- Textualism and Context (Especially as to Terms of Art):
- More on Supposedly "Clear" Texts:
- When Is a Document "Clear"?
- What Does "Free State" Mean in the Second Amendment?
In practice, the litigants provide suggestions in briefing that will tell the court whether the internal dictionary is wrong or not. For something like "dump" which is ambiguous in the broad sense but less ambiguous in certain computer software circles, the parties may even agree on what the term means. If not, then each party will point to where "dump" was used and how, and the court can decide whether the term was used ambiguously or not.
That said, a party wanting a dictionary definition to be considered will put that definition in, and it will be in front of the court for consideration. However, the party wanting a dictionary definition that goes against the use of a term in the patent has a pretty strong burden if the patent is at all helpful in describing the patented invention.
Concatenate these two:
1. How could it make sense for a judge to rely entirely on his likely flawed mental dictionary, coupled with what's in the claim and in the back-and-forth with the patent office, and to simply refuse to consult a written dictionary if the possibly flawed mental dictionary gives a supposedly "unambiguous" answer?
2. a party wanting a dictionary definition to be considered will put that definition in.
BUT - if the parties are at a certain level of expertise (and thus, the jargon of that expertise), it may simply not occur to them that any "dictionary" definition is needed for the judge, nor will it occur to the judge that one should be sought. These differences should become apparent in the "back and forth" and if so the judge should (?) consider it additional evidence - but it may not be self-evident, or there may be reasons for the judge to not add it to the reasoning on the case.
Are you sure it is wise to assume that the policies of the courts make sense? Although jurisprudence is certainly in most respects a rational process, there are some rather glaring exceptions, e.g. the prohibition on juror's discussing the case among themselves until they begin deliberations, the prohibition on jurors' taking notes (in some jurisdictions), or the treatment of translations as original testimony (so that it is problematic to address inadequate translation on appeal). That there should be rules of interpretation that are not sensible but owe their origin to different situations in earlier times or to pressures on the courts other than the need to reach the most accurate conclusion, would not be surprising.
First, the patentee can be their own lexicographer, therefore giving priority to internal definitions of terms. This freedom to define terms would be undermined by using extrinsic sources as a primary means of finding ambiguity.
Second, the patent contains a great deal of material beyond the claims, and frequently includes definitions and other components that inform the construction of claim terms.
Third, patent rights are granted on review by the USPTO, making their understanding as shown in the file wrapper to be the key source of information in defining the scope of a patent right.
The entire patent, and its file wrapper, will contain the definitions of the terms as contemplated by the patentee and the understanding that the USPTO based its grant of the patent. There's more material in the patent than you seem to assume, and certain aspects of the patent system demand that intrinsic evidence have priority in claim construction.
I also wonder how different it would be for those interpretive methodologies which prioritize original intent over objective "reasonable" understanding. (Contract law and Patent law currently tend to follow the former, but in Constitutional law there seems to be an dichotomy among originalists between these alternative versions...) It would certainly alter the kinds of extrinsic evidence that could be considered...
As an aside, I'm about to publish an article that describes why this is not necessarily the best way to determine what patent claims mean.
In short, no.
To quote the en banc Federal Circuit in their most recent attempt to set forth a detailed set of claim construction guidelines:
Phillips v. AWH Corp., 415 F.3d 1303, 1322-23 (Fed. Cir. 2005) (en banc). (Phillips is worth a read if you want a good introduction to the controversies and state of the law about how patent claims are construed.)
Commenter von is correct that Federal Circuit case law does generally privilege the intrinsic evidence over resort to dictionaries, but the test for when to look outside the intrinsic record isn't quite whether the judge (using his/her internal dictionary) perceives the claim to be ambiguous. To oversummarize a bit (which may be dangerous in this field), it's more about whether the four corners of the patent document explicitly define the term or appear to implicitly define it through specialized use in context, or whether they just use it generically (in which case one might assume that a dictionary definition reflecting the "ordinary meaning" might apply).
Keep in mind also, as AZ points out, the patent claims are only a small portion of the patent document; there's also an extensive written description of the invention that is supposed to tell people of ordinary skill in the art how to "practice" the invention. Sometimes there will be explicit definitions of key terms there, and there will almost always be some uses of the terms that are an important source of context. There isn't really a parallel in statutory or constitutional interpretation; legislative history is a lot squishier, less certainly relevant, and not a part of the document itself.
In my humble opinion, the CAFC's biggest error was in Cybor where it read the Supreme Court's decision in Markman to mean that claim constructoin is a "questio of law." That is clearly NOT what the Supreme Court held --- it held that claim construction is a question to be determined by the Judge. It did not hold that it was necessarily a question of law. There are many issues that are questions for the Judge (e.g.., equitable issues), which are not questions of law. The CAFC has NEVER dealt with that issue, and has merely assumed (despite the fact that numerous dissents and law review articles have pointed this out) that since Markman held that claim construction is for the Judge then it follows necessarily that it is questio of law to be reviewed de novo by an appellate court. Very wrong, and the CAFC is a court that is basically out of control and shows why it is usually good to have more than one circuit address a particular field of law. The Supreme Court has finaly seemed to take notice of this Circuit's problems (thanks to CJ Roberts -- and I am a liberal but praise him for concentrating the Court on this issue). Hopefully, it will address the Cybor issue soon.
Thanks for the note. Undoubtably, I agree with you that Judges do apply their own internal dictionaries to try to understand patent documents. How could they stop themselves? But the question I adressed was the legal reliance that may be placed on the construction of a patent claim on the definition provided. There are a lot of factors that may influence a Judge's decision even though they may not carry substantial (or any) legal weight.
On that subject, MR's view generally accords with my own.
David Krinsky is correct to note that Phillips still permits reference to dictionaries, although Phillips is generally read as a walk-back cases that indicted a preference for dictionary definition (e.g., Texas Digital).*
I think David and I are talking past one another to a certain extent, however. David is talking about when a judge may consult extrinsic evidence. It was established (prior to Phillips) that a Judge may consult any extrinsic evidence at any time in the claim construction process. But that doesn't answer the question of under what circumstances those dictionary definitions have significance.
Here's a section of the Phillips that gives a little context to the part cited by David:
To rephrase, one relies on a dictionary definition where "the ordinary meaning of claim language as understood by a person of skill in the art [is] .... readily apparent even to lay judges[.]" Of course, this begs Professor Volokh's question -- which is how does one know that "the ordinary meaning of claim language as understood by a person of skill in the art [is] .... readily apparent even to lay judges". Unfortunately, the Federal Circuit has been decidedly unclear on that point.
Finally, I strongly endorse reviewing Phillips if you're interested on this point: even though it is at points frustratingly unclear, it does do a very good job at structuring the claim construction process. I would also suggest a look to Autogiro Co. of America v. United States, 384 F.2d 391, 155 U.S.P.Q. 697 (Ct. C1. 1967), however. Autogiro is one of the best, early cases at describing the perils of patent claim construction. It continues to be very relevant today (indeed, the best parts of Justice Kennedy's decision in Festo are taking directly from the Autogiro opinion.
von
*Phillips is also considered a walk-back, of course, from the "specification-supreme" view of the SciMed line of cases.
"But the question I adressed was the legal reliance that may be placed on dictionaries as the basis for a claim construction."
Hope that clarifies things.
Not to force you to pre-publish your article in the Volokh conspiracies comments, but will you article be directed to only the prosecution history or does it generally address the claim construction process -- and declare, "this is madness"?
Very rarely, however, will a judge define a term without some reference to external authority (dictionary definition, case defining the term, statute defining the term, etc.). Not citing to external authority, and relying instead on the judge's own knowledge, would (as you indicated Eugene) seem arbitrary.
Keep in mind that patent claims are generally accompanied by many pages of technical description -- which tries to explain to one of skill in the art what the invention is and how to apply it -- and, in most cases, a substantial amount of back-and-forth with the Patent Office trying to distinguish the claimed invention from prior inventions ("prior art"). In addition, there is some support for the preposition that priorart cited during the prosecution of a patent are part of the patent's "intrinsic evidence." So there is usually a lot for a judge to look at, including a substantial amount of background material, all of which trumps extrinsic evidence (dictionaries, experts, etc.).
One way to think of patent claim construction is to think of it as statutory construction, where the statute comes with a lengthy, binding explanation regarding its purpose and intent and the Judge is also required to review and rely upon, as best she can, the entire legislative history.
It is well settled that the patentee can define his own terms (thus he can call "black" "white" if he so chooses). But that needs to be clear. This is why the best source for finding the meaning of the patent term is the patent document itself.
However, and this is where confusion creeps in, the Court also commands that the patent term be examined "in light of their understanding to those skilled in the art." Thus, if "widget" means something quite specific to mechanical engineers, and the patent is in that field, the courts should understand "widget" to mean that (unless otherwise specifically defined) as opposed to whatever the general meaning of "widget" may be.
As you can see then, the logic is somewhat circular. You must first look at the patent itself to see if there is any ambiguity and if there is none, end there. But in order to know if there is ambiguity you must know what one skilled in the art understands the terms to mean. That means you must look at extrinsic evidence. But you are not allowed to look to extrinsic evidence until you decided that the patent is ambiguous. Which you cannot do without looking at extrinsic evidence.
I do think we're talking past each other a little bit; I wasn't responding to your thoughtful posts, but rather to Eugene's question. I maintain that a (competent) district judge simply would not refuse to consider a technical dictionary definition of a term like "dump" on the basis that "dump" has an ordinary meaning of "to throw out."
In response to your comments to me (at 8:16 am), however, I don't think you can fully separate the "what" from the "when." The passage you quote describes the fact that one should sometimes apply a general definition (whether in one's head or in a general-purpose dictionary) and should at other times apply a technical definition (which one might find in, among other places, a technical dictionary). This is self-evident; the hard question is which to do when. Phillips says that the language of the claims themselves, the written description, the prosecution history, and extrinsic sources (especially dictionaries and treatises) all have a role to play in determining the answer, as well as in settling upon a meaning for a claim term. There is no suggestion that a district judge can or should consult her own internal dictionary rather than looking to these sources of context.
In this sense, patents fit nicely into Eugene's observations about construing texts. Although sometimes it seems as though you can look only to the plain language of the claim terms themselves (and sometimes you even say that's what you're doing) you do need to make sure your internal view of what they mean is consistent with the context. The unique feature of patents is that the most relevant context--the written description--is published in the same document with the claims.
Except that this is why I, and I know others, find this area of the law interesting. From a case law point of view, I would suggest that it is one of the more dense areas of the law. Yet, it can also be quite challenging from a technical point of view, trying to understand and explain an invention well enough that you can claim it well and overcome the rejections that you get from
mostly bozoshighly skilled examiners in the USPTO.I like it as a profession because it lets you play lawyer when you are tired of playing an engineer (or scientist), and it lets you play an engineer when you are tired of playing a lawyer.
Not exactly... There's a difference -- which the Court readily recognized in Phillips -- between: (a) looking at extrinsic materials for background information, which seems to be always permissible, and can be done prior to looking at the intrinsic record for the purposes of better understanding it; and (b) looking at extrinsic evidence as an authoritative definition.
The latter is only available if there is neither an idiosyncratic/otherwise limited definition, and that meaning is evidence (but not dispositive evidence, contrary to intrinsic evidence) of the meaning of the disputed claim term to a POSITA.
------
Also, I think that the big difference between patents and statutes is that patents are supposed to be inherently time-limted. The issue is what the inventor "possessed" as of the effective filing date of the patent. Even after-arising equivalents are not relevant to claim construction itself (though they are, quite obviously, to infringement under the DoE, once the claims have been construed).
Again, to analogize to real property, you're trying to figure out what the original property line was at the time the right to title was first recognized, to determine its metes and bounds, even if the river may have shifted course.
So in essence, claim construction really is an Originalist inquiry, and perhaps one that lends itself even more to Originalism (in the context of "original public understanding" for a specialized subset of the public) than do statutes and the Constitution, which are at least somewhat forward looking. I keep meaning to write a paper on that, but bar review is really sapping my will to live, if anybody wants to co-write in the Fall...
"In some cases, the ordinary meaning of claim language as understood by a person of skill in the art may be readily apparent even to lay judges, and claim construction in such cases involves little more than the application of the widely accepted meaning of commonly understood words. In such circumstances, general purpose dictionaries may be helpful." So in these situations dictionaries clearly may be consulted (we agree on this).
"In many cases that give rise to litigation, however, determining the ordinary and customary meaning of the claim requires examination of terms that have a particular meaning in a field of art. Because the meaning of a claim term as understood by persons of skill in the art is often not immediately apparent, and because patentees frequently use terms idiosyncratically, the court looks to 'those sources available to the public that show what a person of skill in the art would have understood disputed claim language to mean.' Those sources include 'the words of the claims themselves, the remainder of the specification, the prosecution history, and extrinsic evidence concerning relevant scientific principles, the meaning of technical terms, and the state of the art.'" So even in these situations, courts can consider "extrinsic evidence concerning ... the meaning of technical terms" -- which would presumably include technical dictionaries, whose job it is to provide the meaning of technical terms, no?
It's also worth emphasizing GMUSL 3L's point that the law recognizes a difference between "between: (a) looking at extrinsic materials for background information, which seems to be always permissible, and can be done prior to looking at the intrinsic record for the purposes of better understanding it; and (b) looking at extrinsic evidence as an authoritative definition." I'd only note that this can frequently be a difference without a distinction, because it's not at all clear where (a) ends and (b) begins.
I can't quite agree with GMUSL 3L's claim about this being an originalist undertaking, however, but he is right that notice is one (and, in my view, the primary) function of the claims. The difficulty is that the notice function of the claims has always been in tension with the concept that one should protect the actual invention. This secondary concept finds expression in the way the patent office construes claims (both during prosecution and in interference proceedings), which is different from the way District Court's construe claims. It's also reflected in the doctrine of equivalents (DOE), which, as Justice Black recognized in Graver Tank, cannot survive if you really believe that the notice function is paramount. (Of course, he was writing in dissent.) Indeed, the process of applying the DOE via a "hypothetical claim" analysis (a la Wilson Sporting Goods) very much goes to the heart of the invention inquiry.*
Bruce Hayden, I share your fascination. I actually don't have a technical background (my degree is in economics). I came to patent law because I happened to get assigned a patent case as a first year -- and was quite happy when patent litigation eventually squeezed out the young associate drudgery associated with class action defense. I stay with patent litigation -- it's about 66% of my practice, with the balance being a mix of civil fraud, RICO, and "softer IP" -- because I'm continually interested in both its linguistic and economic aspects. I also think it's great fun to learn new things -- and patent law certainly makes you do that, particularly if you don't start out with a lot of technical knowledge. (Even ignoring its convoluted legal precedents.)
von
*I recognized that a "heart of the invention" analysis is wholly improper as a rule -- but it is a rule that is frequently observed by its sub silento breach.
but doesn't the very excerpt you quote suggest that consulting dictionaries is permissible in all circumstances (except perhaps when the claim expressly defines some term in an idiosyncratic way, which I agree would generally trump extrinsic definitions)?
My answer is the same as I gave to David — there's a distinction between when a dictionary can be looked at and whether it can be considered. Almost any extrinsic evidence can be consulted at any time by a Judge. The question is when the judge can rely on the extrinsic evidence to construe a claim. Phillips indicates the preferences, in descending order of applicability:
(1) the claim terms themselves, with reference to, if anything, a judge's "internal" dictionary; (2) the patent specification, both as originally filed and as issued*; (3) the patent's prosecution history and cited prior art, to determine whether meanings were expressly abandoned or defined; and (4) everything else, including experts and ordinary and technical dictionaries. (1), (2), and (3) (the instrinsic record) always trump (4). Moreover, if (1), (2), and (3) provide a clear definition, it is improper to cite (4) — specifically including dictionaries — except to get "background information" (whatever that means, and it's not clear). There was a recent time when I would have given you a different answer on this one, but Phillips resolved a tension in the law (the Texas Digital approach). A quick Google search revealed an article address this very fact (not from my firm, and I don't wholly endorse it): Milbank paper on Phillips.
All of this really raises the question: dictionaries can be consulted, but under what circumstances? The best answer is provided in the section on which you focus — and, as I tried to point out, this is no answer at all. Indeed, it does no good to say that a Judge can consult a dictionary when "the ordinary meaning of claim language as understood by a person of skill in the art may be readily apparent even to lay judges" without providing some mechanism to determine when "the ordinary meaning of claim language as understood by a person of skill in the art may be readily apparent even to lay judges".
In other words, as much as I'd like to agree with your statement that we agree, I can't.
If this is unclear — and it may very well be, because it's unclear to patent practicioners — drop me a line at vonofobsidianwings@hotmail.com. If bona fides would be helpful, I think I can get a decent reference from folks at your law firm affiliate — something along the lines of, von is mostly not-stupid.
*I note this difference because there's good authority that the patent abstract can be consulted as part of the specification, and it's generally not added until the end.
Huh? In the US, the abstract needs to be included with the original filing. If it's not, OIPE will send out a notice that the application is incomplete. MPEP 608.01(b)
I was unclear. I didn't mean to say that the abstract isn't filed; I meant that the abstract is frequently changed or resubmitted during prosecution, because it's meant to given insight into the invention as claimed -- and that may not be known until the prosecution is nearly finished.
It's much less common to see changes to the spec outside of the context of a continuation in part.
It's much less common to see changes to the spec outside of the context of a continuation in part.
Interesting - my experience has been the opposite. Do work in a particular field - software, biotech etc.? I don't think I've ever seen an abstract changed, except to cut down the length, and changes to the spec seem fairly common to me.
It also seems to me that changes to the abstract intended to give insight into the invention would raise new matter issues. Do you have any examples handy of patents which illustrate your point?
Now you've lost me. You generally can't change the spec in any substantive way, or you will get a new matter rejection. The changes that do occur tend to reflect typographical corrections and/or the submission of formal figures. Sometimes a spec gets changed (particularly in biotech) to change a prophetic example to a real one, but this almost always gets filed as a continuation in part.
It also seems to me that changes to the abstract intended to give insight into the invention would raise new matter issues. Do you have any examples handy of patents which illustrate your point?
No, you misread me. I said that some changes changes to the abstract are to give "insight into the invention as claimed". It's similar to choosing what figure will appear on the front of the patent. Cutting down the length of the abstract is also a change that can have substantive impact -- and it's a change that's practically unthinkable in the context of the rest of the spec. As a general rule, one never adds to, deletes from, or substantively alters the written spec after it's filed -- or else you must file a new patent (usually, a continuation-in-part tho' sometimes a divisional). But the abstract (and, to a lesser extent, the figures) get changed all the time.
I have several particular examples in mind, but, since each is or may be involved in litigation or another dispute, I'd rather not identify them. My apologies.