Claim Construction: A Structured Framework*

Guest post by Professor Peter S. Menell (UC-Berkeley School of Law); Matthew D. Powers (Weil, Gotshal & Manges LLP); and Steven C. Carlson (Fish & Richardson PC) 

The construction of patent claims plays a critical role in nearly every patent case. It is central to evaluation of infringement and validity, and can affect or determine the outcome of other significant issues such as unenforceability, enablement, and remedies. Yet jurists and scholars have long lamented the challenges of construing patent claim terms. The Federal Circuit’s en banc decision in Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005), stands as the most authoritative decision on claim construction doctrine. But while putting to rest various controversies, many core tensions in claim construction persist. Moreover, the decision itself does not provide a step-by-step approach to construing claims. This commentary provides a structured road map.

Sources for Deriving Claim Meaning

Prior to the en banc Phillips decision, the Federal Circuit doctrine on whether extrinsic evidence could be considered and what role it should play shifted significantly. From 1996 until 2002, consideration of extrinsic evidence beyond educating the court about the technology was heavily disfavored. See Vitronics Corp. v. Conceptonics, Inc., 90 F.3d 1576, 1583 (Fed. Cir. 1996). But nearly contemporaneous decisions cautioned against such a strong reading. In 2002, the Federal Circuit appeared to elevate dictionaries, a special category of extrinsic evidence, to a central role in claim construction. See Texas Digital Systems, Inc. v. Telegenix, Inc., 308 F.3d 1193 (Fed. Cir. 2002). Within a short time, however, the limitations of this approach became apparent.

Phillips shifted attention back toward the intrinsic record while recognizing that extrinsic evidence can be considered, albeit with healthy skepticism. Extrinsic evidence may be considered if the court deems it helpful “to educate [itself] regarding the field of invention . . . [and to] determine what a person of ordinary skill in the art would understand claim terms to mean.” The court emphasized, however, that extrinsic evidence must be considered “in the context of the intrinsic evidence[,]” but is “less reliable than the patent and its prosecution history in determining how to read claim terms.” Since Phillips, the law is clear that intrinsic evidence serves as the principal source for claim construction and that it trumps any extrinsic evidence that would contradict it.

A Structured Approach to Claim Construction: Two Stages of Analysis

There are two distinct stages of analysis: (1) is construction of a claim term required?; and (2) interpretation of a claim term.

Step 1: Is Construction of a Claim Term Required?

The following chart decomposes this inquiry.


On the latter issue, it is useful to distinguish among three potentially overlapping general types of claim terms:


As reflected below, the three types of claim terms are not mutually exclusive and the question of which category is most appropriate will not always be evident based solely on a reading of the claim. The intrinsic record is critical to making this assessment.

Step 2: Interpretation of Claim Language

a. General framework

Once it is determined that claim language must be construed and is ripe for construction, the court must then apply the various substantive rules to the claim language to arrive at the proper construction. Before discussing the disputes that commonly arise in claim construction, it will be useful to state the principles that are generally not in dispute.

A “bedrock principle” of patent law is that “the claims of a patent define the invention to which the patentee is entitled the right to exclude.” Courts must interpret claims from the perspective of “how a person of ordinary skill in the art understands a claim term,” in the context of the patent. This frame of reference “is based on the well-settled understanding that inventors are typically persons skilled in the field of the invention and that patents are addressed to and intended to be read by others of skill in the pertinent art.” Often, other evidence will provide context for characterizing the person having ordinary skill in the art. Courts look to the meaning that the term would have to a person of ordinary skill in the art “at the time of the invention, i.e., as of the effective filing date of the patent application.” The “effective filing date” is the earlier of the actual filing date or the filing date of an application from which priority is accorded. This is quite significant because the meaning of scientific and technical terms can change significantly during the life span of a patent. Litigation over patent claims can occur multiple technological generations after the patent claim term was drafted.

Claim interpretation is highly context-dependent. The person of ordinary skill in the art “is deemed to read the words used in the patent documents with an understanding of their meaning in the field, and to have knowledge of any special meaning and usage in the field.” The meaning that this person would give to claim language, after having considered the intrinsic and extrinsic evidence, is the “ordinary meaning” of the claim terms. This ordinary meaning is considered to be the “objective baseline” for claim construction. Interpreting patent claims thus requires the court to consider “the same resources as would that person, viz., the patent specification and the prosecution history.” The patent and its prosecution history “usually provide[] the technological and temporal context to enable the court to ascertain the meaning of the claim to a person having ordinary skill in the art at the time of the invention.” Thus, patent claims are to be interpreted in light of this “intrinsic” evidence (i.e., the patent specification and its prosecution history) as well as pertinent “extrinsic” evidence (i.e., evidence showing the usage of the terms in the field of art).

b. Claim Construction Methodology

As noted above, Phillips holds that the “ordinary meaning” of a claim term is the “objective baseline” for construing patent claims. The phrase “ordinary meaning” is deeply engrained in the caselaw, but it is a slippery concept. The “ordinary meaning” of a term is what a court arrives at after doing the work of reviewing the specification, the other claims, the file history, the cited prior art, and the pertinent extrinsic evidence. Thus, the “ordinary meaning” is not the first step in the analysis. Nor is it the endpoint, as Phillips and its progeny have confirmed – the proper construction is frequently not a term’s ordinary meaning. Thus Phillips‘ identification of ordinary meaning as the “objective baseline” puts tremendous emphasis on this term, which can create unfortunate confusion and error.

Focusing on “ordinary meaning” has other shortcomings. The term “ordinary meaning” tends to drive the claim construction analysis to the meaning of a single word, or at most to a short phrase. But atomizing the dispute down to a word, or a short phrase, often does not make sense. Most patent disputes go to the overall approach of a patent claim, and focusing on a single word tends to lose the forest for the trees. When the overall approach of a patented invention is the central issue in a patent case, there may be no “ordinary meaning” that attaches. Trying to boil down the overall approach of an invention to a few selected words often misses the point of the dispute. There is a real danger that resolving a dispute over the meaning of a particular claim term will be mistaken for a resolution on the merits of a more fundamental infringement or validity dispute.

A more simple and useful description of the claim construction process starts with the “initial understanding” of claim language. This is the understanding that comes from the first reading of the claims, and from getting a sense as to what the patentee is trying to claim. This “initial understanding” may be focused on a particular claim term of interest, or may take into consideration larger blocks of claim text. The endpoint of the analysis is the “proper construction.” Between this starting point and this ending point, is an analytical framework represented below by a black box below:

If the claim language employs common, non-technical language, its scope will immediately begin to take on meaning. If the claim language term is technical, the court may ascribe little if any meaning to the term without further review of the patent and surrounding evidence.

The ultimate destination for this process is the “proper construction.” Arriving at the proper construction requires filtering the claim language at issue through a number of rules of claim construction, taking into consideration the pertinent statements in the intrinsic and extrinsic evidence. This process requires that the court view the evidence from the appropriate perspective of a person of ordinary skill in the art, from the relevant time period. The rules to be taken into consideration include the doctrine of claim differentiation, the rules for reviewing the specification for meanings of claim terms, prosecution history estoppel, a review of related patents, etc. The various rules that the court must take into analysis are sometimes contradictory, and typically involve a balancing of considerations. The following chart illustrates the principal points of analysis:

The various factors that govern claim construction are vertically aligned in roughly the order of persuasiveness, with intrinsic evidence at the top, and extrinsic evidence below. The Federal Circuit has often noted, and the Phillips decision affirms, that the specification is the “primary basis for construing the claim” and is in most cases “the best source for understanding a technical term.” However, there is no fixed hierarchy of claim construction rules — “there is no magic formula or catechism for conducting claim construction. Nor is the court barred from considering any particular sources or required to analyze sources in any specific sequence, as long as those sources are not used to contradict claim meaning that is unambiguous in light of the intrinsic evidence.”

Note that the term “ordinary meaning” is not reflected in this chart. This viewpoint is not the first step in the analysis, and it is not the endpoint. It is a helpful reference point, and probably occurs somewhere along the path. The “ordinary meaning” might be determined after doing the work of reviewing the pertinent intrinsic and extrinsic evidence, but before the final construction is rendered.

This ordinary meaning then might be found to be the proper construction, or the proper construction may be broader, or narrower, than the ordinary meaning based on the application of the various claim construction doctrines. Some of these doctrines tend to narrow claim scope, while others broaden it. These doctrines push and pull on the concept of “ordinary meaning,” and drive the final construction. The following chart reflects this dynamic. The principles set forth at the top of the chart are foundational principles of claim construction which ground the inquiry. The factors on the left tend to narrow the construction (but may in some cases broaden it), which the factors on the right tend to broaden the construction:


* The article from which this commentary derives is available at and will be published in the Berkeley Technology Law Journal in 2010. The full article provides substantially more detail and explicates not only the substantive rules governing claim construction but also judicial management of the claim construction process. This works grows out the Patent Case Management Judicial Guide project, a treatise developed for federal judges published in 2009 by the Federal Judicial Center. It can be downloaded from and will soon be available in an updated edition from Lexis. We worked with leading patent jurists, practitioners, and academics in developing this guide. We thank the Berkeley Center for Law & Technology and the Federal Judicial Center for their generous support of these projects. We owe special thanks to Judge Ronald Whyte, Judge Kathleen O’Malley, Lynn Pasahow, James Pooley, Mark Lemley, George Pappas, Nick Brown, Carolyn Chang, Tom Fletcher, Jeff Homrig, Marc David Peters, and Sue Vastano Vaughan for their contributions to this project. We also thank Laura Rocheloios and By Design Legal Graphics for their assistance with illustrations.


30 thoughts on “Claim Construction: A Structured Framework*

  1. 29

    The patent prosecutor’s reference “Landis on Mechanics of Patent Claim Drafting” on page 8 has been supplanted by “Faber on Mechanics of Patent Claim Drafting”.

  2. 28

    “Claim meaning” is something I questioned early on without the knowledge of the law in patent matters. Being a non-law individual, it is a challenge to understand and write my thoughts about these questions on patent claims, however I find the questions important enough to try. Claims, as I see it, should be in some agreed upon terms, terms that the exam and inventor understand each other. As it is now I feel there is no way to know if an exam “understands” the inventor’s claims, that results in a grant made in error if not. The understanding is not a matter of how well something is written, it IS the thought, concept, that must get through to make the best eval and grant. The terms must be clear, not just the judgement of the exam, this is where terms play the most important role. The question is how does one write claims within a “set” of terms without the important details of the idea being lost because of what the terms lack. I take an approach to the question from law, when looking at all the writings done in law, it is not only the terms, but also form that gives clear understanding to those that understand the terms and forms.

  3. 27

    Reflecting on how it is done in other jurisdictions will violate the fact that those jurisdictions do not rule here MaxDrei.

    You have once again stepped over from proslytizing to conflating.

    I forgive you – you have been trying hard to think before posting and I want to encourage you to continue.

  4. 26

    I’m struggling to follow you Bread. There is more than one way to skin a cat, or define a patentable invention, of course. That’s trite. You surely don’t suppose I’m denying that, do you?

    Out of the infinity of different independent claims one could write, one is actually written. If that one claim is “unclear” it’s “bad” right?

    If several independent claims use different words to define the “same” thing, that can’t help but do violence to the “clarity” that is delivered by a single independent claim. There is no sufficient reason to permit the practice.

    It is thought by lawyers that a scattergun slew, and a Doctrine of Equivalents, are needed, because no drafting can be good enough to secure a fair scope of protection for the inventor. But that’s pure parochial ignorance. See how claims are construed by technical judges and patent agents (as opposed to lawyers) in England, or how it’s done in mainland Europe, and reflect on that.

  5. 25


    You may have prematurely jumped to the “only” conclusion concerning “diversity” in claims (claiming different inventions).

    Explaining is not the same as enabling – no need to walk down that path.

    Let’s stick to your statement: In any one patent there can be claimed only one “inventive concept”. You then incorrectly limit the number of claims with an assumed one claim – one concept mindset.

    My question explores the claim as an explanation of the concept, so I will ask again (yes or no question), can a concept be explained in only one way?

  6. 24

    Don’t understand the question. To enable is not the job of the claim, is it.

    An inventive concept that is defined by the technical feature combination XYZ can have more than one aspect. Examples are a machine that comprises means for X-ing, means for Y-ing and means for Z-ing, or a process comprising the steps of X-ing, Y-ing and Z-ing. But WYZ (different claim/definition/words) is a different concept. How to put XYZ into effect is the job of the specification. Which part of the patent application has the burden to communicate the underlying scientific explanation to the PHOSITA reader?

  7. 22

    Lindsay writes “Given that there may still be some kind of dice roll involved in the end, trying a variety of different claim structures and strategies will be needed. Let there be diversity in our claims.”

    Which makes sense to me only on the basis that “diversity” has a warm touchy-feely aura to it, so diverse claims must be “good”.

    I say to you that there is a “dice roll” only when the claims are “diverse”.

    In any one patent there can be claimed only one “inventive concept”. If the claims are “diverse” then they are claiming different inventions, or using different words to define the same invention. Either way, clarity is damaged. Only trouble is, in any definition that’s competent, clarity is indispensible.

  8. 21

    Loved Breadcrumbs’ comment on the conflicted nature of PHOSITAs. Thanks!

    Given that there may still be some kind of dice roll involved in the end, trying a variety of different claim structures and strategies will be needed. Let there be diversity in our claims.

    FYI, you’re getting some spam comments from Portable Storage and Portable Containers which may be machine generated, extracting part of the text to make it seem relevant but saying nothing. Or maybe they hired a campaign speech writer.

  9. 20

    Paul, actually it’s your easily readable and compact book on patent drafting that does wonders for my understanding of patent law in England, Europe and the USA. And it’s so surprisingly cheap. Would that all readers here had read it too. Then these threads would be even more fun.

  10. 18

    Thanks Paul. Have you yet seen page 91 of the Menell Paper? I quote:

    “Cross-examination will usually be a sufficient mechanism to expose bias and unreliability, and conversely, to confirm that an expert’s opinions are sound.”

    Thank you, Menell et al.

  11. 17

    It pleases me to see how well Max Drei has understood our case law in England.

    The real difference from the US is this. We give a broad liberal interpretation to the claims and then stick rigidly to that interpretation. As Max says, we look at what the patent means by the words, not what the words mean. In the US, the courts look more narrowly at the words and then mitigate the harshness of over-strict construction by the doctrine of equivalents.

    Both systems should end in much the same place. But ours gets there with less fuss.

  12. 16

    Ha, good laugh there Breadcrumbs. thanks for responding.

    Of course the professor is not exactly the PHOSITA. Nobody is, given the requirements for this mythical being, to know all the art yet be incapable of a scintilla of inventive thought.

    The good professors strive to educate the court about the common general knowledge enjoyed by the PHOSITA, back on the relevant date. Do you doubt that they have the knowledge, and the communication skills, to explain that to the judge?

    And if they over-reach, there’s a learned QC, primed by the equally proud professor on the other side, who will pounce on their error, make them look foolish, and destroy the probative power of their expert report.

    Typically an English patent trial lasts a week. No evidence in chief, just opening statements, rival expert witnesses cross-examined on their respective written reports, then closing statements.

    What do readers say, who are Chief Patent Counsels with recent experience of full-blown patent litigation in England?

    What “sword” are you on about, pray?

    Have any readers ever read an English patent decision, with its copious quoting, verbatim, of what the experts said, under cross-examination, to justify the judicial claim construction?

  13. 15

    As you say MaxDrei, “someone” says cross-examining a PHOSITA is a hopeless task.

    Perhaps it is because PHOSITA is not a real person.

    Having Ordinary Skill does not mean having expert skill – your professor’s may not qualify, or may be over-qualified.

    On the other hand, PHOSITAs through KSR have been granted nigh god-like powers to know ALL published works in the art field and ANY art field that may be linked to solving the problem, which opens the omniscience to possibly ALL art fields – in which case your professors won’t qualify as they may be woefully under-qualififed.

    The sword is sharp.

  14. 14

    If a claim is to be construed by a PHOSITA, then why not ask the PHOSITA what it means to him/her? Would that not help the judge to decide?

    That’s why, in an English patent action, cases seldom go to appeal. The likelihood of the judge being reversed on claim construction is tiny.

    But, you say, cross-examining a PHOSITA is a hopeless task.

    Not in England. Usually Plaintiff puts up somebody like the Professor of Chemistry at Oxford University, and Defendant puts the Cambridge Professor on the stand.

    After that, the judge asks himself one simple question:

    What would the PHOSITA think the claim author was using the language of the claim to mean.

    And, with that, claim construction is done, just once, definitively.

    Seems so simple to me. where is England getting it wrong. any ideas?

  15. 13

    Excellent legal analysis, Storage Containers.

    On an unrelated note, the PTO revealed their proposed count system restructuring plan today. The key elements appear to be:
    a) Within each round of examination, the allotment of counts is rearranged to give more weight to the first action. Also, a partial count is given for a final action. For example, for a new case, the weights are FAOM – 1.25, Final – 0.25, disposal 0.5.
    b) The amount of counts for each round of examination decreases with subsequent RCEs. The first round gets 2 counts (same as now), the 1st RCE round gets 1.75, and subsequent RCE rounds get 1.5.
    c) Examiners get an across the board 2 hours extra per BD (equivalent to 1 extra hour per count), which may be adjusted higher on a per art unit basis depending on the net reduction in counts due to the reduction in RCE counts.

  16. 12

    Thanks for the article.It is very informative and helpful. I agree the construction of patent claims plays a critical role in nearly every patent case. It is central to evaluation of infringement and validity, and can affect or determine the outcome of other significant issues such as unenforceability, enablement, and remedies.The notes with particularly openly distrustful and unwilling to confide.I agree with,Phillips decision affirm that the specification is the primary basis for construing the claim and is in most cases the best source for understanding a technical term.

    Storage Containers,

  17. 11

    Chris makes an interesting point.

    Please further note with particularly suspicious interest, that if issues of language were subject to easy resolution, appellate courts, and patent attorneys for that matter, would not be necessary. While Malcolm Mooney would love this result, (even though, ironically, it would put him out of a job bothering real patent attorneys and making it even more difficult for him to afford a pair of Birkenstocks), it hardly seems plausible even in today’s fast paced world.

    If the Federal Circuit has legitimized bastardizations of the language such as “comprised of,” it is really a waste of time in my opinion to spend any time, academically or otherwise, to try to ferret out any modicum of reason or rhyme to how language elements will be construed.

    You might as well just get out your magic eight ball and hope for the best. Or you can at least make sure that you carefully define any terms whose meanings are likely to be disputed.

  18. 10

    Chris makes a great point.

    To add to the complexity, parties are leary of proposing any human-friendly way of “construing” claims that isn’t in the form of “element A” means “some longer description of element A” such that you can replace the term “element A” with the longer description.

    Ideally, the term should be “element A” but then the parties should be able to present the court with the “bounds of element A:”

    Element A can be any of these things:…
    Element A cannot be these things: …
    Element A must be able to do this:…
    Element A cannot do this:…

    This actually would allow parties to look at something mroe than just a simple word. THey can look at an entire clause:
    Clause A must do …
    Clause A cannot do …

    This would at least be a path to simplifying claim construction.

  19. 9

    The final graphic reproduced above illustrates why claim construction is such a boondoggle.

    When the applicant engages in conduct that the system should be encouraging, i.e. explicitly defining claim terminology, using terms consistently, and clearly differentiating the invention from the prior art, what happens? The applicant is penalized with a patent that will be given limited claim scope.

    On the other hand, the applicant who engages in gamesmanship and sharp practice by using vague and/or poorly defined claim terms or uses such claim terms inconsistently or without antecedent basis is rewarded with greater claim scope despite engaging in tactics that should be discouraged and are inconsistent with the spirit–if not the letter–of the law.

    Given these perverse incentives, is it any wonder that such a high reversal rate exists for claim construction decisions?

  20. 8

    “Substantially”, as an example, is not per se a technical term, but neither is it per se a lay term”

    I think substantially is legitimately a lay term, but that it simply makes no sense to analyze the term separately from the term to be modified. Are there really any strictly technical terms of degree”? All I could come up with was made ones by tacking “-ier” onto the end of a technical term.

  21. 7

    Am hardly started on the Paper yet, but what has tickled me so far is:

    page 49, Table B, how the CAFC sets about construing the word “to”. Reference Cybor v FAS

    page 56, the Chef America case. The one all about baking “to” or “at”.

    Not sure that the CAFC in Chef did in fact handle “to” as p 49 would have it.

    Which rather makes Paul’s point, I think.

  22. 6

    What Paul has described is “Art”. There is some greater whole to the music of a claim than just the notes that make up the claim. This article is like a statistical analysis of a composition by Mozart, though such an analysis is probably useful in court, just like if Mozart was trying to show his music was copied… we can examine the parts, or the overall sound.

  23. 4

    Great comment Paul. I haven’t downloaded and read the Paper yet but, when I do, I will have your observations in mind.

  24. 3

    A very nice post, but directed to the construction of particular words or expressions within a claim.

    But how do you construe the claim as a whole? Is it part of the court’s task to construe the claim as a whole?

    You can view a painting in a museum and inspect the dabs of paint with a magnifying glass. But stand back a little. You might see a meaningful message e.g. an image of sunflowers. Or you might see no message, just random dabs of paint.

    Similarly with a patent claim. If you stand back from the individual words a little what do you see? A set of elements. Defined relationships between the elements. Do they contribute anything when collected together, and if so what is the contribution or alleged contribution?

    Are the elements a true combination? Is it enough to make them a combination that an inventor and his or her patent attorney has grouped them together, or is something more required? If so is that essential additional ingredient present in the claim or not? Are they a single combination, ot two or more smaller combinations spuriously grouped together?

    Very relevant to novelty and inventive step, but mostly overlooked.

  25. 2

    A good post to get us all thinking — which is the definition of “good post.”

    I disagree with your Venn, which proposes that there are no terms of degree that are within the universe of technical terms but not in the universe of lay terms. In fact, I would say that all terms of degree must be understood within their context as lay terms or technical terms.

    “Substantially”, as an example, is not per se a technical term, but neither is it per se a lay term. The layperson will likely not understand the metes and bounds of “substantially” as it is used in “substantially axial” or “substantially equimolar” but they would get the idea of the term “substantially clueless,” (not referring, necessarily, to Moonbeams or 6 here).

  26. 1

    So how does your framework work within the context of the Phillips dispute? I take it that “baffle” would be considered a “lay” term. It has an established meaning in the field, one that doesn’t specify any angles. But no one can argue that the term in Phillips is not subject to reasonable disagreement that requires a court construction.

    It seems to me that your framework fails to address the fact that you can pick virtually any claim term and reasonably say that the term is “ambiguous” in view of the specification, at least under current doctrine.

Comments are closed.