Teva v. Sandoz and Claim Construction Deference

By Dennis Crouch

In Teva v. Sandoz, the Supreme Court is focused-in on the appellate rules that should apply to claim construction decisions. Namely, if a district court makes factual conclusions that underpin a claim construction determination, should those factual conclusions be given deference on appeal or instead reviewed de novo on appeal. The Federal Circuit’s practice is to follow the latter approach while Federal Rule of Civil Procedure 52(a)(6) suggests the former.

The issue as presented to the Supreme Court:

Whether a district court’s factual finding in support of its construction of a patent claim term may be reviewed de novo, as the Federal Circuit requires (and as the panel explicitly did in this case), or only for clear error, as Federal Rule of Civil Procedure 52(a) requires.

Here, the district court may several explicit factual findings, including resolution of conflicting expert testimony regarding how a figure should be interpreted and how one of ordinary skill in the art would understand the term “average molecular weight” in the context of size-exclusion chromatography.  These findings served as precursors to the lower court’s ultimate clam construction determination. On appeal, the Federal Circuit expressly reviewed and rejected those determination — but gave them no deference in the appeal. I should note here that there is disagreement about what should count as a factual finding — the PTO suggests that anything that relies at all upon patent or prosecution related documentation is legal textual analysis and not fact finding. Teva agrees that deference should be reserved to “classic fact­finding,” but that would include factfinding that also draws some inferences from patent documents in conjunction with other evidence.

The Supreme Court recently heard oral arguments, and the transcript is available online. In the case, Rising Star William Jay of Goodwin Procter argued for the patentee Teva who won at the district court but lost on appeal at the Federal Circuit. Carter Phillips of Sidley argued for Sandoz and Ginger Anders represented the Solicitor’s Office as Amicus.

Much of the of debate focused on the best analogy for patent claims — should they be interpreted as we interpret statutes (fully de novo review)? Or should they instead be interpreted as we do contracts (deference for underlying factual conclusions)? Jay attempted to argue that statutes are different because they are always designed to be read by the general public whereas patents are focused on those skilled in the art.

JUSTICE ALITO: If a patent is like public law, if it’s like a statute or like a rule, then factual findings regarding the meaning of that patent are not entitled to clear error review. . . Now, on the other hand, if a patent is private law, if it’s like a deed or if it’s like a contract, then Rule 52(a) comes into play. . . .

MR. JAY: Terms of art in statutes, Justice Alito…are nonetheless written to be read by the general public. And … when they have a legal meaning, the determination of that legal meaning is still a question of law. . . .

I think when you’re interpreting a statute that it s generally clear at least what the right frame of reference is Now in the in the patent case what the frame of reference is is itself a question of fact as this Court said in Graham v John Deere. Ascertaining the level of skill in the art who is the skilled artisan who is this patent written for that is itself a factual question and then figuring out what that what that person knows is also a factual question. . . .

MR. PHILLIPS: But actually, I think Markman
answered that, because I think Markman recognized that this is a public document that is going to be binding on third parties, and that therefore ought to be construed as a matter of law in order to ensure the stare decisis component of it.

Justice Breyer seemingly announced his position in favor of deference:

JUSTICE BREYER: And [the direct viewing of witness testimony], I think, is a very powerful reason for saying in a technical case, don’t overturn the judge’s factual findings whether they are ­­ particularly scientific matters, but no particularly here ­­ unless those three judges who will not even read the whole record normally and certainly won’t hear those witnesses, don’t let them do it unless they are convinced that it is clearly erroneous. Now, that’s the argument, and I would like to say that’s different from a statute.

Of course, in this particular case there was no live testimony for the district court Markman “hearing”, but rather only submissions in briefs and declarations.

A second analogy proffered is to the patent issue of obviousness that is also a question of law but is generally based upon underlying factual findings.  To that, Sandoz response is largely that claim construction should be considered unique.  On that point, a number of Justices seemed to question the fundamental Markman decision and its impact on Seventh Amendment rights.

Policy goals at stake: Uniformity and Consistency?

JUSTICE SOTOMAYOR: If you and the government can’t agree, why should we defer to a district court? Why don’t we defer, as has been done now forever, to the Federal Circuit and let them review these things de novo? . . .

CHIEF JUSTICE ROBERTS: [With deference,] two different district courts construing the same patent could come out to opposite results based on a subsidiary factual finding, and neither of those would be clearly erroneous, and yet on a public patent that is going to bind a lot of other people, people won’t know what to do.

. . . .

MR. PHILLIPS: It’s not ­­ it’s not worth the candle because all you’re going to do is create a cottage industry of trial lawyers fighting with the judge about which bucket some particular evidence fits into.

In my view, Mr. Phillips is right that the relative impact of deference for factual findings is relatively small.  The larger potential impacts will be seen if the court remarks more generally on what should count as a factual finding; whether claim construction methodology should be more aligned with statutory construction vs contract construction; and how the factual findings for claim construction might impact later parallel jury determinations.

62 thoughts on “Teva v. Sandoz and Claim Construction Deference

  1. 12

    Why not simply remove patent cases from district courts in the first place? Do them all in Washington, do them once, and automatically grant cert for patent appeals, with deference of course. The Fed Circuit can handpick its machinery and build the expertise that some propose is lacking at the district level. I doubt many district courts would get a sad about it…

    Judge Young’s words on this subject are timeless:

    “…any “victory” is pyrrhic. Given the monetary stakes involved and a Federal Circuit reversal rate exceeding forty percent, this Court is no more than a way station – an intermediate irritating event – preliminary to the main bout in the Federal Circuit. Whatever the merits of such a system, it is undeniably slow and extraordinarily expensive. The most this Court can say is, “Good luck and Godspeed.”

    1. 12.1

      There is indeed something very, very wrong with a circuit court that reverses district court’s claim construction at such a high rate. Legal errors by district courts should not be anywhere near 40%.

      1. 12.1.1

        The circuit court does not review all the constructions of the districts courts, but just a small portion of them.

        We cannot know for sure if the constructions that are not reviewed would have been affirmed by the circuit court, but I doubt that 40% of those would be reversed. If we assume they would not have been reversed (because not appealed), one can estimate the “real” rate using the percentage of cased reviewed in addition to the 40% reversal rate.

      2. 12.1.2

        Ned: There is indeed something very, very wrong with a circuit court that reverses district court’s claim construction at such a high rate. Legal errors by district courts should not be anywhere near 40%.

        Right. But there’s nothing wrong with a circuit court that frequently reverses incorrect district court claim constructions in cases where the claim construction is disputed. That’s always going to be more true of the fraction of cases that are appealed to the Federal Circuit relative to those cases are not.

    2. 12.2

      Mr. Snyder,

      Not only should Congress take a look at eliminating the role of district courts, but they should also remove the Supreme Court as well.

      As cleverly noted in the comments at link to – the appeal power of the Supreme Court is NOT an original jurisdiction of the Court, and Congress – fully within their constitutional authority could remove patent cases from the Court’s ability to review.

      I would note for the lawyer types that such an action would not violate Marbury which merely holds for the ability to have Article III judicial review – and nowhere does it say in that case that such Article III judicial review must be by the Supreme Court.

      Congress reacted once before in the past to an over-activist and anti-patent Court. This time, Congress should act more forcibly (while fully still within their constitutional authority).

  2. 11

    I wonder if any CAFC judges attended the oral argument just to see how the S.Ct. was going to act/behave? I mean, its only down the street.

    1. 11.1

      Hughes and Taranto were guzzling bourbon and didn’t make it that far.

  3. 10

    There should be a requirement for a bachelors of science degree to be eligible for appointment to the Federal Circuit.

    1. 10.1

      Definitely. Ridiculous to stack the Fed. Cir. with judges that are ignorant of science.

      1. 10.1.1

        Yeah, Night. These judges one ruled that a one had structure, that price was tangible, that a programmed computer was new.

        Their lack of technical education was clearly on display.


          Uh….now Ned, the judges that made those rulings had a science background. And, you have never told me where a “1” is? Where is that “1” you keep mentioning? And tell me how does a computer transform information without structure? Etc……

          More Nedsense.

  4. 9

    Sorry OT, but it seems the PTO e-notifications are sometimes going amiss. Most of the time they get through our firewall just fine, but randomly just a few of them didn’t from late last year/ early this year. Next thing we know we get a Notice of Abandonment. Somebody (not at the PTO) told us the PTO is aware of this issue and is working on a fix, but some people apparently are considering switching back to paper delivery due to the lack of e-notice reliability (including us). Is anyone else having issues with this?

  5. 8

    I find this discrepancy interesting. The Feds seem to quote the prosecution history exactly,

    “Teva stated in its response that “[o]ne of
    ordinary skill in the art, upon reviewing the specification,
    would understand that ‘average molecular weight’ refers
    to the molecular weight at the peak of the molecular
    weight distribution curve shown in Figure 1,” i.e., Mp

    To the Supreme Court Teva stated,

    “Teva responded that the patent
    specification (including Figure 1) would make clear
    to a skilled artisan that average molecular weight is
    peak average.”

    Everyone seems to agree that the peaks in Fig. 1 are not located at Mp, and that Mp is directly measured from a plot of an SEC measurement, which is a bell curve, and not like Fig. 1 at all.

    I think the Feds could have reversed the district court for clear error.

      1. 8.1.1

        RH, if you are in fact right, then everything that Teva says about SEC is wrong. For starters, Figure 1 is not an SEC graph according to Teva, but something else entirely. An SEC graph is depicted on page 8 of their brief.

  6. 7

    I am not sure I understand Chief Justice Roberts concern that two different district courts could construe the same patent in two different ways, and both ways would not be clearly erroneous. I agree that this may happen, but if it does, it is a good indication that the patent in question is indefinite. Does he think the attorney representing the defendant would have missed that and not raise that concern in the district court? Or that the appeal court will not be able to overturn an a verdict that the patent claims are not indefinite? Something may go wrong, I don’t see what it is.

    1. 7.1

      PiKa, the Supreme Court has taken cases from two different circuits where the “construction” of the patent in each was different — in one circuit, the same product was found to infringe and in the second, not.

      Conflict of circuits.

      I think the Court does want a common construction if possible. It is also the single best argument that the Federal Circuit must decide these cases de novo.

      1. 7.1.1

        I thought that unless the patent claims are indefinite, there would not be two different constructions that are both not clearly erroneous. I see now this is where my assumptions must be wrong. I will make some research to see an actual example of this. Thanks Ned.

  7. 6

    Regarding the 7th Amendment, I thought Newman’s dissent in Markman was right.

    If the resolution of the battle of experts about what a claim term means to one of ordinary skill in the art is case dispositive, I don’t know how one cannot say that such is not a disputed fact for the jury.

    It certain was at common law in patent cases.

      1. 6.1.1

        All disputed facts were submitted to a jury when the validity of a patent was at issue.

        John Paxton Norman, The Law and Practice Relating to Letters Patent for Inventions (London: Butterworths, 1853) at 194-211, discussing the procedures involved in revoking a patent at common law.

        The trial by jury in the common law courts is described beginning in Section 28, p. 203. “The Chancellor, though a common law judge, has no power to summon a jury. Therefore, if there are issues in fact, the Court of Chancery cannot try the issues, but the Lord Chancellor delivers the record by his proper hands into the common law court … to be tried there.”

    1. 6.2

      Patents didn’t have claims at the time of the adoption of the constitution. Time to re-read Markman.

    2. 6.3

      The court, methinks, is getting confused between instructing the jury on what a claim means for the purposes of infringement and letting a jury decide disputed facts when the validity of a patent is at issue.

      At common law, when the validity of a patent was at issue, disputed facts were given to a jury.


          the ‘confusion’ is generated when a different ‘ends’ is desired.

          Here again, the fundamental question question resolves down to a “means and ends” discussion.


          iwathere, you might be right given the comments of Ginsberg.

          They may sense a distinction between pure claim construction and holding a patent invalid as a matter of law.

  8. 5

    Phillips may be right that the impact is minimal (as it nearly surely is in nearly every case imaginable) but never the less it seems to me that we should just be doing what we’ve been doing in this realm. It doesn’t seem to have gone too awry as of yet. And besides, allowing factual findings into claim construction this way may skew something like summary judgements or some other bizarre things like that. Maybe someone could address that in the comments. Whether or not creating deference here would mess around with the summary judgement procedure in some cases or not.

    1. 5.1

      It would destroy Markman hearings. Because you could not construe the claim before trial without a jury determining, first, what the ordinary skill in the art would understand a term.


          Dearest 6, let me first say I appreciate the examiner input. Please never take me as a rude person but time is limited.

          Here is the enforcement reality. There are three kinds of markman orders- no infringement, invalidity and – a jury trial on the markman definition. The last on point wherein the markman definition (the jury charge instruction) is no more dispositive, definitive or definitional than the original claim language itself. In other words, a jury trial on the claim definition coupled with the contested issue of fact concerning the accused product. Here even Breyer had it right questions today as where do you stop – because if a claim is pure law – it becomes the legal blob that swallows all issues of fact – and abrogates the 7th.


            Is there a nuance missing here?

            By that I mean, let’s look at the situation from more than just the district court judge perspective.

            Players along the continuum:
            – jury
            – district court judge
            – appeals court judge
            – supreme court justice

            Whatever the outcome – that outcome must be objectively set for the entire continuum. Otherwise, you have chaos.

            does not any aspect of “fact” then demand that the jury level of the continuum carry through – and we eliminate the judicial override of that fact item (and that constrains the district court judge power)?

            does not any aspect of “law” then demand de novo review at the appeal level of the continuum carry through – and we ALSO constrain the district court judge power?

            Is it not true that to any level of legal deference that the ends be driven to in order to give district judges more power, that that same level of legal deference necessitates de novo review at the appellate level?

            Looking at the situation from this perspective, are we really not asking to what level should the judiciary be involved in setting facts? Is there not an attempt here to create a different continuum of fact and law? You have “multiple types of facts” – or do you? Is that not just a scriviner’s attempt at creating a new level of “fact?”

            It appears that the new level would be more clearly stated as:
            – facts (pure bottom level)
            – factual inferences (the new item)
            – legal inferences (the subsumed-but-not-subsumed mongrel)
            – law (the truly subsuming level)

      1. 5.1.2

        J, IF the disputed facts are material in determining the validity of the patent, as they are here, then should the jury get involved. We have a statutory presumption of validity, and a Supreme Court ruling that patents can be rendered invalid only by Clear and Convincing evidence. Yet the Federal Circuit held the patent invalid as indefinite as a matter of law?

        From page 9 of the slip opinion:

        “On de novo review of the district court’s indefiniteness
        holding, we conclude that Dr. Grant’s testimony does not
        save Group I claims from indefiniteness. ”

        I don’t see how this validity case can be resolved, de novo, by the Federal Circuit and be consistent with the statutes, with i4i, or with the 7th Amendment, albeit, this case was not tried to a jury — I am not sure the patent holder even requested a trial by jury, or if he did, why he did not preserve that issue. (The issue of whether there is a right to a trial by jury for validity in ANDA trials has never reached the Federal Circuit, and I, for the life of me, do not know why patent owners are not asserting their rights.)


            anon, let me commend you. When we stay away from software and 101, you do tend to post well.


          I suspect fear of jury nullification either in liability or damages is the reason they would not do it.

  9. 4

    Re: “MR. PHILLIPS: It’s not ­­ it’s not worth the candle because all you’re going to do is create a cottage industry of trial lawyers fighting with the judge about which bucket some particular evidence fits into.”

    Isn’t it even more likely to lead to lots of arguments about requested special jury verdicts that can be treated as specific fact-findings? And more JMOL motions against them?

  10. 3

    What would happen if a Statute regulating a financial instrument X, were written such that it explicitly included reference to the “understanding” of a “kind of person” “at a point in time”. Would facts regarding what the understanding of that particular kind of person at that point in time be considered factual findings to which deference would be accorded?

    Statute H

    “Regulations 1 through 10 will apply to a financial instrument X, in accordance with an interpretation of instrument X consistent with what meaning would be understood therefrom by a specialist of type Z at the time it was written…”

  11. 2

    Speaking of deference to trial court factual findings, Vringo just filed its petition for en banc review.

    Pretty strong. Makes Judges Wallach and Mayer look like they went rogue. This issue is way more important than claim construction given recent examples where the CAFC ignored trial court determinations to simply hand the defendants a win (see Soverain Software as a notable example).

    Until 103 law is clarified to reign in rogue appeals court judges, defendants will view the CAFC as a chance for de novo review on almost every issue in the case.

    1. 2.1

      “Until 103 law is clarified to reign in rogue appeals court judges, defendants will view the CAFC as a chance for de novo review on almost every issue in the case.”

      Wasn’t it intended to be such on issues of law? I mean, I thought that was why “they” (congress?) made all these things issues of law, so that the courts could do their de novo jig all over them.

      That’s a serious question btw, I’m not just being snarky.

      1. 2.1.1

        Obviousness is a mixed question of law and fact. The ultimate conclusion of obviousness is a question of law. The underlying factors are questions of fact.

        not being snarky. this is basic hornbook law.


          I know that brosef. I’m asking didn’t they (congress?) make the ultimate conclusion o obviousness an issue of law in that manner on purpose so that the courts could finangle with it?

          Or is this perhaps more of an area where graham and the supremes or some other court kinda mussed things up in trying to clarify it?



            I am not sure I am understanding the first part of your comment.

            When you ask “ didn’t they (congress?) make the ultimate conclusion o obviousness an issue of law in that manner on purpose so that the courts could finangle with it?” you seem to misapprehend how statutory law works.

            It seems that you infuse common law “finangling” where no such authority exists.

            Also, as Congress has made obviousness an issue of law (this is the top level per my post at, is there any other option for Congress? Surely, Congress could not make such a matter of fact, could they? Congress may recognize facts, but they only write law.


              I agree that they don’t seem to have had any other choice, at least in so far as they wanted to implement obviousness as a condition. Likewise you cannot make obviousness into an issue purely of fact so far as I can see.

              But that seems to lead us to the inescapable conclusion that congress did make it the way it was, knowingly, and likewise knowing that it would be reviewed de novo (unless de novo review of such came after 1952, which I have forgotten).

              If they wrote it knowing such, and did that knowingly then, it seems to me, defendants RIGHTFULLY AND BY DESIGN (fixed from your original comment) view the CAFC as a chance for de novo review on almost every issue in the case.

              Which seems to me to imply that the original topic you broached is actually going smoothly and as intended.

    2. 2.2

      Mr. Prophet, I am not sure what case you are talking about. Do you have full cite or a case number?

      1. 2.2.1

        The case is I/P Engine v. AOL.

        Here’s a link to the petition with the opinion as an exhibit: link to

        Even respected academics that I’ve talked to think the majority went off the reservation on this one.


          Thanks for the link.

          It seems the jury found the prior art did not disclose a particular limitation. Nevertheless, the majority found that that limitation would have been obvious based on common sense.

          Chen in dissent says that the failure of the majority was to provide reason and evidence to show the missing limitation, not just rote use of the words, common sense.

          The petition argues that the use of common sense is for the trier of fact, which they argue the jury used when finding the prior art failed to disclose a limitation.

          But I am not so sure the jury was asked that question, so its findings do not seem to preclude the district court and the Feds from making that assessment.


            “But I am not so sure the jury was asked that question, so its findings do not seem to preclude the district court and the Feds from making that assessment.”

            Even under your argument, the common sense has to have some factual basis. The CAFC cannot simply apply its own view without evidentiary support.

            Further, the trial record was clear. The Google expert simply said that the missing limitations would result from application of common sense. From trial transcript of Google Expert: “Because – just think about it. If you ask a query of a search engine, you get a result, you just have the query sitting there with the result, why not use that also for filtering?”

            The juries verdict was an implicit rejection of the CAFC’s theory of obviousness.


              Implicit in the sense that they must have used common sense to determine that the prior art did not disclose a limitation?

              No, I just do not think the question was ever put to them.

              As to reasoning and evidence, Google’s argument seems to be a lot of hindsight. Lin was of the view that the majority decision did not connect the dots.

              But the bottom line to me is that the trier of fact should have been asked that question, and here, the trier of fact was the jury. Since the question was never put to them, someone seems to have waived the use of common sense to supply the missing element. Perhaps the party having the burden of proof?

    3. 2.3

      Indeed. And that is the major source of criticism of CAFC and de novo review, i.e., CAFC is result oriented – because when it wants to find ‘the’ result – it is always under the guise of claim construction driving the desired outcome.

      1. 2.3.1

        The CAFC is not the only court (or Court) that plays loose with finding the desired result under some less-than-certain guise.

  12. 1

    JUSTICE ALITO: Mr. Phillips, can I try this
    10 out and see if you agree with me.
    11 If a patent is like public law, if it’s like
    12 a statute or like a rule, then factual findings
    13 regarding the meaning of that patent are not entitled to
    14 clear error review.
    15 MR. PHILLIPS: Right.

    That section jumped out at me. Sounds like the day of CAFC claim trump is ending. Alito is impressive IMHO.

    1. 1.1

      not so sure about that – while the process here may be segregated, and the factual predicates determined to be treated along factual grounds, it is difficult to see the end result treated along factual grounds – not unless the Court is prepared to undo the entire line of claim construction doctrine per Philips and Markman. As pointed out, the end result is still a question of law and I just do not see that changing.

      1. 1.1.1

        To be sure, I would not take that bet either. Pragmatically, having claim construction as a matter of law – allows courts to take patent claims away from juries. I would say that generally, the scotus cases of recent vintage are about empowering the judge and stripping away the right to a jury. Drip, drip, drip – more and more issues being deemed matters of law.

        That said, having read the transcript, it appears to me that scotus is viewing markman – as a jury charge issue and not dispositive of claims being pure matters of law. In other words, the courts obligation to provide clear instructions to a jury. And this is not that same issue as deeming a claim as pure law – only for the court. Also, I thought Carter overstated the instance of experts being used for markman. Prompting the ‘why use experts at all’ – should the judge even have let the expert evidence in the record – remarks from the bench. The lower courts have expressed this frustration – you can go through the whole exercise of hearings, expert testimony, cross examination, etc – to create a record for claim construction and just have cafc flip you anyways. To be sure, if the lower court goes into this effort to determine claim meaning – they should get some deference on those findings. Conversely, if you just get some naked bench memo, well that probably is entitled to less deference. That appeared to me as the way the court was leaning. Just my 2 cents.


          See my post at – shining sunlight on what the Court is attempting here – gisting their scrivining as it were – clarity is better achieved if we call a spade a spade. The problem of course is that the Court must be consistent in whatever they decide and their decision on “fact” has no sticking point at the district court judge level. If they go “fact” the whole shebang drives to empowering juries – if they go “law” – or any mongrel degree of law to rise above the jury – to the exact same degree they so do, they must rest on de novo review.

          There is no “little bit pregnant” option.


            (I should add, there is no such option if the Court wishes to be consistent – and this is not something the Court has always done)

Comments are closed.