Concurring opinion filed by LOURIE, Circuit Judge.
Dissenting opinion filed by O’MALLEY, Circuit Judge, with whom RADER, Chief Judge, and REYNA and WALLACH, Circuit Judges, join.
ON REHEARING EN BANC
NEWMAN, Circuit Judge.
The court en banc granted the petition filed by patentee Lighting Ballast Control, in order to reconsider the holding in Cybor Corp. v. FAS Technologies, Inc., 138 F.3d 1448 (Fed.Cir.1998) (en banc) establishing the standard of appellate review of district court decisions concerning the meaning and scope of patent claims — called “claim construction.” Implementing the Supreme Court’s decision in Markman v. Westview Instruments, Inc., 517 U.S. 370, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996) (Markman II), aff'g Markman v. Westview Instruments, Inc., 52 F.3d 967 (Fed.Cir.1995) (en banc) (Markman I), this court in Cybor held that patent claim construction receives de novo determination on appeal, that is, review for correctness as a matter of law. Such review is conducted on the administrative record and any additional information in the record of the district court, and is determined without deference to the ruling of the district court.
In the case now before us, a panel of this court followed the Cybor standard and revised the district court’s claim construction, applying de novo the statutory requirements of 35 U.S.C. § 112 ¶ 6 and § 112 ¶ 2.1 Briefly, the panel held that the claim term “voltage source means” is a means-plus-function term requiring corresponding structure in the specification. On this claim construction, the panel reversed the district court and held the claims invalid for indefiniteness. The pat-entee requests rehearing, stating that on deferential appellate review the district court would not or should not have been reversed. This court undertook rehearing en banc for the purpose of reconsidering the standard of appellate review of claim construction.
For the reasons we shall discuss, we apply the principles of stare decisis, and confirm the Cybor standard of de novo [1277]*1277review of claim construction, whereby the scope of the patent grant is reviewed as a matter of law. After fifteen years of experience with Cybor, we conclude that the court should retain plenary review of claim construction, thereby providing national uniformity, consistency, and finality to the meaning and scope of patent claims. The totality of experience has confirmed that Cybor is an effective implementation of Markman II, and that the criteria for departure from stare decisis are not met.
I
The Reheaeing Arguments
Lighting Ballast argues that de novo plenary determination of claim construction is improper appellate practice, stating that the interpretation of documents is fundamentally factual in nature, and that the district court’s interpretation of patent claims requires deference on appeal. Lighting Ballast states that on deferential review the district court’s claim construction for the patent in suit would be sustained, along with the ensuing judgment that the claims in suit are valid and infringed.
This en banc court agreed to reconsider the principle of de novo review of claim construction, and invited supplemental briefing and amicus curiae participation on the following questions:
(1) Should this court overrule Cybor?
(2) Should this court afford deference to any aspect of a district court’s claim construction?
(3) If so, which aspects should be afforded deference?
The parties as well as the amici curiae were not of one mind, but divided among three general views, all thoughtful and well presented.2 The general positions are summarized:
The first view
The view favored by Lighting Ballast is that the Cybor decision is incorrect and should be entirely discarded. Lighting Ballast argues that this court in Cybor misapplied the Supreme Court’s decision in Markman II, in that the Court had focused only on whether questions of patent claim construction are subject to jury trial, or whether this issue should be decided solely by a judge. These proponents state that the Court in Markman II, in deciding the judge-jury question, did not change the traditional distinction between fact and law, recognized that there are factual aspects of claim construction, and [1278]*1278did not address the standard of appellate review.
These proponents argue that the Court left intact the protocol of appellate deference to a district court’s fact-based rulings, whether the facts relate to claim construction or any other issue, and whether the ruling is by a judge or by a jury. They state that the Cybor standard of plenary appellate review is incorrect, and remind us that the Court in Markman II described claim construction as a “mongrel practice” of law and fact with “evidentiary underpinnings,” 517 U.S. at 378, 390, 116 S.Ct. 1384. They argue that although the Court stated that “the interpretation of a so-called patent claim ... is a matter of law reserved entirely for the court,” id. at 372, 116 S.Ct. 1384, the Court did not strip claim construction of its essentially factual nature.
These proponents point out that in construing patent claims, expert testimony and documentary evidence may be presented to the district court. They argue that restoration of deferential appellate review on the clear error standard would not only respect the traditional trial/appellate relationship, but also is more likely to give weight to aspects involving credibility of witnesses. They point out that Federal Rule of Civil Procedure 52(a)(6) requires that the district court’s factual findings receive review on the deferential clearly erroneous standard,3 citing Pullman-Standard v. Swint, 456 U.S. 273, 287, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982) for its statement that Rule 52(a) “does not divide findings of fact into those that deal with ‘ultimate’ and those that deal with ‘subsidiary’ facts.”
These proponents argue that patent claim construction is most reasonably classified as a question of fact, and that the Court’s Markman II description of claim construction as better suited to determination by a judge rather than a jury does not affect the requirement of appellate deference to findings of fact made at the trial level. Thus Lighting Ballast urges that the Cybor standard of de novo review is incorrect and should be entirely discarded.
The second view
The second approach, favored by some amici curiae including the United States, may be viewed as a fusion or hybrid of de novo review and deferential review. These proponents acknowledge that the Court in Markman II described patents as “legal instruments” and stated that interpretation of patent claims is a “purely legal” matter, 517 U.S. at 391, 116 S.Ct. 1384, but argue that the correct appellate approach is for the factual aspects of claim construction to be reviewed on the clearly erroneous standard, while the final conclusion receives review as a matter of law. See, e.g., Brief of Amicus Curiae United States at 4 (“Because Cybor fails to acknowledge that claim construction may involve factual findings entitled to deferential review under Rule 52(a), it should be overruled. But this Court should reaffirm that the ultimate construction of a patent claim is a legal conclusion subject to de novo review.”). The United States draws analogy to the ruling in the regulatory tariff case of Great Northern Railway Co. v. Merchants’ Elevator Co., 259 U.S. 285, 292, 42 S.Ct. 477, 66 L.Ed. 943 (1922), that when ambiguity arises in the construction of certain legal instruments, the ambiguity is resolved as a question of fact.
Some of these amici recognize that difficulties may arise in practice, in knowing [1279]*1279which aspects of the district court’s claim construction are subject to deferential review and which aspects receive de novo determination. They suggest a solution whereby the standard of review would depend on whether the district court’s claim construction drew solely from the record of the patent and its prosecution history (called “intrinsic evidence”), or whether external information or witness testimony was presented in the district court (that is, “extrinsic evidence”). Applying this distinction, some amici propose that claim constructions based on extrinsic evidence would receive clearly erroneous review, for such evidence may entail credibility or reliability findings, while constructions based solely on the patent document and prosecution history would receive de novo review.
The proponents of a hybrid form of appellate review argue that this approach comports with the Court’s position in Markman II, yet respects the traditional roles of trial and appellate courts. Thus it is proposed that the standard of review of claim construction should vary with the source and purpose of the evidence, drawing analogy to review of the determination of obviousness under 35 U.S.C. § 103. See Brief of Amicus Curiae American Bar Association at 12 (“it is well settled that Rule 52(a) governs appellate review of the factual inquiries underlying the ultimate legal issue of obviousness”).
The third view
The third view, supported by some amici curiae, is that Cybor is both reasonable and correct in view of the Court’s rulings in Markman II. These proponents stress the Court’s statements that claim construction is a “purely legal” matter, 517 U.S. at 391, 116 S.Ct. 1384, and that “the interpretation of a so-called patent claim ... is a matter of law,” id. at 372, 116 S.Ct. 1384. They argue that de novo review of the scope and meaning of patent claims conforms to the rule that applies in all areas of law, that “interpreting a set of legal words ... in order to determine their basic intent” is a “purely legal matter.” Buford v. United States, 532 U.S. 59, 65, 121 S.Ct. 1276, 149 L.Ed.2d 197 (2001). They state that sufficient reason has not been shown to change this established and effective precedent in patent cases.
In Markman II the Court placed claim construction in historical context, quoting Professor Robinson’s treatise:
The duty of interpreting letters-patent has been committed to the courts. A patent is a legal instrument, to be construed, like other legal instruments, according to its tenor.... Where technical terms are used, or where the qualities of substances or operations mentioned or any similar data necessary to the comprehension of the language of the patent are unknown to the judge, the testimony of witnesses may be received upon these subjects, and any other means of information be employed. But in the actual interpretation of the patent the court proceeds upon its own responsibility, as an arbiter of the law, giving to the patent its true and final character and force.
517 U.S. at 388, 116 S.Ct. 1384 (quoting 2 W. Robinson, Law of Patents § 732, pp. 481-83 (1890)) (emphasis the Court’s). Proponents of retaining the Cybor standard point to the Court’s emphasis that the judge is responsible for “the actual interpretation of the patent.” Id. These proponents also point to the Court’s citation to Miller v. Fenton, 474 U.S. 104, 106 S.Ct. 445, 88 L.Ed.2d 405 (1985) in support of the Court’s ruling in Markman II that:
[Wjhen an issue “falls somewhere between a pristine legal standard and a simple historical fact, the fact/law distinction at times has turned on a deter[1280]*1280mination that, as a matter of the sound administration of justice, one judicial actor is better positioned than another to decide the issue in question.” So it turns out here, for judges, not juries, are the better suited to find the acquired meaning of patent terms.
Markman II, 517 U.S. at 388, 116 S.Ct. 1384 (quoting Miller, 474 U.S. at 114, 106 S.Ct. 445). These proponents argue that Miller confirms that the Court in Mark-man II intended to decide — and effectively did decide — both the judge/jury question and the fact/law question, as well as the related question concerning appellate review.
The Court’s reliance on Miller in Markman II illustrates the Court’s recourse to general jurisprudence and practical considerations, suitably applied, in resolving patent issues. Proponents of adherence to Cybor point out that Miller reiterated the Court’s recognition in Pullman-Standard that no principle will “ ‘unerringly distinguish a factual finding from a legal conclusion.’ ” Miller, 474 U.S. at 113, 106 S.Ct. 445 (quoting Pullman-Standard, 456 U.S. at 288, 102 S.Ct. 1781). Proponents apply that observation to criticize departure from Cybor that would add, to the already complex laws of claim construction, a new and uncertain and contentious inquiry into which aspects of a particular construction fall on which side of the fact-law line. These proponents state that Cybor is not subject to these difficulties, and that there is not sufficient reason to impose this new area of dispute and peripheral litigation upon the trial and appeal of patent cases.
The proponents of stare decisis point to the courts’ and patent community’s fifteen years of experience with Cybor, and argue that this experience supports retention of the Cybor principle. Emphasizing the potential multi-case and multi-forum litigation of patents on today’s technologies, they argue that it is particularly important that this court be able to resolve claim construction definitively as a matter of precedent, rather than allow differing trial court constructions of the same patent, as may result from deferential review of close questions. As the Court observed in Markman II, “treating interpretive issues as purely legal will promote (though it will not guarantee) intrajurisdictional certainty through the application of stare decisis on those questions not yet subject to interju-risdictional uniformity under the authority of the single appeals court.” 517 U.S. at 391, 116 S.Ct. 1384.
Thus it is argued that the meaning and scope of a patent claim, which sets the boundaries of an exclusionary right good against the world at large, rather than only for parties to a voluntary transaction or only for the plaintiff and defendant in a particular case, should be construed based on publicly available materials in the record, and resolved for uniform application throughout the nation, as a matter of law. No party or amicus disputed that the only way to achieve uniform construction of the same claim, a goal recognized in Markman II, is by de novo appellate construction of the claim as a matter of law. There is no dispute as to the importance of national uniformity and finality of claim construction, for it is not unusual for different district courts to litigate the same patent against different parties and different assertions of infringement.
In sum, these proponents argue that the Cybor standard of review of claim construction reasonably and appropriately implements the Court’s ruling in Markman II, and urge this court to stand by Cybor and its fifteen years of experience. Proponents of stability through the principles of stare decisis stress that consistency of legal analysis and reliability of judicial process are foundations of not only legal [1281]*1281systems generally, but also of the technological advance and industrial commitment that are goals of the patent system. See Brief of Amici Curiae Cisco et al. at 15 (“Competing and inconsistent interpretations of patent claims obscure the boundaries of patents and deeply undermine their important notice function, inevitably resulting in more — rather than less — litigation.”); id. at 19 (“Clear scope is important to all potential market entrants. This kind of horizontal certainty is important to the entire industry.”).
The criticism of Cybor is not based on any demonstration that de novo claim construction is likely to be incorrect, but rather on concerns for judicial roles and relationships. We do not ignore these important concerns. However, as proponents of stare decisis point out, Cybor is narrowly focused on the threshold construction of a legal document, and does not affect the traditional deference to district court findings of infringement or validity or damages or any other question of fact in patent litigation. The proponents remind us that Cybor was adopted in implementation of the Markman decisions of this court and the Supreme Court, and is in accord with these rulings on the nature of the patent grant and the judicial obligation for correct determination of the legal scope of the patent grant.
Thus it is urged that the Cybor standard should not now be abandoned for a more costly and litigious standard, with diminished stability of procedure and diminished reliability of outcome, and no greater likelihood of correctness of result. These proponents point out that those who would change Cybor’s system of plenary review of claim construction have not shown any benefit or advantage to the law or those served by the law. Thus it is argued that the values of stare decisis counsel against overturning Cybor.
II
StaRE Decisis
The question now before this en banc court is not the same question that was before the en banc court in 1998 when Cybor was decided. The question now is not whether to adopt a de novo standard of review of claim construction, but whether to change that standard adopted fifteen years ago and applied in many hundreds of decisions. There has been extensive experience of Cybor in action, in the district court and on appeal. “Claim construction” has become the gateway issue in patent litigation, often decided in preliminary proceedings before trial and before discovery, and often subject to immediate appeal on summary judgment or injunction grounds. Such experience enriches the principle that courts will “stand by things decided” so that prior rulings may be relied upon.
Stare decisis is of “fundamental importance to the rule of law.” Hilton v. S. Carolina Pub. Ry. Comm’n, 502 U.S. 197, 202, 112 S.Ct. 560, 116 L.Ed.2d 560 (1991) (quoting Welch v. Texas Dep’t of Highways & Pub. Transp., 483 U.S. 468, 494, 107 S.Ct. 2941, 97 L.Ed.2d 389 (1987)). The doctrine of stare decisis enhances predictability and efficiency in dispute resolution and legal proceedings, by enabling and fostering reliance on prior rulings. CSX Transp. Inc. v. McBride, — U.S. —, 131 S.Ct. 2630, 2641, 180 L.Ed.2d 637 (2011). By providing stability of law that has been decided, stare decisis is the foundation of a nation governed by law. The Supreme Court has said: “we will not depart from the doctrine of stare decisis without some compelling justification.” Hilton, 502 U.S. at 202, 112 S.Ct. 560 (citing Arizona v. Rumsey, 467 U.S. 203, 212, 104 S.Ct. 2305, 81 L.Ed.2d 164 (1984)); see Dickerson v. United States, 530 U.S. [1282]*1282428, 443, 120 S.Ct. 2326, 147 L.Ed.2d 405 (2000) (“special justification” is needed to overrule precedent).
Stability in procedural as well as substantive law, on which the public and the courts can rely, guards against the expenditure of time and resources on aspects that have been resolved. These values come to the fore when a court undertakes to reexamine its own precedent, for stare decisis implements the “prudential and pragmatic considerations designed to test the consistency of overruling a prior decision with the ideal of the rule of law, and to gauge the respective costs of reaffirming and overruling a prior case.” Planned Parenthood of S.E. Pa. v. Casey, 505 U.S. 833, 854, 112 S.Ct. 2791, 120 L.Ed.2d 674 (1992). The principles and policies of stare decisis operate with full force where, as here, the en banc court is considering overturning its own en banc precedent.
The presumption that a court will adhere to its prior rulings has “ ‘special force’ ” for precedents that resolve non-constitutional issues, for ‘“Congress remains free to alter what we have done.’ ” J.R. Sand & Gravel Co. v. United States, 552 U.S. 130, 139, 128 S.Ct. 750, 169 L.Ed.2d 591 (2008) (quoting Patterson v. McLean Credit Union, 491 U.S. 164, 172-73, 109 S.Ct. 2363, 105 L.Ed.2d 132 (1989)). In Patterson the Court observed that the same issue had previously divided the Court and that “[s]ome Members of this Court believe that [the precedent] was decided incorrectly”; the Court discussed the principles of stare decisis, and concluded that “no special justification has been shown for overruling” the prior decision, for neither “the growth of judicial doctrine or further action taken by Congress.... have removed or weakened the conceptual underpinnings from the prior decision.” 491 U.S. at 171-173, 109 S.Ct. 2363. The Court observed that no “later law has rendered the decision irreconcilable with competing legal doctrines or policies.” Id. at 173, 109 S.Ct. 2363.
In Hohn v. United States, 524 U.S. 236, 253, 118 S.Ct. 1969, 141 L.Ed.2d 242 (1998) the Court discussed its precedent in light of stare decisis, and stated that “[o]nee we have decided to reconsider a particular rule, however, we would be remiss if we did not consider the consistency with which it has been applied in practice.” In Global-Tech Appliances, Inc. v. SEB S.A., — U.S. —, 131 S.Ct. 2060, 2067-68, 179 L.Ed.2d 1167 (2011) the Court observed that although its prior decision on the same issue was “badly fractured,” the majority holding “has become a fixture in the law,” warranting application of the doctrine of stare decisis.
The purposes of consistency and stability that underlie stare decisis led to the formation of the Federal Circuit, now thirty years past, to provide consistency and stability to the patent law: “The central purpose is to reduce the widespread lack of uniformity and uncertainty of legal doctrine that exist in the administration of patent law,” H.R. Rep. 97-312, at 23 (1981), in view of the importance of technology-based advance to the nation’s economy, id.; S. Rep. 97-275, at 6 (1981) (same). Legal doctrine in patent law starts with the construction of patent claims, for the claims measure the legal rights provided by the patent.
In the Federal Circuit decision that led to the Supreme Court’s Markman II ruling, this court explained that reviewing claim construction as a matter of law assures “a true and consistent scope of the claims.” Markman I, 52 F.3d at 979. The Court did not adjust or criticize that position, which forms the foundation of Cybor. Cybor, in carrying forward Markman I in [1283]*1283light of Markman II, embodies the view that de novo review will “help institute a simplified and clarified method by which both trial and appellate courts address claim construction issues.” Cybor, 138 F.3d at 1463 (Plager, J., concurring).
We do not ignore that neither Cybor nor the Markman decisions were free of contention, then as now. As we undertake this review of Cybor, we recognize that stare decisis is not an “inexorable command,” Agostini v. Felton, 521 U.S. 203, 235, 117 S.Ct. 1997, 138 L.Ed.2d 391 (1997). Thus we have considered whether there are sound reasons for this court now to depart from this precedent. We proceed with the guidance of history, experience, and the many amici curiae, while retaining awareness that to overturn an en banc ruling that has had long and wide application, there must be more than controversy about the prior rule. See Watson v. United States, 552 U.S. 74, 82, 128 S.Ct. 579, 169 L.Ed.2d 472 (2007) (“A difference of opinion within the Court ... does not keep the door open for another try....”). As observed in Morrow v. Balaski 719 F.3d 160, 181 (3d Cir.2013) (Smith, J., concurring), “the very point of stare decisis is to forbid us from revisiting a debate every time there are reasonable arguments to be made on both sides.”
However, departure from precedent may be appropriate when “subsequent cases have undermined [its] doctrinal underpinnings,” Dickerson, 530 U.S. at 443, 120 S.Ct. 2326; or when the precedent has proved “unworkable,” J.R. Sand & Gravel Co., 552 U.S. at 139, 128 S.Ct. 750; or when “a considerable body of new experience” requires changing the law, Pearson v. Callahan, 555 U.S. 223, 234, 129 S.Ct. 808, 172 L.Ed.2d 565 (2009).
Stare decisis embraces procedural as well as substantive precedent. Vasquez v. Hillery, 474 U.S. 254, 266, 106 S.Ct. 617, 88 L.Ed.2d 598 (1986) (considering whether procedures have so far developed as to have left the old rule “outdated, ill-founded, unworkable, or otherwise legitimately vulnerable to serious reconsideration”); Swift & Co. v. Wickham, 382 U.S. 111, 116, 86 S.Ct. 258, 15 L.Ed.2d 194 (1965) (overruling procedural precedent where “unworkable in practice,” among other problems). Procedures in the litigation-prone arena of patent rights can affect the cost, time, and uncertainty of litigation, and in turn affect economic activity founded on the presence or absence of enforceable patents. Courts should be “cautious before adopting changes that disrupt the settled expectations of the inventing community.” Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722, 739, 122 S.Ct. 1831, 152 L.Ed.2d 944 (2002).
Applying these premises, we have reviewed the arguments for changing the Cybor procedure of de novo review of claim construction. First, we have looked for post Cybor developments, whether from the Supreme Court, from Congress, or from this court, that may have undermined the reasoning of Cybor. None has been found, or brought to our attention. There has been no legislative adjustment of the Cybor procedure, despite extensive patent-related legislative activity during the entire period of Cybor’s existence.
We have looked for some demonstration that Cybor has proved unworkable. No proponent of change has shown that de novo review of claim construction is unworkable' — nor could they, after fifteen years of experience of ready workability. Nor has anyone shown that Cybor has increased the burdens on the courts or litigants conducting claim construction.
To the contrary, reversing Cybor or modifying it to introduce a fact/law distinction has a high potential to diminish work[1284]*1284ability and increase burdens by adding a new and uncertain inquiry, not only on appeal but also in the trial tribunal. No consensus has emerged as to how to adjust Cybor to resolve its perceived flaws. Despite probing questioning at the en banc hearing, and despite the extensive amicus curiae participation, there is no agreement on a preferable new mechanism of appellate review of claim construction; there is no analysis of how deference would be applied to the diversity of old and new technologies and modes of claiming, no clear exposition of fact or law as could be applicable to the millions of unexpired patents, each on a different new technologic advance. As will be discussed, no one, including the dissent, proposes a workable replacement standard for Cybor, no workable delineation of what constitutes fact and what constitutes law.
Disentangling arguably factual aspects, some in dispute and some not, some the subject of expert or other testimony and some not, some elaborated by documentary evidence and some not, some construed by the district court and some not, some related to issues to be decided by a jury and some not — and further disentangling factual aspects from the application of law to fact — is a task ripe for lengthy peripheral litigation. We are not persuaded that we ought to overturn the en banc Cybor decision and replace its clear de novo standard with an amorphous standard that places a new, cumbersome, and costly process at the gate, to engender threshold litigation over whether there was or was not a fact at issue. The principles of stare decisis counsel against such an unnecessary change.
No critic of Cybor has provided any analysis of specific claims or cases to show, or even to suggest, that deferential appellate review is more likely to achieve the correct claim construction. The principles of stare decisis counsel against overturning precedent when there is no evidence of unworkability and no clearly better resolution. The amici curiae agreed that modification of Cybor is unlikely to change many results, even if it could be defined well (which it has not been, by any amicus or by our colleagues in dissent).
Claim construction is a legal statement of the scope of the patent right; it does not turn on witness credibility, but on the content of the patent documents. The court may indeed benefit from explanation of the technology and the instruction of treatises, but the elaboration of experts or tutorial explanation of technical subject matter does not convert patent claim construction into a question of fact. The type of evidence that may assist a lay judge in determining what a technical term meant to one of skill in the art does not transform that meaning from a question of law into a question of fact. Reference to technical understanding and usage at the time of enactment does not convert statutory interpretation from law to fact. See, e.g., Corning Glass Works v. Brennan, 417 U.S. 188, 202, 94 S.Ct. 2223, 41 L.Ed.2d 1 (1974) (applying “the language of industrial relations” to statutory interpretation).
Courts routinely look to dictionaries and treatises to determine the meaning of a statute at the time it was written. See, e.g., Sandifer v. U.S. Steel Corp., — U.S. —, 134 S.Ct. 870, 187 L.Ed.2d 729 (2014) (looking to dictionaries to determine the meaning of “changing clothes” in the Fair Labor Standards Act at the time of enactment); Comm’r v. Soliman, 506 U.S. 168, 113 S.Ct. 701, 121 L.Ed.2d 634 (1993) (interpreting the tax code by looking to dictionary definitions at the time of enactment); Chapman v. United States, 500 U.S. 453, 111 S.Ct. 1919, 114 L.Ed.2d 524 (1991) (using dictionaries to ascertain the [1285]*1285ordinary meaning of “mixture” in the drug statute).
Similarly, experts in the science or technology may assist the court in understanding the meaning and usage of a claim term, but this does not morph the question into one of fact. Cf. United States v. Stone & Downer Co., 274 U.S. 225, 47 S.Ct. 616, 71 L.Ed. 1013 (1927) (relying on expert testimony on the meaning of the tariff term “clothing wool” in the custom of the trade). The Court stated in Markman II:
in theory there could be a case in which a simple credibility judgment would suffice to choose between experts whose testimony was equally consistent with a patent’s internal logic. But our own experience with document construction leaves us doubtful that trial courts will run into many cases like that. In the main, we expect, any credibility determinations will be subsumed within the necessarily sophisticated analysis of the whole document, required by the standard construction rule that a term can be defined only in a way that comports with the instrument as a whole.
517 U.S. at 389, 116 S.Ct. 1384. This expectation has proven accurate. The presentation of expert testimony on the meaning of a claim term does not transform the question from one of law to one of fact.
We have carefully considered the arguments for discarding or modifying Cybor, and conclude that they do not justify departing from the now well-established principles and procedures. Under any standard of review consistent with Mark-man II, most issues of claim construction are indisputably matters of law, and would receive de novo review. Even the critics of Cybor agree that any change would affect only a small number of claim construction disputes. Statements at the en banc hearing are edifying; e.g., Tr. at 1:14:10-1:14:35 (Appellant) (“This court could, in whatever it does with Cybor, make very clear that this battle of competing experts is rarely ’productive, rarely going to influence claim construction.... If you said to me, why doesn’t stare decisis carry the day? I don’t have a good answer to that.”); id. at 55:15-55:20 (United States) (“this court’s law on claim construction requires very little modification”); id. at 1:06:50-1:07:30 (United States) (whether a term has special meaning in the art could be “meaningless ultimately” if inconsistent with the intrinsic record).
In response to a question at the hearing, amicus curiae United States could not identify any case that would have come out differently under the modified (hybrid) standard of review it proposed. Tr. at 1:07:30-1:08:00. Certainly stare decisis counsels against overturning en banc precedent where doing so would change our known, workable de novo standard to an undefined alternative, sure to engender peripheral litigation, and which most agree could affect the outcome of very few, if any cases. See Pearson, 555 U.S. at 236-37, 129 S.Ct. 808 (criticizing precedent that “sometimes results in a substantial expenditure of scarce judicial resources on difficult questions that have no effect on the outcome of the case”).
Claim construction is often a preliminary proceeding in the district court, before trial of infringement, validity, damages, etc. At the threshold, the court establishes the metes and bounds of the claims that define the patent right. The questions of claim construction are not questions of weight of evidence or credibility of witnesses, but of the claim scope as set forth in the patent documents.
Claim construction is the interpretation of a legal document that establishes [1286]*1286a property right that applies throughout the nation. The question under review is whether this court, of nation-wide jurisdiction, should continue to review claim construction de novo with national effect, or whether to change to a system whereby a district court’s claim construction is reviewed on the deferential standard appropriate to findings of fact, with or without some sort of hybrid deference to the ultimate determination. The insistence of some amici curiae that some form of deferential review is required as well as superi- or, is not only contrary to the Markman holdings, but to the experience of fifteen years of Cybor.
In the increasingly frequent situation where the same patent is litigated in different forums against different defendants, differing district court rulings on close questions of claim construction could well warrant affirmance on deferential review. Because differing claim constructions can lead to different results for infringement and validity, the possibility of disparate district court constructions unravels the “uniformity in the treatment of a given patent” that the Court sought to achieve in Markman II. 517 U.S. at 390, 116 S.Ct. 1384. It would restore the forum shopping that the Federal Circuit was created to avoid. Just as the Court in Markman II counted such consequences as negatives that its ruling overcame, they count as negatives in the stare decisis analysis.
The question that this court has now reconsidered is whether we should continue to review claim construction as a whole and de novo on the record, or whether we should change to a different system that at best would require us to identify any factual aspects and how the trial judge decided them, and review any found or inferred facts not for correctness but on a deferential standard, with or without also giving deferential review to the ultimate determination of the meaning of the claims. We conclude that such changed procedure is not superior to the existing posture of plenary review of claim construction.
Over these fifteen years this court has applied Cybor to diverse subject matter, and the body of precedent has grown large. Deferential review does not promise either improved consistency or increased clarity. We have been offered no argument of public policy, or changed circumstances, or unworkability or intolera-bility, or any other justification for changing the Cybor methodology and abandoning de novo review of claim construction.
The proponents of overruling Cybor have not met the demanding standards of the doctrine of stare decisis. They have not shown that Cybor is inconsistent with any law or precedent, or that greater deference will produce any greater public or private benefit. We conclude that there is neither “grave necessity” nor “special justification” for departing from Cybor.
Ill
REMARKS ON THE DlSSENT
Our colleagues in dissent offer a few arguments that warrant response. First, referring to “the materials submitted to the court,” the dissent states that “a substantial proportion of the legal community” believes that Cybor was “wrongly decided.” Diss. at 1298. The materials tell a different tale.
As listed ante, at n. 2, thirty-eight organizations and individuals filed twenty-one amicus briefs. Contrary to the dissent’s statements, all of the technology industries that offered advice to the court, urge retention of Cybor’s standard. These amici curiae include the largest technology companies in the nation, all involved with the system of patents, all frequent patent [1287]*1287litigants both as plaintiffs and as defendants — unlike all of the other amici.4 The dissent dismisses these voices as merely “some amici” who support retention of Cy-bor, Diss. at 1298, and offers no response to their concerns for stability, national uniformity, and predictability in claim construction.
The dissent appears unconcerned that the major industrial amici urge retention of the Cybor standard, and instead announces that “no one in the legal community — except perhaps the members of the majority — has come to believe that either the wisdom or vitality of Cybor is settled,” Diss. at 1298. This conclusion is curious. For example, the amicus brief of Google, Amazon, Hewlett-Packard, Red Hat and Yahoo! states that departing from Cybor would “make worse” the uncertainty of claim construction:
[T]he root causes of uncertainty in claim construction are vaguely drafted claims and contradictory claim-construction methodologies, not appellate review. Deference would not ameliorate those causes of uncertainty; it would make them worse.
[Tjreating claim construction as a factual question subject to dear-error review would only aggravate the uncertainty and cost issues plaguing our patent-litigation system.
Brief of Amicus Curiae Google et al. at 4, 5.
The industrial amici also respond to the argument, pressed by the dissenters, that treating claim construction as a matter of law negates settlement and increases litigation cost. Diss. at 1313. These litigants advise that the contrary is true:
[Classifying claim construction as being at least partly factual would make patent litigation even more costly by discouraging courts from resolving claim construction disputes at the outset. Early claim construction is essential to permit the parties to file summary judgment motions, or to engage in informed settlement discussions, before they have to incur potentially unnecessary discovery and other pre-trial costs — costs that force many defendants to settle even meritless cases solely because the exorbitant cost of litigating a case would exceed the settlement amount demanded by the plaintiff.
Google et al. Br. at 4-5. Even Appellant’s counsel at oral argument contradicted the cost-of-litigation and settlement arguments:
A lot of commentators have said Cy-bor is preventing settlements: I don’t believe that. I settle cases all the time. No one has ever focused primarily or even significantly on the standard of review on appeal. They’re focused on the jury. They’re focused on the cost of litigation. So a lot of what’s been put up as reasons to change Cybor I don’t think are there.
Tr. 1:14:40-1:15:05.
In the brief filed by Cisco, Dell, EMC, Intel, SAP, and the SAS Institute, these amici curiae suggest that the proponents of overturning Cybor incorrectly conflate concepts of uncertainty with appellate reversal rates. These amici explain that any possible uncertainty of affirmance on appeal is not the issue in claim construction; [1288]*1288rather, the issue is how to generate accuracy and uniformity in claim construction, that is, how to construe claims correctly and predictably.
Clear scope is important to all potential market entrants. This kind of horizontal certainty is important to the entire industry. By contrast, the concern that de novo review increases the “duration” of a single patent litigation until a final decision is reached in that particular case (Cybor, 138 F.3d at 1476 (Rader, J., dissenting)) — what might be called vertical uncertainty — matters only in the small fraction of cases that reach an appeal. Vertical uncertainty is more visible than horizontal uncertainty, but, as often is the case, here it is the unseen effects that are greater. Cf. Frederic Bastiat, What Is Seen and What Is Not Seen (1848), available at http:// www. econlib.org/library/Bastiat/basEssl.html.
For this reason, it is not merely the overarching principles of claim construction, but their application, that must be consistent. In claim construction as elsewhere, “the relevant legal principle can be given meaning only through its application to the particular circumstances of a case.” Miller, 474 U.S. at 114, 106 S.Ct. 445.
Brief of Amicus Curiae Cisco et al. at 19.
The Cisco amici emphasize the Supreme Court’s recognition of “the importance of uniformity” in Markman II, 517 U.S. at 390, 116 S.Ct. 1384, and stress that treating claim construction issues as “purely legal,” id. at 391, 116 S.Ct. 1384, enables appellate review to supply that uniformity:
Federal Circuit review, and in particular this Court’s application of stare deci-sis, is critical to such uniformity. As the Supreme Court explained, “treating interpretive issues as purely legal” will allow stare decisis to be applied to those interpretive questions and thus promote uniformity among decisions of this Court (so-called “intrajurisdictional certainty”).
Cisco et al. Br. at 5 (citing and quoting Markman II, 517 U.S. at 391, 116 S.Ct. 1384). These amici also discuss the “hybrid” proposal of some theorists, and state:
“treating interpretive issues as purely legal” — not as a mixed question — is the proper approach.
Id. at 5 (quoting Markman II, 517 U.S. at 391, 116 S.Ct. 1384).
Amicus Microsoft discusses another aspect of the Cybor treatment of claim construction, advising that de novo review
works well in practice because it allows the parties to address questions concerning claim scope in pre-trial hearings well in advance of the actual trial. Having a jury decide these redesignated factual matters could eliminate the current practice of pre-trial Markman hearings. This would exacerbate the expense and uncertainty in patent litigation and create new opportunities for forum shopping.
Brief of Amicus Curiae Microsoft at 3.
In sum, the amici curiae record of the nation’s major innovators is contrary to the dissent’s representation that “no one in the legal community — except perhaps the members of the majority — has come to believe that either the wisdom or vitality of Cybor is settled.” Diss. at 1298. The amici curiae record is contrary to the dissent’s pronouncement that “the interests of stability and predictability are disserved” by Cybor’s access to final, uniform, national claim construction. Diss. at 1310. The nation’s major innovators do not agree with the criticisms of Cybor, and so advised the court. These amici direct the court to the pragmatic value of Cybor as they have experienced it, and urge retention of that value.
[1289]*1289Rather than respond to the concerns of the nation’s technology industries, the dissent chastizes certain judges of this court for not “acting on their long-term convictions.” Diss. at 1296. While it is comforting to know that our golden words of the past are not forgotten, those of us with the majority today who have questioned aspects of Cybor in the past, now decide this case on the record of the present and with an eye to the future. The dissent would discard the experience of the past fifteen years. However, the court is not now deciding whether to adopt a de novo standard in 1998. Today we decide whether to cast aside the standard that has been in place for fifteen years.
The dissent offers no superior alternative to de novo review, nor any workable standard for distinguishing between legal and factual components of claim construction. The dissent does not appear to adopt the proposal of several amici that appellate deference should depend on whether the district court relied on intrinsic or extrinsic evidence in construing a claim term. It is surely doubtful that such a distinction controls whether claim construction is fact or law. See Phillips v. AWH Corp., 415 F.3d 1303, 1317 (Fed.Cir.2005) (en banc) (discussing sources of evidence in claim construction, and stating “while extrinsic evidence ‘can shed useful light on the relevant art,’ we have explained that it is ‘less significant than the intrinsic record in determining the “legally operative meaning of claim language” ’ ”) (citing cases). We have come upon no rationale for denominating an issue of claim construction as one of fact or law depending on the source of the information considered by the judge.
The dissent seems to embrace (then expand upon) the “historical fact” notion proposed by Appellant, who states that the meaning a person of ordinary skill in the art would give a term at the time of the filing of the patent application ought to be treated as a question of fact. This is not a question of fact; it is the very inquiry which determines the claim construction in nearly all cases. Claim terms are given their ordinary meaning to one of skill in the art, unless the patent documents show that the patentee departed from that meaning. See Phillips, 415 F.3d at 1314; Thorner v. Sony Computer Entm’t Am., LLC, 669 F.3d 1362, 1365 (Fed.Cir.2012). Treating the ordinary meaning to a skilled artisan as requiring deference would mean deference on the controlling question of claim construction in nearly every case. All the more so with the dissent’s proposed treatment of still additional issues, such as prosecution disclaimer, as warranting deference. Diss. at 1316.
Under the dissent’s approach, and even under the “historical fact” approach, deference would become of central significance in controlling the determination of claim construction, and hence of patent scope. The consequence would be heightened forum-shopping and the inability of the judicial system to arrive at a uniform, settled meaning for a patent’s scope. Those problems are grave ones given the increasingly common situation of multiple cases involving the same patent.
The dissenters do not explain why they choose to abandon the benefits foreseen by the Court in Markman II, that “treating interpretive issues as purely legal” would have benefits of “intrajurisdictional certainty.” 517 U.S. at 391, 116 S.Ct. 1384. Nor have they successfully explained away the analysis in Miller v. Fenton, 474 U.S. at 113-14, 106 S.Ct. 445, that:
[T]he practical truth that the decision to label an issue a “question of law,” a “question of fact,” or a “mixed question of law and fact” is sometimes as much a matter of allocation as it is of analysis.... [T]he fact/law distinction at [1290]*1290times has turned on a determination that, as a matter of the sound administration of justice, one judicial actor is better positioned than another to decide the issue in question.
(cited in Markman II, 517 U.S. at 388, 116 S.Ct. 1384). The Court in Miller explained that the fact/law distinction is not immutable, and may invoke “the sound administration of justice,” id. at 114, 106 S.Ct. 445, leading to a similar acknowl-edgement in the Markman II ruling that claim construction is “a matter of law reserved entirely for the court.” 517 U.S. at 372, 116 S.Ct. 1384.
In addition, the dissent downplays the gravity of overturning a previous en banc court in the absence of intervening Supreme Court or legislative action. Of the several decisions of the Federal Circuit that the dissent cites as setting a pattern of overturning precedent, all involve en banc review of panel precedents, with the arguable exception of Nobelpharma AB v. Implant Innovations, Inc., 141 F.3d 1059, 1068 & n. 5 (Fed.Cir.1998) (en banc in relevant part), where the court clarified that this court’s law would govern the question of patent/antitrust immunity.
The major thrust of the dissent is that Federal Rule 52(a)(6) requires deferential review of district court decisions. But Rule 52(a) does not answer the question here. Rule 52(a) prescribes the standard of review of questions of fact, but courts must look outside the Rule to decide if a question is properly characterized as one of fact. As the Court stated in Pullman-Standard, 456 U.S. at 288, 102 S.Ct. 1781, “Rule 52(a) does not furnish particular guidance with respect to distinguishing law from fact. Nor do we yet know of any other rule or principle that will unerringly distinguish a factual finding from a legal conclusion.” The dissent’s theory that Rule 52(a) demands abandonment of de novo review of claim construction is a simplistic disregard of the Markman II guidance that “treating interpretive issues as purely legal will promote (though it will not guarantee) intra-jurisdictional certainty.” 517 U.S. at 391, 116 S.Ct. 1384.
The dissent argues that de novo review produces a high reversal rate, although it is established that this is no longer true. The reversal rate indeed was a matter of concern, for in the early years of Cybor, this court’s promulgation of claim construction law led to a higher rate of appellate adjustment. However, as consistency evolved and experience grew, rates of appellate reversal for claim construction came to match the norm for other grounds. We observe that every amicus brief that complains about high reversal rates relies on data that are seven to ten or more years old, while the author of a recent study writes in his amicus brief that the data “document a significant drop in the claim construction reversal rate” since 2005, Brief of Amicus Curiae Professor Peter Menell at 15, and explains in his study that “[n]ow the reversal rate for claim construction appeals is much closer to that of other patent-related issues.” J. Jonas Anderson & Peter S. Menell, Informal Deference: An Historical, Empirical, and Normative Analysis of Patent Claim Construction, 108 Nw. U.L.Rev. (forthcoming), Sept. 9, 2013 manuscript at 37.
Our colleagues in dissent, citing the obsolete data, argue that the de novo standard “adds considerable uncertainty and expense to patent litigation,” Diss. at 1297, stating that this standard increases appeals, discourages settlement, and increases the length and cost of litigation. No evidence of this effect is offered, and all of the amici curiae who are frequent litigants state the contrary position. The data published by the Administrative Office of the [1291]*1291United States Courts point the other way. These data show a long, noticeable decline in the percentage of district court patent cases that are appealed, belying the argument that appeals have increased. The following data are from the Annual Reports of the Administrative Office for the years 1994 to 2013. The graph shows the ratio, as a percentage, of the number of patent appeals filed per year (Report Table B-8), against the number of district court patent cases filed in that year (Report Table C-2):
[[Image here]]
The Annual Reports also show the trend data from Table C-4 in the Reports show in the percentage of patent cases that pro- the percentage declining from 5.9% in 1994 ceed to trial in the district courts. The to 2.8% in 2013:
The data do not support the dissent’s theory that Cybor has increased patent litigation and inhibited settlements. In contrast, the industrial amici curiae advise the court that the Cybor review procedure assists in resolving litigation before full trial or extensive discovery, for it often leads to the grant of summary judgment and an immediate appeal. These amici stress that settlement is facilitated by final resolution [1292]*1292of the scope of the claims. They also point out that if the claim construction is definitively resolved, any ensuing trial is on the final claim construction. The dissent does not comment on these values.
Conolusion
We have again considered the standard of review of district court claim construction rulings, in light of experience with the Cybor standard. The ever-enlarging importance of technology-based industry in the economy has reinforced the need for an optimum patent system.
On thorough review, we are not persuaded that discarding de novo review would produce a better or more reliable or more accurate or more just determination of patent claim scope. Those who urge change in the Cybor standard have identified no pattern of error, no indictment of inferior results. No ground has been shown for departing from the principles of stare decisis. Review of claim construction as a matter of law has demonstrated its feasibility, experience has enlarged its values, and no clearly better alternative has been proposed. There has arisen no intervening precedent, no contrary legislation, no shift in public policy, no unworkability of the standard.
We conclude that the criteria are not met for overruling or modifying the Cybor standard of de novo review of claim construction as a matter of law.
PANEL DECISION REINSTATED
No costs.