dissenting.
I respectfully dissent. I believe the majority has reached a wrong conclusion because it has ignored basic principles of contract law. While the majority references the guiding principles of contract law, the majority opinion discusses them in only the most limited detail. In another very brief section, which cites only generally inapposite constitutional law, the majority opinion attempts to explain why the state should be treated differently from any other contracting party.
While the district court below devoted substantial analysis to this problem and the relevant doctrine, Doe-SORA IV, 427 F.Supp.2d at 4111 (citing, inter alia, United States Trust Co. v. New Jersey, 431 U.S. 1, 17, 97 S.Ct. 1505, 52 L.Ed.2d 92 (1977)),2 on appellate review the majority has eschewed this approach, and instead focused on discerning the intentions of the parties. This approach is problematic, since as even defendants-appellants had initially argued, the stipulation is not ambiguous, and we cannot ignore the lack of ambiguity merely because its plain meaning does not correspond "with the parties’ alleged intentions.
This court cannot merely point to the absence of additional language specifying the allocation of risk in the event of changed circumstances as a basis for shifting that risk onto the obligee. This approach, along with the imposition of a new clear statement rule retroactively, so as to shift the burden of changed circumstances away from the party who would traditionally bear it, are radical departures from our earlier jurisprudence and the longstanding principles of the common law. Furthermore, if the agreement was ambiguous, the only prudent approach would be to remand to the district court for further fact finding, given the paucity of the record. Instead, the majority opinion relies on the absence of evidence, inapplicable principles of contract interpretation, and faulty analogies to inapposite constitutional law to establish a basis for reformation of the stipulation.
I. Under New York Law and the Principles Embodied in this Circuit’s Case-law, the Stipulation at Issue is Not Ambiguous
A dispute over the meaning of a provision in a written instrument after it is executed does not generate ambiguity. It is the text of the agreement itself which *81must be ambiguous before a court can construe the contract in view of the objective evidence of the parties’ shared intentions. However, the majority opinion does not identify the words, syntax or punctuation that render the stipulation ambiguous, stating only that while “there is no dispute as to the meaning of the words in these recitations,” “there is a dispute [as to their] operative effect.” The first question before us is whether this dispute is legitimate. The majority opinion does not address this question. Rather than pointing to internal textual inconsistencies that might illustrate the purported ambiguity, it proceeds immediately to the determination of how the stipulation should be construed.
This court has held — repeatedly and consistently — that before resorting to indi-cia of intent, we must determine that an ambiguity is present on the face of the contract; it is not sufficient for the contract to merely appear ambiguous in the light of the previous dealings between the parties. This foundational principle applies equally to consent decrees: “[w]hen the language of a consent decree is unambiguous, the scope of a consent decree must be discerned within its four corners.” United States v. Broad. Music, Inc., 275 F.3d 168, 175 (2d Cir.2001).
This “plain meaning” approach to contract construction is the well-established law of the circuit. Roberts v. Consolidated Rail Corp., 893 F.2d 21, 24 (2d Cir.1989) (“[ajbsent an ambiguity in a written contract, courts will not look to the underlying intent of the parties in executing the contract”); see also O’Neil v. Ret. Plan for Salaried Employees of RKO Gen., Inc., 37 F.3d 55, 58-59 (2d Cir.1994) (holding that the question of “[wjhether contract language is ambiguous ... is resolved by reference to the contract alone”) (internal quotation marks omitted); accord Collins v. Harrison-Bode, 303 F.3d 429, 433 (2d Cir.2002) (noting that “[ujnder New York law, the question of ambiguity vel non must be determined from the face of the agreement, without reference to extrinsic evidence”).3
Conversely, the majority reasons, without even mentioning the question of ambiguity, that the “[pjroper construction of the recitations at issue, we believe, depends not only on the words of the Stipulation, but also, and more importantly, on whether the parties intended to place these words into the agreement as part of a resolution of disputed matters.” Majority Op. [76] (emphasis added). This approach is contrary to our precedent and the principles of contract law, which the majority agrees should govern. Majority Op. [75].
The importance of our steadfast adherence to the plain meaning rule becomes clear when we consider the unbridled freedom an appellate court has when it is abandoned; we could then rewrite any contract to correspond with what we believe to be the intentions of the parties, utilizing the exceptional equitable remedy of reformation under the guise of contract construction. However, “[t]here is no power at common law to reform a written instrument.” 76 C.J.S. Reformation of In*82terests § 2; see also e.g. Ivinson v. Hutton, 98 U.S. 79, 82, 25 L.Ed. 66 (1878). Accordingly, “[a] court may neither rewrite, under the guise of interpretation, a term of the contract when the term is clear and unambiguous, nor redraft a contract to accord with its instinct for the dispensation of equity upon the facts of a given case.” Terwilliger v. Terwilliger, 206 F.3d 240, 245 (2d Cir.2000).
Despite this governing precedent, the reformation of the stipulation is precisely what the majority opinion effects. Here, the majority’s recharacterization of the “operative effect” of these provisions — into a mere reference to the state of the existing law — effectively inserts a new provision of critical importance into the agreement. The contract, as rewritten by the majority, now contains a new provision in the sections referring to SORA’s notice and duration of registration requirements that reads: “these terms are included only to provide an indication of the existing statutory provisions; the State of New York reserves the right to modify the scope and term of your registration as a sex offender,” despite the fact that “[a] promise ... with a reserved right to deny or change the effect of the promise, is an absurdity.” Murray v. Charleston, 96 U.S. 432, 445, 24 L.Ed. 760 (1877).
II. The District Court’s Factual Findings Were Not Clearly Erroneous
Even if the stipulation is ambiguous, we would be required to address the district court’s factual findings, to which we owe considerable deference, in far more detail to be confident that vacatur is required. However, the majority opinion does not address this issue squarely, relegating to a footnote the finding that “to whatever extent the District Court might have made ... finding[s] ... [they] would be clearly erroneous.” Majority Op. [77 n. 9], However, since “[w]here there are two permissible views of the evidence [of the parties’ intent], the court’s pure findings of fact cannot be termed clearly erroneous,” United States Naval Institute v. Charter Communications, Inc., 875 F.2d 1044, 1049 (2d Cir.1989), the majority opinion implicitly holds in the alternative that the district court’s well-reasoned and thoughtful opinions below contained findings that were not even plausible. “The reviewing court oversteps the bounds of its duty under Rule 52(a) if it undertakes to duplicate the role of the lower court. In applying the clearly erroneous standard to the findings of a district court sitting without a jury, appellate courts must constantly have in mind that their function is not to decide factual issues de novo.... [T]he court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently.” Anderson v. Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985).
The district court found that defendants-appellants had “agreed to, and indeed bargained for” the notification procedures applicable to each class of SORA registrants, as embodied in the consent decree. Doe-SORA V, 439 F.Supp.2d at 326, see also Doe-SORA TV, 427 F.Supp.2d at 408-09 (considering how the circumstances of the litigation illuminate the parties’ intentions). While the majority opinion contains a discussion of why the external evidence of the parties’ intentions can be cast in a different light so as to reach a different finding, it does not address the pertinent question, of whether the district court’s factual findings are so implausible as to constitute an impermissible view of the same record. Majority Op. [15-18]. Consequently, even if this court could demonstrate that the stipulation was ambiguous on its face, va-catur would not be warranted, because the *83district court’s conclusion is a permissible view of the evidence.
The majority opinion notes that there was no “objective evidence that the parties intended the duration and scope recitations to have the effect of prohibiting the State from applying future legislation on these topics to the Plaintiffs.” Majority Op. [77 n. 9]. However, the district court relied on the very same type of evidence that the majority argues is dispositive: the context provided by the earlier litigation. See, e.g. Doe-SORA IV, 427 F.Supp.2d at 408-09. If this is not the sort of “objective evidence” that is adequate to support the district court’s holding, it is not apparent why it would provide a more sufficient basis for this court’s fact finding, particularly as “few persons are in a better position to understand the meaning of a [settlement] than the district judge who oversaw and approved it.” United States v. Local 359, United Seafood Workers, 55 F.3d 64, 68 (2d Cir.1995).
The majority opines further that “there is no indication that the District Judge participated in developing any of the terms of the Stipulation that resulted in the decree.” Majority Op. [76 n. 7]. That may be true, but it does not establish that the district court was not familiar with the specifics of the dispute between the parties and their attempts at resolution. The District Judge below supervised — over the span of ten years — various challenges plaintiffs-appellants made to SORA, litigation which the majority believes to provide it with the necessary context for its construction of the consent decree. See Doe-SORA III, 3 F.Supp.2d 456 (Chin, J.); Doe-SORA I, 940 F.Supp. 603 (Chin, J.); Doe v. Pataki, 919 F.Supp. 691 (S.D.N.Y. 1996) (Chin, J.). Accordingly, the principle of deference this court announced in Local 359 applies a fortoiri in the instant case; there is no compelling argument in the majority opinion why it should be set aside.
III. The Majority Relies on the Fact that One Party is a State, Without Specifying Precisely Why this Is Central to Construing the Contract: Winstar is Inapplicable
The majority believes that the most important features of the context of the agreement are that one of the parties is a state, and that the designated forum for its enforcement is a federal court. While this is not unreasonable in itself, it leads the majority to another departure from this circuit’s precedent by inspiring the creation of an unprecedented clear statement rule. While the majority opinion presents this new rule as a straightforward application of Supreme Court precedent, it misconstrues the leading case.
The majority notes that federal courts exhibit a “traditional coneern[ ] ... for restricting] a state’s inherent powers.” Majority Op. [77]. Accordingly, it reasons that a court construing a consent decree must have “a clear indication that a state has intended to surrender its normal authority to amend its statutes.” Id. at [78]. However, what is at issue here is not a wholesale surrender of the police power or an implicit disavowal of the right to amend the statutes that exercise it, but rather whether a state may make a binding promise to certain of its citizens not to subject them to those future amendments.
The majority relies on United States v. Winstar Corp. to support the novel clear statement rule, which it retroactively applies to plaintiffs-appellants, but the opinion does not engage in the analysis deemed critical by the Supreme Court in that case: whether the particular agreement at issue is of the type to which the unmistakability doctrine — i.e. the doctrine that requires that a contract purportedly binding a state *84to contractual obligations despite later legislative amendments to clearly indicate its intent to be so bound — is properly applicable:
Injecting the opportunity for unmistaka-bility litigation into every common contract action would, however, produce the untoward result of compromising the Government’s practical capacity to make contracts, which we have held to be of the essence of sovereignty itself. From a practical standpoint, it would make an inroad on this power, by expanding the Government’s opportunities for contractual abrogation, with the certain result of undermining the Government’s credibility at the bargaining table and increasing the cost of its engagements.
518 U.S. 839, 884, 116 S.Ct. 2432, 135 L.Ed.2d 964 (1996) (plurality opinion).4 Accordingly, the majority’s invocation of the unmistakability doctrine here in support of deference to state sovereignty is perplexing. The state also exercised a sovereign power when executing the stipulation; the majority opinion does not consider whether a federal court’s declaration that this sovereign act was implicitly limited is consistent with the principles of federalism.5
Winstar establishes that when no clear statement rule such as the unmistakability doctrine is properly applicable, a sovereign may be held liable for its abrogation of a contractual agreement, because in that event, a bedrock principle of the common law dictates this outcome. The applicable principle is that a party assumes the risk that it will not be able to perform its obligation under the contract. Day v. United States, 245 U.S. 159, 161, 38 S.Ct. 57, 62 L.Ed. 219 (1917) (Holmes, J.) (“One who makes a contract never can be absolutely certain that he will be able to perform it when the time comes, and the very essence of it is that he takes the risk within the limits of his undertaking”). Although the majority opinion states that “there is no indication, however, that the Plaintiffs sought to have the duration-of-registration or the scope-of-notification requirements remain unaltered by subsequent state legislation,” Majority Op. [77], this absence merely indicates that there was no provision in the contract stating that the government will be hable in the event that it cannot perform, because it would have been superfluous.
The Supreme Court reaffirmed in Wins-tar that the government cannot ignore the *85common-law principle that a party responsible for the changed circumstances that make performance of its contractual obligations impossible cannot evade responsibility by pointing to the changed circumstances, or by asserting that the contract did not specify what should occur in that event. 518 U.S. at 904-910, 116 S.Ct. 2432. Put simply, “[i]f the adverse event is due to the fault of the obligor, the offending party cannot be heard to cry that performance is impracticable.” 30 Williston on Contracts § 77:1 (4th ed.); see also Restatement (Second) of Contracts § 264 cmt. a (1981).
It follows directly from this principle that a state which amends its laws, such that it cannot perform its contractual obligation, is barred from asserting the defense of impracticability. See, e.g., 14 Cor-bin on Contracts, § 76.1 (2006) (“[t]he rule favoring discharge [for legal impossibility] following government prohibition does not apply, however, when the government action affects the contractual duties of the government itself.”) The majority opinion turns this principle on its head. More problematically, it manufactures a wide-ranging exception to the common law, that: “the normal rules of construction may vary depending on the nature of the parties and the effect of enforcement,” Majority Op. [76] and applies this exception for the first time here to create an additional hurdle for those attempting to vindicate their contractual rights against the government.
IV. If the Majority Believes the Contract is Ambiguous, The Proper Disposition is Remand
The majority opinion also fails to mention that the usual evidence that a court ordinarily uses to discern the parties’ intentions is absent from the record before us. If the majority believes that the intentions of the parties are central to the proper construction of the contract, surely these intentions should be probed under oath, rather than relying on the representations of their attorneys at oral arguments. The majority opinion does not advance any reasons why this court should not undertake the normal course of action when we hold that a contract is ambiguous, which is to remand to the district court for an evidentiary hearing on the issue of the parties’ intentions, where the parties might then submit parol evidence and evidence of custom and usage, among any other forms of evidence that might be probative of the parties’ intent. See Int’l Multifoods Corp. v. Commercial Union Ins. Co., 309 F.3d 76, 87 (2d Cir.2002). To make findings about the intentions of the parties according to the averments of counsel when we can easily remand for reopening of the record and further fact finding is contrary to our precedent that the “resolution of [factual] problems is not the province of an appellate tribunal, but of the trial court.” M.W. Zack Metal Co. v. S.S. Birmingham City, 291 F.2d 451, 454 (2d Cir.1961).
In addition to what might be discovered if we were to allow for the introduction of parol evidence on remand, a hearing might reveal that the parties’ intentions may even have been memorialized, and in any event, what is testified to under oath might be more illuminating than what was merely averred. At present, we also do not know if either party had any reason to be aware that the other was operating with a different understanding of the provisions now at issue, a point of critical importance. We also have no evidence as to which party chose the language at issue, so we cannot invoke the doctrine of contra prof-erentem.
Finally, since the government’s performance under the stipulation was frustrated by the legislation, any court that would decide the outcome of this dispute should *86have evidence of whether that legislation was foreseeable — a question this court probed at oral argument without success— since “[i]f the risk of impossibility of performance was foreseeable, that contingency should have been addressed in the contract. The absence of such a contractual provision gives rise to an inference that the risk was assumed [by the obligor,]” in this case, the government. 30 Williston on Contracts, § 77.95 (4th ed.).
Instead of allowing for the introduction of the evidence necessary for a well-reasoned resolution, the majority opinion reasons from the absence of evidence. The majority opinion mentions that there is “no indication” that certain events occurred— twice. If it not clear that opening the record would not reveal any such indications, there is no reason why this court should not allow the introduction of what might well prove to be crucial to our determination.
While the question of whether or not the contract is ambiguous is one of law, determining the intent of the parties is not only factual, but often requires a painstaking reconstruction of their motivations and strategies and how these changed over time. “Words and conduct used in the process of making a contract — offers, acceptances, modifications, preliminary communications not themselves operative in any way — all these need interpretation ... before we can determine the operative effect that should be given to the contract.” 5 Corbin on Contracts, § 24.1 (2006). The majority opinion states that “what is in dispute is the operative effect of these recitations,” Majority Op. [76], but resolves this question without even seeing any evidence of the words and conduct of the parties’ negotiations, never mind attempting to interpret this evidence. This is a difficult endeavor in the best of circumstances: to attempt it with nothing but a cold and bare record as a guide is obdurate.
CONCLUSION
I cannot join in the majority’s opinion. It dispenses with the plain meaning rule of contractual interpretation by proceeding directly to the question of the parties’ intent without identifying an ambiguity visible on the face of the stipulation; it effectively reforms the stipulation, by means of what is admittedly construction — which this court has no authority to do in the absence of such an ambiguity; it does not adhere to several foundational principles of the common law of contracts; and it resolves the ambiguities as to the parties’ intentions on the basis of an absence of evidence, where it would be a simple matter to reopen the record — which is entirely bare — in order to obtain vital information on several key issues. For these reasons, I respectfully dissent.
. I have adopted the numeration of the cases brought challenging SORA by the plaintiffs-appellants as set forth in the majority opinion.
. As will be discussed in Section III, infra, the question of whether a state can modify its contractual obligations through the amendment of legislation is a vexing one. While states possess a police power which the federal courts should avoid hindering, "[t]he United States Constitution prohibits the impairment of the obligation of contracts by the states.... [Mjost courts have held that statutes enacted subsequent to the making of a contract are not incorporated in the contract. ... [Accordingly,] [t]o hold that such statutes are incorporated would possibly be an unconstitutional impairment of contract rights, because the parties assented to be bound to a different set of rights and obligations.” 5 Corbin on Contracts, § 24.26 (2006).
. This court has implicitly held that the state law of the forum applies to stipulations settling federal claims. Torres v. Walker, 356 F.3d 238, 245-46 (2d Cir.2004). This court has also determined that the principle of whether the existence of ambiguity in a written agreement must be determined from the face of the agreement is embodied in New York law. Collins, 303 F.3d at 433 Accordingly, by looking instead to the attendant circumstances to reveal the ambiguity, the majority opinion not only disregards the circuit's precedent, but binding New York law, a violation of the very principles of federalism that the majority argues should inform our decision.
. While it is virtually certain that the stipulation at issue in the instant case would be governed by the unmistakability doctrine if the United States were one of the parties, it far from certain whether it applies here, since the Contract Clause applies directly to the States, whereas Congress is merely the subject of analogous but less exacting limitations on its ability to abrogate its contractual obligations derived from the Due Process Clause. U.S. Const, art. I, § 10, cl. 1; U.S. Const, amend. V; see also United States Trust Co. of N.Y. v. New Jersey, 431 U.S. 1, 17 n. 13, 97 S.Ct. 1505, 52 L.Ed.2d 92 (1977) (contrasting Contract Clause with Due Process Clause); Pension Benefit Guar. Corp. v. R.A. Gray & Co., 467 U.S. 717, 732-33, 104 S.Ct. 2709, 81 L.Ed.2d 601 (1984) (similar).
. Since the majority opines that the unmistak-ability doctrine is “more pertinent to the issue” here than the clear statement rules for the surrender of sovereign immunity, Majority Op. [79], the defects in the analogy drawn by the majority between this case and our Eleventh Amendment jurisprudence need not be explored, except to note that the surrender of immunity from suit in federal court vests jurisdiction where a federal court would otherwise have none, see Seminole Tribe v. Florida, 517 U.S. 44, 73, 116 S.Ct. 1114, 134 L.Ed.2d 252 (1996), but there is no such "jurisdictional bar” to a district court’s construction of a stipulation executed under that court’s supervision, especially where the parties explicitly bargained for that court’s supervision of the resulting consent decree. Cf. Geller v. Branic Int’ l Realty Corp., 212 F.3d 734, 737 (2d Cir.2000).