with whom AGEE, J., joins, dissenting.
In my opinion, the majority correctly reads our prior panel decisions, including MacNelly v. MacNelly, 17 Va.App. 427, 430, 437 S.E.2d 582, 584 (1993), and Gayler v. Gayler, 20 Va.App. 83, 86 n. 2, 455 S.E.2d 278, 280 n. 2 (1995), to have
extrapolated the phrase “unless otherwise provided” in Code § 20-109(D) to apparently require explicitly stating the words “remarriage” or “death” in a property settlement agreement in order to overcome the statutory presumption that spousal support payments terminate upon the death of a spouse or the payee spouse’s remarriage.
Hardesty v. Hardesty, 39 Va.App. 102, 107, 570 S.E.2d 878, 880 (Agee, J., concurring), en banc granted, 39 Va.App. 253, 572 S.E.2d 493 (2002).
Under a fair reading of the MacNelly-Gayler extrapolation, it can be said that “[b]y inference, our decisions forbid any other language, no matter how clear or express, from rebutting the statutory presumption unless the specific words ‘remarriage’ or ‘death’ appear in the operative contract provision.” Id. And that is true “regardless of how clear and unmistakable” the meaning of a contractual provision “may appear to a reader of the English language.” Id.
The extrapolation of MacNelly and Gayler evolved into the ratio decidendi of Langley v. Johnson, 27 Va.App. 365, 376, 499 S.E.2d 15, 20 (1998). Relying on Gayler footnote 2, *680Langley held that the separation agreement “failed expressly to state that the husband’s support obligation would not terminate upon the wife’s remarriage,” id. at 376, 499 S.E.2d at 20, and any contractual language less than that would not suffice.9 In doing so, Langley successfully refuted the maxim, “Breath spent repeating dicta does not infuse it with life.” Metro. Stevedore Co. v. Rambo, 515 U.S. 291, 300, 115 S.Ct. 2144, 2149, 132 L.Ed.2d 226 (1995).
The dissent’s effort to distinguish these cases away does not persuade me. For that reason, if I were sitting on a three-judge panel deciding this case, I too would take the majority’s view — being bound by prior panel precedent under the inter-panel accord component of the stare decisis doctrine. See Hardesty, 39 Va.App. at 108, 570 S.E.2d at 881 (Agee, J., concurring) (“Accordingly, bound by the doctrine of stare decisis, I concur in the result.”). “This principle applies not merely to the literal holding of the case, but also to its ratio decidendi — the essential rationale in the case that determines the judgment.” Clinchfield Coal Co. v. Reed, 40 Va.App. 69, 73-74, 577 S.E.2d 538, 540 (2003).
In view of this principle, the issue now is whether the en banc court should overrule the intended (per the majority) or unintended (per the dissent) consequences of these prior panel decisions. We certainly have the authority to do so. See Armstrong v. Commonwealth, 263 Va. 573, 581, 562 S.E.2d 139, 143 (2002) (recognizing that prior panel decisions remain “subject to review by the Court of Appeals sitting en banc”); Code § 17.1-402(D) (“The court sitting en banc shall consider and decide the case and may overrule any previous decision by any panel or of the full court.”). We should exercise that authority if a “detailed inquiry” demonstrates that “a mistake exists in our prior decisions.” Armstrong v. Commonwealth, 36 Va.App. 312, 321, 549 S.E.2d 641, 645 (2001) (citations and *681internal quotation marks omitted). For two reasons, I believe a mistake of that magnitude exists here.10
A.
First, the extrapolation introduced into the law by MacNelly and Gayler, when carried to its logical extreme, as this case illustrates as well as any, produces a truly anomalous result. The majority holding boils down to a simple proposition: An agreement stating that spousal support “cannot be terminated for any reason” does not include termination for the reason of remarriage. Under the majority’s reasoning, the agreement should have said something along these lines: “Support cannot be terminated for any reason, and, by saying ‘for any reason,’ we mean to include for the reason of remarriage.”
The majority argues that even though “on its face the language of the PSA seems to clearly reflect the intent of the parties,” the agreement’s failure to use talismanic words creates ambiguity — the kind that could lead to further litigation and thereby “undermine the statute’s purpose.” Ante, at 670, 581 S.E.2d at 217; see also Hardesty, 39 Va.App. at 107, 570 S.E.2d at 880 (also noting that “on its face the language of the *682PSA seems to clearly reflect the intent of the parties”). The hunt for the statutory purpose of Code § 20-109(D), however, has led to a self-defeating irony: But for the magic words requirement, the contractual provision “on its face,” ante, at 669, 581 S.E.2d at 217; see also Hardesty, 39 Va.App. at 107, 570 S.E.2d at 880, provides sufficient clarity to discern the intent of the parties. The alleged ambiguity exists only because the majority discards the obvious and unqualified meaning of the word “any” in the property settlement agreement.
Had this case been decided by the application of the plain-meaning rule (which governs most other contracts in Virginia and, except for this one topic, still governs separation agreements), the trial court, the majority, and the dissent would be in full accord. Each would agree that the phrase “for any reason” means what it says and thus includes remarriage. No inductive inference is required here. The meaning comes directly from the words themselves.
In no other context have Virginia courts struggled over whether the term “any” means any. See, e.g., Sussex Cmty. Servs. Ass’n v. Va. Soc’y for Mentally Retarded Children, 251 Va. 240, 243, 467 S.E.2d 468, 469 (1996) (“The word ‘any,’ like other unrestrictive modifiers such as ‘an’ and ‘all,’ is generally considered to apply without limitation.”); Rubin v. Gochrach, 186 Va. 786, 797, 44 S.E.2d 1, 5-6 (1947) (“The words ‘any renewal’ are comprehensive, and logically include more than one renewal. The meaning of the word ‘any’ in the connection in which it was used seems to be clear and deliberate.”); Cox v. Cox, 16 Va.App. 146, 148, 428 S.E.2d 515, 516 (1993) (“The plain and unambiguous meaning of the word ‘any’ is ‘one or more indiscriminately from all those of a kind.’ ” (quoting Webster’s Third New International Dictionary 97 (1981))). In this case, therefore, the quest for greater clarity and less litigation has produced neither.
B.
Second, when the legislature requires that a verbatim contractual term be used, its enactments usually say exactly that. *683See, e.g., Code § 8.2-316(2) (requiring disclaimers of an implied warranty of merchantability to “mention” the word “merchantability”); Code § 8.01-433.1 (requiring contracts authorizing confession of judgment to include specific statutory verbiage). Because Code § 20-109(D) includes no such requirement, it should not be added by interpretative accretion even if we could be persuaded that doing so would improve upon the basic point of the statute. “Courts are not allowed to write new words into a statute plain on its face.” Shenk v. Shenk, 39 Va.App. 161, 171, 571 S.E.2d 896, 901 (2002) (quoting Flanary v. Milton, 263 Va. 20, 23, 556 S.E.2d 767, 769 (2002)); see also SIGNAL Corp. v. Keane Fed. Sys., Inc., 265 Va. 38, 46, 574 S.E.2d 253, 257 (2003) (courts are “not free to add language” to statutes under the guise of interpretation); Woods v. Mendez, 265 Va. 68, 75, 574 S.E.2d 263, 267 (2003) (courts cannot “add words to the statute”).
The majority’s suggestion that today’s decision represents a cautious, and thus commendable, deference to the legislature is unconvincing. See Ante, at 669, 581 S.E.2d at 216-17. We are asked to believe that the General Assembly has implicitly endorsed the majority’s holding because the legislature has not amended the statute to repeal our prior panel decisions. From that inaction, the majority reasons, we should infer the General Assembly has by silence put its legislative imprimatur on the majority’s interpretation of Code § 20-109(D) because the legislature is “presumed to know the law.” Ante, at 669, 581 S.E.2d at 217.
I concede the value of the general principle underlying this view. See, e.g., Burns v. Bd. of Supervisors, 227 Va. 354, 360, 315 S.E.2d 856, 860 (1984). But I would not apply it to cases where, as here, the interpretation that the legislature has allegedly endorsed by inaction is both unsettled and a product of piling dicta upon dicta. See Metro. Stevedore Co., 515 U.S. at 299-300, 115 S.Ct. at 2149-50. Legislative inaction “lacks persuasive significance,” Brown v. Gardner, 513 U.S. 115, 121, 115 S.Ct. 552, 557, 130 L.Ed.2d 462 (1994) (citation omitted), *684when it can be shown that “several equally tenable inferences may be drawn from such inaction,” Cent. Bank of Denver, N.A. v. First Interstate Bank of Denver, N.A., 511 U.S. 164, 187, 114 S.Ct. 1439, 1453, 128 L.Ed.2d 119 (1994) (quoting Pension Benefit Guar. Corp. v. LTV Corp., 496 U.S. 633, 650, 110 S.Ct. 2668, 2678, 110 L.Ed.2d 579 (1990)). See also Zuber v. Allen, 396 U.S. 168, 185-86 n. 21, 90 S.Ct. 314, 324 n. 21, 24 L.Ed.2d 345 (1969) (“The verdict of quiescent years cannot be invoked to baptize a statutory gloss that is otherwise impermissible.”).
We need go no further than the differing views of the majority and the principal dissent in this case to see that neither interpretation of our prior precedents has claimed a strong consensus. That being true, how confident can we be that the General Assembly previously guessed the interpretation the majority adopts today and then silently decided to endorse that particular interpretation by not rewriting the statute? Under these circumstances, legislative silence is a “poor beacon to follow in discerning the proper statutory route.” Zuber, 396 U.S. at 185, 90 S.Ct. at 323. The only sure guide is the altogether ordinary task of reading the statutory text and applying its plain meaning. See Patterson v. Commonwealth, 39 Va.App. 610, 617, 575 S.E.2d 583, 586-87 (2003). Because the magic-words extrapolation of our prior panel opinions cannot be grounded in the plain meaning of Code § 20-109(D), and therefore reflects that a mistake exists in our prior decisions, I respectfully dissent.
. "Indeed, the language of the parties’ agreement is virtually identical to 'the language of the [Gayler] addendum standing alone [which] would not be sufficient to evince an intent of the parties to avoid the operation of Code §§ 20-109 and 20-109.1.’ 20 Va.App. at 86 n. 2, 455 S.E.2d at 280 n. 2.” Langley, 27 Va.App. at 374, 499 S.E.2d at 20.
. We sit today as an en banc court reviewing prior three-judge panel decisions. We are not reconsidering a prior en banc decision, which would be analogous to the Virginia Supreme Court reconsidering its own precedent (the situation addressed in the stare decisis citations listed by the majority). In this respect, multi-panel appellate courts are structurally different from unitary appellate courts. While the determinacy concerns underlying stare decisis still play an important role when an en banc court reviews a panel decision, the doctrine cannot be of such force that it binds the en banc court or in any way undermines the en banc court’s duty under Code § 17.1-402(D) to provide full-court review of prior three-judge panel decisions. See generally United States v. American-Foreign S.S. Corp., 363 U.S. 685, 689-90, 80 S.Ct. 1336, 1339, 4 L.Ed.2d 1491 (1960) ("The principal utility of determinations by the courts of appeals in banc is to enable the court to maintain its integrity as an institution by making it possible for a majority of its judges always to control and thereby to secure uniformity and continuity in its decisions, while enabling the court at the same time to follow the efficient and time-saving procedure of having panels of three judges hear and decide the vast majority of cases as to which no division exists within the court.”).