filed a concurring opinion
in which Justice ENOCH joined, except for Part IV and the discussion of Inglish and Bandera, and in which Justice HANKINSON joined, except Part IV.The Court granted these petitions in Lehmann and Harris to solve the Mafrige problems. The Court fails to do so. Thus, while I concur in the result the Court reaches, I cannot agree with the reasoning it uses to reach that result.
In March 1993, we granted writ in Mafrige v. Ross to resolve the inherent problems in determining finality of summary judgments for purposes of appeal. 866 S.W.2d 590 (Tex.1993). There we recognized that determining finality had “been a recurring and nagging problem throughout the judicial history of this state.” Mafrige, 866 S.W.2d at 590. Thus, in a major departure from our prior jurisprudence, we created a new rule providing: “If a summary judgment order appears to be final, as evidenced by the inclusion of language purporting to dispose of all claims or parties, the judgment should be treated as final for purposes of appeal.” Mafrige, 866 S.W.2d at 592.
Despite the certainty we intended this bright-line rule to provide, the last seven years have proved that the Mafrige rule has created more problems than it solved — confusing the lower courts, operating as a trap for unwary litigants, and consistently bringing about arguably unjust and oftentimes absurd results. So, in November 1999, we granted the petitions in these cases to resolve the Mafrige problems. Inexplicably, the Court begins its opinion by chronicling the evolution of the rules and presumptions governing finality of orders following a conventional trial on the merits from the middle of the last century to the present.1 Then, with very little discussion of the problems Mafrige and its progeny created in determining summary judgment finality, the Court concludes that the solution is to maintain the principle of the Mafrige legal fiction — with only slight modification.
However, rather than solve, the Court merely perpetuates the problems Mafrige created. The cases grappling to apply Mafrige illustrate that there is but one real solution. We should return to the principle we announced in Teer v. Duddles-ten — that a Mother Hubbard clause simply “has no place in a partial summary judgment,” and that a summary judgment order is not an appealable, final judgment unless it actually disposes of all parties and issues. 664 S.W.2d 702, 703-04 (Tex.1984).
The Court states: “[W]e do not write rules by opinion.” 39 S.W.3d at 205. The Court is right; we should not establish rales by judicial fiat. We should not have done so in Mafrige and we should not have perpetuated the Mafrige problems with Inglish and Bandera. Any new summary judgment finality rale should be achieved by this Court’s formally promulgating a new procedure rule. The Court should recognize this, overrule Mafrige and its progeny, and await a recommendation by *209our rules advisory committee. Because the Court refuses to take this path, I concur in the judgment only.
I. MAFRIGE AND ITS PROGENY
Before Mafrige, courts determined summary judgment finality by reviewing the five pleadings, the summary judgment motion, and the summary judgment order. Harris County v. Nash, 22 S.W.3d 46, 49-50 (Tex.App.—Houston [14th Dist.] 2000, pet. filed); Kaigler v. General Elec. Ins. Mortgage Corp., 961 S.W.2d 273, 275 (Tex.App.—Houston [1st Dist.] 1997, no pet.). A summary judgment was deemed final and appealable only if it expressly disposed of all parties and issues or if it was severed from the remainder of the suit. Pan Am. Petroleum Corp. v. Texas Pac. Coal & Oil Co., 159 Tex. 550, 324 S.W.2d 200, 200 (1959) (“[A] summary judgment which does not dispose of all parties and issues in the pending suit is interlocutory and not appealable unless a severance of that phase of the case is ordered by the trial court.”).
With Mafrige, this Court attempted to simplify this process by holding that the “magic language” of a Mother Hubbard or similar finality clause conclusively transforms an interlocutory summary judgment into a final, appealable order. Mafrige, 866 S.W.2d at 592. We have twice revisited Mafiige to clarify its scope. See Inglish v. Union State Bank, 945 S.W.2d 810, 811 (Tex.1997) (holding that the Mafrige rule applies even when neither party appeals the erroneous summary judgment); Bandera Elec. Coop., Inc. v. Gilchrist, 946 S.W.2d 336, 337 (Tex.1997) (explaining that when the Mafrige rule renders a partial summary judgment final for purposes of appeal, the appellate court should reverse and remand only the erroneously disposed claims). Unfortunately, Mafrige did little towards alleviating the lower courts’ confusion — and Inglish and Bandera only compounded it. The Court’s opinion suffers the same problem. Namely, its slightly-modified Mafrige rule falls far short of remedying the myriad of problems the Mafrige fiction and its progeny created,
A. Finality Language
One source of confusion under Mafrige has been uncertainty about what language triggers its finality rule. In Mafrige, we held that a partial summary judgment is treated as final for appeal purposes when the order contains a Mother Hubbard clause stating that “all relief not expressly granted is denied” or other language “purporting to dispose of all claims or parties.” 866 S.W.2d at 590 & n. 1, 592. We further clarified that “other” finality language includes “a statement that the summary judgment is granted as to all claims asserted by the plaintiff, or a statement that the plaintiff takes nothing against defendant.” Mafrige, 866 S.W.2d at 590 n. 1.; see also Inglish, 945 S.W.2d at 811 (holding statement that “[defendant is entitled to summary judgment in this case,” and that plaintiff should “take nothing on account of his lawsuit” rendered partial summary judgment final for purposes of appeal); Springer v. Spruiell, 866 S.W.2d 592, 593 (Tex.1993) (holding that summary judgment order reciting plaintiffs “have and recover nothing” purported to dispose of all parties and issues).
Despite these examples, some lower courts have refused to hold orders containing this exact language final for purposes of appeal. E.g., Carey v. Dimidjian, 982 S.W.2d 556, 558 (Tex.App.—Eastland 1998, no pet.) (holding that order containing Mother Hubbard clause was not final and appealable where the motion was labeled “Partial Summary Judgment” and the parties treated the order as interlocutory); Hinojosa v. Hinojosa, 866 S.W.2d 67, 69-70 (Tex.App.—El Paso 1993, no writ) (holding that order containing Mother Hubbard clause did not render judgment final because it did not dispose of counterclaim). Other courts have struggled with what “other” language purports to render a judgment final — often reaching opposite conclusions about identical clauses. Com*210pare Postive Feed, Inc. v. Guthmann, 4 S.W.3d 879, 881 (Tex.App.—Houston [1st Dist.] 1999, no pet.) (holding that order granting defendant’s summary judgment “in all things” purported to be final), with St. Paul Ins. Co. v. Mefford, No. 05-96-01581-CV, 1998 WL 821537 (Tex.App.—Dallas Nov. 30, 1998, no pet.) (not designated for publication), 1998 WL 821537, at *22 (holding that order granting defendant’s summary judgment “in all things” did not purport to be final).
While the Court recognizes that the “routine inclusion of [a Mother Hubbard clause] in otherwise plainly interlocutory orders and its ambiguity in many contexts have rendered it inapt for determining finality,” 39 S.W.3d at 192, it ignores the obvious problems courts have faced interpreting other language “purporting to dispose of all claims or parties.” Mafrige, 866 S.W.2d at 592. In fact, despite the Court’s extensive analysis and discussion, its holding represents but a minor departure from Mafrige.
Its modified rule has two parts. The first represents no change in Texas law. It simply reiterates that a summary judgment order that actually disposes of all parties and issues is final for purposes of appeal. 39 S.W.3d at 192. The second part provides that a Mother Hubbard clause is no longer enough to invoke the fiction that an otherwise interlocutory order is treated as final for purposes of appeal. Instead, to invoke the Mafrige fiction, an interlocutory order must now “clearly and unequivocally state[ ] that it finally disposes of all claims and all parties.” 39 S.W.3d at 205. The Court further explains that the statements “plaintiff take nothing by his claims in the case” and “[t]his judgment finally disposes of all parties and all claims and is appeal-able” clearly and unequivocally state that an order is final. 39 S.W.3d at 205. In essence, the Court’s rule does no more than replace one set of magic language with another — while ignoring the reality that courts will likely face the same challenges deciding what language “clearly and unequivocally states” that an order is final, 39 S.W.3d at 205, as they did deciding what other language clearly “purport[s] to dispose of all claims or parties” under Mafrige. 866 S.W.2d at 592.
B. Omitted Parties
Applying Mafrige to omitted parties, like those in both Lehmann and Harris, has also troubled the lower courts. Specifically, they have struggled with deciding when finality language operates to render a summary judgment final against omitted parties. This issue often surfaces when both the summary judgment motion and the resulting order omit any specific reference to one or more parties.3 In this situation, several courts have held that Mafrige applies, reasoning that issues and parties are co-extensive and thus if “an order disposes of all issues in a case, then it necessarily disposes of all parties to a case, and vice versa.” Kaigler, 961 S.W.2d at 276; see also Lehmann v. Har-Con Corp., 988 S.W.2d 415, 416-17 (Tex.App.—Houston [14th Dist.] 1999, pet. granted); Harper v. Newton, 910 S.W.2d 9, 12 n. 1 (Tex.App.—Waco), rev’d sub nom. on other grounds, Dallas County v. Harper, 913 S.W.2d 207 (Tex.1995).
In contrast, other courts have interpreted Mafrige more narrowly, reasoning that an “order that explicitly grants a summary judgment in favor of less than all the defendants does not clearly evidence an intent to dispose of all claims against all defendants, especially those against whom *211summary judgment was not sought, regardless of the inclusion of a Mother Hubbard clause.” Lowe v. Teator, 1 S.W.3d 819, 823-24 (Tex.App.—Dallas 1999, pet. filed); see also Midkiff v. Hancock E. Tex. Sanitation, Inc., 996 S.W.2d 414, 416 (Tex.App.—Beaumont 1999, no pet.); Vanderwiele v. Llano Trucks, Inc., 885 S.W.2d 843, 845 (Tex.App.—Austin 1994, no writ).
Here the Court summarily dismisses this omitted parties problem:
Nothing in the order in Lehmann indicates that it is a final judgment, and it did not dispose of all pending claims and parties. The order in Hams states that plaintiff take nothing as to “one of the defendants”, but that language does not suggest that all of the plaintiffs’ claims were denied. As the order recites and as the record demonstrates, the defendant named in the order was not the only defendant remaining in the case. Thus, we conclude that a final appeal-able judgment was not rendered in either case.
39 S.W.3d at 206. Despite the presence of a Mother Hubbard clause, the trial court and parties in Lehmann continued treating the order as interlocutory-even in the face of this Court’s admonishment that a Mother Hubbard clause indicates finality.4 988 S.W.2d at 416. The Court now holds that the order did not purport to be final based solely on its new rule discounting the dis-positive effect of Mother Hubbard clauses.
However, the Court’s resolution merely sidesteps the real problem. What happens in the next case when, on facts identical to Lehmann, a trial court signs an interlocutory summary judgment -with the Court’s new magic language rather than a Mother Hubbard clause? We are right back where we started. Substituting one magic phrase for another leads nowhere.
The reality is simply that omitted parties oftentimes do not believe that a summary judgment order that they have not seen, that does not mention them, and that results from a hearing in which they did not participate will operate to dispose of them or their claims. But, under the Court’s standard, if these parties do not perfect a timely appeal from the erroneous judgment, their right to appeal is forever lost. This result elevates form over substance and hinders parties’ rights to have the merits of their claims considered. See, e.g., Rodriguez v. NBC Bank, 5 S.W.3d 756, 763 n. 4 (Tex.App.—San Antonio 1999, no pet.) (recognizing this Court’s “express goal of reaching the merits of a cause of action, instead of dismissing actions on procedural technicalities”).
C. Omitted Cross-Claims and Counterclaims
The courts of appeals have also treated omitted cross-claims and counterclaims inconsistently — despite our holding in Band-era. In Bandera, the trial court signed an order with a Mother Hubbard clause that did not mention the defendant’s counterclaims. 946 S.W.2d at 337. This Court explained that “[bjecause the order contained a Mother Hubbard clause denying all other relief, it also purported to dispose of [the defendant’s] counterclaims.” Bandera, 946 S.W.2d at 337. But several courts have refused to apply Mafrige in this situation, maintaining that a summary judgment that does not mention counterclaims or cross-claims cannot purport to be final-regardless of whether it contains finality language. E.g., Sommers v. Concepcion, 20 S.W.3d 27, 33 (Tex.App.— Houston [14th Dist.] 2000, pet. denied); Hervey v. Flores, 975 S.W.2d 21, 25 (Tex.App.—El Paso 1998, pet. denied); cf. Coleman Cattle Co., Inc. v. Carpentier, 10 S.W.3d 430, 433 n. 2 (Tex.App.—Beaumont 2000, no pet.). Other courts have followed Bandera’s mandate, holding that finality language — such as “plaintiff takes noth-*212mg” — renders a judgment final for appeal purposes, despite omission of any reference to defendant’s counterclaims. In re Monroe, No. 05-99-01758-CV, 2000 WL 378519 (Tex.App.—Dallas Mar.31, 2000, orig. proceeding) (not designated for publication), 2000 WL 378519, at *1-2; see also Kaigler, 961 S.W.2d at 275-76.
The Court’s rule does not provide a satisfactory remedy for this situation either. The Court states:
An order that adjudicates only the plaintiff’s claims against the defendant does not adjudicate a counterclaim, cross-claim, or third party claim, nor does an order adjudicating claims like the latter dispose of the plaintiffs claims. An order that disposes of claims by only one of multiple plaintiffs or against one of multiple defendants does not adjudicate claims by or against other parties. An order does not dispose of all claims and all parties merely because it is entitled “final,” or because the word “final” appears elsewhere in the order, or even because it awards costs. Nor does an order completely dispose of a case merely because it states that it is appealable, since even interlocutory orders may sometimes be appealable. Rather, there must be some other clear indication that the trial court intended the order to completely dispose of the entire case.
39 S.W.3d at 205.
Under its modified finality rule, the lower courts’ disagreement in this area will continue because too many questions are left unanswered. For example, should a “final” summary judgment order stating that defendant is granted summary judgment “in all things” dispose of a cross-claim by another defendant as well as the claim by the plaintiff that brought the original claim? In this situation, there is no doubt that the order is unambiguous. However, it is likewise clear, but not from the order, that the third party’s claim against the defendant was never considered. Should an order granting summary judgment for a plaintiff that recites it is a final and appealable order be final for counterclaims not mentioned in the motion or order? The order unequivocally states that it is a final, appealable order. Nonetheless there is a counterclaim that has not been considered. The Court states that a summary judgment granted for a plaintiff “does not adjudicate a counterclaim” and then goes on to say that to make the order final there must be “some other clear indication that the trial court intended the order to completely dispose of the entire case.” 39 S.W.3d at 205. In the example above, does the additional statement that “this is a final, appealable order” provide this “other clear indication”? These very issues are repeatedly raised in the courts of appeals, and the Court’s modified rule simply does not resolve them.
D. Trial Courts’ and Parties’ Intent
Differing philosophies about the effect the trial courts’ and parties’ intent should have on how Mafrige applies has created the most confusion and inconsistency. The courts of appeals have taken three approaches. Some courts apply a bright-line test, holding that a Mother Hubbard clause or other finality language always renders an order final for appeal purposes, regardless of any evidence of contrary intent. E.g., Preston v. American Eagle Ins. Co., 948 S.W.2d 18, 20-21 & n. 1 (Tex.App.—Dallas 1997, no writ) (holding that summary judgment purported to be final despite fact it was entitled “partial summary judgment”); cf. In re Cobos, 994 S.W.2d 313, 315 (Tex.App.—Corpus Christi 1999, orig. proceeding) (“As Maf-rige and Inglish make clear, the intent of the trial court is not the controlling consideration in determining whether a judgment is final.”). Other courts modify this approach, looking only within the four corners of the order and giving effect to any evidence of contrary intent found there. E.g., Rodriguez, 5 S.W.3d at 763-64 (Tex.App.—San Antonio 1999, no pet.) (“Looking within the four corners of the summary judgment order, the plain language of the *213Mother Hubbard clause did not, and could not, purport to grant or deny any more relief than the relief which [the defendant] sought.”); Midkiff, 996 S.W.2d at 416 (looking to order “as a whole” to conclude that summary judgment order containing Mother Hubbard clause did not purport to be final).
Finally, despite our holding in Inglish that the trial court’s intent is irrelevant in this context, other courts still refuse to apply Mafrige if there is evidence of contrary intent anywhere in the record. This usually occurs when the parties and court treat an order as interlocutory by continuing with the litigation rather than appealing the erroneous order. E.g., Lowe, 1 S.W.3d at 823-24 (holding that summary judgment could not be final where the record reflected that there were parties who did not participate in the summary judgment proceeding); Carey, 982 S.W.2d at 558 (relying, in part, on court’s and parties’ treatment of order containing Mother Hubbard clause as interlocutory to conclude judgment was not final).
The Court’s solution to this problem is as confusing as the rule it seeks to supplant. It appears to reject the bright-line approach Mafrige espouses and instead adopt a rule combining the second and third approaches. First, the Court notes that an order is final for appeal purposes if it “unequivocally states that it finally disposes of all parties and all claims and is appealable.” 39 S.W.3d at 205. It also explains that “[i]f the language of the order is clear and unequivocal, it must be given effect despite any other indications that one or more parties did not intend for the judgment to be final.” 39 S.W.3d at 206. From these statements, the Court’s new rule walks and talks a lot like a bright-line Mafrige rule, with magic language establishing finality.
However, the Court also states that “[t]o determine whether an order disposes of all pending claims and parties, it may of course be necessary for the appellate court to look to the record in the case.” 39 S.W.3d at 205. This sounds more like a pre-Mafrige rule, where a court must look to the record and the order to determine if an order actually disposes of all pending parties and issues.
Because of the lower courts’ confusion and disagreement about the role of intent in determining finality, I am convinced that the Court has not provided a workable rule that clearly defines that role as it applies to determining summary judgment finality.
E. Applying Mafrige To Non-SummaRY Judgment ORDERS
Finally, the question of whether Mafrige applies outside the summary judgment context has confused the lower courts. Courts of appeals have applied Mafrige to a plea to the jurisdiction, Webb v. HCM Mgmt. Corp., No. 07-96-0369-CV, 1998 WL 16033 (Tex.App.—Amarillo Jan. 12, 1998, pet. denied) (not designated for publication) 1998 WL 16033, at *1; an agreed judgment, In re 'Cobos, 994 S.W.2d at 315-16; a directed verdict, e.g., Polley v. Odom, 957 S.W.2d 932, 943 (Tex.App.—Waco 1997, judgm’t vacated); and a severance order, Harris County Flood Control Dist. v. Adam, 988 S.W.2d 423, 427 (Tex.App.—Houston [1st Dist.] 1999, pet. filed). In contrast, at least one court has declined to apply Mafrige to a dismissal for want of jurisdiction. In re Tejas, Nos. 01-98-00688-CV, 01-98-00689-CV, 01-98-00690-CV, 1998 WL 394562 (Tex.App.—Houston [1st Dist.] July 13, 1998, orig. proceeding) (not designated for publication), 1998 WL 394562, at *1 n. 1. And another has expressly refused to extend Mafrige to any order that is not a summary judgment. Biltmore Swim & Racquet Club Recreational Ass’n v. McAbee, No. 05-98-00252-CV, 1998 WL 459819 (Tex.App.—Dallas Aug.10, 1998, no pet.) (not designated for publication), 1998 WL 459819, at *1.
In Aldridge, this Court held that a presumption of finality exists when an order is signed following a traditional trial on the *214merits. Aldridge, 400 S.W.2d at 897-98. But we specifically noted that such a finality presumption would not be appropriate in other contexts. Aldridge, 400 S.W.2d at 897. Then in Mafrige we carved out an exception to what we had said in Aldridge by holding that an irrebuttable finality presumption applies to summary judgments containing a Mother Hubbard or similar finality clause. Mafrige, 866 S.W.2d at 592. Here again, just as we had limited Aldridge to conventional trials on the merits, we expressly limited Mafrige to summary judgments. Mafrige, 866 S.W.2d at 591 (“[T]he issue is whether ... a summary judgment, which purports to be final by the inclusion of Mother Hubbard language or its equivalent, should be treated as final for purposes of appeal”). Unfortunately, several courts of appeals have erroneously applied Mafrige in other contexts, causing confusion over how to determine finality of various other types of orders.
Mafrige and its progeny are limited to summary judgments — with good reason. No good can come of interjecting additional uncertainty into (1) conventional trials on the merits, to which the majority acknowledges the Aldridge presumption has “proved a fairly workable” rule, 39 S.W.3d at 200, or (2) numerous other types of orders, when even the majority acknowledges that “the ordinary expectation” supporting a finality presumption “simply does not exist when some form of judgment is rendered without such a trial” because “it is quite possible, perhaps even probable these days ... that any judgment rendered prior to a full-blown trial is intended to dispose of only part of the case.” 39 S.W.3d at 200.
However, the Court’s opinion here implicates finality of all judgments. This expansion into issues not before the Court today can only cause mischief in areas already plagued by confusion. If the Court persists in adhering to Mafrige’s principles, it should at least limit its holding, as we did in Mafrige, to summary judgments.
II. POLICY CONSIDERATIONS
Not surprisingly, the post-Mafrige era has given rise to considerable analysis by courts and commentators of both the competing policies Mafrige implicates and suggestions for reform. A few have applauded the bright-line rule. See Kaigler, 961 S.W.2d at 275-76 (recognizing that the rule provides harsh results, but emphasizing that uniform enforcement “encourage[s] attentiveness to correct judgments”); Boyce, Mafrige v. Ross and the Pitfalls of Presumptions, Appellate Advooate, Nov. 1997, at 7 (opining that Mafrige “resolved the confusion created by prior contradictory language and flatly inconsistent holdings”).
However, praises have been few and far between. Criticism has been the rule and the comments call for this Court to reconsider our decision:
What began as a benign growth allowing review of unripe claims on appeal, in Mafrige, became a malignant cancer cutting off causes of action before trial, in Inglish. If it were up to me, I would lock Mother Hubbard in the cupboard and return to the rule before Aldridge that a judgment is final and appealable only if it expressly disposes of all parties and all claims in the case. That appellants can even cite authority for the absurd result they seek, illustrates how wrong a turn the law has taken in this area — and how strong the need to right it.
Harris County Flood Control Dist., 988 S.W.2d at 427-28 (Taft, J., concurring in denial of rehearing en banc); see also, e.g., Lehmann, 988 S.W.2d at 418 (“Mafrige is not as clear to litigants as the supreme court believes it is.... In short, Mafrige has created several problems: 1) it is catching the parties by surprise ...; 2) it exalts form over substance; and 3) in more than a few situations, it ignores common sense.”); Carlson & Dunn, Navigating *215Procedural Minefields: Nuances in Determining Finality of Judgments, Plenary Power, and Appealability, 41 S. Tex. L.Rev. 953, 971 (2000) (“[Djespite the appeal of the certainty provided by this bright-line rule, the reality is that still, after seven years, it continues to operate as a trap for unwary litigants, bringing about arguably unjust and oftentimes draconian results.”); Swanda, Summary Judgment, Mother Hubbard Clauses, and Mafrige v. Ross, Appellate Advocate, May 1997, at 3 (complaining that the questions Mafrige raises “are just as elusive” as the questions it sought to resolve).
Strong policies support our practice of adhering to settled rules of law “unless there exists the strongest reasons for chang[e].” Benavides v. Garcia, 290 S.W. 739, 740-41 (Tex. Comm’n App.1927, judgm’t adopted). But we have also recognized the “doctrine of stare decisis does not stand as an insurmountable bar to overruling precedent.” Gutierrez v. Collins, 583 S.W.2d 312, 317 (Tex.1979). “Generally, we adhere to our precedents for reasons of efficiency, fairness, and legitimacy.” Weiner v. Wasson, 900 S.W.2d 316, 320 (Tex.1995). However, when adherence to a judicially-created rule of law no longer furthers these interests, and “the general interest will suffer less by such departure, than from a strict adherence,” we should not hesitate to depart from a prior holding. Benavides, 290 S.W. at 740. The lower courts’ application of Mafrige over the last seven years illustrates undeniably that this is just such a case.
We intended Mafrige, Inglish, and Bandera to provide certainty to litigants. Instead, they have bred chaos. Most disturbing is that the casebooks are now replete with examples of dismissed cases where the parties and courts clearly intended an order containing finality language to be interlocutory.5 E.g., Inglish, 945 S.W.2d at 811; In re Cobos, 994 S.W.2d at 315-16; Pena v. Valley Sandia, Ltd., 964 S.W.2d 297, 298-99 (Tex.App.—Corpus Christi 1998, no pet.); Kaigler, 961 S.W.2d at 275-76. Even the Court acknowledges:
[T]he ordinary expectation that supports the presumption that a judgment rendered after a conventional trial on the merits will comprehend all claims simply does not exist when some form of judgment is rendered without such a trial. On the contrary, it is quite possible, perhaps even probable these days in cases involving multiple parties and claims, that any judgment rendered pri- or to a full-blown trial is intended to dispose of only part of the case. Accordingly, the finality of the judgment must be determined without the benefit of any presumption.
39 S.W.3d at 200. Because of this reality, it is difficult to understand why the Court persists in adhering to Mafrige’s principles.
The author of the Court’s opinion recently opined: “Appellate procedure should not be tricky. It should be simple, it should be certain, it should make sense, and it should facilitate consideration of the parties’ argument on the merits.... ” Lane Bank Equip. Co. v. Smith Southern Equip., Inc., 10 S.W.3d 308, 314 (Tex.2000) (Hecht, J., concurring). This Court has repeatedly refused to adopt positions which elevate form over substance. See, e.g., Phillips v. Beaber, 995 S.W.2d 655, 658 (Tex.1999); Nueces Canyon Consol. Indep. Sch. Dist. v. Central Educ. Agency, 917 S.W.2d 773, 775-76 (Tex.1996). The Court here even recognizes that “[simplicity and certainty in appellate procedure are nowhere more important than in determining the time for perfecting appeal.” 39 S.W.3d at 205. Unfortunately though, the Court declines to embrace this opportunity *216to effectuate meaningful change and provide certainty for courts and litigants. Instead the Court leaves them as it found them, grappling with determining whether summary judgment orders are fictitiously made final.
III. THE SOLUTION
The Court notes: “[W]e do not write rules by opinion. We must decide what Texas law requires for finality, given the present rules .” 39 S.W.3d at 205. Yet, the Mafrige finality rule this Court created represented such a major departure from prior Texas law. In fact, but for the judicially-created Mafrige rule, no one would dispute that “what Texas law requires for finality” of summary judgments is an order actually disposing of all parties and issues.
Rather than simply amend the Mafrige finality rule and perpetuate the problems the unworkable system Mafrige and its progeny created, the Court should focus on shaping a real solution — one providing the desired certainty and protecting parties’ right to appellate review. This requires wiping the slate clean. Mafrige created enough problems with its fictional finality and its holding that trial courts can use magic language to create final summary judgments by granting relief not requested. 866 S.W.2d at 591-92. In Inglish we compounded the problem by confirming that Mafrige applies even when the parties continue litigating rather than appealing a partial summary judgment made final under Mafrige. 945 S.W.2d at 811. We completed the trilogy in Bandera, holding that when a party appeals a summary judgment granting more relief than requested, the court of appeals should address the merits of the appeal, remanding only the part of the judgment that exceeds the relief requested in the summary judgment motion. 946 S.W.2d at 337. Undeniably, these rules were designed to simplify summary judgment finality. But, in application, these cases only demonstrate that we should have adhered to our own admonishments that this Court simply should not make rules by opinion. E.g., Alvarado v. Farah Mfg. Co., 830 S.W.2d 911, 915 (Tex.1992) (explaining that we should not revise rules by opinion); see also Verburgt v. Dorner, 959 S.W.2d 615, 619 (Tex.1997) (Baker, J., dissenting) (noting that this Court’s jurisprudence forbids rule amendments by judicial fiat).
Thus, we should overrule Mafrige, Inglish, and Bandera — to the extent they created new rules by judicial fiat — and instead tackle the problems of summary judgment finality through our rulemaking process. Accordingly, we should return to our prior position that a Mother Hubbard clause (or other magic language) has no place in any summary judgment order— final or partial — and that a trial court may not sua sponte grant more relief than the parties request simply by adding concluso-ry finality language to a summary judgment order. Further, a summary judgment should be entitled to no presumption at all about whether it is final.
Returning to the law as it was pre-Mafrige requires determining the state of the law before Mafrige. Mafrige actually held two things: (1) that “ ‘Mother Hubbard’ language or its equivalent in an order granting summary judgment makes an otherwise partial summary judgment final for appeal purposes;” and (2) that if a summary judgment “grants more relief than requested, it should be reversed and remanded, but not dismissed.” 866 S.W.2d at 590, 592.
Before Mafrige, this first holding was not the law. In Teer v. Duddlesten we held that:
There is no presumption in partial summary judgments that the judgment was intended to make an adjudication about all parties and issues. The Mother Hubbard clause that “all relief not expressly granted is denied” has no place in a partial summary judgment hearing. The concepts of a partial summary judgment on the one hand, and a judgment *217that is presumed to determine all issues and facts on the other, are inconsistent.
664 S.W.2d at 704. In Majrige we recognized this earlier statement in Teer, but rejected it and held that finality language could render a partial summary judgment final for purposes of appeal. 290 S.W.2d at 592.
Mafrige’s second holding — that a summary judgment granting more relief than requested should be reversed and remanded, but not dismissed — does not appear to be an entirely new rule. In both Teer and Chessher, another pre-Mafrige case, we reversed and remanded (rather than dismissed) summary judgment orders after determining that they were interlocutory because they granted more relief than requested. See Teer, 664 S.W.2d at 705; Chessher v. Southwestern Bell Tel. Co., 658 S.W.2d 563, 564 (Tex.1983). But see Ross v. Arkwright Mut. Ins. Co., 834 S.W.2d 385, 393 (Tex.App.—Houston [14th Dist.] 1992) (opining that these cases are “in direct contravention of Tex.R. Civ. P. 166a(e)” and discussing disagreement in the courts over whether summary judgment orders granting more relief than requested were interlocutory or appealable, but erroneous, judgments), rev’d sub. nom. Mafrige, 866 S.W.2d at 590. Thus, while the courts were not entirely in agreement, it appears we had already established the rule that a summary judgment order granting more relief than requested is not interlocutory — it is simply erroneous. For this reason, I agree with the Court that if an order actually does dispose of each claim and every party, it is an appealable judgment, even if it grants more relief than requested. This is consistent -with the long-standing rule that if an order actually disposes of all parties and issues, it is final for appeal purposes. E.g., Houston Health Clubs, Inc. v. First Court of Appeals, 722 S.W.2d 692, 693 (Tex.1986). However, consistent with my view that we should overrule Majrige and its progeny and recognize no presumption for or against finality, I do not believe any type of conclusory finality language can ever be read to grant more relief than requested by the parties.6
We should determine summary judgment finality by comparing the live pleadings and the summary judgment order. A summary judgment order should only be final if it matches the contents of the pleadings. And, as was the law before Bandera, a court of appeals should summarily reverse any summary judgment granting more relief than requested, without any sua sponte severance of some issues while others are remanded.
Wiping the slate clean by overruling the rules created in Mafrige, Inglish, and Bandera while we study the best method of tackling summary judgment finality through our formal rule-promulgation process is the better solution for several reasons. First, this approach strikes a more reasonable balance between the competing policies of promoting certainty and preserving parties’ rights to appellate review. And, under this approach, the trial court and the parties drafting summary judgment orders would have the burden, and the incentive, to ensure that the pleadings, summary judgment motions, and the summary judgment orders match. If a premature appeal is taken, the court of appeals need only compare the pleadings, motions, and order. If the order does not dispose of parties or issues raised in the pleadings, then it is interlocutory and the court must dismiss the appeal.7 If the order explicitly *218disposes of issues and parties not raised in the motion, it is erroneous and the court must reverse the entire order.
Most importantly, this approach alters the consequences of poorly-drafted orders. Specifically, the consequence flowing from a poorly drafted order becomes the risk of a premature appeal rather than an untimely one. This eliminates the greatest risk Mafrige created — that an interlocutory order, contrary to the trial court’s and (at least one party’s) intent, will be fictitiously made final, starting the appellate and plenary power timetables even while the litigation continues. No one would ai’gue that conducting a trial after the trial court’s plenary power has expired is not a waste of judicial resources. Moreover, because overruling Bandera eliminates the benefits of a premature appeal, taking such an appeal would not be a cost-efficient mistake for litigants to make, increasing the incentive to ensure orders are more clearly drafted. If a premature appeal is nonetheless taken, it would not create an onerous burden for the appellate court. The opposing party need only file a brief pointing out that the pleadings, motion, and order do not match, leading to automatic remand or dismissal.
No one disputes that rules governing summary judgment finality could be helpful to the bench and bar and facilitate judicial efficiency. But history, as well as our own precedent, has shown that judicial opinions are not the place to achieve this. Any attempt to adhere to the Mafrige principle or retain parts of it while rejecting others can only lead to more problems. Instead, this Court should overrule Mafrige and its progeny and start anew. As the Court even notes, our rules advisory committee is currently studying summary judgment finality. 39 S.W.3d at 216. Retaining parts of Mafrige, Inglish, Bandera as modified by the Court’s less-than-clear opinion today — only to follow with promulgation of a concurrent finality rule — will only lead to more confusion.
I agree that the cases here should be reversed. But, because the Court refuses to fix the problems its judicial rulemaking in Mafrige caused and allow our rulemak-ing process to work, I cannot join the Court’s opinion.
IV. RECOMMENDATION
I recognize that the Supreme Court of Texas Advisory Committee on Rules of Civil Procedure has been studying the problem of summary judgment finality. It has proposed an amendment to Rule 166a of the Texas Rules of Civil Procedure:
(j) Statement of Grounds. An order granting summary judgment must state the ground or grounds on which the motion was granted. No judgment may be affirmed on other grounds stated in the motion unless they are asserted by appellee in the appellate court as alternative grounds for affirmance.
I do not believe this proposed amendment goes far enough.
First I would suggest to the committee that they consider requiring each summary judgment order specifically identify: (1) the claims each party brings; (2) the grounds upon which each party seeks summary judgment; (3) each ground upon which the trial court granted summary judgment; and (4) each ground upon which the trial court denied summary judgment.
This solution is intuitive. In the vast majority of cases, this formality, rather than including magic language, would provide notice to parties about what has actually happened. In practice, this procedure alleviates many problems Mafrige's finality rule has caused.
Under this approach, a summary judgment is not final unless the order specifically identifies each claim for relief, the grounds upon which each party seeks summary judgment, and the court’s disposition *219of each claim and party. The appellate court’s jurisdiction is determined only by looking at whether the trial court rendered an order expressly disposing of all remaining parties and issues. If the trial court errs by omitting certain claims or parties from the order, as happened in Lehmann and Hams, it is not a final order for purposes of appeal. Under this approach a party never loses its right to appeal based upon the finality of a summary judgment order that is silent about the party or its claims or that sua sponte grants relief no party requested without mentioning the parties or claims — regardless of how clearly it states that it is a final judgment disposing of all parties and issues.
Most significantly, in practice this would lead to better drafting and fewer erroneous appeals. Specifically, if required to expressly list each ground upon which summary judgment is requested, trial courts are not likely to add grounds to their order that the summary judgment motion did not raise.
Second, I would suggest the committee consider a rule requiring that the prevailing party, who is charged with drafting the court’s order, serve copies on all other parties at least ten days before the trial court is to sign and enter the order. Consistent with this suggestion, I agree with the Court’s suggestion that the clerk send copies of all the actual signed orders— rather than just a postcard indicating that the court has signed an order.
The majority’s author criticizes my first recommendation, asserting that there is a “very real risk” that requiring judges to be explicit in them summary judgment orders would result in “thousands of judgments intended to be final ... remaining] interlocutory.” 39 S.W.3d at 196. He contends that “[t]his is precisely what has happened in the federal system even though the federal rules impose far fewer requirements on final judgments than the dissent would.” 39 S.W.3d at 208. Federal Rule 58, to which he refers, requires that all final judgments “be set forth on a separate document” and be entered by the clerk on the docket. Fed.R.Civ.P. 58.
This criticism only serves to amplify the real dangers of straying outside the summary judgment context in these cases. How finality of different types of judgments is determined must be governed by the nature of the judgment. Houston Health Clubs, Inc., 722 S.W.2d at 693 (“In determining whether a judgment is final, different presumptions apply depending on whether the judgment follows a conventional trial on the merits or results from default or a motion for summary judgment.”). Cognizant of this, my recommendation, unlike Federal Rule 58, is limited to summary judgment finality.
The live pleadings define the issues in a case. The issues tried do not always mirror these pleadings. See Vance v. Wilson, 382 S.W.2d 107, 108 (Tex.1964). Nonetheless, we have repeatedly recognized that a presumption should exist that all issues presented by the pleadings are disposed of in a conventional trial on the merits. See Aldridge, 400 S.W.2d at 897-98; Vance, 382 S.W.2d at 108. This presumption can be rebutted by a contrary showing in the record. See Richey v. Bolerjack, 589 S.W.2d 957, 959 (Tex.1979). But absent such a rebuttal, this presumption prevents judgments from languishing after trial based solely on variations in the pleadings and judgment. This presumption has saved us from the types of problems the federal system has experienced.
However, we sensibly limited this presumption to judgments “not intrinsically interlocutory in character.” Aldridge, 400 S.W.2d at 897. We have also explained that summary judgments are intrinsically interlocutory and thus they should not be presumed final. Houston Health Clubs, Inc., 722 S.W.2d at 693. Thus, there is nothing illogical about requiring that finality language be explicit. And I respectfully disagree that my recommendation, limited to summary judgments, will cause such *220major havoc in the court system. Further, I believe the additional formality in this context is worth the certainty and protections such a rule provides.
V. CONCLUSION
In Texas, the test for determining summary judgment finality has always been whether the judgment disposes of all parties and all issues raised in the pleadings. In Mafrige we created a legal fiction to simplify the process of determining finality. But Mafrige created more problems than it solved. It is beyond me why the Court insists on struggling through pages and pages of history about presumptions, magic language, and Mother Hubbard clauses instead of squarely considering the problems Mafrige caused and providing a solution. Its willingness to cling to this legal fiction, while refusing to recognize that our rulemaking in Mafrige and its progeny was not the correct solution, will only create more problems.
I concur in the judgment in these cases. But, because the Court declines to overrule Mafrige, Inglish, and Bandera, and await our promulgation of a rule governing summary judgment finality, I do not concur in its reasoning.
. These rules and presumptions are irrelevant to the issues before the Court today. As we have repeatedly admonished — in Mafrige, in Aldridge, and even in the Court’s opinion today — the rules governing finality after a conventional trial are wholly inappropriate for determining finality of summary judgments. See Mafrige, 866 S.W.2d at 592; North East Indep. Sch. Dist. v. Aldridge, 400 S.W.2d 893, 897-98 (Tex.1966); Lehmann, 39 S.W.3d 191.
. The unpublished opinions cited in Part I are cited only as examples, not as precedent. See Tex.R.App. P. 47.7.
. This issue also arises when a trial court expresslyNmentions and disposes of a party even though that party was not mentioned in the motion for summary judgment. Here, the lower courts have been more willing to apply Mafrige and hold that the order purports to dispose of all parties and issues. See, e.g., Mikulich v. Perez, 915 S.W.2d 88, 91-92 (Tex.App.—San Antonio 1996, no writ).
. In fact, the district clerk sent all the parties (including those omitted from the summary judgment order) a postcard indicating that an' "Order for Interlocutory Summary Judgment” had been signed. Lehmann, 988 S.W.2d at 416.
. Oftentimes in these cases litigation continues to move forward. Any error in including magic finality language in a summary judgment is not discovered until it is too late; the appellate timetable has expired and the trial court has lost plenary power to act. The litigants have forever lost their right to complain of the judgment.
. It would not be enough for a court to generally state "plaintiff takes nothing,” "defendant is granted summary judgment in all things,” or "this is a final appealable judgment.” Conclusory finality clauses (i.e. "magic language”) do not indicate that a trial court actually granted relief not requested for or against parties or issues are not mentioned in the order.
. Of course, this procedure would not apply if the order fell within the category of cases for which there can be more than one final judgment, or the category of orders for which a *218court of appeals has been granted statutory authority to review interlocutory orders.