dissenting:

I respectfully dissent.

I believe that the majority errs in holding that the district court abused its discretion by not abstaining from deciding the merits. There are no uncertain state law questions that must be resolved here. Rather, the case turns principally on a factual issue with a painfully obvious answer. Moreover, I believe that the majority’s disposition of the case illustrates how procedural rules intended to aid the federal judiciary in performing its duties can, if misapplied, serve to deprive parties of a fair and adequate resolution of their dispute. By unnecessarily and improperly invoking the abstention doctrine, the majority does a disservice to both parties, and no favor to the federal courts. Rather than vacate the district court’s decision on abstention grounds, we should reverse its state law decision on the merits and remand the case to the district court for resolution of Kolls-man’s constitutional claims.

In January 1977, Paul Kollsman submitted to the City of Los Angeles an application for the development of a seventy lot *838 subdivision. While that application was pending, he filed an action in district court containing both state and federal claims. 1 After considering the arguments for abstaining, the district court ruled that abstention was not appropriate. 2 The district court went on to dispose of the action on state law grounds and never reached the federal claims.

The district court held, in what must have been a surprise to all concerned, that, under California law, Kollsman’s application was accepted as complete on November 22, 1978 and was deemed to be approved as a result of the City’s failure to act by November 22, 1979. Apparently finding it difficult to accept the district court’s singular approach to the issue, the majority concluded that there are too many uncertain questions of state law involved and abstention is therefore warranted. I, on the other hand, think it clear that the district court wholly ignored the plain requirements of the California Government Code as well as the uncontrovertible facts when it decided that the City had accepted Kollsman’s application as complete. In my opinion, there are no unclear or uncertain questions of state law that require resolution in this case.

The district court, intentionally or otherwise, circumvented the provisions of California law by finding as a matter of fact that the City had accepted Kollsman’s application as complete. Because its critical finding of fact is so clearly erroneous, the district court’s decision cannot stand. It is obvious from the record before us that the City did the precise opposite of that which the district court says it did. The City expressly found that the application was incomplete, not complete. Moreover, even if, under California law, Kollsman’s application could have been deemed complete, that event would not have occurred until several days late on November 26, 1978. Even in that case, the City could not be deemed to have approved the application, for the City subsequently explicitly disapproved the application in a timely manner. Thus, contrary to the district court’s assertion, it is absolutely clear that the City never approved the application, explicitly or by operation of law.

The majority unnecessarily raises many state law questions that it says have unclear answers. However, we need not resolve any of those state law issues in order to decide this case. We can and must view the issue of the appropriateness of abstention differently after the district court has decided the case on the merits than we would at the outset of the litigation. When an abstention motion is made early in the proceedings, the court will often not be in a position to know which issues will prove determinative or even which issues it will be necesary to consider. Much may depend on how the parties decide to develop facts that are not yet before the court and how the court then resolves particular factual questions: whether or not a court must proceed down a particular road often will be determined by the decisions that are made at the preceding crossroad. However, after the trial is over and the facts and the law have both been fully explored, we can determine with certainty what, if any, state law questions must be decided and whether the answer to any such questions is unclear. The majority therefore errs in justifying abstention by raising many state law questions simply because at first blush they may appear to be of some relevance; in fact, a careful review of the fully developed record demonstrates that those questions are wholly irrelevant to the proper resolution of the dispute before us.

The majority errs in another basic respect in its approach to the abstention doc *839 trine. It is not enough simply to say, as the majority does: the district court says X and the City says Y; therefore, there is a dispute as to state law; accordingly, the law is unclear and we must abstain. In order to conclude that abstention is appropriate, a court must at least identify which issues are critical to the outcome of the litigation and which involve questions to which the answer is, in its opinion, actually unclear. The majority does neither here.

THE INAPPLICABILITY OF THE ABSTENTION DOCTRINE

The general rule concerning the exercise of federal jurisdiction was once simple:

Cohens v. Virginia, 19 U.S. (6 Wheat.) 264 , 404, 5 L.Ed. 257 (1821) (Marshall, C.J.).

Over time, however, the Supreme Court has made limited inroads on that rule. One example is the abstention doctrine, originally created in Railroad Commission of Texas v. Pullman Co., 312 U.S. 496 , 61 S.Ct. 643 , 85 L.Ed. 971 (1941). “[T]he Supreme Court has stated that the doctrine of abstention ‘is an extraordinary and narrow exception to the duty of a District Court to adjudicate a controversy properly before it.’ ” Hillery v. Rushen, 720 F.2d 1132, 1137 (9th Cir.1983) (quoting County of Allegheny v. Frank Mashuda Co., 360 U.S. 185, 188 , 79 S.Ct. 1060, 1063 , 3 L.Ed.2d 1163 (1959)); see, e.g., Colorado River Water Conversation District v. United States, 424 U.S. 800, 813 , 96 S.Ct. 1236, 1244 , 47 L.Ed.2d 483 (1976); C-Y Development Co. v. City of Redlands, 703 F.2d 375, 377 (9th Cir.1983).

Designed for reasons of federalism and for the purpose of making it possible to avoid the unnecessary resolution of constitutional questions, the abstention doctrine allows a district court to exercise its equitable powers in extraordinary instances to abstain from resolving state law issues. See, e.g., Pue v. Sillas, 632 F.2d 74, 78 (9th Cir.1980) (citing Baggett v. Bullitt, 377 U.S. 360 , 84 S.Ct. 1316 , 12 L.Ed.2d 377 (1964)). In reviewing a district court’s decision whether “compelling reasons” justify application of the abstention doctrine, we apply an abuse of discretion standard. See, e.g., Midkiff v. Tom, 702 F.2d 788 , 789 n., 1, 799 (9th Cir.1983), reversed on other grounds, sub nom. Hawaii Housing Authority v. Midkiff, — U.S.-, 104 S.Ct. 2321 , 81 L.Ed.2d 186 (1984). We have been extremely hesitant to find that a district court abused its discretion in deciding state law questions. For example, in a case involving a municipal zoning decision, we refused to hold that the district court abused its discretion in deciding not to abstain; we did so even though it “was a case in which abstention would have been advisable, and well within the discretion of the court____ State law questions clearly predominated over federal questions. But we have no basis for holding that it was an abuse of discretion to retain the case.” Shamrock Development Co. v. City of Concord, 656 F.2d 1380, 1385 (9th Cir.1981). 3

As the majority correctly restates, we have relied on three criteria in evaluating *840 the propriety of the application of the abstention doctrine. See, e.g., Canton v. Spokane School District No. 81, 498 F.2d 840, 845 (9th Cir.1974). The first criterion is that the complaint must involve a “sensitive area of social policy” that is best left to the states to address. 4 The second criterion is that a ruling on the state law issue may render unnecessary the resolution of a constitutional question. The final factor, which I believe is dispositive here, is whether the proper resolution of the possibly determinative state law issue is uncertain. See, e.g., Pue v. Sillas, 632 F.2d at 78-81 .

. If the proper resolution of the state law question is certain, abstention is improper. Even if the particular issue has never been addressed by the state courts, the proper resolution of a state law question may be certain. See, e.g., Harman v. Forssenius, 380 U.S. 528, 534-35 , 85 S.Ct. 1177, 1188-89 , 14 L.Ed.2d 50 (1965); Meredith v. Winter Haven, 320 U.S. 228, 232, 234-35 , 64 S.Ct. 7, 10-11 , 88 L.Ed. 9 (1943); Hillery v. Rushen, 720 F.2d 1132, 1137 (9th Cir.1983); Midkiff v. Tom, 702 F.2d at 789 -90 n. 1; see also Field, Abstention in Constitutional Cases, 122 U.Pa.L.Rev. 1071, 1092 (1974) (arguing in favor of “[l]imiting abstention to cases in which the state law issue is extremely unclear”; (emphasis added)). Thus, even though the district court said that “[t]here have been few cases which interpret the California code sections which are involved in this case,” 565 F.Supp. 1081, 1090 , that alone would not mandate abstention.

Although several interrelated provisions of California law are raised by one or the other of the parties, the facts of this case, and the express language of the provisions that determine its results, are such that it is not necessary to resolve any unclear questions of state law. The one legal theory that the district court seems to be suggesting may readily be determined to be wholly without foundation. Then, when the simple critical factual question raised by the district court is isolated and given even the most cursory consideration, the outcome of the case becomes absolutely clear; whatever uncertain state law questions may otherwise be said to exist are no longer of any possible relevance. In short, the majority’s entire list of “uncertain” state law questions consists of questions' that we need not, and should not, resolve in this case. Accordingly, the majority’s invocation of the abstention doctrine is unnecessary and inappropriate.

THE DISTRICT COURT’S DECISION

It is clear that under California law the City was required to determine whether Kollsman’s application was complete and notify him in writing of that determination by November 26, 1978; in addition, if the City determined that the application was not complete it was required to specify the deficient parts of the application in the written notification. There can be no dispute as to the state law on those points. 5

*841 The City informed Kollsman in a letter dated November 22, 1978, that his application had “been determined to be incomplete,” and stated what information was missing from the application. The district court seemingly relied on two grounds in holding that by sending the letter the City accepted the application as complete on November 22, 1978. First, the district court concluded that the application in fact was complete under California law; it then apparently decided that, for that reason, the letter must be read as having the opposite effect of its stated purpose. Second, the district court apparently read a subsequent letter of November 14, 1979 as somehow reversing, with retroactive effect, the plain meaning of the letter of November 22, 1978. Neither of these arguments can withstand any level of scrutiny, and neither purports to rely on any principle of California law.

The district court’s unarticulated legal theory that if an application is in fact complete, a determination that it is incomplete constitutes a determination that it is complete is highly ingenious but wholly unmeritorious. Moreover, even assuming that the application did in fact contain all the information required by California law 6 and that the City’s request for further information was erroneous or made in bad faith, the City’s one year period for approving a complete application still could not have commenced to run earlier than November 26, 1978, if at all.

Assuming the answer to every theoretically unclear state law issue to be favorable to Kollsman’s position, the one year period for approval or disapproval would have begun on November 26, 1978 because that was the date on which, under the California statute, the City was required to make its determination as to the completeness of Kollsman’s application. 7 The only possible way the one year period could have commenced to run before November 26, 1978 would have been not only for the City’s finding of incompleteness to be deemed a finding of completeness, but even more critical, for the City to have somehow obtained notice that such was the paradoxical effect of its request for further information. While it may be argued that under some circumstances a City may have notice that an application will be deemed complete if it fails to act on a date by which it is required to decide whether or not the application is complete, I am aware of no theory under which the City of Los Angeles could *842 be said to have been afforded notice that its preliminary determination as to incompleteness and its request for further information would trigger the running of a one year statute that commences, according to its terms, only when the application is “accepted as complete.” 8

Thus, the district court’s decision must rest on its second theory — its apparent resolution of a factual issue it perceived to be raised by the two letters sent by the City to Kollsman. The November 22, 1978 letter was, as we have noted, clear and unambiguous. It found Kollsman’s application incomplete and requested further information. Almost one year later, in a form letter to Kollsman dated November 14, 1979, the City stated that under a recent statutory amendment it “is required to act on certain projects ... within one year of the date of acceptance of an application as complete” and that it planned to disapprove the application on November 21, 1979. The letter contained no specifics about Kollsman’s application, except for the date that it was originally filed. It did not state that Kollsman’s application had ever been accepted as complete; nor did it state the date on which any such action might have occurred, nor the date on which the City was required to approve or disapprove his application. In fact, the letter stated specifically that the application “has been held pending receipt of requested additional materials.” Thus, on its face the November 14, 1979 letter in no way reverses, or even contradicts, the November 22, 1978 one; in fact, it is in some important'respects consistent with the earlier letter.

The district court read the letter as “a stroke of conscience,” 565 F.Supp. at 1096 , and concluded that it demonstrated that the City had accepted the application as complete on November 22, 1978. This factual conclusion is wholly contrary to the evidence and to all rules of logic. It is inconceivable that the November 14, 1979 letter — a form letter that was devised to comply with a recent amendment to a statute and was in significant respects consistent with the November 22, 1978 letter — could serve to reverse completely the plain meaning of the earlier letter so that an express statement that the application was incomplete must be read retroactively as stating that the application was complete. The majority says that “[t]he district court’s rationale for concluding that the application must be deemed complete on November 22, 1978, is somewhat opaque.” Majority opinion p. 835 n. 12. I disagree. Rather than merely opaque,, the district court’s factual conclusion is, in my opinion, bizarre, as well as wholly and clearly erroneous.

There is, in fact, only one possible interpretation of the November 22, 1978 letter: Kollsman was notified on November 22, 1978 that the City had determined that his application was incomplete; the City’s letter meant and continued to mean precisely what it said and not the opposite. That meaning could not be changed, retroactively or otherwise, by a subsequent letter. Moreover, there are various' possible explanations for the City’s action in sending the November 14, 1979 form letter. The letter could have been sent to Kollsman (1) because the City was making one final attempt to obtain information it had been seeking for some time, (2) because the City had decided to disapprove all applications in which the parties failed to supply requested information, despite the fact that no formal disapproval was required with respect to applications previously found to *843 be incomplete, (3) out of bureaucratic confusion over whether the recent amendment to the statute required new disapproval letters in cases of longstanding incomplete applications, or (4) simply out of an excess of caution. It is of no significance which explanation is correct, nor is it significant that the City could have sent an individualized letter that more clearly specified its purpose and that was more narrowly tailored to Kollsman’s situation. What matters for the purposes of the case before us is that, as I have said, the 1979 letter could not under any theory serve to reverse, retroactively, the plain meaning of the 1978 letter. Of equal importance, even if a district court did have the discretion to give such effect to the 1979 letter, there is no means by which retroactive notice to the City could serve to afford it the one-year statutory period within which to approve or disapprove a completed application.

After an application is accepted as complete, the City is required to approve or disapprove it within one year. Cal.Gov’t Code § 65950 (Deerings 1979). 9 Failure to do so is deemed to constitute approval. 10 Even assuming that Kollsman’s application could be deemed complete as a result of the City’s finding it incomplete, as I have already explained, the earliest date on which it would have been deemed complete would have been November 26, 1978. Thus, at worst, the City had until November 26, 1979 to approve or disapprove the application. It acted within that deadline when it sent a letter of disapproval on that date.

In sum, despite the majority’s statements to the contrary, there are no uncertain state law questions that we need to resolve. Rather, to reach the proper result, we need only resolve a factual question with an obvious answer.

Abstention was improper for another reason as well. There is no “sensitive area of social policy” implicated here. Although zoning may generally involve sensitive areas of social policy, this narrow dispute does not.

The specific issues raised here are far from sensitive if for no other reason than that they will almost certainly never be raised again. As the district court recognized, there is “little likelihood that the future will bring State Court decisions on applications which were filed prior to January 1, 1978 and remain undetermined. In this respect, the Kollsman case may be unique. At least the possibility of additions to the species has long been extinct.” 565 F.Supp. at 1090 .

In short, the majority errs in stating that Kollsman’s claim “raises difficult state law issues of great importance to the people of California.” Majority Opinion slip op. p. 3053, p. 833. We need not decide any difficult or unclear questions of state law, but only a simple factual issue — the meaning of an unambiguous letter. ■ Moreover, because most of the issues the parties discuss in the briefs probably will never be raised again, it is wrong to say that they are of great, or even any, importance to the people of California.

In conclusion, because two of the three requirements for abstention have not been satisfied, the majority errs in invoking the doctrine.

THE ADVERSE EFFECTS OF ABSTENTION

The majority’s holding that the district court abused its discretion in not abstaining will unnecessarily delay resolution of Kollsman’s claims for many years. Assuming that the plaintiffs have the necessary funds and stamina to pursue this litigation indefinitely, the state law issues will at some point be resolved in favor of the City, on one ground or another, and the federal courts will ultimately, if belatedly, be required to resolve the constitutional questions. In a decision holding that a district court properly decided the state law issues before it, the Supreme Court Con *844 demned the delay resulting from unnecessary application of the abstention doctrine. Meredith v. Winter Haven, 320 U.S. 228, 237 , 64 S.Ct. 7, 12 , 88 L.Ed. 9 (1943); see C. Wright, Law of Federal Courts § 52, at 305 (recognizing that Pullman abstention has “[i]n a number of well-known cases ... led to delays of many years before the case was finally decided on its merits or limped to an inconclusive end” (footnotes omitted)).

Here, the district court had jurisdiction over both the federal and state law claims; absent extraordinary circumstances, the plaintiff should not be deprived of the right to have his claims resolved in a federal forum. Furthermore, the delay that will be caused by unnecessary and improper abstention in this case is far longer than the two year delay that the Supreme Court criticized in Winter Haven .

In January 1977, Kollsman submitted the development proposal to the City. In March 1977, he filed this complaint. The district court decided the case in February 1983. We heard the appeal in January 1984. The majority’s decision will effectively render moot the district court proceedings and send the state issues to the California courts for a new beginning. This will, of course, delay for many years the final resolution of this dispute.

The suit has already outlived the original plaintiff, Paul Kollsman. The Kollsman estate will be most fortunate indeed if it is able to resolve even the simple and unimportant state law aspect of this litigation by 1987 — ten years after the time the original application was filed. The case will then return to the federal courts for resolution of the federal issues unless by that time the plaintiffs have abandoned their rights in despair. In my opinion, we have a legal duty to exercise our jurisdiction now.

THE MERITS OF THE STATE LAW ISSUES

In order to discuss the abstention questions fully, it has been necessary to point out why I believe the district judge decided the merits of the state-law questions erroneously. I would, for the reasons I have already discussed, reach those merits and reverse the district court’s decision.

CONCLUSION

Because there are no unclear questions of state law that require resolution, because a sensitive social policy is not implicated, and because the state law issues are in any event of little or no importance to the people of California, the majority errs in invoking the abstention doctrine. To dispose of this case, we need only resolve a simple factual question with an obvious answer, By attempting to make abstention the absolute rule in zoning cases rather than the exception in all cases, the majority does a disservice to the litigants, who deserve, at a minimum, a reasonably prompt resolution of the federal issues that are determinative of this lawsuit. We should allow the already long delayed dispute to proceed to an orderly resolution. The abstention doctrine and its accompanying delay should be reserved for eases involving extraordinary circumstances.

The district court’s decision on the state law questions should be reversed on the merits and the case remanded so that Kollsman’s constitutional claims may be resolved. I believe that’the majority errs in vacating the district court’s decision on the basis of the abstention doctrine. I therefore dissent.

. Although the majority does not specify the precise basis of the district court’s jurisdiction, the district court stated that it had both diversity and federal question jurisdiction. We have jurisdiction over the appeal of the district court’s final decision under 28 U.S.C. § 1291 (1982).

. In deciding not to abstain, the district court suggested that its decision turned on the conclusion that it was capable of interpreting the state law "with minimum interference in state processes.”

. The majority’s decision is based entirely on its view that the district court should have abstained. Thus, to reach its result, the majority was required to apply an abuse of discretion standard. It is far from clear why the majority believes that the district court's decision was not only wrong but that it constituted an abuse of discretion. I, on the other hand, believe the district court was right not to abstain but that it based that decision on the wrong reasons. I first review the district court's substantive decisions regarding state law issues under the standard we normally use to review such decisions. Having concluded that the district court was wholly incorrect regarding the merits of those state law issues, I believe it is then appropriate to review the abstention issue under a de novo standard. However, I would reach the result I do here regardless of the standard of review applicable to any of these questions.

. I agree with the district judge that the effect of any decision here as to the state law questions, whatever it might be, will for practical reasons be so limited that the first criterion is not met. See infra p. 843.

. As enacted in 1977, with an effective date of January 1, 1978, Cal.Gov't Code § 65943 (Deer-ings 1979) stated that:

However, Cal.Gov’t Code § 65924 (Deerings 1979) as enacted in 1977 and amended in 1978 established a different time limitation for applications, like Kollsman’s, filed before, but not determined to be complete by, January 1, 1978. Section 65924, in relevant part, states that

Because section 65924 became effective on September 26, 1978, the City had 60 days from September 26, 1978, or until November 26, 1978, to determine whether Kollsman’s applica *841 tion was complete. Neither the district court nor any party disputes the applicability of this section to Kollsman’s appeal.

. The letter indicated that, along with other information, an environmental impact report (EIR) was missing from the application. The parties disagree as to whether under California law Kollsman was required to file an EIR. Given the proper resolution of the case, see infra pp. 841-843, I find it wholly unnecessary to resolve the question of whether an EIR was required to be submitted with the application.

. We may assume, for example that under California law Kollsman’s application would have been deemed complete if the City had failed to act by November 26, 1978. We can also assume that Kollsman's application contained all the information required by applicable statutes. We can even assume that under California law a bad faith response is treated in the same manner as a failure to respond at all and that in this case the City's response-was made in bad faith. I make these assumptions arguendo although in my opinion there is no statutory basis for deeming Kollsman’s application to be complete under any circumstances. Cal.Gov't Code § 65943 (Deerings 1984) was amended in 1979 (and the amendment became effective on October 2, 1979) to add a new second sentence: "If such written determination is not made within 30 days after receipt of the application, the application shall be deemed complete for purposes of this chapter.” Because his application was filed in January 1977, this new "deemed complete” provision to section 65943 did not apply to Kollsman’s application. Section 65924 did, instead. But even if the amendment to section 65943 were applied retroactively and to applications subject to the separate time limitation of Cal.Gov't Code § 65924 (Deerings 1979), see supra note 5, under the "60 days after the effective date” provision of the latter section the application could not have been deemed complete until November 26, 1978. Therefore, the City’s disapproval on November 26, 1979, would have fallen within the one year limitations period of Cal. Gov’t Code §§ 65950, 65956(b) (Deerings 1979). Accordingly, we need not address any of the "unclear" state law questions.

. Cal.Gov’t Code § 65950 (Deerings 1979) states, in relevant part, that "any public agency which is the lead agency for a development project shall approve or disapprove such project within one year from the date on which an application requesting approval has been received and accepted as complete by such agency." (emphasis added). Section 65950 was amended in 1983. The amendments become effective on January 1, 1990 and therefore do not bear upon Kolls-man’s application. Even making all the assumptions set forth in note 7, supra there is no way the City could have known that all the statutory construction issues, including the obviously insuperable bar of retroactivity, would be decided in an arbitrary and erroneous manner and that in addition its sending of its November 22, 1978 letter of incompleteness would be treated by the district court as constituting a determination of completeness.

. See supra note 8.

. Cal.Gov’t Code § 65956(a) (Deerings 1979) provides that "[i]n the event that a lead agency or a responsible agency fails to act to approve or to disapprove a development project within the time limits required by this article, such failure to act shall be deemed approval of the development project.”