dissenting.
Applying Lemon v. Kurtzman, 403 U.S. 602, 91 S.Ct. 2105, 29 L.Ed.2d 745 (1971), to religious symbols on city seals is no cakewalk. Lemon’s “three-part test” is not a test. It is a triad of questions, the answers to which conflict in all interesting cases. Conflict is what makes a case worth litigating. How to resolve conflicts? It is easy to list things to inquire about and hard to establish weights that produce outcomes congruent with the Constitution. Lemon does the former but not the latter. You need an understanding of the religion clauses, not a list of factors, to resolve tough cases such as ours. Before giving a constitutional opinion, however, we should be sure that there is a case or controversy. I have my doubts, and it is on this ground that the majority and I part company.
I
Plaintiffs say that they must go out of their way to avoid seeing religious symbols. This sort of allegation we deemed sufficient in ACLU v. City of St. Charles, 794 F.2d 265, 267-69 (7th Cir.1986), and Doe v. Village of Crestwood, 917 F.2d 1476, 1478 (7th Cir.1990). Standing sometimes depends on an “identifiable trifle”. Claims about detours are about as trifling as one can produce. They sidle up against the line drawn by Valley Forge Christian College v. Americans United For Separa*1420tion of Church and State, Inc., 454 U.S. 464, 102 S.Ct. 752, 70 L.Ed.2d 700 (1982), which holds that offense taken at the government’s complicity in religion does not create standing. Cf. Allen v. Wright, 468 U.S. 737, 104 S.Ct. 3315, 82 L.Ed.2d 556 (1984). If offense is not enough, why is a detour attributable to that offense enough? The answer cannot be, as my colleagues say, that the willingness to take a detour “validates the existence of genuine distress” (927 F.2d at 1406); Valley Forge tells us that dismay does not establish standing, and therefore new and better ways to prove its existence cannot create standing. St. Charles holds instead that the detour narrows the class of plaintiffs and scope of dispute. If that is its function then we must ensure, when the plaintiffs’ claims are challenged, that they have changed their daily constitutionals, that we are indeed dealing with injured rather than ideological plaintiffs. Both cities contested the plaintiffs’ claims, yet the district court decided the cases without resolving these disputes.
My colleagues say that in determining standing to sue the allegations of the complaint must be taken as gospel. Some language in St. Charles looks in this direction, distinguishing the allegations needed to establish standing from the proof needed to prevail, 794 F.2d at 269, but the defendants there conceded the plaintiff’s detour. These cities do not concede that the plaintiffs have rearranged their daily routines to avoid exposure to religious symbols. Quite the contrary, Zion contends that Harris moved in precisely because of the imagery in its seal. Harris related during his deposition that he had been recruited by the Society of Separationists as a potential plaintiff (apparently none of the existing residents of the city objected to the seal). He moved to a boarding house within the city’s limits and lent his name to the litigation. This looks like a put-up job, distinct from the claim in St. Charles and Crest-wood that long-time residents had been compelled by religious scruples to change the patterns of their lives. Harris has not incurred a “tangible, albeit small cost that validates the existence of genuine distress" (927 F.2d at 1406); he has incurred a tangible cost to expose himself to the seal. For Harris the seal is an attractive nuisance, not an irritant compelling changes of the sort we discussed in St. Charles and Crest-wood. At the end of the day, what matters is that St. Charles recognized that the plaintiff could not •prevail without proving the “detour” allegations of the complaint. 794 F.2d at 269. Today’s opinion dispenses with that requirement.
Ought we render constitutional judgment on the theory that federal courts are bound to accept whatever the scrivener put in the complaint? Then parties could confer jurisdiction on a court by consent, and collusive litigation would burgeon. Impositions on federal courts could flourish: one citizen of Illinois could sue another under the diversity jurisdiction by alleging that he is a citizen of Indiana, and despite the defendant’s protests the court would not examine the veracity of the allegation. Yet courts do examine the jurisdictional allegations of complaints — that is what 28 U.S.C. § 1359 and Fed.R.Civ.P. 12(b)(1), among other tools, are for.
Lujan v. National Wildlife Federation, — U.S. -, 110 S.Ct. 3177, 111 L.Ed.2d 695 (1990), furnishes an example. The D.C. Circuit allowed an inference of standing from the general allegations of the complaint, despite a challenge by the defendants and an inability of the plaintiffs to substantiate their allegations with proof. The Supreme Court reversed, holding that plaintiffs must do more than allege; in response to a motion for summary judgment, they must produce affidavits showing the essential kind of injury. So it should be in our cases. So it has been in the past in this circuit. Plaintiffs in Freedom from Religion Foundation, Inc. v. Zielke, 845 F.2d 1463 (7th Cir.1988), contended that a monument displaying the Ten Commandments violated the establishment clause. Most of the plaintiffs said that they took offense; we deemed this insufficient under Valley Forge. One of their number, Phyllis Grams, alleged that she lived nearby. This, too, we held insufficient:
Although Grams lives in the City of La Crosse, the appellants did not demon*1421strate that she lives anywhere near Cameron Park, that the monument is visible in the course of her normal routine, or that her usual driving or walking routes take her past the park. The appellants also failed to establish that Grams suffered any injury simply because of her close proximity to the ... allegedly unconstitutional display. We therefore conclude that Grams cannot establish Article III standing simply on the basis of her alleged but unproven proximity to the offending conduct.
845 F.2d at 1469. If Phyllis Grams had to prove her claims when challenged, why not Clint Harris and Theodore Kuhn? See also O’Hair v. White, 675 F.2d 680, 696 & n. 35 (5th Cir.1982) (in banc) (plaintiff must prove the factual allegations used to establish standing); Citizens Concerned for Separation of Church and State v. Denver, 628 F.2d 1289, 1298 (10th Cir.1980). A requirement of proof is essential to avoid unnecessary constitutional adjudication, and I do not see why we should go out of our way (take a detour?) to undermine that safeguard.
Doubtless there is a sense in which failure of proof does not abolish standing. A plaintiff suing a driver who, he says, ran him down in the street, loses outright if the evidence shows that the injury was suffered falling down stairs at home; we do not retroactively dismiss for want of standing. It is easy to marshal cases for this proposition; the majority does so, citing not only St. Charles but also South East Lake View Neighbors v. Department of Housing and Urban Development, 685 F.2d 1027, 1034 (7th Cir.1982). Little beyond nomenclature is at stake, because both of these cases recognize that the plaintiff must prove the allegations in order to prevail, just as Lujan held. See also Crawford v. United States, 796 F.2d 924, 928-29 (7th Cir.1986).
Failure of proof is not the only problem. I doubt that the allegations establish standing even if they can be proved. The cities put their seals on vehicle stickers, garbage bags, public buildings and cars, and public documents. Plaintiffs object to all of these, but most of these uses do not lead to an alteration in conduct. Harris’s complaint, for example, protests the appearance of the seal on mail Zion sends him. But he does not allege that these symbols lead him to alter his conduct in any fashion. So too with the seal on the stickers that demonstrate payment of the vehicle tax: Harris contends that he avoids looking at the sticker, but this is hardly a detrimental alteration in conduct; drivers are supposed to look at the road ahead of them, not at tax stickers. Even this tiny inconvenience could be eliminated. Someone with a simple self-help remedy for his problem suffers no “injury in fact” and is not entitled to a judicial remedy that adversely affects the interests of his neighbors who may like the seals. Plaintiffs have such a remedy. Wooley v. Maynard, 430 U.S. 705, 97 S.Ct. 1428, 51 L.Ed.2d 752 (1977), allows plaintiffs to cut out or cover up the objectionable portions of the seals. Drivers may not be conscripted as billboards for ideas of which they disapprove. So too with the imprinting of seals on the garbage bags Rolling Meadows supplies. If it tried to stop Kuhn from using bags without seals he would have standing — but only to claim his right under Wooley to edit the bags. For all we can tell, the city has no objection to redaction. Counsel for Kuhn said at oral argument that Rolling Meadows does not pick up garbage from bags that do not bear the seal. But the record does not reveal whether this is because of the absence of the seal or because the substitute bags were the wrong size or material. Silence on questions of this nature is not a firm basis for reaching a difficult constitutional question. At all events, our plaintiffs lack standing to stop other persons from having or using the seals. Could Maynard have obtained an injunction blocking New Hampshire from putting “Live Free or Die” on any license plates? I doubt it.
The public seals (e.g., on the water tower in Zion) give the best case for standing, because residents then may say: “I take a longer route to avoid viewing that seal.” There are two problems. First, despite what the majority says, Kuhn does not allege that he goes out of his way to avoid public displays of the seal. His affidavit *1422omits any such assertion, contending instead that he is embarrassed by the seal, and that he does not enter public buildings where it appears. (Would he have used these buildings in the absence of the seal?) Second, allegations of this kind, if proven, challenge only the display of the seal in a particular place, not its composition. St. Charles allowed the plaintiff to object to a particular, visible cross; we did not say that because the city erected a 35-foot, lighted cross in public, the plaintiff could obtain an order forbidding, say, a crucifix on the mayor’s desk. The constitutional injury could be eliminated by taking down the cross; here any injury could be eliminated by removing the seals from the water tower and other public places.
Knowing that someone, somewhere, is using a seal with a cross yields purely abstract injury, a demand that the first amendment be enforced for its own sake. Under Valley Forge neither Harris nor Kuhn has standing to challenge internal use, any more than they would have standing to challenge an icon in the mayor’s office, or a creche being stored in the basement of city hall. Even if they had standing, they could not prevail on the merits. Bowen v. Roy, 476 U.S. 693, 699-701, 106 S.Ct. 2147, 2152-53, 90 L.Ed.2d 735 (1986). Perhaps one could say that they have “standing” once they come across a single seal but that the relief should be limited to getting the seals out of the public eye, but this would be a detail. What matters is that they have asked for — and the majority gives them — relief more extensive than that necessary to eliminate their injuries. Call it standing, call it redressability, call it (following Roy) the limited scope of permissible relief on the merits, the point is the same. They have at most an entitlement to the obliteration of the symbols that cause detours; any more extensive relief is unwarranted. We should dismiss the complaints to the extent they challenge anything other than placements that lead to inconvenient detours, and remand the residue with instructions to put plaintiffs to their proof.
My colleagues’ unwillingness to do this stems in part from their belief that if the plaintiffs lack standing, then “there is simply no way to test the constitutionality of the city seals” (927 F.2d at 1409). This misunderstands two things: my position, and the Constitution. I do not say that the plaintiffs lack standing utterly; I contend, rather, that the district court should determine whether the plaintiffs’ factual assertions are true. If these plaintiffs cannot prove their allegations, perhaps some other residents of Zion and Rolling Meadows could. One who says that he was run down by a truck must prove it, and the demand for proof no more obliterates tort litigation than my position would wipe out establishment clause litigation. But if it did, so what? Article III of the Constitution limits judicial power to cases and controversies. If because no one is injured there is no controversy, then the Constitution demands that the court dismiss the suit. There is no exception for subjects that as a result cannot be raised at all. “ ‘The assumption that if [plaintiffs] have no standing to sue, no one would have standing, is not a reason to find standing.’ Schlesinger v. Reservists Committee to Stop the War, 418 U.S. [208,] 227, 94 S.Ct. 2925, 2935, 41 L.Ed.2d 706 [ (1984) ]. This view would convert standing into a requirement that must be observed only when satisfied.” Valley Forge, 454 U.S. at 489, 102 S.Ct. at 767. Constitutional decisions may be necessary in the course of deciding real cases. Marbury v. Madison, 5 U.S. (1 Cranch) 137, 2 L.Ed. 60 (1803). Courts do not create cases so that they will have the opportunity to render constitutional decisions. Adjudication is a byproduct of a real controversy, and not the other way around. If there is no case, then there is no occasion for deciding a constitutional question, and we should not mourn or struggle against this allocation of governmental powers. Hewitt v. Helms, 482 U.S. 755, 761-62, 107 S.Ct. 2672, 2676-77, 96 L.Ed.2d 654 (1987); Alliance to End Repression v. Chicago, 820 F.2d 873, 876 (1987).
II
I comment briefly on the merits because we may not be the final decision-makers.
*1423The establishment clause of the first amendment, as I understand it, forbids taxation and coercion in support of religion but does not forbid the display of religious symbols. Allegheny County v. ACLU, 492 U.S. 573, 109 S.Ct. 3086, 3134-46, 106 L.Ed.2d 472 (1989) (Kennedy, J., dissenting); American Jewish Congress v. Chicago, 827 F.2d 120, 128-40 (7th Cir.1987) (Easterbrook, J., dissenting). I omit the line of argument that produces this conclusion and refer readers to my earlier opinion and Justice Kennedy’s (which Chief Justice Rehnquist and Justices White and Scalia joined). Religious imagery on city seals does not raise taxes (seals with crosses are no more costly to produce than seals using other symbols) or penalize non-observance. The Constitution distinguishes threats from shadows. Zion’s seal has been in use for 89 years without stifling religious diversity or slowing the conversion of a religious enclave to a secular city. Not one resident of Zion other than Harris has expressed concern. Instead of driving non-believers away, the seal attracted Harris to it.
Perhaps a seal could penalize non-belief by striking fear into the hearts of viewers. See Friedman v. Board of County Commissioners, 781 F.2d 777 (10th Cir.1985) (in banc). “Con esta vencemos”, Bernalillo County’s slogan, was one the Jesuits used when burning heretics. When the police cars bearing it toured a predominantly Indian population, which had been mistreated by the Christians, there was reason to fear that the “Christian police” were at hand, perhaps with the Grand Inquisitor in tow. 781 F.2d at 779. Zion’s seal carries no such message to an objective (or any other) observer, and Rolling Meadows’ is pallid.
Justice Kennedy’s views did not prevail in Allegheny. Three justices (Brennan, Marshall, & Stevens, JJ.) took a strict separatist position, and two (Blackmun & O’Connor, JJ.) reached the conclusion that a local government may use religious symbols if it does not “endorse” religion. 109 S.Ct. at 3100-01. That two-justice position, the most narrow offered in support of the judgment, is the “holding” of Allegheny. Marks v. United States, 430 U.S. 188, 193, 97 S.Ct. 990, 993, 51 L.Ed.2d 260 (1977). Any legal position rejected by seven justices is shaky at best, and we cannot be confident that it will survive. Stability has not been a notable feature of establishment clause jurisprudence. No one should be surprised if the view held by four justices, one with deeper historical roots, see AJC, 827 F.2d at 135-37, eventually obtains a majority.
Applying the two-justice “no endorsement” view to city seals produces nothing but headaches. What is endorsement in a world pervaded by religious imagery, from the eye in the Great Seal of the United States (the eye of God in a pyramid representing the Christian Trinity) to “In God We Trust” on the coinage to Thanksgiving Day (to Whom are thanks being given?) to the religious stamps the Postal Service issues at Christmas and Easter to the names of our cities (Los Angeles, San Francisco, Corpus Christi) and submarines (the same list; there is even a “Los Angeles class” of submarines)? See Aronow v. United States, 432 F.2d 242 (9th Cir.1970); O’Hair v. Murray, 588 F.2d 1144 (5th Cir.1979). Did the United States endorse a particular religion when putting “Corpus Christi” on a submarine? But perhaps the name’s significance has been forgotten, or is overlooked in daily use. Why not say the same about the imagery on seals?
Does the city “endorse” everything on its seals? The majority answers yes and on this ground gives judgment for the plaintiffs. But Justices Blackmun and O’Con-nor did not cleave to such a view. Allegheny involved religious symbols on public property. The justices could have said that a city necessarily endorses whatever appears in its halls. Had they done this, the Court would have held unconstitutional the placement of a huge menorah in front of Pittsburgh’s city hall. Yet the Court sustained the placement, because a nearby Christmas tree withdrew any implication that the city endorsed Judaism. Should one not say the same thing for at least the Rolling Meadows seal, in which the cross is surrounded by secular symbols, including a leaf, a school, and a water tower? The majority’s riposte — whatever the city *1424prints, it endorses — is inconsistent with the views of six justices in Allegheny.
Cases such as ours show some of Lemon ’s problems. The majority immediately throws away one “prong” of its test for both cases, and a second tine (“purpose”) sloughs off for Rolling Meadows. Everything collapses to the no-endorsement rule, one more compelling if divorced from the three-part test (or as advanced thinkers call it these days, “tri-partite inquiry”). Why use three parts when only one matters?
A majority of the Court has reached this conclusion, making our reference to Lemon outdated. The four dissenting justices in Allegheny withheld support, 109 S.Ct. at 3134, and Justice O’Connor, who provided the essential fifth vote for the majority, did so without relying on Lemon, a case of which she has been critical. Aguilar v. Felton, 473 U.S. 402, 426-30, 105 S.Ct. 3232, 3245-47, 87 L.Ed.2d 290 (1985) (O’Connor, J., dissenting). Thus only four justices relied on Lemon in Allegheny. Lynch v. Donnelly, 465 U.S. 668, 104 S.Ct. 1355, 79 L.Ed.2d 604 (1984), the Court’s other recent case dealing with religious symbols, also bypasses the Lemon “tests”. We should take this cue and do likewise.
To see why Lemon has lost its tang one need look no further that the majority’s application of its purpose prong. My colleagues say that Zion’s seal violates the first amendment because it was designed by Rev. Dowie and adopted by the city council in 1902 to carry out the religious mission of the Christian Catholic Church. Indeed so; Zion was founded as a religious utopian community, and the Theocratic Party governed for more than two generations. In 1902 no one would have given this a second thought: the speech clause of the first amendment was not applied to the states until Gitlow v. New York, 268 U.S. 652, 45 S.Ct. 625, 69 L.Ed. 1138 (1925), and the religion clauses were not extended to the states for another 27 years. Zorach v. Clauson, 343 U.S. 306, 72 S.Ct. 679, 96 L.Ed. 954 (1952). Given our commitment to cultural and religious diversity, Wisconsin v. Yoder, 406 U.S. 205, 92 S.Ct. 1526, 32 L.Ed.2d 15 (1972), the establishment of religious communities is unexceptionable. Many of this Nation’s founders sought to establish religiously identifiable communities, and their value has been shown recently by the Mormons, who needed separate communities to survive hostility to their beliefs, and by the Amish, whose way of life is tightly integrated with their religious beliefs. The purpose behind Zion’s creation — of which its seal is only an incidental part — does not count against it.
Why should a purpose to promote religion ever condemn an act of government? Often it does not: all of the accommodation-of-religion cases flunk this purpose inquiry, yet many accommodation rules are sustained and, when doing this, the Court ignores Lemon. Michael W. McConnell, Accommodation of Religion, 1985 Sup.Ct. Rev. 1. Other laws have religious purposes (e.g., the Sunday closing laws) but are sustained because they also have secular purposes.
Suppose Zion originally built schools because Rev. Dowie said they made for better missionaries. Would the schools in Zion have to be torn down? Fra Angelica painted to glorify the Roman Catholic Church; do we doubt that the display of his paintings in the National Gallery has a secular purpose? Why would the purpose of maintaining the seal in Zion be different? If there is something wrong with the seal, how about the city’s name? This, too, was picked for a religious purpose and conveys a religious message — more potent than the seal. Must the city change its name?
The majority allows that a genuine change of heart can cleanse the improper motive. Yet why has not Zion had a change of heart? Are we looking into the heads of the members of the city council who voted in 1986 to retain the seal for (they assert) historical reasons? The Theocratic Party lost power in Zion more than 40 years ago and is defunct. So far as the record shows, not one member of Zion’s city council belongs to the Christian Catholic Church or owes office to constituents who support that faith. Before declaring that the current purpose of the seal is religious, we would need a trial that has *1425not yet occurred. Zion can’t lose a motive question on an empty record. (There were fat records in Edwards v. Aguillard, 482 U.S. 578, 107 2573, 96 L.Ed.2d 510 (1987), and Wallace v. Jaffree, 472 U.S. 38, 105 S.Ct. 2479, 86 L.Ed.2d 29 (1985).)
So we should bypass Lemon as Justice O’Connor did in Allegheny and ask her question directly: would an objective observer view these seals as endorsements of Christianity? Whether the observer would have this reaction gives me more trouble than it does the majority. The imagery in Zion’s seal is overwhelming, but the historical context may prevent an informed observer from inferring contemporary endorsement. As for Rolling Meadows: its seal is not powerfully Christian. The object of everyone’s attention looks more like a telephone pole than a cross; its arms are too short to crucify anyone. The original plaintiff dropped out because he came to conclude that the symbol is a utility pole in front of an industrial building. We know, from the affidavit of the artist, that it was meant to be a cross, but an objective observer lacks this knowledge. I do not think we can say, as the majority does (927 F.2d at 1403), that it is “no doubt” a cross. It is “obviously” a cross only after you know the genesis of the seal, which is to say that to the outside observer it is not obviously a cross. It is enough, however, that Rolling Meadows concedes that an observer would take this as a cross, and we ought to accept the concession rather than treat the debatable as obvious.
How, then, are we to carry out the tasks set for us by Justices Blackmun and O’Con-nor? There is no straightforward way— perhaps no way, period — to do so. Which may indicate that we have been sent on a wild goose chase, that the proper answer lies along the path of the separatists (Justices Brennan, Marshall, and Stevens) or along the path marked by Justice Kennedy. The best I can do is to predict that Justices Blackmun and O’Connor would condemn the use of any religious imagery not nestled in a secular context. What kind of context? We discussed the subject in Crestwood, 917 F.2d at 1478-79:
Two contexts mattered — first the season, for in each case [.Allegheny and its predecessor Lynch ] the government was displaying the symbols appropriate to the time of year; second the immediately surrounding symbols, for in each case the government was displaying an assortment of symbols appropriate to all aspects of the holidays. Christmas and Hanukkah are secular as well as religious holidays; to use symbols appropriate to all aspects of the occasion is not to endorse a particular religion. Mather v. Village of Mundelein, 864 F.2d 1291 (7th Cir.1989). If Christmas may be a secular holiday, the state may recognize whose birthday is being commemorated.
Zion’s symbols do not have a secular context of any sort. Rolling Meadows is not trying to put the cross in a seasonal context and is not displaying all of the symbols for an occasion. It selected a few, which are unrelated except by being within the borders of the city. Governance is not a “context” that drains the significance of a religious image.
So long as the two-justice position in Allegheny rules, we must condemn these seals (if we reach the merits). This approach avoids the need to apply judicial aesthetic judgment, or perhaps recur to the gestalt psychologists. I don’t know how to determine what observers think, if they think at all about these seals. Line drawing in this area will be erratic and heavily influenced by the personal views of the judges. We ought to use bright line rules, and Allegheny (while it lasts) admits of this one.