concurring in the opinion in part, and in the judgment.
Judge Flaum is rightly concerned with the district court’s failure to probe the adequacy of the settlement embodied in the consent decree that the court approved. But Judge Fairchild and I believe that there is no one before us who is entitled to challenge the consent decree, because the denial of the motion to intervene by the two fathers of fetuses must be affirmed and the women’s appeal from the consent decree must be dismissed because they lack standing to appeal even though they were parties in the district court. On the first issue, however — the propriety of the denial of intervention to the fathers — Judge Fair-child’s and my grounds differ, while on the second (with which I shall begin) I write in amplification of his discussion.
The female appellants, being Illinois women of child-bearing age, are members of the plaintiff class because it is defined as “all Illinois women of child-bearing age who desire or may desire an abortion sometime in the future” (emphasis added). Members of a class whose rights will by operation of res judicata be extinguished by the settlement of the class action of course have standing to object to the settlement and to press their objections on appeal, but like other litigants they must show that the order to which they object actually harms them. Ordinarily the harm arises from the fact that the settlement does not provide the class with as much relief as the objecting member had wanted. He claims that the named plaintiffs prematurely and excessively bargained away his rights. That is not the objection of the female appellants in our case. Their objection is the opposite: that the settlement is too favorable to the plaintiff class. Like the would-be intervenors, these women oppose abortion and want the statute enforced a outrance. To object to a settlement on the ground that you shouldn’t have done as well in the settlement as you did identifies you as an ideological litigant; and an affront to one’s ideology is not an interest that will support standing to sue. Diamond v. Charles, 476 U.S. 54, 66-67, 106 S.Ct. 1697, 1705-06, 90 L.Ed.2d 48 (1986); Valley Forge Christian College v. Americans United for Separation of Church & State, Inc., 454 U.S. 464, 473, 102 S.Ct. 752, 759, 70 L.Ed.2d 700 (1982); United States v. SCRAP, 412 U.S. 669, 687, 93 S.Ct. 2405, 2416, 37 L.Ed.2d 254 (1973).
Maybe the class was misdefined, and women who, though of child-bearing age, are opposed to the availability of inexpensive abortions should not have been included because they lack the relevant community of interest with women who consider themselves harmed by the statute. If the female appellants had not been members of the class they would have had to seek leave to intervene in the lawsuit, like the men, In re Fine Paper Antitrust Litigation, 695 F.2d 494, 499 (3d Cir.1982), and their claim to intervene would have fared no better than the men’s. But even if the class was properly defined to include these women, which would make them parties in the district court without their having to intervene, a member of a class does not have standing to challenge a settlement *507that favors him in a tangible sense, though it offends him in an ideological sense. No one has such standing.
Because the women lack standing to maintain this appeal even though by virtue of being class members they were entitled to the rights of parties in the district court, it is critical to the maintenance of the appeal to decide whether the men should have been allowed to intervene in the district court either to represent the interests of the fetuses affected by the consent decree or to represent their own interests as fathers of such fetuses. Judge Fairchild believes that these would-be intervenors had failed to show that the State of Illinois is an inadequate representative of the fetuses who will be affected by the consent decree and therefore had failed to satisfy an essential requirement of Rule 24(a)(2) of the Federal Rules of Civil Procedure for intervention as a matter of right. I have my doubts, as I also do about Judge Fairchild’s alternative ground for upholding the denial of intervention — that the motion to intervene was untimely. On November 22, 1989, the named plaintiffs and the Illinois attorney general jointly requested the Supreme Court to postpone oral argument pending the submission of a proposed consent decree to the district court. The district judge fixed February 13, 1990, as the deadline for submitting objections to the decree. The motion to intervene was filed that day. The timing is suspicious. It may well indicate a desire to disrupt the schedule that the judge had set for consideration of the proposed decree. Yet the motion was filed only three months or so after the movants first learned (sometime in November) that the attorney general had decided in effect to throw in the towel. It is true that this was years after the suit had been brought. But a motion for intervention filed back then would have been dismissed out of hand. There was every reason to expect the Attorney General of Illinois to defend the statute as vigorously as he could. It was not until November of 1989 that this expectation was shattered. The petition for intervention was filed shortly afterward. It may still have been untimely (a question on which we usually defer to the district judge, United States v. City of Chicago, 897 F.2d 243 (7th Cir.1990)), but I find Judge Fairchild’s analysis of the point unconvincing because he focuses on the delay from the inception of the suit, and that is the wrong focus. United States v. City of Chicago, 870 F.2d 1256, 1263 (7th Cir.1989). See also United States v. South Bend Community School Corp., 710 F.2d 394, 396 (7th Cir.1983).
The proposition that the attorney general is an adequate representative of the fetuses that will be aborted if the consent decree is approved fictionalizes the notion of “adequacy” of representation, and I am not a fan of legal fictions. The consent decree guts a statute that was (to speak realistically) designed to limit the number of abortions performed in Illinois by making abortion more expensive. The statute imposed onerous requirements, in the name of health, on abortion providers. In invalidating it this court found the health rationale less than persuasive, so that the statute stood exposed as an attempted end run around Roe v. Wade. Ragsdale v. Turnock, 841 F.2d 1358, 1371-75 (7th Cir.1988). The statute’s basic requirements are that abortions may not be performed in doctors’ offices and that abortion clinics have to be so well equipped to deal with possible medical emergencies as to be the equivalent of small hospitals. The joint effect of these requirements, had they not been enjoined from the outset, would have been to increase the cost of abortion in Illinois substantially, because few public hospitals are willing to perform abortions.
The consent decree preserves very little of the statute. Ragsdale v. Turnock, 734 F.Supp. 1457 (N.D.Ill.1990). The “small hospital” requirements are confined essentially to facilities that perform abortions after the eighteenth week of pregnancy. So first-trimester abortions — the vast majority (more than 90 percent nationwide, though I don’t have figures for Illinois)— are unaffected by the statute. And the decree abandons the statute’s ban on abortions performed in a physician’s office.
The Attorney General of Illinois thus knuckled under to this court’s divided panel decision even though the Supreme Court had scheduled the case for oral argument of his appeal, and by doing so he doomed *508the first-trimester fetuses who would have been saved if the statute had been saved. He traded these fetal interests for other goods, such as an end to a costly lawsuit that he feared losing, or more likely (for he might well have won the case in the Supreme Court) for political advantage. But traded them he has. Perhaps the beneficiaries of the trade include a few second- or third-trimester fetuses, but almost all abortions are performed by the eighteenth week of pregnancy.
I know the government is presumed to be an adequate representative of a proposed intervenor when it is “charged by law with representing [the proposed inter-venor’s] interests.” American National Bank & Trust Co. v. City of Chicago, 865 F.2d 144, 148 (7th Cir.1989); United States v. South Bend Community School Corp., 692 F.2d 623, 627 (7th Cir.1982); see United States v. Hooker Chemicals & Plastics Corp., 749 F.2d 968, 984-90 (2d Cir.1984) (Friendly, J.). And since the case law treats the state as the guardian of the interests of fetuses carried in the wombs of women in the state, Roe v. Wade, 410 U.S. 113, 150, 93 S.Ct. 705, 725, 35 L.Ed.2d 147 (1973), it is an easy step to the conclusion that the state is a presumptively adequate representative of those interests. The step was taken in Keith v. Daley, 764 F.2d 1265, 1270 (7th Cir.1985), and Roe v. Casey, 623 F.2d 829, 832 n. 7 (3d Cir.1980). But a real presumption-is rebuttahle-f ‘conclusive presumption” is an oxymoron), and if this one is not, why did the court in Keith bother to point out that the attorney general was defending the abortion statute at issue in that case adequately? 764 F.2d at 1270. In negotiating the consent decree in the present case, the attorney general pretty much abandoned the fetuses to the abortionist’s knife.
To speak in this dramatic fashion is of course to treat fetuses as people — as holders of interests — rather than as inanimate objects; and the status of fetuses is controversial, to say the least. But for many purposes the law does treat them as people. In Illinois, if you shoot a pregnant woman in the abdomen and kill the fetus, you are guilty of the crime of intentional homicide of an unborn child, and the penalty is almost as severe as for first-degree murder; the only difference is that the death penalty may not be imposed. Ill.Rev.Stat. ch. 38, ¶ 9-1.2. This is true regardless of the age of the fetus. Illinois has also extended its tort statute for wrongful death of a human being so that it covers fetuses from the moment of conception. Ill.Rev.Stat. ch. 70, II 2.2. No one suggests that in extending to first-trimester fetuses legal protections originally designed for children, the state is violating Roe v. Wade, so long as it doesn’t try to use these legal protections to interfere with abortions privileged by that decision — in other words, so long as it protects fetuses against third persons but not against mothers upon whom Roe v. Wade confers a constitutional right of abortion, or against their agent, the abortionist. It follows, I should think, that fetuses are (in Illinois anyway) persons for purposes of a decision on what weight to give their interests in applications for intervention made under Rule 24(a)(2). That rule doesn’t create or define interests; it takes those interests as it finds them in state law or other sources of legally protected rights. A fetus’s father is not entitled under existing law to prevent the mother from having an abortion, Planned Parenthood of Central Missouri v. Danforth, 428 U.S. 52, 70-71, 96 S.Ct. 2831, 2841-42, 49 L.Ed.2d 788 (1976), but who is better fitted to represent the fetus if the state’s attorney general abandons the protective responsibility that law places on him? “Abandonment” of fetal interests is a fair description of the attorney general’s action in negotiating this consent decree.
Judges have a natural inclination to fictionalize “adequacy” of representation in order to prevent the courts from being swamped by multiparty litigation. (In this ease intervention would lead to the nullification of the consent negotiations and the restoration of the case to the Supreme Court’s docket, since a case can’t be settled without the consent of all the named parties and these would-be intervenors are not prone to compromise.) If a state in defending an environmental statute gave less weight to preserving forests than to preserving wetlands, it would not follow that arborealists should be permitted to inter*509vene to present evidence on behalf of their beloved trees. This is the sense of such decisions as United States v. Hooker Chemicals & Plastics Co., supra, and of our own United States v. 36.96 Acres of Land, 754 F.2d 855 (7th Cir.1985). More than docket pressures are at work in these cases. Environmental litigation involves tradeoffs among human health, aesthetic and ecological concerns, and commercial values, and the government agencies charged with administering these statutes should not be presumed to be incompetent to balance the competing considerations in a reasonable way. This case, too, could be thought to involve tradeoffs between “statistical lives” (i.e., low probabilities of illness or death) and other valuable goods, inasmuch as the abortion statute does not forbid abortions but merely makes them more costly. That is one way to look at the case but another is that a government official has decided to allow a class of what for purposes of this suit we must treat as human beings to die because the official lacks the stomach, political or otherwise, to litigate the case in the Supreme Court. In such a case the presumption that he is representing the class adequately might be thought rebutted.
I need not pursue the question. Intervention was properly denied, regardless of adequacy of representation, simply because to be allowed to intervene as a party you must have standing to litigate and these movants do not. It is true that Diamond v. Charles is noncommittal on the question whether an intervenor must have standing. 476 U.S. at 69, 106 S.Ct. at 1707; see also id. at 73-74, 106 S.Ct. at 1709 (concurring opinion); Chiles v. Thornburgh, 865 F.2d 1197, 1212 (11th Cir.1989) (reviewing split in the courts of appeals). But this court has held that he must. Keith v. Daley, supra, 764 F.2d at 1268. He wants to be a party, in major part so that he can litigate if the other parties with whom he is aligned fall out of the case, which is just what has happened here. The state has dropped the torch; the fathers of potentially affected fetuses want to pick it up. They must therefore demonstrate that they have standing to litigate this case as plaintiffs and appellants.
They do have standing in the barebones Article III sense. As fathers of unborn children (as the State of Illinois chooses to regard fetuses from the moment of conception), they are harmed, albeit in merely a probabilistic sense — but that is enough for standing, North Shore Gas Co. v. EPA, 930 F.2d 1239, 1242 (7th Cir.1991), and cases cited there — by a consent decree that by reducing the cost of abortions makes it more likely (though not highly likely) that their unborn children will be aborted. They are also the natural representatives of a group of inarticulate and helpless persons whose lives are at stake, and the extinction of those lives is the sort of tangible injury that is the stuff of actual cases in the Article III sense.
But there are other criteria of standing besides whether the plaintiff or persons represented by him have suffered an actual injury, which is the Article III criterion. The pertinent one here is whether the person seeking party status is someone upon whom the statute confers a right of enforcement. North Shore Gas Co. v. EPA, supra, 930 F.2d at 1243. The movants want to intervene in order that the abortion statute may be enforced in its pristine form, uncontaminated by the attorney general’s compromises. But it is a regulatory statute, and in general and also in regard to this particular statute private persons have no right (in our legal system, unlike for example the British) to enforce criminal or other regulatory statutes, unless of course the statutes also create private rights of action, which this one does not. Private persons can complain to the enforcement authorities and badger them to bring enforcement actions but they cannot force them to do so or stand in their place and bring the actions themselves. Some regulatory statutes, it is true, are interpreted to create implied private rights of action, but no one argues that the Illinois abortion statute should be so interpreted. And while I can imagine an argument that a woman who suffers a medical injury as a result of the failure of an abortion clinic to comply with the statute should be able to use the violation to establish medical malpractice, the argument is not made and *510anyway is not available to these inter-venors; they are not within the hypothetically protected class. (The female appellants are, and their party status — as members of the plaintiff class — is unquestioned. But they have as I said earlier no standing to appeal a settlement on the ground that it gives them more than they want or are entitled to. Nor have they asked that they be realigned as parties defendant — in which event their lack of standing would rest on the same ground as the fathers’.) The would-be intervenors are persons distressed by the attorney general’s refusal to enforce a regulatory statute to the hilt. Such persons have no standing to sue, and therefore they have no standing to intervene in this suit to prevent adoption of a consent decree that will disable the attorney general from enforcing the statute effectively.
Any doubt is dispelled by the Supreme Court’s decision in Diamond v. Charles, supra. Diamond was a physician who supported another Illinois abortion statute held unconstitutional by this court. He had been permitted (rightly or wrongly we need not decide today) to intervene in the lower-court proceedings in order to defend the statute, and he tried to appeal the case to the Supreme Court, the state attorney general having acquiesced in our decision and refused to appeal.. The Supreme Court held that Diamond lacked standing to maintain the appeal. A private citizen lacks a legally protected interest in the prosecution of another person. 476 U.S. at 64-65, 106 S.Ct. at 1704-05. See also Linda R.S. v. Richard D., 410 U.S. 614, 619, 93 S.Ct. 1146, 1149, 35 L.Ed.2d 536 (1973); Board of Trade v. SEC, 883 F.2d 525, 529 (7th Cir.1989). Hence no individual may “assert any constitutional rights of the unborn fetus. Only the State may invoke regulatory measures to protect that interest, and only the State may invoke the power of the courts when those regulatory measures are subject to challenge.” Diamond v. Charles, supra, 476 U.S. at 67, 106 S.Ct. at 1706.
Any other conclusion would interfere with the separation of powers within state government, and that separation is not a proper matter of federal judicial concern. Risser v. Thompson, 930 F.2d 549, 552 (7th Cir.1991). Illinois has given its attorney general the exclusive responsibility to enforce the abortion statute. It has not parceled out that responsibility between him and a host of self-appointed private attorneys general. It is his decision to make whether and how vigorously to enforce the abortion statute, cf. Heckler v. Chaney, 470 U.S. 821, 831, 105 S.Ct. 1649, 1655, 84 L.Ed.2d 714 (1985), and part of this deci-sional responsibility is deciding how vigorously to defend the statute in suits challenging its constitutionality. “The concerns for state autonomy that deny private individuals the right to compel a State to enforce its laws apply with even greater force to an attempt by a private individual to compel a State to create and retain the legal framework within which individual enforcement decisions are made.” Diamond v. Charles, supra, 476 U.S. at 65, 106 S.Ct. at 1705. We may not use Rule 24(a)(2) to subvert the state’s separation of powers by preventing the Attorney General of Illinois from exercising the responsibilities that the state has assigned to him, and to him alone.
Concern has been expressed that state or federal officials might use federal consent decrees to tie the hands of their successors or to disrupt the allocation of power between branches of state government. Frank H. Easterbrook, “Justice and Contract in Consent Decrees,” 1987 U. Chi. Legal Forum 19; Michael W. McConnell, “Why Hold Elections? Using Consent Decrees to Insulate Policies From Political Change,” id. at 295. A lame-duck state attorney general might agree to embody in a federal consent decree restrictions on the operation of the attorney general’s office that might cripple his successor, or (as charged in this case) that might gut an arguably constitutional statute that had been duly enacted by the state legislature. By asserting that no one has standing to appeal the consent decree in this case, I may seem to be placing such gambits beyond possibility of appellate correction. That would be paradoxical, since one of my grounds for denying that any of the appellants has standing was that permitting the appeal would interfere with the internal *511state allocation of governmental powers. Concern with such interference has been held a proper ground for declining to approve a consent decree. Kasper v. Board of Election Commissioners, 814 F.2d 332 (7th Cir.1987).
The paradox is dispelled by reflection that if Attorney General Hartigan’s successor should seek to enforce the statute in defiance of the consent decree and be met with a charge that he is violating the decree, he will be able to challenge its lawfulness on appeal, since the fact that the decree could not have been appealed by the usual route would make that a proper form of collateral attack. Cf. Martin v. Wilks, 490 U.S. 755, 768, 109 S.Ct. 2180, 2187-88, 104 L.Ed.2d 835 (1989). Or should the state legislature condition appropriations for the attorney general’s office on his enforcing the statute notwithstanding the decree and the condition be challenged as a violation of the decree, the legislature can by this route, and for the same reason, obtain appellate review. (And in federal court, notwithstanding the intragovernmen-tal character of such a suit. Cf. Coleman v. Miller, 307 U.S. 433, 438, 59 S.Ct. 972, 975, 83 L.Ed. 1385 (1939); Risser v. Thompson, supra, 930 F.2d at 550-51.) The appellants before us cannot obtain appellate review only because the decree has not done them the type of harm that the law requires as a predicate for mounting a legal challenge, whether in a trial or in an appellate court. They are the wrong appellants.