The Hope Clinic v. James Ryan, 195 F.3d 857

The Hope Clinic v. James Ryan

195 F.3d 857 | 7th Cir. | 1999

195 F.3d 857 (7th Cir. 1999)

The Hope Clinic, et al., Plaintiffs-Appellees,
v.
James E. Ryan, Attorney General of Illinois, and Richard K. Devine, State's Attorney of Cook County, Illinois,Defendants-Appellants.
Dennis D. Christensen, et al., Plaintiffs-Appellants,
v.
James E. Doyle, Attorney General of Wisconsin, and Diane M. Nicks, District Attorney for Dane County, Wisconsin, Defendants-Appellees.

Nos. 98-1726, 99-2528, 99-2533

United States Court of Appeals, Seventh Circuit

Argued September 22, 1999
Decided October 26, 1999

Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 97 C 8702--Charles P. Kocoras, Judge.

Appeals from the United States District Court for the Western District of Wisconsin. No. 98-C-0305-S--John C. Shabaz, Chief Judge.[Copyrighted Material Omitted][Copyrighted Material Omitted]

Before Posner, Chief Judge, and Coffey, Flaum, Easterbrook, Manion, Kanne, Rovner, Diane P. Wood, and Evans, Circuit Judges.*

Easterbrook, Circuit Judge.

We must decide whether state laws prohibiting partial-birth abortions are unconstitutionally vague or unduly burden women's rights. Acting without an evidentiary hearing, a district court held the Illinois statute unconstitutional and entered a permanent injunction. Hope Clinic v. Ryan, 995 F. Supp. 847 (N.D. Ill. 1998). But after a trial, another district court concluded that the Wisconsin statute is valid. Planned Parenthood of Wisconsin v. Doyle, 44 F. Supp. 2d 975 (W.D. Wis. 1999). A panel of this court earlier had ordered preliminary relief against Wisconsin's law, see Planned Parenthood of Wisconsin v. Doyle, 162 F.3d 463 (7th Cir. 1998), and plaintiffs say that this decision, applying Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), justifies permanent injunctions against both states' rules. To ensure consistency, we heard the appeals en banc. We conclude that both laws can be applied in a constitutional manner. Whether that will occur depends on state courts, which alone can settle questions about the construction of the statutes. To ensure that physicians are not deterred from performing other medical procedures while issues wend their way through state tribunals, we hold that both sets of plaintiffs are entitled to injunctive relief that will limit the statutes' application to the medical procedure that each state insists is its sole concern.

* Induction, suction curettage, and dilation and evacuation (D&E) are the principal methods of performing abortions in the United States. Prohibiting any one of these would conflict with the right of abortion recognized by cases such as Casey, 505 U.S. at 877 (plurality opinion) (adopting "undue burden" as the constitutional standard), and Planned Parenthood of Central Missouri v. Danforth, 428 U.S. 52, 75-79 (1976) (holding that a state may not forbid saline amniocentesis, at the time the principal means of induction). Our cases involve an uncommon procedure known to the medical community as "intact dilation and extraction" or just "dilation and extraction (D&X)," and to the public as "partial-birth abortion."

Some medical background is essential to understanding the issues. Induction means inducing preterm labor, which causes the expulsion of the conceptus. Methotrexate or mifepristone (ru-486, now in clinical trials), in combination with misoprostol, can be used for this purpose early in pregnancy; saline amniocentesis and injected prostaglandins serve the same function in the second trimester. Suction curettage (vacuum aspiration), the most common surgical method of abortion early in pregnancy, refers to evacuation of the uterine cavity; the embryo or fetus is separated from the placenta either by scraping or vacuum pressure, then is removed by suction. When these methods are inappropriate, or do not work, physicians employ the D&E procedure. To perform a D&E, the physician dilates the cervix and dismembers the fetus inside the uterus using forceps. Fetal parts are removed with forceps or by suction.

A D&X is a variant of a D&E in which the fetus is removed without dismemberment. The American College of Obstetricians and Gynecologists (ACOG) defines D&X as follows: "1. deliberate dilatation of the cervix, usually over a sequence of days; 2. instrumental conversion of the fetus to a footling breech; 3. breech extraction of the body excepting the head; and 4. partial evacuation of the intracranial contents of a living fetus to effect vaginal delivery of a dead but otherwise intact fetus." Martin Haskell, the physician who developed the D&X procedure, see Dilation and Extraction for Late Second Trimester Abortion (1992), reprinted in 139 Cong. Rec. E1092 (Apr. 29, 1993), believes that how the head is diminished in size so that it can pass through the cervix is not important: mechanically crushing the skull serves the same end as evacuating its contents, which causes its collapse. It is this combination of coming so close to delivering a live child with the death of the fetus by reducing the size of the skull that not only distinguishes D&X from D&E medically but also causes the adverse public (and legislative) reaction. Opponents deem the D&X procedure needlessly cruel and bordering on infanticide, and all three states in this circuit have enacted statutory restrictions.

The statute in Illinois has three sections with legal significance:

720 ILCS sec.513/5. Definitions

In this Act: "Partial-birth abortion" means an abortion in which the person performing the abortion partially vaginally delivers a living human fetus or infant before killing the fetus or infant and completing the delivery. The terms "fetus" and "infant" are used interchangeably to refer to the biological offspring of human parents.

720 ILCS sec.513/10. Partial-birth abortions prohibited

Any person who knowingly performs a partial-birth abortion and thereby kills a human fetus or infant is guilty of a Class 4 felony. This Section does not apply to a partial-birth abortion that is necessary to save the life of a mother because her life is endangered by a physical disorder, physical illness, or physical injury, including a life-endangering condition caused by or arising from the pregnancy itself, provided that no other medical procedure would suffice for that purpose.

720 ILCS sec.513/15. Civil action

The maternal grandparents of the fetus or infant, if the mother has not attained the age of 18 years at the time of the abortion, may in a civil action obtain appropriate relief unless the pregnancy resulted from the plaintiff's criminal conduct or the plaintiff consented to the abortion. The relief shall include money damages for all injuries, psychological and physical, occasioned by the violation of this Act and statutory damages equal to 3 times the cost of the partial-birth abortion.

The Indiana statute defines "partial-birth abortion" in identical terms. Ind. Code sec.16- 18-2-267.5. Like Illinois it forbids partial- birth abortions unless that procedure is necessary to save the mother's life, and no other procedure would suffice. Ind. Code sec.16-34-2- 1(b). The Indiana statute has never been challenged and has been in effect since July 1, 1997.

Wisconsin has taken a slightly different approach. Its statutes provide:

Wis. Stat. sec.895.038 Partial-birth abortions; liability.

(1) In this section:

(a) "Child" has the meaning given in sec.940.16 (1) (a).

(b) "Partial-birth abortion" has the meaning given in sec.940.16 (1) (b).

(2) (a) Except as provided in par. (b), any of the following persons has a claim for appropriate relief against a person who performs a partial-birth abortion:

1. If the person on whom a partial-birth abortion was performed was a minor, the parent of the minor.

2. The father of the child aborted by the partial-birth abortion.

(b) A person specified in par. (a) 1. or 2. does not have a claim under par. (a) if any of the following apply:

1. The person consented to performance of the partial-birth abortion.

2. The pregnancy of the woman on whom the partial-birth abortion was performed was the result of a sexual assault in violation of sec.940.225, 944.06, 948.02, 948.025, 948.06 or 948.09 that was committed by the person.

(3) The relief available under sub. (2) shall include all of the following:

(a) If the abortion was performed in violation of sec.940.16, damages arising out of the performance of the partial- birth abortion, including damages for personal injury and emotional and psychological distress.

(b) Exemplary damages equal to 3 times the cost of the partial-birth abortion.

(4) Subsection (2) applies even if the mother of the child aborted by the partial-birth abortion consented to the performance of the partial-birth abortion. Wis. Stat. sec.940.16 Partial-birth abortion.

(1) In this section:

(a) "Child" means a human being from the time of fertilization until it is completely delivered from a pregnant woman.

(b) "Partial-birth abortion" means an abortion in which a person partially vaginally delivers a living child, causes the death of the partially delivered child with the intent to kill the child, and then completes the delivery of the child.

(2) Except as provided in sub. (3), whoever intentionally performs a partial-birth abortion is guilty of a Class A felony.

(3) Subsection (2) does not apply if the partial-birth abortion is necessary to save the life of a woman whose life is endangered by a physical disorder, physical illness or physical injury, including a life-endangering physical disorder, physical illness or physical injury caused by or arising from the pregnancy itself, and if no other medical procedure would suffice for that purpose.

Wisconsin's definition of partial-birth abortion is substantially the same as Illinois', but the mental-state elements differ, as do the maximum penalties: Illinois makes unjustified partial- birth abortion a Class 4 felony, with a maximum penalty of three years; in Wisconsin the offense is a Class A felony, for which Wis. Stat. sec.939.50(3)(a) provides a penalty of life imprisonment.

Because Wisconsin and Illinois use similar language to define partial-birth abortions, we illustrate plaintiffs' concerns with the latter's statute:

"Partial-birth abortion" means an abortion in which the person performing the abortion partially vaginally delivers a living human fetus or infant before killing the fetus or infant and completing the delivery. The terms "fetus" and "infant" are used interchangeably to refer to the biological offspring of human parents.

This legal definition is an imperfect match for the medical definition of D&X. It is easy to see why a legislature would be chary of the ACOG's specification: then any small variation (such as a change in the method of reducing the head size, or snipping off a toe to defeat the "otherwise intact" specification) would take the abortion outside the prohibition, even though the reasons why the technique has been deemed objectionable would be unaffected. But, as is common with legislation, the price of avoiding loopholes is generality. Section 513/5 captures the idea, central to the D&X procedure, that an intact fetus moves from uterus to vagina before death occurs. But it also uses the words "delivers" and "delivery," which many physicians understand to refer to any removal of fetal material from the uterus. The law might be read to prohibit the extraction of dismembered parts following a D&E, or to prohibit abortion by induction if by chance the fetus survives until it reaches the birth canal. Moreover, physicians performing a D&E sometimes do not complete the dismemberment inside the uterus, and some fear that this could lead the procedure to be characterized as a partial delivery under the statute.

The possibility that sec.513/10 might discourage risk-averse physicians from performing the D&E procedure led the district court in Illinois to declare that it is too vague to be enforced, and a similar concern animated our panel's decision to require preliminary relief against Wisconsin's law. Moreover, if these laws needlessly raise the effective costs of induction, suction curettage, or D&E, then they would unduly burden the right of abortion, another concern expressed not only by the district court in the Illinois case but also by the panel in the Wisconsin case. But after a trial on remand, the district court in Wisconsin concluded that physicians who perform abortions recognize that the statutory formula, however vague or broad these words appear to lay eyes, refers to the D&X procedure alone. The court held that, as so limited, the statute does not substantially burden any woman's right to an abortion, because D&X is never the only safe procedure.

Litigation elsewhere has produced results mirroring the divergence between the outcomes in Illinois and Wisconsin. The sixth circuit held that Ohio's ban on partial-birth abortions is unconstitutional, Women's Medical Professional Corp. v. Voinovich, 130 F.3d 187 (6th Cir. 1997), but the fourth circuit issued a stay preventing interference with Virginia's statute. Richmond Medical Center for Women v. Gilmore, 144 F.3d 326 (4th Cir. 1998), motion to vacate stay denied, 183 F.3d 303. Both of those decisions were rendered over dissents. A single panel of the eighth circuit recently held three states' partial-birth-abortion laws unconstitutional, but only after first concluding that all three statutes forbid the D&E procedure. Carhart v. Stenberg, 192 F.3d 1142 (8th Cir. Sept. 24, 1999) (Nebraska); Little Rock Family Planning Services, P.A. v. Jegley, 192 F.3d 794 (8th Cir. Sept. 24, 1999) (Arkansas); Planned Parenthood of Greater Iowa, Inc. v. Miller, 195 F.3d 386 (8th Cir.1999) (Iowa).

Plaintiffs would like us to employ the approach of our panel's opinion in the Wisconsin case and declare both states' laws unconstitutional without ado. But different statutes have different language, which may be important. Litigants also have made distinctive arguments-- for example, Illinois relies on United States v. Salerno, 481 U.S. 739, 745 (1987), for the proposition that, except in first amendment cases, a law may be held unconstitutional only when "no set of circumstances exists under which the Act would be valid." Moreover, two events since the panel wrote call for fresh analysis. First, that opinion dealt with preliminary relief and observed that "the full trial may cast the facts in a different light". 162 F.3d at 466. That trial has been held in Wisconsin, and its results must be considered. Second, both the panel's decision in the Wisconsin case and the district court's opinion in the Illinois case conclude that the statute as written is vague. A more recent decision of the Supreme Court, Chicago v. Morales, 119 S. Ct. 1849 (1999), stresses that state courts are entitled to construe state laws to reduce their ambiguity, and that federal courts should evaluate state laws as they have been construed, not just as they appear in the statute books. See also, e.g., Parker v. Levy, 417 U.S. 733 (1974); Civil Service Commission v. Letter Carriers, 413 U.S. 548 (1973). This implies that Illinois and Wisconsin are entitled to interpret their own laws--that federal courts should not enjoin all application before enforcement and thus prevent the state courts from having a chance to save their statutes.

Salerno offers a potential ground for giving the states that chance. It would be hard to say that every possible instance of the D&X procedure is protected by the Constitution. Along similar lines are the many cases saying that outside the domain of the first amendment, vagueness challenges must be assessed "as applied." See, e.g., Maynard v. Cartwright, 486 U.S. 356, 361(1988); Hoffman Estates v. Flipside, Hoffman Estates, Inc., 455 U.S. 489, 494-95 (1982). But we are reluctant to rely too heavily on this approach, because Morales shows that three Justices believe that this language is too sweeping, see 119 S. Ct. at 1858-59 & n.22 (Stevens, J., joined by Souter & Ginsburg, JJ.), and the Supreme Court has on occasion resolved constitutional challenges to statutes that do not involve speech without asking whether every conceivable application would be unconstitutional. E.g., Romer v. Evans, 517 U.S. 620 (1996); Kolender v. Lawson, 461 U.S. 352, 358-59 n.8 (1983); Nash v. United States, 229 U.S. 373 (1913). Only one Justice offered any support for Salerno's approach in Morales, and he did so in a dissenting opinion. 119 S. Ct. at 1867-72 (Scalia, J., dissenting). Justice Scalia identified many cases that articulate Salerno's requirement, and many others that bypass the inquiry. Courts of appeals are divided on the question whether Salerno applies to abortion legislation. Compare Jane L. v. Bangerter, 102 F.3d 1112, 1116 (10th Cir. 1996), with Barnes v. Moore, 970 F.2d 12, 14 n.2 (5th Cir. 1992). The safest course for a court of appeals, when confronted with inconsistent lines of precedent, is to decide on other grounds if at all possible. Here it is possible, for we think that the Supreme Courts of Illinois and Wisconsin could read their laws in ways that comport with the Constitution.

1. One means of doing this would be to assimilate the statutory definitions to the medical definition of D&X, with allowance for different ways of reducing the head size and other immaterial variations. Both states are concerned about the D&X procedure and did not set out to forbid any other. The Attorneys General of Illinois and Wisconsin, the principal defendants, tell us that their statutes are concerned only with the D&X procedure and will be enforced only against its use. That assurance might be enough by itself, in the absence of any contrary indication from the state judiciary, to resolve immediate vagueness concerns. See Frisby v. Schultz, 487 U.S. 474, 483 (1988); cf. Arizonans for Official English v. Arizona, 520 U.S. 43, 76- 80 (1997). Although local prosecutors (States' Attorneys in Illinois, District Attorneys in Wisconsin) initiate criminal prosecutions, representatives of these prosecutors have been named as parties in each case, and they agree with the Attorneys General. Plaintiffs object that this approach would rewrite rather than interpret the law, but we doubt this. Both medical and popular literature equate "partial- birth abortion" (the statutory term) with the D&X procedure. See Janet E. Gans Epner, Harry S. Jonas & Daniel L. Seckinger, Late-term Abortion, 280 J. Am. Medical Ass'n 724 (Aug. 26, 1998); M. LeRoy Sprang & Mark G. Neerhof, Rationale for Banning Abortions Late in Pregnancy, 280 J. Am. Medical Ass'n 744 (Aug. 26, 1998). The district court in Wisconsin found that this is the statute's aim. 44 F. Supp. 2d at 984 (relying on the Attorney General's concession). Using a medical definition to supplement a vague lay definition does not strike us as revisionism or an exercise in deconstruction.

But if this approach would nonetheless be an example of brute force used to save a statute-- well, courts do it all the time. Florida had a law forbidding "the abominable and detestable crime against nature". Placement in the code showed that this crime had something to do with sex, but what? Did it forbid incest? Necrophilia? Bestiality? A legal historian might give an answer, but to lay readers of the statute, and even to most lawyers, the words are Delphic. The state's highest court filled in the blank by saying that the object was sodomy--and the Supreme Court of the United States rebuffed a charge of unconstitutional vagueness, given the state court's prestidigitation. Wainwright v. Stone, 414 U.S. 21 (1973). Only if vagueness remains after judicial interpretation is there a constitutional problem, the Court held. See also, e.g., United States v. Lanier, 520 U.S. 259 (1997); Rose v. Locke, 423 U.S. 48 (1975); Lanzetta v. New Jersey, 306 U.S. 451, 455-57 (1939); Kolender, 461 U.S. at 355-57. Federal courts have performed equally dramatic feats. As written, the Federal Election Campaign Act is hopelessly vague and overbroad. Buckley v. Valeo, 424 U.S. 1 (1976), saved the bulk of the statute by adding details. In two recent cases we predicted that state courts would follow suit with comparable state laws. See Wisconsin Right to Life, Inc. v. Paradise, 138 F.3d 1183 (7th Cir. 1998); Brownsburg Area Patrons Affecting Change v. Baldwin, 137 F.3d 503 (7th Cir. 1998). In the latter case, because we were not completely confident that the state judiciary would follow Buckley's path of freewheeling interpretation, we certified the issue to the Supreme Court of Indiana-- which responded that it indeed possesses, and is inclined to use, that power. Brownsburg Area Patrons Affecting Change v. Baldwin, 714 N.E.2d 135 (Ind. 1999).

Relying on Colautti v. Franklin, 439 U.S. 379 (1979), the sixth circuit concluded in Women's Medical Professional Corp. v. Voinovich that any statute concerning abortion is unconstitutionally vague, no matter how precise an interpretation state courts eventually develop, unless it requires proof that the physician knows that a given act has been forbidden. If this were so, then the approach to avoiding vagueness that we have just sketched would be untenable. But it is not so; never has the Supreme Court held that all criminal laws (or even just all criminal laws affecting abortion) depend on scienter. Cases we have mentioned already (and more that we cite later) hold that a statute may be sustained against a charge of vagueness if, as construed, it gives reasonable notice of the forbidden conduct. That the notice does not sink in--that some people close their eyes (or minds) and thus do not learn of the law's contents or appreciate its application to their conduct--does not prevent a state from enforcing its rules. See United States v. Wilson, 159 F.3d 280, 288-89 (7th Cir. 1998). This subject received sustained attention in Karlin v. Foust, 188 F.3d 446, 460- 64 (7th Cir. 1999). For the reasons laid out in Karlin we respectfully disagree with the sixth circuit's understanding of this question. See also Note, 112 Harv. L. Rev. 731 (1999) (criticizing Women's Medical Professional Corp.).

2. Although the Constitution does not compel states to ameliorate vagueness problems by using a mental-state requirement to limit prosecutions to situations in which the defendant knows that his acts are forbidden, they are free to do so. This is the path the Supreme Court took in Screws v. United States, 325 U.S. 91 (1945), and United States v. Kozminski, 487 U.S. 931 (1988), to save 18 U.S.C. sec.sec. 241 and 242. These companion statutes, the poster children for a vagueness campaign, make it a crime to deprive anyone of any constitutional right (or conspire to do so). These statutes are written entirely in terms that have been difficult to pin down, such as "under color of any law" and "rights, privileges, or immunities secured or protected by the Constitution . . . of the United States". Section 242 contains the word "willfully", and in Screws the Court read this to mean that a person commits the offense only if he knows that his acts deprive someone of a constitutional right. A person who actually recognizes his legal obligation can't complain that he lacked notice. Many cases since Screws use a knowledge requirement to prevent unfair surprise when the statute is vague or complex. E.g., Cheek v. United States, 498 U.S. 192 (1991) (tax laws); Staples v. United States, 511 U.S. 600 (1994) (gun-control laws).

Both Illinois and Wisconsin put mental-state elements in their statutes. Attorneys General of both states contend that under their laws a procedure may be deemed a "partial-birth abortion" only if at the outset of the procedure the physician intends to perform all of the steps that mark the D&X. We think that the Supreme Courts of both states could go even farther and read the statutes to require knowledge of the legal rules, and thus to follow the trail blazed in Screws.

The critical language in Illinois reads: "Any person who knowingly performs a partial-birth abortion and thereby kills a human fetus or infant is guilty of a Class 4 felony." 720 ILCS sec.513/10 (emphasis added). If the Supreme Court of Illinois reads "knowingly" in sec.513/10 the same way the Supreme Court of the United States read "willfully" in sec.242, then there is no vagueness problem with sec.513/10. And we can't see any impediment to giving "knowingly" such a reading. In context, the word sounds distinctly like a requirement that the physician know that the medical procedure being performed is a "partial-birth abortion" and not simply that the physician know that he is performing particular physical acts. The most natural reading is one that requires knowledge of the law's application to the medical procedure being performed. Federal courts have been creative with knowledge requirements as a means of saving statutes against constitutional infirmity. E.g., United States v. X-Citement Video, Inc., 513 U.S. 64 (1994). Section 513/10 readily supports a construction that eliminates all concern about vagueness, which makes it inappropriate for a federal court to enjoin its application. (In Morales, the Supreme Court's most recent vagueness opinion, the lead opinion stressed that the statute lacked any mental-state requirement.)

Wisconsin uses the word "intentionally" rather than the word "knowingly." Its statute has two intent elements: a partial-birth abortion is defined as one in which the physician acts with "the intent to kill the child", Wis. Stat. sec.940.16(1)(b), and the section defining the offense reads: "[W]hoever intentionally performs a partial-birth abortion is guilty of a Class A felony." Wis. Stat. sec.940.16(2). The reference to "intent" in sec.940.16(1)(b) must mean "that the actor either has a purpose to do the thing or cause the result specified, or is aware that his or her conduct is practically certain to cause that result." Wis. Stat. sec.939.23(4) (defining "with intent to"). The reference to "intentionally" in sec.940.16(2) has at least potentially a broader scope. Although sec.939.23(3) defines "intentionally" as "a purpose to do the thing or cause the result specified", a "partial-birth abortion" is neither a thing nor a result; it is a defined legal term. It is therefore a distinct possibility that the Supreme Court of Wisconsin will read this reference as equivalent to "wilfully" or "knowingly" notwithstanding the caution in sec.939.23(5) that criminal intent "does not require proof of knowledge of the . . . scope or meaning of the terms used in" a statute. The question here is not whether a physician understands the terms of sec.940.16(1)(b) in the abstract, but whether the physician intends that the plan of action add up to a "partial-birth abortion." In cases such as X-Citement Video, Cheek, and Staples, the Supreme Court of the United States was exceptionally creative with statutory allusions to mental states; plaintiffs have not established that the Supreme Court of Wisconsin would be unwilling to save its statutes the way the Supreme Court of the United States saved 18 U.S.C. sec.242.

3. Still a third interpretive approach is open to the state courts. They may elect to apply the statute to its central core of meaning, the D&X, while working out in common law fashion its outer boundaries. See United States v. Pitt-Des Moines, Inc., 168 F.3d 976, 986-90 (7th Cir. 1999). This is how the Supreme Court has chosen to proceed with respect to the Sherman Antitrust Act, 15 U.S.C. sec.sec. 1, 2, a statute much less precise than the partial-birth-abortion laws. Long before the development of a body of antitrust rules, or even of a clear distinction between the domains of the per se rule and the rule of reason, the Court rebuffed a vagueness challenge, remarking that "law is full of instances where a man's fate depends on his estimating rightly, that is, as the jury subsequently estimates it, some matter of degree." Nash, 229 U.S. at 377 (Holmes, J.). The Court took the same approach in Parker v. Levy for Article 133 of the Uniform Code of Military Justice, which makes it a crime for a commissioned officer to engage in "conduct unbecoming an officer and a gentleman." That statute, compared to which sec.513/5 and sec.940.16(1)(b) are paragons of specificity, was sustained against a vagueness challenge because the military courts had added details over many years, and the Supreme Court anticipated that they would continue to do so as new situations arose. Letter Carriers applies the same approach to vague provisions of the Hatch Act. Parker cannot be passed off as a sport of military law; it is just one illustration of the proposition that vagueness may be overcome incrementally, as well as by an all-at-once construction. Consider United States v. Powell, 423 U.S. 87 (1975), a civilian case dealing with 18 U.S.C. sec.1715, which makes it a crime to mail a firearm capable of being concealed on the person. A Derringer unquestionably qualifies, and although larger guns pose tough questions, the combination of a core meaning plus size as a metric for adjudication enabled the Court to hold that borderline cases could be decided later, as they arose.

Notice the difference between this approach and Salerno. In cases such as Nash, Parker, and Powell, the Court did not ask simply whether there was one constitutional application. It asked whether there was a core of meaning that people could grasp. In Nash the common law of restraint of trade supplied that core; in Parker the accumulated body of military precedents did so; in Powell the core comprised handguns. Occasionally the Court fortifies this approach by combining it with a scienter requirement, as United States v. United States Gypsum Co., 438 U.S. 422 (1978), did for the Sherman Act. Illinois and Wisconsin can follow the same path. For their statutes, the medical definition of the D&X procedure supplies the core of understandable application. Indeed, the district court in Wisconsin found as a matter of fact that physicians who are likely to perform an abortion already understand that this is the point of partial-birth-abortion statutes. 44 F. Supp. 2d at 978. That finding cannot be set aside as clearly erroneous.

Plaintiffs respond to this core-plus-common-law approach by contending that a statute is unconstitutionally vague until the interstices have been filled in. But then all of the cases we have cited were wrongly decided, for the interstices of antitrust law were not filled in by 1913, nor were all of the gun-mailing issues worked out before 1975: Powell involved a sawed- off shotgun, not a pistol. To say that the law cannot be applied to anyone, even a person who commits a violation within the core of the statute, until all issues about its reach have been resolved would prevent those issues from coming before the court to be resolved. How, precisely, are the interpretive wrinkles to be smoothed if the law can't be invoked? By advisory opinions? Anyway, there is always a periphery, no matter how many situations judges have evaluated. Resolving each new question moves the area of uncertainty but can't abolish it. "Whenever the law draws a line there will be cases very near each other on opposite sides. The precise course of the line may be uncertain, but no one can come near it without knowing that he does so, if he thinks, and if he does so it is familiar to the criminal law to make him take the risk." United States v. Wurzbach, 280 U.S. 396, 399 (1930) (Holmes, J.).

Protection for events outside the core is provided by the principle that surprising and unanticipated applications of a criminal statute violate the due process clause. See, e.g., Bouie v. Columbia, 378 U.S. 347, 351 (1964); Marks v. United States, 430 U.S. 188, 196 (1977); Lanier, 520 U.S. at 266. Given the assurances made in this litigation by the Attorneys General of Illinois and Wisconsin, application of these states' partial-birth-abortion statutes to a normal D&E or induction (including one that began as a D&E but turned out to have elements of a D&X because of inadvertence or developments during surgery) would be surprising and constitutionally impermissible. Plaintiffs therefore can practice medicine without undue concern on the legal front. The Constitution provides the assurance they crave--but without the need to prevent application of the state law to the D&X procedure itself. States have a powerful interest in working out the details of their criminal laws in their own courts. See Younger v. Harris, 401 U.S. 37 (1971); Nelson v. Murphy, 44 F.3d 497 (7th Cir. 1995). On remand the district judges should enter precautionary injunctions, limited to implementing the conclusion of this paragraph that the state laws may not be applied to a normal D&E or induction until after the state has provided additional specificity, by statutory amendment, regulations, or judicial interpretation (which could arise either from civil litigation, see Part IV below, or from criminal prosecutions in which the parties disagree about whether the medical procedure properly may be labeled a D&X). With that assurance in hand, plaintiffs would not face any substantial threat of prosecution.

A skeptic might respond: what's the basis for an injunction without a finding of actual constitutional violation? The basis is the risk that plaintiffs face. See Ex parte Young, 209 U.S. 123 (1908). Whenever suit is brought against a statute that has yet to be enforced, there is a chance that what the federal court does will be unnecessary or even advisory--for in the absence of relief the statute might never be enforced in a way that plaintiffs fear, or that would violate their rights. Everything depends on probabilities. Plaintiffs believe that the probability of improper prosecution is so high that the statutes must be enjoined root and branch. We think that the probability of improper prosecution is low, but non-zero; and a prosecution for performing a D&E not only raises vagueness concerns but also could unduly burden women's rights (see Part III below). That risk can be reduced to a constitutionally acceptable level with a simple injunction based on Bouie.

Persons who believe that the risk of improper prosecution is so high that all prosecution under the statute should be enjoined are poorly situated to contend that the risk also is so low that the plaintiffs lack standing under Los Angeles v. Lyons, 461 U.S. 95 (1983). When this case began the plaintiffs faced a real risk and therefore had standing; and no one thinks that defendants' representations about their projected course of enforcement have made the controversy moot. If as plaintiffs believe the Constitution requires an injunction against everything Illinois and Wisconsin have enacted, despite these assurances, then the Constitution can't simultaneously forbid an injunction against potential misuses of state power in the enforcement of these laws. Article III does not limit a federal court's choice to enjoining nothing, or enjoining everything. The path we choose allows the states to interpret their laws and supply more concrete rules, see Bellotti v. Baird, 428 U.S. 132, 146-47(1976); the path the plaintiffs (and our dissenting colleagues) prefer would foreclose now and forever all interpretation by the state judiciary. By filing suit against a novel and ambiguous law in federal court, litigants can't preclude interpretation by the state judiciary. Our precautionary approach preserves the state judiciary's role while protecting plaintiffs' (and their patients') legitimate interests in the interim.

4. In the Wisconsin case our panel expressed concern that even if the probability of unanticipated application is low, the penalty is so high (life imprisonment) that physicians would avoid performing the D&E procedure. 162 F.3d at 469. This cannot be said of Illinois, where the maximum penalty is three years' imprisonment. Although the threat of life in prison has a greater potential to induce caution than does the threat of three years in prison, the precautionary injunction discussed above should reduce the risk of improper or unanticipated application to a trivial level. The right constitutional objection to a Draconian penalty is not vagueness (a statute's scope is equally clear or uncertain whether the penalty is a $10 fine or 20 years in prison) but a contention that severe criminal punishment for a medical misunderstanding would inflict cruel and unusual punishment. Plaintiffs themselves do not make such an argument, however; they object to the substance of the state laws, not to the punishment available on conviction. The penalty is an afterthought, scarcely mentioned in their briefs. If Wisconsin should some day impose a lengthy prison term for a violation of sec.940.16(2), then it will be time to consider arguments based on the eighth amendment.

5. For the last two years, a natural experiment has been conducted in the United States. Thirty states enacted laws forbidding most partial-birth abortions. Judges prohibited the application of these laws in two-thirds of these states; in the other third the statutes have been in force. One way to perform a reality check on the district court's conclusion in the Wisconsin case that physicians know that these statutes cover only the D&X is to see what has happened in the states where the laws have been permitted to take effect. Vagueness could affect physicians in either or both of two ways. First, they might send women seeking second-trimester abortions to other states that do not regulate partial-birth abortions. The result would be a decrease in second-trimester abortions as a percentage of all abortions in the state. Second, physicians might select a different procedure-- principally substituting induction for a D&E--even though the replacement was medically inferior. This would produce a change in the ratio of inductions to D&E procedures in the affected state.

Indiana supplies a nice test, because its law took effect on July 1, 1997, so it is easy to compare the first six months of 1997 to the second six. Moreover, Indiana is surrounded by states that lack an effective prohibition on partial-birth abortions (because of the Illinois injunction and the sixth circuit's ruling in Women's Medical Professional Corp.), so it was relatively easy for physicians to send their patients elsewhere. Although the maximum penalty in Indiana is not steep (an unjustified partial- birth abortion there is a Class C felony, Ind. Code sec.16-34-2-7(a), for which the punishment is four years' imprisonment, Ind. Code sec.35-50- 2-6(a)), no state or federal court has construed the statutory term, so whatever uncertainty the raw text of the statute engenders is unabated.

During 1997 a total of 13,208 abortions were performed in Indiana. (These data come from tables prepared by Indiana's State Epidemiologist.) During the first six months, 74 late-second-trimester abortions were performed: 72 by D&E and 2 by intra-uterine prostaglandin injection. During the second six months (that is, after the partial-birth-abortion law took effect), 87 late-second-trimester abortions were performed, all by D&E. Thus physicians did not substitute induction for D&E in order to reduce risk to themselves. Nor did they avoid the D&E procedure by sending women to other states. Abortion by D&E during the first six months represented 1.03% of all abortions performed in Indiana (72 of 6,950). Abortion by D&E during the second six months represented 1.39% of all abortions in the state (87 of 6,258). These data are incompatible with plaintiffs' a priori belief that the partial-birth-abortion statutes will discourage the performance of the D&E procedure or cause the physician to substitute an inferior procedure. Perhaps things have turned out differently in other states, or perhaps physicians behaved differently in 1998 or 1999. We do not mean to preclude plaintiffs from submitting the data in Indiana and other states to rigorous analysis. A few procedures were classified in the tables as "unknown," which may mask some effects--though the number of "unknown" procedures did not change much from the first to the second half of 1997. Our point is only that a simple inquiry reinforces the findings of fact made in the Wisconsin case and implies that partial-birth-abortion statutes need not have the baleful effects that plaintiffs foresee.

III

Plaintiffs in both cases contend that, even if the statutes are precise enough to be enforced, they create an undue (and therefore unconstitutional) burden on abortion. The eighth circuit reached this conclusion in Carhart and its two companion cases, but only after first holding that the state laws effectively prohibit the D&E procedure. If we thought that the Illinois or Wisconsin laws forbade D&E, then Planned Parenthood of Central Missouri v. Danforth would require us to agree with the eighth circuit. For reasons we have already given, however, we believe that state courts are entitled to accept the view of both states' Attorneys General that their laws do not forbid, or even affect, the D&E procedure. The question we must address, then, is whether a statute limited to D&X unduly burdens abortion.

One reason why this might be so is that neither state's law contains an exception for situations in which the D&X procedure is necessary to protect the woman's health. A second argument is that any prohibition of any medical procedure unduly burdens a woman's right to abortion. The first line of argument finds some support in Casey, which held that any regulation of abortion must make an exception for procedures that protect "the woman's life or health", 505 U.S. at 846 (emphasis added), as well as our panel's decision in the Wisconsin case, 162 F.3d at 467-68, 470- 71. But we do not think that the plurality in Casey meant this as a universal rule, one applicable even when the procedure in question lacks demonstrable health benefits. The point that the plurality made was that a statute that lacks a "health exception" may unduly burden the woman's right to obtain an abortion before the fetus has reached viability; when state law offers many safe options to that end, the regulation of an additional option does not produce an undue burden.

Section 513/10 of the Illinois statute contains an exception for "a partial-birth abortion that is necessary to save the life of a mother because her life is endangered by a physical disorder, physical illness, or physical injury, including a life-endangering condition caused by or arising from the pregnancy itself, provided that no other medical procedure would suffice for that purpose." It does not contain a comparable exception for a partial-birth abortion that promotes the woman's health. Wisconsin's law has a similar structure. Both sets of plaintiffs contend that the lack of an exception for a woman's health dooms the statute. Coming from the Illinois plaintiffs, this is a weak argument indeed, for none of the plaintiff physicians or clinics in Illinois uses the D&X, and none asserts any desire to start. Whether the D&X procedure sometimes would protect a woman's health is therefore immaterial to these physicians and the women in their care. But several of the plaintiff physicians in the Wisconsin case do employ the D&X procedure on occasion, so we must consider the merits of the argument.

After a trial, the district court in the Wisconsin case concluded that the D&X procedure is never necessary from the perspective of the patient's health. 44 F. Supp. 2d at 979-82. This conclusion cannot be called clearly erroneous; as things transpired, it was not a seriously contested issue. None of the plaintiff physicians in that case testified to the contrary. The district court summed up:

None of the physicians would state unequivocally that the D&X procedure is safer than the D&E procedure. Broekhuizen conceded that further study of the procedures is required. Smith admitted that he had never encountered a situation where D&X would have been the best procedure to use. Haskell, who invented the procedure, admitted that the D&X procedure is never medically necessary to save the life or preserve the health of a woman. Giles agreed.

Id. at 980. The judge added that, although the D&X procedure has been used for more than a decade, no published study compares the risks of D&X to those of D&E. Id. at 979. Some authors believe that the D&X procedure is more hazardous. See Sprang & Neerhof, supra, 280 J. Am. Medical Ass'n at 746 ("Intact D&X (partial-birth abortion) should not be performed because it is needlessly risky, inhumane, and ethically unacceptable."); Nancy G. Romer, The Medical Facts of Partial- Birth Abortions, 3 Nexus 57 (1998).

Plaintiffs have not identified any data that undercut the district court's finding of fact. Their principal medical reference is the ACOG's 1997 statement and technical bulletin on the D&X procedure, which concluded that, although D&X is never the only medically appropriate option, choice still should be reserved to the physician. Although the statement asserts that D&X sometimes may be the best option, the ACOG did not identify any concrete circumstance under which this would be so or provide support for the assertion. Several of the plaintiff physicians testified to like effect, again without specifics or data. No published medical study fills the gap. Cf. Kumho Tire Co. v. Carmichael, 119 S. Ct. 1167, 1177-79 (1999). Unelaborated avowals do not demonstrate that a finding of fact reached after a trial is clearly erroneous. When there is a conflict in the testimony--and that is the description most favorable to the plaintiffs--the finder of fact is entitled to choose. See Anderson v. Bessemer City, 470 U.S. 564 (1985). We proceed on the assumption, which appears to be shared by the American Medical Association, that the D&X is not the best or safest option in any articulable category of situations. The ama's Policy H-5.982 concludes, among other things, that "there does not appear to be any identified situation in which intact D&X is the only appropriate procedure to induce abortion." The district court in Illinois did not make contrary findings, or indeed any finding; acting without a trial or even an evidentiary hearing, it could not properly have rendered "findings" about disputed issues. District courts in other circuits have disagreed with the conclusions the district judge reached in the Wisconsin case, but this does more to show that the issue is debatable among reasonable persons than to show that a particular view is in error.

The question in the end is not what one or another judge found on a given record; it is whether the state legislatures exceeded their constitutional powers. Factual premises underlying legislation normally are not subject to review by trial courts. FCC v. Beach Communications, Inc., 508 U.S. 307, 315 (1993); Minnesota v. Clover Leaf Creamery Co., 449 U.S. 456, 463-65, 101 S.Ct. 715, 66 L.Ed.2d 659 (1981); Vance v. Bradley, 440 U.S. 93, 111 (1979); National Paint & Coatings Ass'n v. Chicago, 45 F.3d 1124, 1127 (7th Cir. 1995). The district court's findings, the statements of medical groups, and the inferences to be drawn from what has happened in states that have implemented partial-birth-abortion statutes, are informative rather than conclusive on the question whether the underpinnings of the legislation are tenable. See Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L. Rev. 229 (1985). Although "undue burden" analysis does not map neatly to the standard classifications ("rational basis," "strict scrutiny," and the like), making it hard to know the extent to which courts must respect legislative decisions about contestable factual matters, it is significant that none of the Supreme Court's decisions in abortion cases suggests that the same law would be constitutional in one state, and unconstitutional in another, depending on a district judge's resolution of factual disputes. Thus we treat the findings in the Wisconsin case, like the view of the ama, not as replacements for legislative conclusions, but as establishing that there is real, and not just hypothetical, support for a belief that the partial-birth-abortion laws do not pose hazards to maternal health. (This is the approach to evaluation of facts when the constitutional standard prescribes an intermediate level of analysis. See Craig v. Boren, 429 U.S. 190, 199-204 (1976).)

Only if every regulation related to abortions must contain a case-by-case "health exception" is there a problem with these laws. Yet Casey did not say that health effects must be evaluated case by case, rather than procedure by procedure. Abortions are not havens for junk science--so that, for example, the state must let any chiropractor perform an abortion, if the chiropractor believes that manipulation of the spine is safest for the woman. Cases such as Mazurek v. Armstrong, 520 U.S. 968 (1997), which held that states may limit performance of abortions to physicians, permit laws that regulate by class of procedures (or of medical providers). Cf. United States v. Rutherford, 442 U.S. 544 (1979) (fda may ban a generally ineffective drug, even though for a few people it may be the only cure). Doubtless some nurses are better than some physicians at performing abortions, so patients' health would be improved by allowing these nurses to perform abortions; yet the Court held that states may provide otherwise. Similarly, Illinois and Wisconsin concluded that the D&X procedure is not essential to protect the health of any woman, given the availability of other procedures. Compilation of additional information by the medical profession could call this conclusion into question. As things stand, however, interference with the operation of the state laws cannot be justified on the ground that the D&X procedure is necessary to protect women's health.

A requirement of a case-by-case "health exception" to every statute concerning abortion would amount to a rule that no state may regulate any abortion procedure. For a physician will use a particular procedure only if in the physician's judgment the procedure is superior in some way-- faster, safer, more likely to work (and thus indirectly safer because an abortion can be achieved with one procedure rather than two, with cumulative risks), or perhaps less expensive (and thus again indirectly safer, because the lower cost brings it within the reach of additional patients). Can plaintiffs, who object to the vagueness of the states' laws, really welcome a ruling under which every abortion creates the possibility of a prosecution in which the state invites a jury to disagree with the physician's assessment of the procedure's safety? A health exception, where jurors rather than physicians assessed health, would be an order of magnitude worse than the ambiguity plaintiffs perceive in the partial-birth-abortion laws. Perhaps plaintiffs believe that a "health exception" should make the physician's assessment of health consequences incontestable. That would avoid vagueness at the price of vitiating the laws. Some physicians are bound to think that a given procedure is safer or otherwise superior, even if the medical profession as a whole disagrees. A constitutionally based health exception for every procedure, coupled with a prohibition against review of physicians' beliefs about which procedures are safest, would amount to a rule that anything goes. And plaintiffs make this argument explicitly: they contend that any regulation of any medical procedure that pertains to abortion is an "undue burden" on women's rights.

To the observation that D&X is not an important procedure--indeed, that it is such a marginal procedure that few physicians anywhere in the nation (and none in Illinois) express any interest in performing it--plaintiffs rejoin that the question is not whether a law creates an undue burden for women as a group, but only whether it does so for those women it affects. The affected set here is women for whom a physician will think that D&X is the procedure most likely to succeed, or to entail the least cost. For these women, plaintiffs contend, the burden is substantial and therefore "undue." Moreover, the argument continues, if the states are right--if D&X really is a procedure that the medical profession abjures--then the law must be political grandstanding. A restriction imposed for no better reason than to make a point has no benefits and therefore must be "undue," the argument concludes. On plaintiffs' approach, if the statutes actually accomplish something, then the burden is "undue" because the laws have effects; and if the laws accomplish nothing, then the burden is "undue" because any cost is excessive when the benefit is zero. Plaintiffs have the states both coming and going.

But this cannot be a sound understanding of Casey, or Mazurek would have come out the other way. By plaintiffs' lights, Montana's statute restricting the performance of abortions to licensed physicians would have had an "undue burden" on those women it affected--that is, the set of women for whom the extra cost of a physician's fees made obtaining an abortion impractical or burdensome. Yet the Court held that a physician-only rule is not an "undue burden," as the plurality in Casey used that phrase. What plaintiffs in our case really argue is not that limiting the D&X procedure to situations (if there are any) in which it is necessary to save the woman's life is an undue burden, when compared with other burdens (such as waiting periods, information requirements, parental approval or consultation requirements, and physician-only rules) that the Court has sustained; it is, rather, that all "burdens" are "undue" because women should have unfettered choice among medical procedures. But then the phrase "undue burden" is simply redundant; the word "undue" does no work.

It is always difficult for a court of appeals to predict how Justices of the Supreme Court will apply a phrase with as much plasticity as "undue burden." But our best estimate is that "undue" rather than "burden" is the key word, and that "undue" means not only "substantial" (a small cost or inconvenience is not "undue") but also that the burden must be undue in relation to the woman's interests, rather than undue in relation to the court's assessment of society's interests. Plaintiffs' (implicit) argument that every regulation of a medical procedure is "undue" fails because many regulations create small burdens. The physician-only rule was one; a prohibition of D&X is another. Even for the class of women who seek late-second-trimester abortions, there is always one or more other safe methods of abortion in addition to D&X. Let us return to the natural experiment now under way. Plaintiffs do not contend that in any of the states where a partial-birth-abortion law is in effect, even one woman has been injured or denied an abortion because of the law.

As for the second branch of plaintiffs' argument--that because the partial-birth-abortion laws do not diminish the number of abortions, they must be condemned as moralizing rather than utilitarian laws--this calls on judges to recapitulate Lochner's methodology. This line of argument amounts to the contention that Casey so restricts the states' power that states are limited to symbolic gestures; and then the states' very attempt to comply with Casey by refraining from substantially impeding women's access to abortion dooms their laws. We do not read Casey as condemning states to this self- defeating legislative menu. So long as the law does not harm women's legitimate interests, the fact that the law's effects are small and justified by moral rather than utilitarian considerations does not spell unconstitutionality.

One final issue requires attention. The district court in Illinois read 720 ILCS sec.513/15 to establish a parental-consent requirement, which the court enjoined because it lacks a judicial- bypass procedure. 995 F. Supp. at 861, relying on a portion of Casey, 505 U.S. at 899-900, that reaffirmed Hodgson v. Minnesota, 497 U.S. 417 (1990). But what sec.513/15 actually provides is different. It reads:

Nothing here requires parental consent to a partial-birth abortion. Section 513/10 forbids most partial-birth abortions; no one can authorize that procedure in Illinois unless it is necessary to protect the woman's life, and then no parent's consent is necessary.

Section 513/15 provides civil damages for conduct that violates sec.513/10, and it assigns the right to pursue these damages to parents if the person who undergoes the procedure is a minor. A parent who gave consent to the procedure naturally can't recover damages, even though a criminal prosecution remains possible under sec.513/10 if the mother's life was not in jeopardy. The possibility that a parent may consent to an unlawful procedure (and thus insulate the physician from civil damages), or consent to a life-saving procedure for which no consent was necessary in the first place, hardly implies that a judge also must be empowered to authorize a violation of sec.513/10.

Wisconsin's civil-liability provision, Wis. Stat. sec.895.038 (2), is entirely derivative from its criminal-liability provision, and the district court held it valid because the criminal provision is valid. Plaintiffs in the Wisconsin case do not take issue with this approach: they argue only that sec.895.038(2) is unconstitutional because sec.940.16 is unconstitutional. We therefore do not discuss sec.895.038(2) further.

Actually, we have said too much already about the civil-liability provisions. They are enforced in private litigation; the states' Attorneys General and local prosecutors have nothing to do with civil suits. Relief against the public officials therefore would be pointless even if the civil-liability provisions were problematic. See Summit Medical Associates, P.C. v. Pryor, 180 F.3d 1326 (11th Cir. 1999). There is no controversy between the parties to this case that can be resolved by a declaration concerning the civil-liability rules. But see Okpalobi v. Foster, 190 F.3d 337 (5th Cir. Sept. 17, 1999) (purporting to enjoin a private-civil- damages provision in a suit to which only state officers were defendants, without mentioning the eleventh circuit's conclusion in Summit Medical).

The judgments in both cases are vacated. The cases are remanded with instructions to enter the precautionary injunctions discussed in Part II.3 of this opinion, to dismiss the challenges to the civil-liability provisions for want of a case or controversy, and otherwise to enter judgment for the defendants.

Vacated and Remanded

Notes:

Circuit Judge Ripple did not participate in the consideration or decision of these cases.

Posner, Chief Judge, with whom Rovner, Diane P. Wood, and Evans, Circuit Judges, join, dissenting.

Compromise holds seductive allure for a court faced with a hot issue, and there is none hotter than the issue of abortion rights. The Illinois and Wisconsin statutes criminalizing "partial birth" abortion are challenged here both as being unconstitutionally vague and as unconstitutionally burdening the right of abortion. (These are independent grounds; if either is valid, the statutes are invalid.) The court rejects both challenges yet orders the district courts to enjoin enforcement of the statutes against any method of abortion other than the one the medical community refers to as "intact D & E" (that is, intact dilation and evacuation) or, more commonly, "D & X" (dilation and extraction). This is the form of "partial birth" abortion that gave rise to these statutes, although they are not limited to it.

The court's decision is not a real compromise. It leaves intact the core of the statutory prohibitions, which unlawfully burden the right of abortion by outlawing the D & X procedure. The court does toss a bone to the plaintiffs, but at the cost of expanding federal judicial power over the states by a method that the Supreme Court has never countenanced and that violates Article III of the Constitution. It is a bone, incidentally, that the plaintiffs didn't ask for; neither side, in either case, requested this novel form of relief or commented on it in their briefs. We are taking a leap into the unknown without any input from the parties.

The "precautionary" injunctions that the court is directing the district courts to enter will forbid the enforcement of the statutes outside their core prohibition, thus placing the federal contempt power behind this court's interpretation of state statutes. The court does this while accepting the enforcers' assurances that they will not enforce the statutes outside the core, and concludes, therefore, "that the probability of improper prosecution is low." The states are nevertheless to be enjoined because the probability, though low, is greater than zero. The probability of improper prosecution under every criminal statute ever written is greater than zero; and so the court's decision, should it be followed outside the abortion context--and nothing in the decision suggests a principled limitation to that context--implies a radical expansion of the power of the federal courts to superintend the enforcement of state statutes. State officials will be subject to federal contempt sanctions for failing to abide by a federal court's interpretation of the statutes that these officials, not federal judges, are charged with administering.

The court adds that while the probability of improper enforcement now is low, given the assurances that the state law enforcement authorities have made in their briefs in this court and at oral argument, the probability was high enough when these cases were brought to confront the plaintiffs with "a real risk" of being prosecuted for performing constitutionally privileged abortions. But if that is so, it means the statutes are unconstitutionally vague. If the law enforcers' present assurances are not enough to moot the case, how can they be enough to moot the central issue to which those assurances are addressed, the issue the court ducks of whether the statutes are in fact unconstitutionally vague? And if the present probability of improper enforcement is slight, as the court believes, what equity do the plaintiffs have to obtain an injunction?

The court is directing that state statutes be enjoined that it has not found either violate federal law or create a significant danger of such violation. This is unprecedented, and (on the court's view of the facts) violates Article III. If the court is right--I do not think it is, but it is the premise of the decision--that there is only a "non-zero" probability that either state statute will be enforced against any abortion procedure other than the D & X, the threatened injury is too slight to activate the curative powers of federal courts. A nonzero probability could be a probability of one in a thousand, or one in a million. A probabilistic injury can support standing, e.g., Clinton v. City of New York, 118 S. Ct. 2091, 2100-01 (1998); Walters v. Edgar, 163 F.3d 430, 434 (7th Cir. 1998); North Shore Gas Co. v. EPA, 930 F.2d 1239, 1242 (7th Cir. 1991), but the probability must be nontrivial, not merely nonzero. E.g., Murphy v. Hunt, 455 U.S. 478, 482-83 (1982) (per curiam); City of Los Angeles v. Lyons, 461 U.S. 95, 105-07 and n. 8 (1983); Walters v. Edgar, supra, 163 F.3d at 434-35. Otherwise the federal courts will be flooded with cases, since few statutes are altogether free of vagueness; statutes that are too precise allow loopholes.

Lyons, the choke-hold case, makes the basic point. The probability that the plaintiff would again be arrested and subjected to a choke hold was held to be too slight to support standing. But it was no less than the implicit probability that this court today assigns to the threat of prosecution for performing abortions that the state law enforcement authorities concede are constitutionally privileged.

Although the injunctions that the court is ordering ostensibly are temporary, to remain in force only until the state courts have a chance to clarify the statutes, those courts are unlikely ever to get that chance, since the enforcement of the statutes beyond their core prohibition, which in the court's view needs no clarification, will be enjoined. Private civil suits are unaffected, but the provisions for civil enforcement of these statutes are limited and cases are unlikely. The Illinois statute authorizes civil suits only by the parents of a minor who has a "partial birth" abortion, and the Wisconsin statute by either the minor's parent or the fetus's father. Should a prosecutor indict a physician for performing a D & X, and the physician contend that it was really a D & E, the dispute will be resolved by the district court's interpreting its injunction in contempt proceedings. As a practical matter, the courts of Illinois and Wisconsin will never have an opportunity to explore the outer bounds of these statutes.

The court cites no case that authorizes such injunctions. It relies on cases such as City of Chicago v. Morales, 119 S. Ct. 1849, 1861 (1999), and Village of Hoffman Estates v. The Flipside, Hoffman Estates, Inc., 455 U.S. 489, 494 n. 5 (1982), which hold that if a state statute has been interpreted by the state's highest court, the interpretation binds any federal court asked to decide the meaning of the statute. No case holds, however, that until the statute is construed by the state courts, a claim that it is unconstitutionally vague is premature. No case denies the authority of the federal courts to interpret state statutes, which is something we do all the time. No case holds that until a state statute is construed by the state courts its operation may be enjoined without a determination that the statute is unconstitutional. All the cases on which the majority relies involve federal courts either narrowing federal statutes that are fairly susceptible of being narrowed or accepting as authoritative a narrowing interpretation of a state statute by a state court. In no case cited by the majority did a federal court impose its own narrowing interpretation on a state statute by injunction, let alone a state statute the language of which defies a saving interpretation.

Although grounded in a concern with preserving the prerogative of state courts to interpret state statutes, today's decision curtails that prerogative by forbidding the states to enforce state statutes that the court has not found to be invalid. The decision is thus internally inconsistent (the court both rejects the charge that the statutes are unconstitutionally vague and, by enjoining their application outside their small clear core until they are clarified by the state courts, holds that they are too vague to provide fair warning--which means they are unconstitutionally vague, e.g., Bouie v. City of Columbia, 378 U.S. 347, 351 (1964)), as well as being an evasion of judicial duty, a violation of Article III, and an affront to federalism. The content of the injunctions that the court orders entered, moreover, is not specified; disputes over it will undoubtedly give rise to further litigation. And there will be a nice question of whether the plaintiffs can obtain attorneys' fees on the ground that they have won something, with the defendants arguing that it is they who have won the case because the court has both held the challenged statutes to be valid and accepted the defendants' assurances that the statutes will not be enforced beyond their valid scope.

On the merits, the most important issue raised by the appeals is that of undue burden. (I'll come back later to the substantive issue of vagueness, as distinct from the problems raised by the court's effort to evade the issue by means of the "precautionary" injunctions.) To understand this issue requires understanding the peculiar and questionable character of these statutes. They do not protect the lives of fetuses either directly or by seeking to persuade a woman to reconsider her decision to seek an abortion. For the statutes do not forbid the destruction of any class of fetuses, but merely criminalize a method of abortion--they thus have less to recommend them than the antiabortion statutes invalidated in Roe v. Wade, 410 U.S. 113 (1973). If any fetal lives are saved by these statutes, it will only be by scaring physicians away from performing any late-term abortions, an effect particularly likely in Wisconsin, whose statute imposes a punishment of life imprisonment for its violation. In tacit acknowledgment of that statute's especially questionable character, the majority opinion invites an Eighth Amendment challenge should a lengthy prison sentence ever be imposed on a violator of the statute.

The statutes do not seek to protect the lives or health of pregnant women, or of anybody else, as by confining the performance of abortions to licensed physicians, as in Mazurek v. Armstrong, 520 U.S. 968 (1997) (per curiam), or to facilities equipped to deal with emergencies that may arise in the course of an abortion, particularly a late-term one. A legislature can ban quacks from practicing medicine without making an exception for those quacks (and no doubt there are some) who are abler than the worst physicians. Any general health regulation is likely to hurt a few people. But as banning "partial birth" abortions is not intended to improve the health of women (or anyone, for that matter), it cannot be defended as a health regulation.

The statutes make no exception for cases in which pregnancy results from rape or incest, or in which the fetus is profoundly deformed, nonviable, or unlikely to live more than a few hours after birth, or indeed, as I shall show, in which the woman's life would be endangered if she carried the fetus to term. The absence of any such exceptions is particularly surprising because late-term abortions are much less likely than the much more common first-trimester abortions to be motivated by considerations merely of convenience rather than of urgency. A woman who finds herself five months pregnant is unlikely to decide to have an abortion unless advised by her physician either that the fetus is profoundly deformed or that her own health requires that the pregnancy be terminated. Indeed, if the fetus has become viable, she cannot lawfully obtain an abortion unless her life or health is in danger. Wis. Stat. sec. 940.15(3); 720 ILCS 510/5(1). These are constitutionally permissible protections of fetal life. Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833, 879 (1992). It is incomprehensible to me why these states, if acting in good faith, were unwilling to write the same health exception into statutes that criminalize the abortion of nonviable as well as viable fetuses. The "partial birth" statutes do not distinguish between fetuses that are viable, in the sense of being sufficiently developed to be able to survive outside the mother's body, and those that are not. Some fetuses don't become viable until 27 weeks or even later, but the "partial birth" abortion, even defined so narrowly as to cover only the D & X procedure, is performed from 20 weeks on. A physician might be convicted of a felony violation of these statutes even though the Supreme Court has held that a state may not prevent the abortion of a nonviable fetus. Id. at 846, 877.

The statutes do not outlaw a particularly cruel or painful or horrifying mode of abortion. This can be shown with the aid of a simple example. Suppose that the fetus is hydrocephalic, so that its head is too large to pass through the cervix. If the physician performing the abortion crushes the fetus's skull in the uterus, killing the fetus while the fetus is still entirely within the uterus, he is not guilty of violating either of the statutes before us. But if before crushing the fetus's skull the physician turns the fetus around so that its feet are protruding into the vagina, he has committed a felony. In both cases, the fetus is killed by the crushing of its head in the uterus. (The crushing is necessary to enable the fetus to be removed through the birth canal without making the woman go into labor.) From the standpoint of the fetus, and, I should think, of any rational person, it makes no difference whether, when the skull is crushed, the fetus is entirely within the uterus or its feet are outside the uterus. Yet the position of the feet is the only difference between committing a felony and performing an act that the states concede is constitutionally privileged.

The tortured efforts of the states' lawyers to provide guidance to physicians engaged in performing abortions--physicians who already are frequent subjects of picketing and other harassment and occasionally of physical assaults, see, e.g., United States v. Soderna, 82 F.3d 1370 (7th Cir. 1996), and who are now to be threatened with criminal prosecution and professional defrocking-- make one thing clear: it is extremely difficult, indeed probably impossible, to distinguish a "partial birth" abortion from the methods of abortion that are conceded to be privileged. (This is the key to the statutes' vagueness, and so a point to which I'll return.) The principal reason for the difficulty is that, as my example of the hydrocephalic fetus suggests, there is no meaningful difference between the forbidden and the privileged practice. No reason of policy or morality that would allow the one would forbid the other. We should consider therefore why any state would pass such a law. An important part of the answer is found in Judge Manion's opinion in Planned Parenthood of Wisconsin v. Doyle, 162 F.3d 463 (7th Cir. 1998). The states want to dramatize the ugliness of abortion. "[A]ll methods [of abortion] are gruesome. But this is the one method that has been at least partially exposed to the light of day. . . . The exposure of this procedure has cast a bright light on the alternative procedures that are equally gruesome. . . . No doubt when this statute is properly construed--[so] that it bans only the rarely-used D & X procedure--it does not appear to accomplish much in the way of saving babies. But these laws have certainly elevated the public's concern over the fate of the 'potential human lives' that are exposed to abortion." Id. at 477, 479 (dissenting opinion). From this standpoint, the more a "partial birth" abortion is like an abortion that is conceded to be constitutionally privileged, the better.

The wave of "partial birth" abortion statutes that broke over the nation after a description of the D & X procedure was publicized--see Martin Haskell, "Dilation and Extraction for Late Second Trimester Abortion" (1992), reprinted in 139 Cong. Rec. E1092, 1993 WL 135664 (Apr. 28, 1993), and in The Partial-Birth Abortion Ban Act of 1995, Hearing before the S. Comm. on the Judiciary, 104th Cong., 1st Sess. 5 (Nov. 17, 1995)--does not exhibit the legislative process at its best, whatever one thinks of abortion rights. Whipped up by activists who wanted to dramatize the ugliness of abortions and deter physicians from performing them, the public support for the laws was also based--as is implicit in Judge Manion's defense of the laws-- on sheer ignorance of the medical realities of late-term abortion. The uninformed thought the D & X procedure gratuitously cruel, akin to infanticide; they didn't realize that the only difference between it and the methods of late- term abortion that are conceded all round to be constitutionally privileged is which way the fetus's feet are pointing. Opposition to the bills that became these laws was at first muted not only by ignorance of the character of a late- term abortion but also by the fact that few women are likely to be affected by the laws. Circumstances conspired, as it were, to produce a set of laws that can fairly be described as irrational.

This is a harsh verdict, but doubt about its soundness is laid to rest by the absence of any exception for situations in which a "partial birth" abortion is necessary to protect the pregnant woman's health. The right of abortion is unduly burdened by any law that endangers the woman's health. Even a law that limits only the right to abort viable fetuses--a limitation not found in the statutes challenged in this case-- must, to pass constitutional muster, make an exception "for pregnancies which endanger the woman's life or health." Planned Parenthood of Southeastern Pennsylvania v. Casey, supra, 505 U.S. at 846 (emphasis added). Immediately the question arises why the Wisconsin and Illinois legislatures didn't try to come within reach of the cases that permit states to limit the right of abortion by making an exception for pregnancies that endanger the woman's health, the same exception they have written into their other abortion statutes. When pressed at argument, the lawyers for the two states could answer only that the exception is unnecessary. The lawyers were confident that such an abortion is never required to preserve a woman's health. They may be right, though I think not (more on this below). But if so, they are right only for today. Tomorrow, studies may show that, yes, there indeed are cases where a "partial birth" abortion is necessary to protect the mother's health, as many physicians believe. Tomorrow, then, these two statutes may be unconstitutional even by the lights of the majority opinion. Why would a state risk the early obsolescence of its statute by making it wholly dependent on ever-changing medical opinion, when to avoid this risk it need only have excepted those "partial birth" abortions, if any, that are necessary to protect the woman's health? If there are few such cases, the exception will rarely be invoked; if none, never.

The answer is that opponents of abortion do not think there should be an exception for abortions that endanger a woman's health. Life, yes, but not health. These statutes, remember, are not concerned with saving fetuses, with protecting fetuses from a particularly cruel death, with protecting the health of women, with protecting viable fetuses, or with increasing the Wisconsin population (as intimated, surely not seriously, by Wisconsin's counsel). They are concerned with making a statement in an ongoing war for public opinion, though an incidental effect may be to discourage some late-term abortions. The statement is that fetal life is more valuable than women's health.

I do not deny the right of legislatures to enact statutes that are mainly or for that matter entirely designed as a statement of the legislators' values. Milner v. Apfel, 148 F.3d 812, 814 (7th Cir. 1998). Nothing in the Constitution forbids legislation so designed. Many statutes are passed or, more commonly, retained merely for their symbolic or aspirational effect. But if a statute burdens constitutional rights and all that can be said on its behalf is that it is the vehicle that legislators have chosen for expressing their hostility to those rights, the burden is undue. The statutes before us endanger pregnant women-- and not only pregnant women who want to have an abortion. There is no exception for women whose physicians tell them you must have an abortion or die. It is true that if a "partial birth" abortion is necessary to save the woman's life, the statutes permit this. But if her life could be saved by another type of abortion, even one that threatened her health--that threatened to sterilize her or to paralyze her--then the physician would be committing a felony if he performed a "partial birth" abortion. I cannot believe that my colleagues would think the physician could be punished in such a case, and I wonder therefore why the "precautionary" injunctions do not extend to it. The statutes allow a "partial birth" abortion to be performed if "no other medical procedure would suffice for that purpose" (that is, saving the mother's life). Wis. Stat. sec. 940.16(3); 720 ILCS 513/10. Could not "would suffice" be interpreted to mean "would be as good as from the standpoint of the woman's health"? Such an interpretation, no more audacious than the interpretive leaps that underlie the order to enter "precautionary" injunctions, would seem compelled by the court's own reasoning, but goes unremarked. If the interpretation is rejected, it would imply that a D & X would be forbidden even if the only alternative were a hysterectomy. That can't be right.

The court points out that it is unclear how many women are endangered by these statutes. The forbidden procedure is performed only (or almost only) in late-term abortions, which fortunately are rare; and only in some unknown fraction of them is a "partial birth" abortion the medically preferable procedure. But government cannot without a reason forbid a person to obtain medical treatment and if challenged defend by pointing out that most people don't need that treatment. In the context of abortion this principle is implicit in the statement in Planned Parenthood of Southeastern Pennsylvania v. Casey, supra, 505 U.S. at 878, that "purpose or effect" (emphasis added)--and thus purpose alone--"to place a substantial obstacle in the path of a woman seeking an abortion before the fetus attains viability" condemns an antiabortion statute. E.g., Okpalobi v. Foster, 190 F.3d at 354; Jane L. v. Bangerter, 102 F.3d 1112, 1116- 18 (10th Cir. 1996). That is no constitutional novelty; statutes that burden other constitutional rights besides the right of abortion have been invalidated on the basis of improper purpose. E.g., Edwards v. Aguillard, 482 U.S. 578 (1987). The Casey opinion speaks of placing a substantial obstacle in the path of "a woman," not "many women." Imagine a married woman, pregnant, told by her physician that her life depends on her obtaining an abortion. He tells her it would be better from the standpoint of minimizing the risk to her of medical complications from the abortion for her to have a D & X. But, he adds, unfortunately the law prohibits the procedure. It does so not because the procedure kills the fetus, not because it risks worse complications for the woman than alternative procedures would do, not because it is a crueler or more painful or more disgusting method of terminating a pregnancy, but because the state wishes to make a statement of opposition to constitutional doctrine. A legislature may be taken to intend any consequences of its handiwork that are at once natural, highly probable, and wholly foreseeable (and foreseen). Here the intent is to block a woman from seeking an abortion when her doctor advises her that the best procedure for her is criminal.

The statutes' improper purpose is further shown by the terminology they employ. The Wisconsin statute defines a fetus from the moment of conception to the moment of birth as a "child" (thus absurdly implying that when we are born we are no longer children). The Illinois statute defines the fetus, also from the moment of conception, as an "infant." If these definitions are sound, all abortion is infanticide, which is not the Supreme Court's view. A state cannot be permitted to abrogate federal constitutional rights by definition. Line drawing is inescapable but the line between feticide and infanticide is birth. Once the baby emerges from the mother's body, no possible concern for the mother's life or health justifies killing the baby. But as long as the baby remains within the mother's body, it poses a potential threat to her life or health and this threat presents a compelling case (or so at least the Supreme Court believes) for a right of abortion. The Wisconsin statute, whose provision for life imprisonment already marks it as the crazier one, also allows either of the fetus's grandparents to sue in respect of a "partial birth" abortion. One can imagine a father who cares more for having a grandchild than he cares for the health of his daughter threatening to sue in order to block a "partial birth" abortion, while the mother (who may be divorced from the father), valuing the situation differently, sits on the sideline, helpless to protect her daughter.

In concluding that the challenged statutes do not pose any threat to women's health, the court relies heavily on tendentious factfindings by the district judge in the Wisconsin case. He fastened on the fact that a Dr. Harlan Giles believes that the D & X procedure is rarely, maybe never, necessary for the protection of the pregnant woman's health. The judge brushed aside the contrary testimony of reputable physicians--more reputable, perhaps, as we're about to see, than Dr. Giles. Giles's testimony is essentially the only basis of the judge's finding, and the judge's finding is the principal basis on which this court, disregarding the contrary findings in the Illinois case, dismisses concerns about the absence of a health exception. Yet apart from everything else, Judge Shabaz's opinion is internally contradictory. He acknowledged that the D & X procedure, being quicker and easier to perform on fetuses that are more than 18 weeks old, has advantages for the woman's health in reducing the amount of bleeding, the risk of uterine perforations, and the amount of time that she has to be under anesthesia. Planned Parenthood v. Doyle, 44 F. Supp. 2d 975, 979 (W.D. Wis. 1999). Yet without retracting these findings, he concluded that the procedure is never necessary to protect the woman's health.

I daresay reputable physicians can be found who believe that surgery is never necessary in a case of prostate cancer, or that all vitamins are worthless, or that no efforts should be made to resuscitate a person over 80 years old who experiences cardiac arrest, since he is quite likely to be "revived" to a vegetative state. Dr. Giles may not even be wholly reputable. Another district judge rejected his testimony in opposition to the D & X procedure because, among other reasons, Giles was "more focused on the political aspects of the abortion debate than on the medical questions essential to resolution of the issues presented in this action. And, Dr. Giles was evasive when responding to questions directed at crucial issues in the case. Hence the testimony of Dr. Giles was neither credible, reliable nor helpful." Richmond Medical Center for Women v. Gilmore, 55 F.Supp.2d 441, 450 (E.D. Va.1999) (footnote omitted). Another district judge criticized Giles--who appears to be a member of that much-criticized fraternity, the professional expert witness--for pronouncing on an abortion statute with which he was unfamiliar. Evans v. Kelley, 977 F. Supp. 1283, 1309-10 (E.D. Mich. 1997). Another rejected his testimony that the D & X is "a resurrection of an obstetric method discarded in the 1960s, which was used to deliver dead fetuses, and known as craniotomy," and concluded that Giles's criticisms of the D & X procedure were "not persuasive." Women's Medical Professional Corp. v. Voinovich, 911 F. Supp. 1051, 1070 (S.D. Ohio 1995), aff'd, 130 F.3d 187 (6th Cir. 1997).

100

Judge Shabaz, his only support the dubious Dr. Giles, stands alone among district judges who have made factfindings concerning the potential benefits of the D & X procedure for women's health. See Rhode Island Medical Society v. Whitehouse, No. C.A. 97-416L, 1999 WL 683846, at *22 (D. R.I. Aug. 30, 1999) (contrary testimony is "unbelievable"); Richmond Medical Center for Women v. Gilmore, supra, at 451; Carhart v. Stenberg, 11 F. Supp. 2d 1099, 1124-26 (D. Neb. 1998), aff'd, 1999 WL 753919 (8th Cir. Sept. 24, 1999); A Choice For Women v. Butterworth, 54 F. Supp. 2d 1148, 1153, 1156-57 (S.D. Fla. 1998); Hope Clinic v. Ryan, 995 F. Supp. 847, 852 (N.D. Ill. 1998) (the decision reversed today without regard for the district judge's findings); Evans v. Kelley, supra, 977 F. Supp. at 1296; Women's Medical Professional Corp. v. Voinovich, supra, 911 F. Supp. at 1070; see also Causeway Medical Suite v. Foster, 43 F. Supp. 2d 604, 608, 613-14 (E.D. La. 1999); Planned Parenthood of Central New Jersey v. Verniero, 41 F. Supp. 2d 478, 484- 85 (D.N.J. 1998). If this court is to give weight to district judges' assessment of the credibility of the physicians who have testified pro or con the D & X procedure, it ought at least to acknowledge that Judge Shabaz is in a minority of one.

101

In refusing to give any weight to the findings of fact made by Judge Kocoras in the Hope case, the court seems to have forgotten that we are reviewing his judgment too. It is not some unrelated case that we are free to ignore on the ground that it is based on a different record from any before us. His findings, made in a proceeding that resulted in the issuance of a permanent injunction against the Illinois statute, are entitled to as much weight as Judge Shabaz's, Knapp v. Northwestern University, 101 F.3d 473, 478 (7th Cir. 1996); Durasys, Inc. v. Leyba, 992 F.2d 1465, 1471 (7th Cir. 1993)--more, really, because they are not infected by the internal contradiction that weakens Judge Shabaz's findings. If those findings justify our upholding the Wisconsin statute, Judge Kocoras's should lead us to invalidate the Illinois statute.

102

The court denies that Judge Kocoras made findings. I don't understand this. It is true that he based his findings on affidavits rather than on live testimony, but the state did not challenge the facts contained in the affidavits, including the facts about the medical advantages of the D & X procedure. Findings do not lose weight by being based on uncontested facts; admissions are not weaker evidence than facts found after vigorous contest. And if it were true that there were no findings in the Illinois case, this would call for a remand; it would not entitle us to plug Judge Shabaz's findings into a separate litigation with no common parties. (What kind of transplant is that?)

103

Consistent deference to district court factfindings in this pair of cases would lead to an inconsistent result--the upholding of one statute and the condemnation of its sister. This demonstrates that the constitutional right of abortion cannot be made to depend on whether a particular district judge finds a particular physician who disagrees with the consensus of medical opinion to be more credible than the spokesmen for the consensus. The consensus here is found in the statement by the American College of Obstetricians and Gynecologists (ACOG) that the D & X procedure "may be the best or most appropriate procedure in a particular circumstance to save the life or preserve the health of a woman, and only the doctor, in consultation with the patient, based upon the woman's particular circumstances can make this decision." As Judge Shabaz himself acknowledged, the procedure reduces the risk of lacerating the uterus and the cervix with the sharp instruments used in the standard D & E. Such lacerations cause trauma and loss of blood. The D & X also reduces the risk that fetal tissue will remain in the woman's body, where it can cause infection. It reduces the length of time that the woman is under anesthesia. And it facilitates testing for genetic abnormalities, which is particularly important to the woman's decision whether to become pregnant again. (So much for the suggestion of Wisconsin's lawyer that banning the D & X will lead to more births.) Not all physicians agree with these points; and ACOG hedged its opinion by noting that a select panel of physicians had been unable to identify specific circumstances in which there is no satisfactory alternative to a D & X. But the team of Dr. Giles and Judge Shabaz are, with all due respect, not competent to resolve this medical controversy.

104

How often the circumstances sketched in the previous paragraph that led ACOG to recommend retention of the D & X as a treatment option in late-term abortions will be present is not critical. It is slight consolation to be told that while the state has forbidden the optimal treatment of your medical problem, that problem happily is rare. Although we haven't been told of cases in states that have enacted laws forbidding "partial birth" abortion that were not promptly enjoined in which the consequence has been to endanger women, for all we know women in those states go out of state if they are candidates for a D & X. Cf. Jane L. v. Bangerter, supra, 102 F.3d at 1117. That would be the prudent course. My colleagues in the majority perform their own statistical analysis, outside the record, in an effort to refute this possibility. I have no objection to a court's relying on extra-record evidence to determine the health effects of "partial birth" abortion. Those effects should indeed be treated as a legislative fact rather than an adjudicative fact, in order to avoid inconsistent results arising from the reactions of different district judges, sheltered by the deferential "clear error" standard of appellate review of factfindings, to different records--the inconsistency illustrated by the different findings of Judges Shabaz and Kocoras in the two cases before us. Even so, we should hesitate to play statistician. It is incongruous for this court to brush aside the findings of district judges in other cases while bolstering Judge Shabaz's inadequate findings with extra-record evidence of its own.

105

The court tries to shore up Judge Shabaz's findings with two medical papers. One of these, however, Nancy G. Romer, "The Medical Facts of Partial-Birth Abortion," 3 Nexus: A Journal of Opinion 57 (1998), is not a medical paper at all. Although the author is an M.D., she is also a pro-life activist, and the burden of her paper is that viable fetuses should be saved rather than aborted. This is a respectable ethical stance but it has nothing to do with the medical pros and cons of a D & X. The other paper, M. LeRoy Sprang & Mark G. Neerhoff, "Rationale for Banning Abortions Late in Pregnancy," 280 JAMA (Journal of the American Medical Association) 744 (1998), appears in a section of JAMA captioned "controversies," and though it does point out possible dangers to the woman from the procedure, the authors are heavy on the ethical issues involved in abortion. Like Dr. Romer, they come out in opposition to late-term abortion by any method, and opine very unscientifically that "partial birth" abortion is "closer to infanticide than it is to abortion." Id. at 746. Another medical article in the same issue of JAMA concludes that the D & X procedure should, as ACOG concluded, be an option for the physician. See David A. Grimes, "The Continuing Need for Late Abortions," 280 JAMA (Journal of the American Medical Association) 747 (1998). The court does not cite that article.

106

No hypothesis has been proposed by either of the states, or by the amici curiae who have submitted briefs in support of the Wisconsin statute, as to why physicians would perform, or the American College of Obstetricians and Gynecologists support, a method of abortion that benefits no patient. There has been no suggestion of any pecuniary incentive that might cause them to engage in a controversial procedure of this kind.

107

Apart from performing its own statistical analysis and bringing to bear so-called medical papers, the court evinces its queasiness about resting entirely on Judge Shabaz's findings by saying that "the question in the end is not what one or another judge found on a given record" (the question the court had spent so much time discussing). "It is enough that there is real, and not just hypothetical, support for a belief that the partial-birth-abortion laws do not pose hazards for maternal health." This sounds suspiciously like the test for whether a statute has a rational basis, see, e.g., FCC v. Beach Communications, Inc., 508 U.S. 307, 315 (1993), and so may explain the court's one-sided treatment of the facts (the failure to cite Dr. Grimes's paper, for example). Rational basis is not the issue here. The issue is whether the Wisconsin and Illinois "partial birth" statutes impose an undue burden, and specifically whether they are a threat to maternal health. That is rightly an issue of legislative fact, meaning that its resolution is not to be cabined by facts determined in an adjudicative hearing; still it must be resolved in accordance with the weight of the evidence, including such extra-record evidence as the consensus of the relevant expert community, in this case the medical community. It is not enough to note that there is some evidence in support of one side of the issue, when there is more evidence on the other side. In the case the court cites as illustrating the appropriate scope of review, Craig v. Boren, 429 U.S. 190, 199-204 (1976), the Supreme Court critically reviewed statistical evidence underlying the claimed constitutionality of the challenged statute--and found it wanting. When one critically reviews the testimony of Dr. Giles, and the court's own statistical analysis, and the so-called medical papers, and weighs them against the evidence that establishes the medical consensus, it becomes apparent that the court does not take the standard of Craig seriously.

108

Let me suggest an analogy. Suppose that a legislature in the South in the 1950s had listened to several "reputable" cognitive psychologists testify that blacks as a group are slow learners, and hence that they should be educated in separate schools. No one would argue that this legislative finding, because supported by some evidence from the relevant expert community, would have to be accepted in deciding the constitutionality of segregated education. The court, if it thought the issue relevant, would have to decide whether the finding was correct before relying on it. Similarly, the Wisconsin and Illinois legislatures are not entitled to ban an abortion procedure that the medical community believes may be preferable from a medical standpoint for some women, simply because a marginally reputable expert (Dr. Giles) thinks that the set of women for whom the procedure is preferable is actually zero. That might bear on the legislatures' good faith (though there is contrary evidence on that too, as I have pointed out), but it would not decide the issue of the statutes' effects on the right of abortion.

109

Enough said: banning the D & X, the core prohibition in these statutes, imposes an undue burden on women who seek to have an abortion, and is therefore unconstitutional. Let me turn now to the alternative ground of vagueness. Both statutes forbid the partial vaginal delivery of a living fetus, and this describes many instances of the D & E method of abortion--which the court acknowledges cannot constitutionally be banned, see also Carhart v. Stenberg, 192 F.3d 1142 (8th Cir. Sept. 24, 1999); Little Rock Family Planning Services, P.A. v. Jegley, 192 F.3d 794 (8th Cir. Sept. 24, 1999); Planned Parenthood of Greater Iowa, Inc. v. Miller, 195 F.3d 386 (8th Cir. Sept. 24, 1999)--as well as it does the D & X. Judge Shabaz himself acknowledged that "often" in the course of a D & E there is a fetal heartbeat after portions of the fetus have been delivered into the vagina, 44 F. Supp. 2d at 977, making the D & E definitionally a "partial birth" abortion under both statutes. The court's response, as I noted at the outset of this dissent, is not to hold that the statutes are unconstitutionally vague, but instead to hold that they can't be enforced against any procedure except the D & X until the Wisconsin and Illinois state courts have had a chance (the chance that is unlikely ever to come) to "interpret" their statutes to permit the standard D & E method even when it results, as it frequently does, in the partial vaginal delivery of a living fetus.

110

Putting to one side my objections to the court's procedure, which I discussed at the outset of this dissent, I note that the court seems to have bought into the legal-realist view that a statute means whatever the courts say it means, despite the Supreme Court's admonition that a court "may impose a limiting construction on a statute only if it is 'readily susceptible' to such a construction." Reno v. American Civil Liberties Union, 521 U.S. 844, 884 (1997), and cases cited there. If the realists are right, and the contrary intimations of cases such as Reno v. American Civil Liberties Union can be brushed aside, I wish the court would, as I suggested earlier, hold that these statutes can be further narrowed by "precautionary" injunction to make an exception for cases in which the D & X procedure is necessary to protect the woman's health. The interpolation of such an exception would be no greater act of judicial hubris than narrowing the statute to the D & X when the draftsmen of the statutes decided not to use that term, preferring a vaguer term intended to be broader.

111

The constitutional law of vagueness is obviously in turmoil and so I shall suggest a lowest- common-denominator approach with which I think few judges will disagree, however much they may disagree about its application to particular cases. If a criminal statute is so worded that it is highly likely to condemn a constitutionally privileged act, and if the constitutionally privileged actor could not as a practical matter avoid being punished for engaging in such an act by raising a defense in a criminal proceeding, then the statute is enjoinable, in advance of any prosecutions, as unconstitutionally vague. For the very existence of such a statute "deter[s] constitutionally protected activity." Coe v. County of Cook, 162 F.3d 491, 496 (7th Cir. 1998), and cases cited there. When vague statutes prescribe heavy penalties for their violation, rational people avoid any conduct that might be thought to fall within the statute's scope, even if in an error-free litigation the conduct would be sure to be found constitutionally protected. See, e.g., Village of Hoffman Estates v. The Flipside, Hoffman Estates, Inc., supra, 455 U.S. at 494 n. 6; Colautti v. Franklin, 439 U.S. 379 (1979); Charles v. Daley, 749 F.2d 452, 460 (7th Cir. 1984).

112

When the demanding conditions that I have outlined are satisfied, the statute is invalid, and the legislature has to go back to the drawing board. The court cannot rewrite the statute, or tell the state courts how to rewrite it, and, in advance of any further interpretation, uphold the rewritten statute. That procedure, the procedure followed by the court today, in effect nullifies the doctrine of unconstitutional vagueness.

113

These statutes are either absurdly (and unlawfully) overbroad or remarkably vague, and the court's direction to issue "precautionary" injunctions is a tacit acknowledgement of its inability to give the statutes a plausible narrowing interpretation; the confused concessions of the states' lawyers at the oral argument were no help in the endeavor. Although the statutes are addressed to physicians--it is their conduct alone that is regulated--the language employed is not medical language. Instead of denoting the victim of illegal abortion as a live fetus, a fetus with a heartbeat, or a viable fetus (a fetus, remember, that can survive outside the mother's body), the Illinois statutes refers to "a living human fetus or infant," terms that the statute explains are "used interchangeably to refer to the biological offspring of human parents," 720 ILCS 513/5, while the Wisconsin statute refers to "a living child" or "partially delivered child," with "child" defined, as I noted earlier, as "a human being from the time of fertilization until it is completely delivered from a pregnant woman." Wis. Stat. sec.sec. 940.16(1)(a), (b). The use of such language makes the statutes incurably ambiguous. Neither "biological offspring" nor "living" is defined. At argument the Solicitor General of Illinois was emphatic that "biological offspring" is a synonym for "intact fetus," that is, a fetus (however deformed) before the physician performing the abortion begins to work on it. But he retreated from this position when faced with the hypothetical case of a physician who, intending to perform a D & X, removes a limb or other part of the fetus before beginning the procedure. His retraction was consistent with his statement that the Illinois legislature had tried to make the statute "as broad as possible." More extensive dismemberment than merely removing a limb or two, Wisconsin's lawyer assured us, would inevitably kill the fetus before any part of it emerged from the uterus. She offered no basis for this hopeful bit of medical advice, and it is wrong. Carhart v. Stenberg, supra, at *4; Richmond Medical Center for Women v. Gilmore, supra, at *10. By "killing" the fetus, she evidently meant doing something to cause its heart to stop beating, although a fetus is living before it develops a heart. Obviously a one-day old embryo, like the cells that compose a living human body, is alive, not "dead."

114

In a standard D & E, part or all of the fetus often will still have a heartbeat, and so be "living," in the sense of the word apparently intended by the legislatures, when it emerges from the uterus. Therefore it is "partially vaginally delivered," just as in a D & X. "During the D & E procedure, the fetus may be removed from the uterus and brought through the cervix and vagina either intact or disarticulated. . . . [I]t may happen that part of the intact fetus will be in the vagina and part in the uterus or a disarticulated part of the fetus will be in the vagina while the remainder of the fetus is in the uterus. In either of these situations, that part of the fetus which remains in the uterus may still have a heartbeat." Planned Parenthood of Central New Jersey v. Verniero, supra, 41 F. Supp. 2d at 484. If the physician then kills the fetus, he would, one would think, be guilty of having violated these statutes, which punish the killing of a fetus that has been partially vaginally delivered.

115

Against this interpretation, the Illinois Solicitor General, discarding settled principles of criminal law for the sake of a short-term victory, declared that the statute would not be violated unless the physician had planned before beginning the abortion to wait until a part of the fetus had emerged from the uterus and kill the fetus then; the mere fact that the physician knew it was likely that a part of the fetus would be outside the uterus at the moment of death would not show the requisite intent. By this logic--which the state's Solicitor General actually embraced at argument--if you fire a machine gun into a lighted college dormitory at night, reckoning that you have only a 10 percent chance of actually killing anyone, and you do kill one or more of the residents, you are not guilty of murder, provided you didn't want to kill anyone but just wanted to see whether your machine gun was in working order. The law is otherwise. First-degree murder in Illinois requires only that "in performing the acts which cause the death" the defendant "knows that such acts create a strong probability of death or great bodily harm." 720 ILCS 5/9-1; People v. Daniels, 702 N.E.2d 324, 330-31 (Ill. App. 1998); People v. Mitchell, 601 N.E.2d 916, 919 (Ill. App. 1992); Hennon v. Cooper, 109 F.3d 330, 333- 334 (7th Cir. 1997). A physician who performs a standard D & E knowing, as he must, before he begins, that part of the fetus is quite likely to be outside the uterus when he kills the fetus, and then kills it after a part has emerged from the uterus, is guilty of the crime of partially vaginally delivering a living fetus and then killing it.

116

The Illinois statute itself, moreover, requires only that the partial-birth abortion have been performed "knowingly." 720 ILCS 513/10. Under the law of Illinois, a defendant is deemed to know "the result of his conduct, described by the statute defining the offense, when he is consciously aware that such result is practically certain to be caused by his conduct." 720 ILCS 5/4-5(b).

117

Wisconsin authorizes life imprisonment for performing a "partial birth" abortion. That is the state's heaviest penalty, the penalty for murder. Wisconsin deems its statute--as both the penalty and the use of the term "child" to denote the fetus reflect--a law against infanticide. (So does Illinois, by using the term "infant," but it is not serious, because the maximum punishment for killing the "infant" is only three years.) No one would suppose, in the example of shooting into a dormitory, that if only a one-day-old baby were killed the killer would be innocent of murder. If Wisconsin is serious in regarding a fetus that has begun to emerge from the uterus as "a living child" who should be protected to the full extent of the law against murder, it cannot rationally interpret the statute to permit the physician to escape punishment by the plea that when he began the abortion he thought there was only a 50 percent chance that he would be killing "a living child." It would be like arguing that if a nurse hit a newborn over the head with a baseball bat, with all her might, not to kill it but merely to make it stop crying, and the newborn died, the nurse could defend against a charge of first-degree murder by proving that while she thought it quite possible that the newborn would die, she didn't want it to die and thought it had a 50 percent chance of surviving.

118

Any doubts on this score are stilled by Wisconsin's statutory definition of intentional. The statute punishes "intentionally" performing a partial birth abortion, Wis. Stat. sec. 940.16(2), and the term means either acting purposely to accomplish the forbidden end or being "aware that [one's] conduct is practically certain to cause that result." Wis. Stat. sec. 939.23(3). As glossed by the case law, the alternative definition is satisfied by conduct carrying a known high risk of the forbidden result even if that result is not desired. See State v. Gould, 202 N.W.2d 903, 906 (Wis. 1973); State v. McCarter, 153 N.W.2d 527, 529 (Wis. 1967); State v. Webster, 538 N.W.2d 810 (Wis. App. 1995). In Webster, where the defendant was convicted of attempted first-degree murder, it was deemed irrelevant that the jury might have accepted his testimony "that he precisely aimed the shotgun at Hood's armpit, and that the gun was loaded only with shotshells filled with bird shot." 538 N.W.2d at 815.

119

So the state-of-mind requirements of the two statutes turn out to be the same. They are, moreover, the standard state of mind requirements for liability for serious crimes. A criminal defendant acts intentionally, and, even more clearly, knowingly, "when he knows that that result [here, a "partial birth" abortion] is practically certain to follow from his conduct, whatever his desire may be as to that result." 1 Wayne R. LaFave & Austin W. Scott, Jr., Substantive Criminal Law sec. 3.5, p. 304 (1986); see, e.g., United States v. United States Gypsum Co., 438 U.S. 422, 444-46 (1978); Smith v. Farley, 59 F.3d 659, 663 (7th Cir. 1995); United States v. McAnally, 666 F.2d 1116, 1119 (7th Cir. 1981).

120

I have discussed the state of mind requirements of these statutes at such length in order to make clear how little mileage the court can get from the cases that it cites, such as Screws v. United States, 325 U.S. 91 (1945), that save vague statutes by narrow interpretations. Put aside the fact that these were federal statutes given narrowing interpretations by federal courts, rather than state statutes given narrowing interpretations by federal courts, and that the Supreme Court in Bellotti v. Baird, 428 U.S. 132, 146-47 (1976), a decision that is still good law on this point, held that the proper way to obtain a narrowing interpretation of a vague state abortion statute is to abstain in favor of the state courts. The point I wish to stress is that the device used in Screws and many similar cases to save the statute, that of interpolating a scienter requirement, is unavailable here. And this for two reasons. The first is that the Wisconsin and Illinois statutes make clear what the state of mind requirement for guilt under each statute is, leaving no room for interpolating such a requirement in the guise of interpretation; and the second is that the problem with knowing what these statutes require of a physician is not ignorance of legal obligation, but ignorance of the factual premises of liability. The statute challenged in Screws made it a crime to deprive a person of a constitutional right. The Court held that the statute could be saved by requiring that the defendant be shown to have known that he was violating a constitutional right. It was a realistic requirement because the set of constitutional rights at any moment is more or less known and it is a straightforward (though not always easily answerable) factual question whether the defendant knew that his conduct infringed something in that set. No one knows what procedures constitute "partial birth" abortion, and specifically whether many, perhaps most, D & E's do, given that the fetus often has a heartbeat after part of the fetus has emerged from the uterus. Screws, when he beat his victim to death, knew that he was committing the assault; the question was whether he knew that the assault (because he was a sheriff and his victim an arrestee) violated the Constitution. Physicians performing late-term abortions in Illinois and Wisconsin do not know whether they are performing "partial birth" abortions and thus committing crimes.

121

The incurable vagueness of these statutes might seem to be of little moment if a physician prosecuted for infanticide or feticide had a sure defense should the prosecution stray beyond the narrow path to which the court today attempts to confine the statute. He could not avoid prosecution but at least he would know he would be acquitted, and if he were lucky he would be let out on bail pending trial and maybe Planned Parenthood would pay his legal fees. But let us be realistic, and not only about the possibility of legal error. The physician is not the real target of the statute; the pregnant woman is. It is not the physician's pregnancy that is to be terminated. He has no incentive to undergo the agony of a criminal prosecution merely in order to perform an abortion in a particular way. Better for him either to abandon late-term abortions altogether, or to be sure to give the fetus a lethal injection in utero, and wait for it to take effect, before performing the abortion. The in terrorem effect of these statutes, especially the Wisconsin statute with its maximum penalty of life imprisonment, is likely to induce physicians in these states to steer well clear of the forbidden zone. What physician would be fool enough, or hero enough, to risk a criminal prosecution in order to explore the precise meaning and outer bounds of the precautionary injunctions that this court has ordered the district courts in these two cases to enter, even if the risk is small? And now we can see more clearly one of the fallacies in the "precautionary" injunction gimmick. It is the existence of these statutes, because of their extraordinary breadth and the limited incentive of a physician to test their scope, not the risk of prosecution, which may indeed be slight, that burdens those abortion rights that are conceded to be constitutionally protected. A peculiarity of abortion rights is that they require the assistance of a third party, the physician, who has a lesser interest in them than his patient. Because his interest is less, he is more easily deterred, and this should increase our concern with the deterrent effects of these vague statutes.

122

It remains only to make clear that in finding these statutes pernicious and unconstitutional I do not mean to criticize anyone who believes, whether because of religious conviction, nonsectarian moral conviction, or simply a prudential belief that upholding the sacredness of human life whatever the circumstances is necessary to prevent us from sliding into barbarism, that abortion is always wrong and perhaps particularly so in late pregnancy, since all methods of late-term abortion are gruesome. If a woman told by her physician that her fetus will be a Down's baby or a Tay-Sachs baby or will be born without arms or legs, or that her own health or even life will be endangered if she carries the fetus to term, decides nevertheless against abortion, I would be the last to criticize her decision. I might consider her a heroine or a saint. But what is at stake in these cases is whether the people who feel that way are entitled to coerce a woman who feels differently to behave as they would in her situation. The Constitution as interpreted by the Supreme Court in decisions that we are not free to palter with answers this question "no." We should therefore affirm the district court's decision in Hope and reverse the district court's decision in Doyle.