POSNER, Circuit Judge.

Before us are appeals by two Illinois prosecutors, Wharrie and Kelley, who claim absolute immunity from being sued by Nathson Fields under 42 U.S.C. § 1983. Appeals from denial of immunity, though interlocutory because the case against them remains pending in the district court, are immediately appealable provided that, as in this case, the claim (in this case claims) of immunity depends on an issue of law rather than one of fact. Mitchell v. Forsyth, 472 U.S. 511, 527-30, 105 S.Ct. 2806, 86 L.Ed.2d 411 (1985).

Fields' suit charges the defendants with depriving him of liberty in violation of the Fourteenth Amendment's due process clause and committing torts of malicious prosecution, intentional infliction of emotional distress, and conspiracy, in violation of Illinois law. Specifically he accuses the defendants of having coerced witnesses to give testimony that the defendants (as well as the witnesses) knew to be false, resulting in Fields' conviction of two murders and his imprisonment for 17 years until he was acquitted in a retrial; he later received a certificate of innocence from the court in which he had been tried. 735 ILCS 5/2-702.

This is the defendants' second round of appeals. Our decision in round one sets forth the factual details relating to Fields' claims, 672 F.3d 505, 508-09 (7th Cir.2012); we needn't repeat them. Fields accuses Wharrie of two separate acts (one in 1985, the other in 1998) of coercing false testimony from witnesses, and Kelley of similar coercion in 1998. We should clarify the terminology used by Fields to describe his claims. He uses "coerced," "fabricated," and "false" testimony interchangeably to mean testimony procured by a prosecutor who knows it's false. The use of the three terms to denote the same conduct is confusing. For they mean three different things. Coerced testimony is testimony that a witness is forced by improper means to give; the testimony may be true or false. Fabricated testimony is testimony that is made up; it is invariably false. False testimony is the equivalent; it is testimony known to be untrue by the witness and by whoever cajoled or coerced the witness to give it. Much testimony is inaccurate, but not deliberately so and therefore not false or fabricated as we are using these words.

Originally the district court had dismissed, on the ground of absolute prosecutorial immunity, only the federal claim against Wharrie that was based on his alleged misconduct in 1985. Fields did not appeal, but the prosecutors did, challenging the district court's refusal to dismiss the other claims. Our decision in that first appellate round ordered the dismissal of the remaining federal claims against both defendants on grounds of absolute prosecutorial immunity. 672 F.3d at 519. The state law claims, also a subject of the prosecutors' appeal, remained in the case. But we didn't discuss them, because we expected that with the federal claims dismissed the district judge would relinquish jurisdiction over the state law claims; for they were supplemental claims, which normally are dismissed when the federal claims to which they are supplemental drop out of the case before trial. See 28 U.S.C. § 1367(c)(3); Brazinski v. Amoco Petroleum Additives Co., 6 F.3d 1176, 1182 (7th Cir.1993); Musson Theatrical, Inc. v. Federal Express Corp., 89 F.3d 1244, 1254-55 (6th Cir.1996); Rodriguez v. Doral Mortgage Corp., 57 F.3d 1168, 1177 (1st Cir.1995). Indeed we suggested he do that. 672 F.3d at 518-19.

So the case was still alive in the district court, if barely, when unexpectedly the district judge granted the plaintiff's motion to reconsider the dismissal of one of his federal claims — Wharrie's alleged fabrication of testimony by a witness during the investigation in 1985 that led to Fields' indictment and trial — in light of our intervening decision in Whitlock v. Brueggemann, 682 F.3d 567 (7th Cir.2012). The judge also decided to retain supplemental jurisdiction of several of Fields' state law claims. By thus partially stripping the defendants of absolute prosecutorial immunity, the judge's rulings precipitated this, the defendants' second appeal.

The claim that the district judge reinstated against Wharrie is the one concerning Wharrie's investigation of Fields in 1985. Prosecutors, like judges, enjoy absolute immunity from federal tort liability, whether common law or constitutional, because of "concern that harassment by unfounded litigation would cause a deflection of the prosecutor's energies from his public duties, and the possibility that he would shade his decisions instead of exercising the independence of judgment required by his public trust." Imbler v. Pachtman, 424 U.S. 409, 423, 96 S.Ct. 984, 47 L.Ed.2d 128 (1976); see also Gregoire v. Biddle, 177 F.2d 579, 581 (2d Cir.1949) (L.Hand, J.). But the absolute immunity is only for acts they commit within the scope of their employment as prosecutors. Buckley v. Fitzsimmons, 509 U.S. 259, 273-76, 113 S.Ct. 2606, 125 L.Ed.2d 209 (1993); Thomas v. City of Peoria, 580 F.3d 633, 638-39 (7th Cir.2009); Pinaud v. County of Suffolk, 52 F.3d 1139, 1147 (2d Cir.1995). Often their employment duties go beyond the strictly prosecutorial to include investigation, and when they do nonprosecutorial work they lose their absolute immunity and have only the immunity, called "qualified," that other investigators enjoy when engaged in such work. Buckley v. Fitzsimmons, supra, 509 U.S. at 275-76, 113 S.Ct. 2606. Qualified immunity "protects government officials `from liability for civil damages insofar as their conduct does not violate clearly established statutory or constitutional rights of which a reasonable person would have known.' Pearson v. Callahan, 555 U.S. 223, 231, 129 S.Ct. 808, 172 L.Ed.2d 565 (2009) (quoting Harlow v. Fitzgerald, 457 U.S. 800, 818, 102 S.Ct. 2727, 73 L.Ed.2d 396 (1982))." Messerschmidt v. Millender, ___ U.S. ___, 132 S.Ct. 1235, 1244, 182 L.Ed.2d 47 (2012).

Fields was not arrested until June 1985. Wharrie's alleged procurement of false statements from a prospective witness in Fields' forthcoming trial had taken place a month earlier. The trial (the first of Fields' two trials) took place a year later. Wharrie was one of the prosecutors at that trial. Though as we said a prosecutor's absolute immunity is limited to the performance of his prosecutorial duties, and not to other duties to which he might be assigned by his superiors or perform on his own initiative, such as investigating a crime before an arrest or indictment, Wharrie argues that he is insulated from liability for his investigative work by our decision in Buckley v. Fitzsimmons, 20 F.3d 789 (7th Cir.1994). The plaintiff in that case, which like this one was a malicious prosecution suit against prosecutors, claimed that at the investigative stage of the criminal case against the plaintiff the prosecutors had coerced witnesses to testify against him. We held that he could not base a legal claim on those acts because until the evidence obtained by the improper acts was introduced at his trial he had not been injured and therefore no tort had been committed. "Events not themselves supporting recovery under § 1983 do not become actionable because they lead to injurious acts for which the defendants possess absolute immunity," id. at 796 — namely presenting the coerced evidence at trial. Presenting evidence at trial is a core prosecutorial function, protected by absolute prosecutorial immunity and therefore an insuperable bar to an award of damages in a suit for malicious prosecution against the prosecutor.

The analysis in our Buckley decision thus rests on the principle that there is no tort without an actionable injury caused by the defendant's wrongful act, 20 F.3d at 796. That is indeed the law. See, e.g., Jackson v. Pollion, 733 F.3d 786, 790 (7th Cir.2013). But the act that causes an injury need not be simultaneous with the injury (indeed it will never be exactly simultaneous) for the actor to be liable. Think of products liability. The defect that caused a pipe to burst and flood your home may have been present when the pipe was manufactured years earlier. The manufacturer would be liable despite the lapse of time. The statute of limitations would not have begun to run until the pipe burst and caused damage (though if there were a statute of repose, the deadline created by that statute may have passed). It may seem difficult to understand why a prosecutor who, acting in an investigative role before judicial proceedings against a criminal defendant began, coerced a witness or fabricated testimony, intending that it be used against the defendant and knowing that it would be used against him, should be excused from liability just because the defendant was not harmed until the witness testified and, as a result, the defendant was convicted and sentenced. He who creates the defect is responsible for the injury that the defect foreseeably causes later. Nor is the only harm that resulting from the conviction and the sentence. In the present case, as in our recent decision in Julian v. Hanna, 732 F.3d 842, 847 (7th Cir.2013), the fabrication of evidence harmed the defendant before and not just during the trial, because it was used to help indict him.

But consider a case in which a prosecutor, whom we'll call A, acting in a purely investigative role, fabricates evidence against a suspect, but the prosecution of the suspect is handled by a different prosecutor, B, who though knowing that the evidence was fabricated decides to use it. And suppose A has second thoughts about what he did, and tells B not to use the evidence. But B goes ahead and uses it. It can be argued that A should not be regarded as having caused the use of the evidence at trial. Cf. Whitlock v. Brueggemann, supra, 682 F.3d at 583-84. B is protected from suit by his absolute prosecutorial immunity, thus leaving the victim of the fabricated evidence without a complete damages remedy. He would have a partial remedy, against A, if A's fabrication had inflicted harm against him before his trial, as by causing him to be indicted, as in Julian v. Hanna, supra.

But this is not such a case because Wharrie, the alleged fabricator of evidence against Fields, was also one of his prosecutors at trial. Nevertheless both Whitlock and the present case differ from our Buckley decision in a respect that the panel in Whitlock (which did not question the soundness of Buckley) thought critical — namely that the misconduct of the prosecutor (there was only one) in Whitlock consisted not in coercing witnesses, as in Buckley, but in fabricating evidence. (So here is where the terminological distinctions suggested earlier in this opinion bite.) Whitlock observes that "coercively interrogating witnesses, paying witnesses for testimony, and witness-shopping may be deplorable, and these tactics may contribute to wrongful convictions, but they do not necessarily add up to a constitutional violation even when their fruits are introduced at trial," because "evidence collected with these kinds of suspect technique, unlike falsified evidence and perjured testimony, may turn out to be true." 682 F.3d at 584 (emphasis added). The prosecutor who was sued in Whitlock was part of an investigative team that told witnesses what to say knowing that what the team was telling them was false. See id. at 571-72. This is different from coercing a reluctant witness to say what may be true. The point is not that coercion is legally irrelevant; far from it — coercion (which in an extreme case could amount to torture) may be an essential tool in "persuading" a witness to fabricate testimony. The point is only that coercion per se does not make a prosecutor acting as an investigator liable should the coerced evidence be used to obtain a conviction of an innocent criminal defendant. If the evidence obtained by coercion is sound and the defendant would have been indicted, found guilty, and sentenced, without it (though in fact he was innocent), the only victim of government misconduct is the witness who was coerced; and that is not Fields.

In Buckley the plaintiff had alleged that the coerced evidence was not only coerced, but false. But our opinion described the primary victims, and the only ones with standing to complain, as the coerced witnesses themselves. "Coercing witnesses to speak," we said, "rather than loosening their tongues by promises of reward, is a genuine constitutional wrong, but the persons aggrieved would be Cruz and Hernandez [the allegedly coerced witnesses] rather than Buckley. Overbearing tactics violate the right of the person being interrogated to be free from coercion. Buckley cannot complain that the prosecutors may have twisted Cruz's arm, any more than he can collect damages because they failed to read Cruz Miranda warnings or searched Cruz's house without a warrant. Rights personal to their holders may not be enforced by third parties." 20 F.3d at 794-95 (citations omitted). This is analogous to the rule that a criminal defendant will not be heard to object to the admission against him of probative evidence unlawfully seized from a third party. Minnesota v. Carter, 525 U.S. 83, 90-91, 119 S.Ct. 469, 142 L.Ed.2d 373 (1998); Rakas v. Illinois, 439 U.S. 128, 148-50, 99 S.Ct. 421, 58 L.Ed.2d 387 (1978).

This "tough luck" rule doesn't apply when there is no coerced witness but nevertheless fabricated evidence; the evidence might be given by a paid police informant — a compensated witness, not a coerced one. That is the distinction drawn in Whitlock, and Wharrie acknowledges that if Whitlock is the law he loses his defense of absolute immunity because he is accused of fabricating evidence before trial; and so he asks us to overrule Whitlock as inconsistent with our Buckley decision. Fabrication was alleged in Buckley as well, but as the language we just quoted from that opinion reveals, the focus was on coercion, and the primary victims, and only victims held to have standing to sue, were the coerced witnesses.

In asking us to overrule Whitlock, Wharrie ignores the alternative of overruling Buckley as being inconsistent with (the newer) Whitlock. There's no need to embrace either alternative. Because the cases are distinguishable and the present case falls on the Whitlock side of the line, we agree with the district judge's reinstatement of Fields' claim against Wharrie for the latter's alleged fabrication of evidence in 1985, before Fields' indictment and arrest.

Wharrie is asking us to bless a breathtaking injustice. Prosecutor, acting pre-prosecution as an investigator, fabricates evidence and introduces the fabricated evidence at trial. The innocent victim of the fabrication is prosecuted and convicted and sent to prison for 17 years. On Wharrie's interpretation of our decision in Buckley, the prosecutor is insulated from liability because his fabrication did not cause the defendant's conviction, and by the time that same prosecutor got around to violating the defendant's right he was absolutely immunized. So: grave misconduct by the government's lawyer at a time where he was not shielded by absolute immunity; no remedy whatsoever for the hapless victim. In Buckley, in contrast, two victims — the coerced witnesses — had a damages remedy, and the fact that there thus were potential plaintiffs could be expected to have some deterrent effect against future misconduct. To extend Buckley to overrule Whitlock would leave a victim of graver misconduct (because coercion does not always result in fabrication of evidence — even torture must often elicit truthful confessions) completely unprotected.

That's not only an offensive and indeed senseless result, but it doesn't jibe with the Supreme Court's decision in Buckley, where we read that "the [fact that] prosecutors later called a grand jury to consider the evidence this work produced does not retroactively transform that work from the administrative into the prosecutorial. A prosecutor may not shield his investigative work with the aegis of absolute immunity merely because, after a suspect is eventually arrested, indicted, and tried, that work may be retrospectively described as `preparation' for a possible trial; every prosecutor might then shield himself from liability for any constitutional wrong against innocent citizens by ensuring that they go to trial." Buckley v. Fitzsimmons, supra, 509 U.S. at 275-76, 113 S.Ct. 2606 (emphasis added, footnote omitted); see also Zahrey v. Coffey, 221 F.3d 342, 349, 354 (2d Cir.2000); McGhee v. Pottawattamie County, 547 F.3d 922, 932-33 (8th Cir.2008); Moore v. Valder, 65 F.3d 189, 194-95 (D.C.Cir.1995). A prosecutor cannot retroactively immunize himself from conduct by perfecting his wrongdoing through introducing the fabricated evidence at trial and arguing that the tort was not completed until a time at which he had acquired absolute immunity. That would create a "license to lawless conduct," which the Supreme Court has said that qualified immunity is not to do. Harlow v. Fitzgerald, 457 U.S. 800, 819, 102 S.Ct. 2727, 73 L.Ed.2d 396 (1982). Wharrie's interpretation of our decision in Buckley would place that decision in conflict with the Supreme Court's Buckley decision, by giving absolute immunity to prosecutor-investigators who having fabricated evidence make sure that the evidence is used to convict the innocent victim of the fabrication.

So Wharrie has not demonstrated an entitlement to absolute immunity — nor to qualified immunity for the fabrication, either. For it was established law by 1985 (indeed long before), when the fabrication is alleged to have occurred, that a government lawyer's fabricating evidence against a criminal defendant was a violation of due process. See Napue v. Illinois, 360 U.S. 264, 269, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959); Pyle v. Kansas, 317 U.S. 213, 215-16, 63 S.Ct. 177, 87 L.Ed. 214 (1942); Mooney v. Holohan, 294 U.S. 103, 110, 112-13, 55 S.Ct. 340, 79 L.Ed. 791 (1935) (per curiam). It is true that the cases we've just cited involved not merely the fabrication, but the introduction of the fabricated evidence at the criminal defendant's trial. For if the evidence hadn't been used against the defendant, he would not have been harmed by it, and without a harm there is, as we noted earlier, no tort. But when the question is whether to grant immunity to a public employee, the focus is on his conduct, not on whether that conduct gave rise to a tort in a particular case. As the Court said in Harlow v. Fitzgerald, supra, 457 U.S. at 819, 102 S.Ct. 2727, the test for qualified immunity is "the objective legal reasonableness of an official's acts. Where an official could be expected to know that certain conduct would violate statutory or constitutional rights, he should be made to hesitate; and a person who suffers injury caused by such conduct may have a cause of action" (emphasis added) — or may not.

So notice the disjunction: the immunity depends on the official's acts; the existence of a cause of action depends on the illegality of those acts and on whether an injury results, because, to repeat, no injury — no tort. Recall the earlier point derived from the Supreme Court's decision in Buckley that without that distinction a prosecutor who in a pre-prosecution investigative role fabricates evidence obtains immunity simply by turning over the fabricated evidence to the prosecutor, or prosecuting the defendant himself, for in neither case is he the direct cause of the defendant's injury (the wrongful conviction).

It remains to consider two of Fields' state law claims. The first arises from the same 1985 conduct alleged of Wharrie that we've been discussing. The second is false testimony allegedly coerced by defendant Kelley in 1998, while Fields was awaiting the retrial of his murder case.

The district court held, and both parties agree, that the Illinois rule of absolute prosecutorial immunity is the same as the federal rule. Actually that's not entirely clear. The last time the Supreme Court of Illinois spoke to the issue it held that all Illinois officials — including judges — enjoy immunity only "from liability for error or mistake of judgment in the exercise of their duty in the absence of corrupt or malicious motives." People ex rel. Schreiner v. Courtney, 380 Ill. 171, 43 N.E.2d 982, 986 (1942) (emphasis added). That's not absolute immunity. But 71 years on, it is doubtful that this is still the law in Illinois. The Illinois Appellate Court seems not to think so. See, e.g., Frank v. Garnati, 370 Ill.Dec. 931, 989 N.E.2d 319, 322 (Ill.App.2013); White v. City of Chicago, 369 Ill.App.3d 765, 308 Ill.Dec. 518, 861 N.E.2d 1083, 1087-88 (2006); Weimann v. Kane County, 150 Ill.App.3d 962, 104 Ill.Dec. 110, 502 N.E.2d 373, 377 (1986); People v. Patrick J. Gorman Consultants, Inc., 111 Ill.App.3d 729, 67 Ill.Dec. 540, 444 N.E.2d 776, 779 (1982); Coleson v. Spomer, 31 Ill.App.3d 563, 334 N.E.2d 344, 347 (1975). And the state's supreme court has not called a halt to the trend in the intermediate appellate court. E.g., Frank v. Garnati, 374 Ill.Dec. 565, 996 N.E.2d 12 (2013) (denying appeal). Although some cases recite what sounds like the rule from Courtney, in none was immunity denied on the ground of bad faith. See, e.g., Aboufariss v. City of De Kalb, 305 Ill.App.3d 1054, 239 Ill.Dec. 273, 713 N.E.2d 804, 812 (1999).

Since, as just explained, Illinois may not even have absolute prosecutorial immunity, and since Wharrie does not claim that if Illinois does have it, it is more absolute than the federal rule, and since we've already held that Wharrie is not entitled to absolute immunity from being sued on the federal claims against him, there is no basis for giving him absolute prosecutorial immunity from the state law claims for the same conduct alleged as a violation of Illinois tort law.

Kelley's alleged procurement of false testimony took place in 1998, in preparation for Fields' second trial and therefore in the midst of his prosecution. Once prosecution begins, bifurcating a prosecutor's role between investigation and prosecution is no longer feasible. If in the course of a trial a prosecutor were to urge one of his witnesses to lie, it would be arbitrary to describe this as an investigative act separate from prosecution; the prosecutor's conduct would have been "intimately associated with the judicial phase of the criminal process," and he would therefore be entitled to absolute immunity. Imbler v. Pachtman, supra, 424 U.S. at 430, 96 S.Ct. 984; see also Smith v. Power, 346 F.3d 740, 742 (7th Cir.2003); Warney v. Monroe County, 587 F.3d 113, 120-21 (2d Cir.2009); Genzler v. Longanbach, 410 F.3d 630, 638 (9th Cir.2005) (investigation "bound up with the judicial process"). It would be unlike the case discussed in the Supreme Court's Buckley decision of prosecutors held by the Court not to be entitled to absolute immunity for fabricating evidence "during the early stages of the investigation" when "police officers and assistant prosecutors were performing essentially the same investigatory functions." Buckley v. Fitzsimmons, supra, 509 U.S. at 262-63, 113 S.Ct. 2606.

The district judge ruled after our first decision and consistently with the analysis just presented that Wharrie had absolute prosecutorial immunity under state as well as federal law for his allegedly procuring false testimony in anticipation of Fields' retrial in 1998, and Fields has not appealed that ruling. We cannot understand on what basis the district judge distinguished the identical conduct alleged against Kelley in the same stage of the same proceeding against Fields, and so denied his motion to dismiss the claim against him. True, Kelley may not yet have been "officially" part of the trial team at Fields' retrial, but he was already functioning as part of the team and that's all that matters, as we noted in our first opinion. See 672 F.3d at 51112, 515-16; see also Van de Kamp v. Goldstein, 555 U.S. 335, 343-45, 129 S.Ct. 855, 172 L.Ed.2d 706 (2009).

And so, to summarize, the denial of Wharrie's motion to dismiss the federal and state claims against him relating to the fabrication of false statements from a witness during the investigation in 1985 is affirmed, but the denial of Kelley's motion to dismiss the 1998 state law claims against him is reversed with instructions that the district court reconsider that denial in light of the analysis in this opinion.

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.

SYKES, Circuit Judge, concurring in part and dissenting in part.

I agree that Fields's claims against Kelley are barred by absolute immunity. See Majority op. at 1114-16. In our earlier opinion in this case, we held that Wharrie is absolutely immune from suit on similar allegations of misconduct — specifically, that Wharrie coerced Earl Hawkins to falsely implicate Fields in the 1984 murders of Talman Hickman and Jerome Smith during the interval between Fields's first and second trials, while his postconviction proceedings were ongoing and in anticipation of retrial. See Fields v. Wharrie ("Fields I"), 672 F.3d 505, 511-16 (7th Cir.2012). The allegations against Kelley are materially indistinguishable from the allegations against Wharrie stemming from this time period. The only difference is that Kelley is accused of inducing another witness, Randy Langston, to make a false identification of Fields. But this act, like Wharrie's alleged coercion of Hawkins, took place during the judicial phase of the criminal process and was functionally prosecutorial. So Kelley, like Wharrie, is absolutely immune from suit for this conduct. See Van de Kamp v. Goldstein, 555 U.S. 335, 340-45, 129 S.Ct. 855, 172 L.Ed.2d 706 (2009); Imbler v. Pachtman, 424 U.S. 409, 430-31, 96 S.Ct. 984, 47 L.Ed.2d 128 (1976); Fields I, 672 F.3d at 510-15.

Illinois apparently follows the federal law on absolute immunity, see Frank v. Garnati, 370 Ill.Dec. 931, 934, 989 N.E.2d 319 (Ill.App.Ct.2013); White v. City of Chicago, 369 Ill.App.3d 765, 308 Ill.Dec. 518, 861 N.E.2d 1083, 1088-94 (2006); the parties do not argue otherwise. Accordingly, I join the majority opinion to the extent that it reverses the district court's order permitting the state-law claims against Kelley to proceed. See Majority op. at 1114-16.

Unlike my colleagues, however, I would also reverse the district court's decision to reinstate the § 1983 and state-law claims against Wharrie for his conduct in 1985, before Fields was charged. Fields alleges that during the investigation of the Hickman and Smith murders, Wharrie coerced Anthony Sumner, a fellow gang member, to falsely implicate him in the crimes. Fields I, 672 F.3d at 508-09. As we explained in our earlier opinion, Fields was convicted of killing Hickman and Smith based in part on Sumner's testimony, but he won a new trial for unrelated reasons (the trial judge had been bribed), was acquitted on retrial, and obtained a certificate of innocence.1 Id. at 509. His complaint seeks damages under 42 U.S.C. § 1983 for violation of his right to due process; he also asserts several state-law theories of liability for the wrongful convictions.2

To the extent that the § 1983 and state-law claims against Wharrie are based on his alleged coercion of the false statement from Sumner, absolute immunity does not apply. That act took place before there was probable cause to arrest Fields — that is, before the judicial process began — and was not functionally prosecutorial, so Wharrie cannot claim to be absolutely immune from suit for damages. Buckley v. Fitzsimmons, 509 U.S. 259, 273-76, 113 S.Ct. 2606, 125 L.Ed.2d 209 (1993); Buckley v. Fitzsimmons, 20 F.3d 789, 794 (7th Cir.1994).

But Wharrie remains protected by qualified immunity, see Buckley, 509 U.S. at 275-76, 113 S.Ct. 2606, which "is available unless the acts violated `clearly established' constitutional norms," Buckley, 20 F.3d at 794. The Supreme Court's decision in Buckley addressed only absolute prosecutorial immunity; the Court did not have occasion to decide the qualified-immunity question. That is, the Court did not decide whether coercing or otherwise inducing a witness to give a false statement during a criminal investigation violates clearly established constitutional rights. But we decided that very question when the Supreme Court returned Buckley to this court for further proceedings.

In our decision on remand in Buckley, we held that coercing or otherwise soliciting a witness to falsely incriminate a suspect during a criminal investigation does not violate any established constitutional rights — except perhaps the rights of the witness who is coerced. Id. at 794-97. If the suspect is charged, then failing to disclose the false statement's corrupt origins at trial violates his due-process right to a fair trial under the rule of Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), and knowingly using perjured testimony to convict him is a more general violation of his due-process right to a fair trial. See Buckley, 20 F.3d at 794-95; see also Albright v. Oliver, 510 U.S. 266, 273 n. 6, 114 S.Ct. 807, 127 L.Ed.2d 114 (1994) (plurality opinion); Napue v. Illinois, 360 U.S. 264, 269, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959); Mooney v. Holohan, 294 U.S. 103, 112, 55 S.Ct. 340, 79 L.Ed. 791 (1935); Serino v. Hensley, 735 F.3d 588, 592 (7th Cir.2013); Newsome v. McCabe, 256 F.3d 747, 750-52 (7th Cir. 2001). But a prosecutor who commits these acts or omissions at trial is functioning quintessentially as a prosecutor, so under well-established immunity law, he is absolutely immune from suit for damages under § 1983. Van de Kamp, 555 U.S. at 340-45, 129 S.Ct. 855; Imbler, 424 U.S. at 430-31, 96 S.Ct. 984; Fields I, 672 F.3d at 510-15; Buckley, 20 F.3d at 794-96.

In contrast, a prosecutor who coerces or otherwise procures a false statement from a witness during an investigation, before probable cause exists and the judicial process has begun, is not protected by absolute immunity, but he is entitled to qualified immunity because his conduct does not violate clearly established constitutional rights. Buckley, 20 F.3d at 794-98. Due-process rights are not implicated at this stage; they arise later, when judicial proceedings are instituted. Id. Accordingly, extracting a false statement from a witness during a criminal investigation is not independently actionable as a constitutional violation, and Buckley holds that "events not themselves supporting recovery [against prosecutors] under § 1983 do not become actionable because they lead to injurious acts for which [the prosecutors] possess absolute immunity." Id. at 796.

In our earlier opinion in this case, we relied on this qualified-immunity holding from Buckley as an alternative basis for finding Wharrie and Kelley immune from suit under § 1983 for their solicitation of false statements from Hawkins and Langston. Fields I, 672 F.3d at 516-18. Citing Buckley, we noted that "fabricating evidence, including in the form of testimony, is not an actionable constitutional wrong.... The constitutional violation occurs when the means by which the testimony was acquired are not disclosed at trial — or when other information that impeach[es] the testimony's reliability is not shared with the defense." Id. at 516-17. Our alternative holding in Fields I followed the rule, established in Buckley, that even if a prosecutor participates in securing a false statement from a witness during a criminal investigation, his "absolutely immunized prosecutorial decision to proceed to trial and introduce the [witness's] testimony" forecloses suit against him; there is no independently cognizable due-process claim for his investigative misconduct.3 Id. at 517 (discussing Wharrie's qualified immunity); see also id. at 51718 (discussing Kelley's qualified immunity).

Three months after we issued our opinion in Fields I, a new decision of this court, Whitlock v. Brueggemann, 682 F.3d 567 (7th Cir.2012), unsettled Buckley (and by extension, unsettled Fields I as well), which led the district court to do an about-face on remand in this case. Wharrie's new appeal requires us to decide whether Whitlock and Buckley can be reconciled. I think the answer is plainly "no."

As my colleagues have noted, Whitlock drew a distinction, for qualified-immunity purposes, between a prosecutor who coercively interrogates witnesses and a prosecutor who fabricates evidence. See Majority op. at 1112-13. Whitlock observed that coercion and unsavory tactics like paying for testimony and witness shopping "may be deplorable, and ... may contribute to wrongful convictions, but they do not necessarily add up to a constitutional violation even when their fruits are introduced at trial ... [because] [e]vidence collected with these kinds of suspect techniques, unlike falsified evidence and perjured testimony, may turn out to be true." 682 F.3d at 584. This proposed distinction between "coerced" and "fabricated" evidence was the linchpin for distinguishing Buckley and permitting the claim against the prosecutor in Whitlock to proceed. The Whitlock panel thought the prosecutors in Buckley had been accused of coercing witnesses and using other suspect tactics but were not alleged to have fabricated evidence, and on that basis distinguished the case. Id. at 584-85.

Regrettably, that was a mistake. In fact, both Buckley and Whitlock involved allegations that prosecutors had coerced, cajoled, paid for, or otherwise solicited falsified statements from witnesses — in other words, they fabricated evidence. Indeed, Buckley repeatedly refers to allegations that prosecutors coercively obtained "false inculpatory statements" from witnesses and "fabricated" or "manufactured" testimony and evidence from an expert.4 20 F.3d at 794-95. So regardless of whether Whitlock's proposed distinction between "coerced" and "fabricated" witness statements is valid in theory — and makes a difference in the constitutional analysis — it simply was not present as a factual matter and therefore cannot provide a basis on which to distinguish Buckley from Whitlock.5

Yet my colleagues perpetuate the distinction here. See Majority Op. at 1109-10, 1112-13. I appreciate the force of stare decisis; we should try to harmonize the two cases if we can. With respect, however, harmonization is impossible. Whitlock and Buckley are factually indistinguishable and legally irreconcilable. They cannot both be the law. We must decide which one is correct.6

For my part, I think Buckley is correct and Whitlock should be reconsidered. Because mine is the minority view here, any reconsideration of Whitlock must await a petition for rehearing en banc, which Wharrie may choose to pursue or forego. For the record, I'll briefly sketch the conceptual difficulty Whitlock has introduced, which I believe warrants the full court's attention.

I have already described Buckley's qualified-immunity holding, which was and remains sound. Whitlock rejected qualified immunity for a similarly situated prosecutor by using common-law causation principles to find an actionable constitutional violation where one did not otherwise exist. I do not agree with this development in our circuit's law.

As I've explained, Buckley contains two important principles of immunity law that apply in suits alleging prosecutorial misconduct: (1) a prosecutor's use of fabricated evidence at trial may be actionable as a violation of the defendant's right to due process — under the rubric of Brady or perhaps more generally as a violation of the right to a fair trial — but the prosecutor is absolutely immune from suit under Imbler and related cases, see 20 F.3d at 794-95; and (2) a prosecutor's fabrication of evidence against a suspect during an investigation is covered by qualified immunity because it doesn't violate clearly established constitutional rights, see id. at 794-98. Whitlock's innovation is to use common-law causation analysis to eliminate the effect of both forms of immunity. The panel held that "[t]he actions of an official who fabricates evidence that later is used to deprive someone of liberty can be both a but-for and proximate cause of the due process violation."7 Whitlock, 682 F.3d at 583.

As applied to a prosecutor, this reasoning circumvents both qualified and absolute immunity. Both immunities are well established and important. See Van de Kamp, 555 U.S. at 341-43, 129 S.Ct. 855 (absolute immunity); Harlow v. Fitzgerald, 457 U.S. 800, 818, 102 S.Ct. 2727, 73 L.Ed.2d 396 (1982) (qualified immunity); Imbler, 424 U.S. at 424-26, 96 S.Ct. 984 (absolute immunity). Absolute immunity can sometimes produce harsh results, but it has long been thought necessary to encourage and protect the vigorous performance of the prosecutorial function. See Van de Kamp, 555 U.S. at 341-43, 129 S.Ct. 855; Imbler, 424 U.S. at 424-26, 96 S.Ct. 984.

In a shorthand version, the rule announced in Whitlock is basically this: A prosecutor who falsifies evidence during an investigation violates no clearly established constitutional rights and thus has qualified immunity from suit (see Buckley), but his conduct is nonetheless actionable as a non-immune subsidiary "cause" (both but-for and proximate) of a due-process violation that occurs later, when the prosecutor introduces the falsified evidence at trial — even though the prosecutor is absolutely immune from suit for the due-process violation. In other words, the only conduct that can possibly form the basis of a constitutional claim — the prosecutor's trial conduct — is fully protected by absolute immunity, but the prosecutor can be sued anyway, based on the causal link between his nonactionable investigative conduct and his immunized trial conduct.

Aside from destabilizing immunity law, this chain of reasoning overlooks some basic differences between common-law and constitutional torts. Common-law causation rules flow from the nature of duty and breach in tort law. Everyone has a general tort duty to refrain from doing an act or omitting a precaution that creates a foreseeable, unreasonable risk of harm to other persons or property. See generally RESTATEMENT (SECOND) OF TORTS § 282 (1965) (defining negligence). The duty is broad and undifferentiated and is owed to everyone at all times, and anyone who breaches it is liable for harms factually and proximately caused.

Constitutional rights — and the corresponding duties imposed on governmental actors — are not like the generalized rights and duties imposed by negligence law. They are implicated at specific times and in specific circumstances. As relevant here, Fields's due-process rights came into play after he was charged; the Brady disclosure duty is an aspect of the right to a fair trial, as is the broader right not to have the trial process subverted by the knowing introduction of falsified evidence. See Serino, 735 F.3d at 592; Newsome, 256 F.3d at 751-52; Buckley, 20 F.3d at 796-97. So Wharrie's act of extracting a false statement from Sumner during the investigative phase of the case did not violate Fields's due-process rights. A prosecutor who commits this kind of misconduct has behaved deplorably but has breached no constitutional duty and thus committed no constitutional wrong.

Common-law causation analysis cannot be used to transform an act that does not violate the Constitution into one that does. That, in essence, is the effect of Whitlock. It turns the prosecutor's nonactionable investigative misconduct into an actionable constitutional wrong by recharacterizing it as a subsidiary "cause" of a due-process violation that occurs later at trial but is absolutely immunized.

To help explain and justify Whitlock, my colleagues analogize to product-liability causation rules. See Majority op. at 1111-12. I think the analogy exposes a key conceptual problem in the analysis. The pipe maker who negligently manufactures a pipe breaches his general tort duty to refrain from acts that create an unreasonable risk of harm to persons or property. The breach may not cause harm until the pipe bursts, and the statute of limitations does not begin to run until the harm occurs. But the breach of duty occurs at the point of negligent manufacture, and the manufacturer will be liable for harm that flows causally from that breach. (In the law of strict product liability, the manufacturer's sale of the defective pipe starts the causal chain, but the basic point about the order of analysis — duty/breach/cause — remains the same.)

Section 1983 does not impose an analogous generalized duty on governmental agents to refrain from putting people at an unreasonable risk of harm. Rather, § 1983 creates a cause of action for damages against state and local officials who inflict specific deprivations of federal rights. See 42 U.S.C. § 1983 (creating a cause of action against any official who, under color of state law, "subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws"). It is well understood that § 1983 "`is not itself a source of substantive rights,' but merely provides `a method for vindicating federal rights elsewhere conferred.'" Albright, 510 U.S. 266, 114 S.Ct. at 811 (plurality opinion) (quoting Baker v. McCollan, 443 U.S. 137, 144 n. 3, 99 S.Ct. 2689, 61 L.Ed.2d 433 (1979)). In other words, the Constitution supplies the rights and duties and otherwise fills in the content of a § 1983 claim. So "[t]he first step in any such claim is to identify the specific constitutional right allegedly infringed." Id.

Here, the right allegedly infringed is the due-process right to a fair trial, but under the rationale of Whitlock, the asserted act of infringement occurred before any fair-trial duty arose. To put the problem colloquially, Whitlock puts the causation cart before the duty horse. A prosecutor who coerces a witness to falsely incriminate a suspect during a criminal investigation breaches no constitutional duty owed to the suspect. If the suspect is charged, then suborning perjury against him at trial would violate his due-process rights — so too would withholding evidence about the coercion of the witness. But these are trial rights, and a prosecutor's violation of them is absolutely immunized. The prosecutor's investigative misconduct cannot independently support a due-process claim; that conduct violates no due-process duty.

It is true, as my colleagues have noted, that the Supreme Court said the following in Buckley:

Buckley, 509 U.S. at 276, 113 S.Ct. 2606. This passage in the Court's opinion simply drives home the point that prosecutors are not protected by absolute immunity for their pre-charging misconduct. That much is clear from the very next sentence: "When the functions of prosecutors and detectives are the same, as they were here, the immunity that protects them is also the same." Id. That's an obvious reference to qualified immunity.

It's worth repeating that the Court sent Buckley back to us to decide the qualified-immunity question. We unambiguously held — in 1994 — that falsifying evidence during an investigation does not violate established constitutional rights, and "events not themselves supporting recovery [against prosecutors] under § 1983 do not become actionable because they lead to injurious acts for which [the prosecutors] possess absolute immunity." Buckley, 20 F.3d at 796.

Directly contradicting that clear holding of Buckley, my colleagues now conclude that "it was established law by 1985, when the fabrication is alleged to have occurred, that a government lawyer's fabricating evidence against a criminal defendant was a violation of due process." Majority op. at 1114 (emphasis added). For this proposition they cite Napue v. Illinois, 360 U.S. 264, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959); Pyle v. Kansas, 317 U.S. 213, 63 S.Ct. 177, 87 L.Ed. 214 (1942); and Mooney v. Holohan, 294 U.S. 103, 55 S.Ct. 340, 79 L.Ed. 791 (1935), but these cases do not hold that fabricating evidence violates due process. Rather, they hold that the use of falsified evidence or perjured testimony at trial violates the defendant's due-process right to a fair trial and is grounds for habeas or other postconviction relief. See Napue, 360 U.S. at 269, 79 S.Ct. 1173; Pyle, 317 U.S. at 215-16, 63 S.Ct. 177; Mooney, 294 U.S. at 112-13, 55 S.Ct. 340. As my colleagues have noted, the qualified-immunity inquiry asks whether a reasonable public official could be expected to know that the conduct in question violates constitutional rights. See Majority op. at 1114 (citing Harlow, 457 U.S. at 819, 102 S.Ct. 2727). The specific conduct in question here — inducing a witness to tell a lie during an investigation — is clearly wrong, but it does not violate any constitutional rights.

So although my colleagues have formally left Buckley in place, the core of the opinion — its qualified-immunity holding — is effectively overruled.

Finally, a brief comment on my colleagues' assertion that adhering to Buckley would work a "breathtaking injustice" and produce "an offensive and indeed senseless result."8 Id. at 1113, 1113. These are strong words. No one doubts that wrongful convictions are unjust; a person who is convicted and punished for a crime he did not commit has a serious moral claim to a compensatory remedy. Usually the law provides one, commonly in the form of a Brady claim against the officers who were involved in suppressing exculpatory evidence during the prosecution. It's possible that in some cases the effect of absolute immunity — or the combined effect of absolute and qualified immunity — might leave a wrongly convicted person without an actionable damages claim against any of the wrongdoers. I could be wrong, but I don't think that happens very often. Prosecutors do not work alone, and if the police officers working with them withhold exculpatory information about coerced or fabricated evidence, the aggrieved defendant will have a good § 1983 claim against the officers for violation of Brady. The Brady duty is well established, and the claim against the officers is available regardless of whether the prosecutor participated in the "creation" of the fabricated evidence or the cover-up at trial or both. That basically describes this case.9

My colleagues also perceive a need to authorize a damages remedy as a deterrent against rogue prosecutors. Id. at 1113-14. That's a valid concern, but it doesn't justify finding a cognizable constitutional violation where one does not exist. The policy justification for absolute immunity accepts that deterrence can be achieved in other ways. Although a complicit prosecutor escapes civil liability for damages, he remains subject to criminal prosecution and professional discipline for his misdeeds; he is not immune from these consequences for his misconduct. See Imbler, 424 U.S. at 429, 96 S.Ct. 984. The prospect of criminal liability and disbarment are powerful deterrents. See id. ("These [criminal and professional] checks undermine the argument that the imposition of civil liability is the only way to insure that prosecutors are mindful of the constitutional rights of persons accused of crime.").

In the end, the policy behind absolute prosecutorial immunity "reflects `a balance' of `evils'" based on a judgment that it is "`better ... to leave unredressed the wrongs done by dishonest officers than to subject those who try to do their duty to the constant dread of retaliation.'" Van de Kamp, 555 U.S. at 340, 129 S.Ct. 855 (quoting Gregoire v. Biddle, 177 F.2d 579, 581 (2d Cir.1949) (Hand, J.)). Moreover, as we said in Buckley, "[q]ualified immunity does not permit us to recognize ... [new constitutional] right[s] in this litigation." Buckley, 20 F.3d at 797. Although Wharrie's alleged wrongdoing may go unredressed via a federal damages remedy against him, Fields has an ongoing § 1983 claim against the police officers who were allegedly complicit in withholding exculpatory evidence about the circumstances surrounding Sumner's false statement.

In sum, applying Buckley requires a conclusion that Wharrie is entitled to qualified immunity for his investigative misconduct; his act of soliciting Sumner's false statement did not violate clearly established constitutional rights and does "not become actionable because [it led to] injurious acts for which [the prosecutor] possess[es] absolute immunity." Id. at 796. Applying Whitlock, my colleagues conclude that Wharrie's investigative misconduct is actionable as a non-immune subsidiary "cause" of his absolutely immunized due-process violation at trial. I would reconsider Whitlock, restore Buckley, and reverse with instructions to dismiss the § 1983 claim against Wharrie. With no federal claim remaining, the district court ordinarily should relinquish jurisdiction over the state-law claims against him. See Fields I, 672 F.3d at 518-19.

For the foregoing reasons, I respectfully concur in part and dissent in part.