Robert Pressley was the first black police officer in Wheeling, Illinois. For years he remained the only black employee of that village. Occasional unpleasant incidents marred his life, but most of his colleagues were friendly and the rest tractable. When Michael Haeger arrived as chief of police in February 1982, Pressley told him about problems he occasionally had with two officers, including the posting of racially offensive cartoons. Haeger did nothing to help and soon began a series of administrative actions against Pressley. A jury concluded that Haeger discriminated against Pressley on account of race and awarded $40,000 in compensatory damages, to which the district judge added more than $177,000 in attorneys' fees under 42 U.S.C. § 1988.

Ample evidence supports the jury's verdict. Haeger curtailed Pressley's scheduled salary increases despite the consistently excellent reviews he received from his superiors, citing unresolved complaints that citizens had lodged against Pressley. Similar complaints against white officers were disposed of quickly or disregarded at raise time. That the difference was attributable to Haeger's racial views is a natural inference of a comment he made to one of the village trustees soon after Haeger assumed office: "I'm going to have trouble with that nigger cop." That attitude made trouble inevitable.

Pressley often ate breakfast at the Granny Annie restaurant in Wheeling after working the night shift. Several officers told Haeger that Ruby Johnson, a married woman, sometimes joined Pressley, and that the two could be seen together in Johnson's van outside the restaurant. Haeger checked one morning in September 1982 and found the two in the van. Haeger usually dealt with actions that might lead the public to think poorly of the police by private counseling or reprimands; for Pressley he chose suspension without pay. To make the suspension stick Haeger had to press a formal charge; the village's counsel, with Haeger's assent, charged Pressley with engaging in sexual acts in the van. The local media had a field day, to the chagrin and dismay of both Pressley and Johnson. Pressley was exonerated after an administrative hearing. Haeger conceded at trial that he knew that Pressley and Johnson had not engaged in sexual relations. Just why he condoned a false charge he had a hard time explaining; the jury was entitled to infer that Haeger would not have subjected a white officer to such calumny.

Only one among Haeger's many assignments of error at trial requires discussion. Over Haeger's objection, the judge told the jury:

If it appears from the evidence in the case that a person had information which would lead a reasonably prudent person to make inquiry through which he would surely learn certain facts, then this person may be found to have had actual knowledge of those facts, the same as if he had made such inquiry and had actually learned such facts.

That is to say, the law will charge a person with notice and knowledge of whatever he would have learned, upon making such inquiry as it would have been reasonable to expect him to make under the circumstances.

This instruction should not have been given. Its first paragraph nods in the direction of an "ostrich" instruction, which tells a jury that it may infer knowledge from a combination of suspicion plus steps to avoid learning more. See United States v. Giovannetti, 919 F.2d 1223, 1226-29 (7th Cir.1990); United States v. Ramsey, 785 F.2d 184, 188-91 (7th Cir.1986). Suspicion plus deliberate avoidance is a form of real knowledge. Paragraph 1 is not a sound ostrich instruction, because it omits any reference to going out of one's way to avoid additional knowledge. Paragraph 1 turns out to be irrelevant anyway, because p 2 restates the rule (it begins, "[t]hat is to say") and so becomes the only text the jury need consider. And p 2 is a constructive knowledge instruction. It tells the jury to treat a person as knowing what he would have discovered by making prudent inquiries. Constructive knowledge is not a form of knowledge. It states conditions under which knowledge is irrelevant and ignorance no defense--as for example when the rule is designed to induce the person to inquire. United States v. Ross, 917 F.2d 997, 999-1001 (7th Cir.1990). When the rule of law makes consequences depend on actual knowledge, an instruction such as p 2 is impermissible. E.g., Contract Courier Services, Inc. v. Research and Special Programs Administration, 924 F.2d 112 (7th Cir.1991); United States v. Bader, 956 F.2d 708 (7th Cir.1992).

Racial discrimination is an intentional wrong. An empty head means no discrimination. There is no "constructive intent," and constructive knowledge does not show actual intent. Ignorance may be reprehensible, but not because it is racial discrimination. A supervisor who does not find out what is going on in the workplace should be sacked as incompetent, not lumped with bigots.

Knowledge plays a role in forming or ascertaining intent. Haeger could not discriminate against Pressley on account of race if he did not know Pressley's race. But Haeger obviously knew that Pressley is black and the other officers white. Haeger conceded knowing that Pressley and Johnson did not have sexual relations in the van. So just what bit of knowledge was the jury supposed to impute to Haeger using the method described in the instruction, and why would this knowledge show discriminatory intent? Early in the trial there was a dispute about whether Haeger knew about the cartoons and other racially derogatory material that one or more officers put in Pressley's mailbox and tacked on a stationhouse bulletin board. That knowledge would have been relevant to the question whether Haeger and the Village maintained a racially hostile environment. At the close of the case, however, Pressley withdrew his hostile-environment claim, and with it evaporated any rationale for an instruction about knowledge.

The knowledge instruction should have vanished with the hostile-environment claim because of the risk that the jury would confuse knowledge with intent, and hold Haeger liable for what he should have known rather than what he actually intended. Pressley's brief and oral argument evince such confusion. If lawyers cannot keep the two straight, how can we expect jurors to do so?

Realism about the role of instructions in a jury trial inverts the question: if lawyers can't distinguish knowledge from intent, why should we think that fiddling with what a judge says at the end of the trial would have much effect? Jurors consider the evidence and trial as a whole; appellate judges must guard against the error of assuming that jurors dissect instructions as if the sentences were frogs. Needham v. White Laboratories, Inc., 847 F.2d 355, 358-60 (7th Cir.1988). The trial lasted two weeks, followed by lengthy instructions. Most of the evidence and argument concerned Haeger's intent. And the instructions on intent were thorough and correct. The judge told the jury that it could not return a verdict for Pressley unless it found that Haeger imposed discipline or withheld raises because of Pressley's race. To impress the jury with the essence of the inquiry, the court added: "you are not to consider whether the discipline which was imposed upon the plaintiff was reasonable or rational. The defendant can discipline the plaintiff for any reason or no reason at all as long as the reason for the discipline was not the plaintiff's race." See Pollard v. Rea Magnet Wire Co., 824 F.2d 557, 559-61 (7th Cir.1987). Knowledge was a sidelight in both the trial and the instructions. Under the circumstances, the risk that the constructive-knowledge instruction would have led the jury to find "constructive intent" is small. The error did not affect Haeger's substantial rights. Fed.R.Civ.P. 61.

Each side appeals from the computation of attorneys' fees, which turned out to swamp the damages. Haeger's principal submission is that the judge should have reduced the award by 20% across the board in order to approximate the expense Pressley incurred in litigating eight theories that never made it to the jury--not only the hostile-environment claim but also a claim, against the Village itself, that Haeger's decisions represented official policy. Pressley gave up on the hostile-environment claim and others; at the close of the evidence the judge granted judgment in the Village's favor as a matter of law. Pressley thus achieved less than complete vindication and is not entitled to collect from Haeger the legal costs he incurred in charging up blind alleys. Hensley v. Eckerhart, 461 U.S. 424, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983). Of this, the district court wrote only: "Plaintiff prevailed on all claims that went to the jury. In addition, plaintiff reduced his own fee petition to account for all claims raised but not pursued." The court's first observation is irrelevant, and the second misleading. Pressley excluded from the petition claims that were "not pursued" in the sense that he dropped them before trial. He did not exclude time devoted to claims abandoned during trial, or the subject of adverse decision by the court. Pressley raised eleven claims. Two he dropped before trial, six went away during trial, and the jury decided the final three in his favor. The district judge did not discuss what to make of time devoted to the six. Hensley permits the court to award fees for losing arguments in support of prevailing claims, but not for losing claims; the court must decide which is which in this case. Lenard v. Argento, 808 F.2d 1242 (7th Cir.1987).

Three lawyers toiled a total of 1272 hours on Pressley's behalf. If this seems excessive in light of the stakes--as it did to the district judge, who trimmed the allowable hours to 1102--it was not excessive in relation to the need to overcome a defense that Haeger and the Village waged with 2041 hours of lawyers' time. Graham v. Sauk Prairie Police Commission, 915 F.2d 1085, 1109 (7th Cir.1990). Pressley does not ask us to restore the time the district court carved away, but he does contest the court's alteration of counsel's hourly rates. The court wrote:

Plaintiff requests an award based on an hourly rate of $182 per hour for each of his three lawyers. Defendant does not object to paying Burns at the rate of $182 per hour, nor do we. The fair hourly rate for Freedman and Bornstein is more difficult. Neither Freedman nor Bornstein were lead counsel at trial and thus do not deserve to be paid as such. However, during much of the pre-trial proceedings and during discovery Freedman and Bornstein were the lead counsel, and for that period deserve the top rate of $182 per hour. An exact calculation is impossible, but we find that the rate of $150 per hour for all of their time throughout this litigation to be a fair hourly rate.

The court computed the award by multiplying all of Burns's time by $182 per hour, and the other two lawyers' out-of-court time by $150 per hour. Freedman and Bornstein got nothing, not even the $150 per hour, for the trial itself in light of the judge's further finding that having more than one lawyer represent Pressley at trial was wasteful.

This outcome--$182 for Burns and nothing for the other two lawyers during trial, when Burns had the lead; and $182 for Burns and $150 for Freedman and Bornstein during the remainder of the case, when they had the lead--suggests that the district judge may not have carried out the calculation according to his own design. But there is a deeper problem. The judge assumed that he was searching for the "just" or "fair" price of legal services, what the lawyers "deserve." Perhaps, in a just world, first violins would earn more than second fiddles. Frequently, in this world, the two earn the same; sometimes second chairs earn more. Prevailing plaintiffs are entitled not to a "just" or "fair" price for legal services, but to the market price for legal services. Burlington v. Dague, --- U.S. ----, 112 S.Ct. 2638, 120 L.Ed.2d 449 (1992); Missouri v. Jenkins, 491 U.S. 274, 285, 109 S.Ct. 2463, 2470, 105 L.Ed.2d 229 (1989); Blum v. Stenson, 465 U.S. 886, 892-96, 104 S.Ct. 1541, 1545-48, 79 L.Ed.2d 891 (1984). "[I]t is not the function of judges in fee litigation to determine the equivalent of the medieval just price. It is to determine what the lawyer would receive if he were selling his services in the market rather than being paid by court order." In re Continental Illinois Securities Litigation, 962 F.2d 566, 568 (7th Cir.1992). See also Kurowski v. Krajewski, 848 F.2d 767, 776-77 (7th Cir.1988); Kirchoff v. Flynn, 786 F.2d 320 (7th Cir.1986); Bohen v. East Chicago, 666 F.Supp. 154 (N.D.Ind.1987).

All three lawyers submitted affidavits stating that $182 was the market rate for their time, and a fourth lawyer submitted an affidavit verifying that this is the prevailing rate in Chicago, where the case was litigated, for work of the type and quality. Although this is odd (why $182 rather than $180 or $185? why three lawyers with identical rates?), Haeger did not contest these affidavits and submitted no evidence on the point. Instead he argued that there ought to be a difference in the rates charged by leaders and followers. Fee litigation under § 1988 depends on what the market rate is rather than what litigants and judges think it ought to be. All of the evidence in this record shows that the market rate for these three lawyers' time was $182 per hour, and § 1988 requires the judge to use that rate. Nanetti v. University of Illinois, 944 F.2d 1416, 1418 (7th Cir.1991); Henry v. Webermeier, 738 F.2d 188, 193 (7th Cir.1984).

The judgment on the merits is affirmed. The award of attorneys' fees is vacated, and the case is remanded for further proceedings consistent with this opinion.