Anthony Riccardo, an inmate of the Centralia Correctional Center in Illinois, needed a new cellmate after his former cellmate complained about being housed with him. Normally that pairing would have been made by Centralia's placement office, but when the evening of May 28, 1997, arrived and some inmates remained unassigned after the regular placement officers had left, the task fell to Lt. Larry Rausch, who was serving the second shift. Rausch matched Riccardo with Juan Garcia, a pairing that should have lasted only until the placement officers on the day shift could review matters. Two days later Garcia sexually assaulted Riccardo, who sued Rausch under 42 U.S.C. § 1983. A jury concluded that Rausch had subjected Riccardo to cruel and unusual punishment and awarded $1.5 million in compensatory damages. The district court entered judgment on this verdict, and Rausch appeals.

His lead-off argument is that Riccardo failed to use his administrative remedies. If so, then 42 U.S.C. § 1997e(a), part of the Prison Litigation Reform Act, forecloses this suit even though Riccardo challenges a discrete incident and wants a form of relief — money damages — that the administrative process in Illinois does not provide. See Porter v. Nussle, 534 U.S. 516, 122 S. Ct. 983, 152 L. Ed. 2d 12 (2002); Booth v. Churner, 532 U.S. 731, 121 S. Ct. 1819, 149 L. Ed. 2d 958 (2001). Riccardo did file an administrative grievance, but Rausch contends it was too late (in February 1998, while Illinois sets a limit of six months) and asked the state to prosecute Garcia rather than do anything about Rausch and the classification system.

Prisoners must follow state rules about the time and content of grievances. See Pozo v. McCaughtry, 286 F.3d 1022 (7th Cir.2002); Strong v. David, 297 F.3d 646 (7th Cir.2002). Failure to do this means failure to use (and thus to exhaust) available remedies. Yet the state's administrative apparatus did not reject Riccardo's grievance as untimely; it accepted and denied the grievance on the merits. At the time of these events, Illinois permitted a filing after six months when the prisoner had good cause, see 20 Ill. Admin. Code § 504.810 (1997). The official handling the grievance must have found good cause; anyway, we held in Pozo that, when a state treats a filing as timely and resolves it on the merits, the federal judiciary will not second-guess that action, for the grievance has served its function of alerting the state and inviting corrective action. 286 F.3d at 1025.

As for the content of this grievance: true enough, its main objective was to have Garcia prosecuted. (Riccardo deemed inadequate Garcia's punishment within the prison system.) But it also at least hinted at problems in prison administration. Riccardo wrote: "[T]he administration don't [sic] do there [sic] job. [A sexual assault] should've never [sic] happen again." This language is ambiguous. There are two principal ways to reduce the number of sexual assaults in prison: better steps ex ante to separate potential aggressors from potential victims; and harsher penalties ex post in order to deter future assaults. Riccardo did not distinguish between the two, and a prison administration receiving such a grievance should have considered both. Illinois has not adopted any rule governing the level of detail required of prisoners' grievances. "When the administrative rulebook is silent, a grievance suffices if it alerts the prison to the nature of the wrong for which redress is sought. As in a notice-pleading system, the grievant need not lay out the facts, articulate legal theories, or demand particular relief. All the grievance need do is object intelligibly to some asserted shortcoming." Strong, 297 F.3d at 650. The document that Riccardo filed is at the border of intelligibility; it is hard to imagine much less that a prisoner could do and still alert the prison; yet this grievance did complain that Garcia had committed a rape and that "the administration don't do there job." A generous construction of this grievance would have induced the prison to consider the possibility that the guards could have prevented this assault. So we conclude that Riccardo exhausted the administrative process. If Illinois wants grievances to be more detailed, it must adopt appropriate regulations and inform prisoners what is required of them. Riccardo is entitled to a decision on the merits of his constitutional claim.

Because Riccardo prevailed at trial, we recap the facts in the light most favorable to his position. Riccardo was anally raped by his cellmate at Cook County Jail, shortly after his conviction for aggravated assault. When he arrived at Centralia Correctional Center in November 1996, he told the prison psychologist that he did not feel safe. After a stint in segregation for violating prison rules, Riccardo declined to return to the general population. He told guards that a cellmate in the segregation unit had stolen some of his property and objected to spending another day with that cellmate. He believed that the responsible inmate belonged to the Latin Kings gang and that the Latin Kings may have been preparing to kill him — though he did not say why he believed this. But Centralia allows inmates to veto housing with persons they declare to be enemies, so the prison found Riccardo a new cellmate. When, after a few days, that cellmate objected to spending more time with Riccardo, another pairing was required. (The segregation unit was too crowded to permit Riccardo a cell of his own.)

During the afternoon of May 28, Garcia had offered to help Riccardo retrieve his stolen property. Riccardo took this as an ill omen rather than as a genuine offer of assistance and told Lt. Alemond that he feared for his life if celled with Garcia. Although Alemond said that he would "take care of it," he did nothing — he did not either find a cellmate for Riccardo or alert Lt. Rausch, Alemond's replacement on the next shift. About 9:30 that evening, Rausch brought Garcia to Riccardo's cell and told him that Garcia was his new cellmate. Before the cells were locked for the night, Riccardo sought out Rausch in private and told him that he believed that the Latin Kings had a "hit" out on him, and that he feared for his life if celled with Garcia. Rausch replied that there was no place else to put Garcia (or Riccardo) that evening, and that he could not refuse housing while in segregation. Rausch then brought Riccardo and Garcia back together and asked each, in turn, if he had a problem with the other. Riccardo shook his head in the negative. Rausch took that as agreement to the assignment. That was Riccardo's last contact with Rausch. As we have mentioned, nothing untoward happened that evening or the next, but during the evening of May 30 Garcia compelled Riccardo to perform oral sex. The record does not suggest that this assault had any connection to the Latin Kings. During the time between assignment and assault, Riccardo did not ask for a different cellmate (though he did file two grievances on May 29 about other matters). Circumstances brought out at trial suggest that other guards should have recognized on May 30 that problems had developed between Riccardo and Garcia; their failure to intervene may be culpable but cannot be attributed to Rausch, whose liability depends exclusively on his actions the evening of May 28.

Rausch did not assault Riccardo and is not vicariously liable for Garcia's crime. Like other guards, however, Rausch was required to refrain from placing Riccardo in harm's way gratuitously. The qualification "gratuitously" is important, because prisons are dangerous places. Inmates get there by violent acts, and many prisoners have a propensity to commit more. Guards cannot turn away persons committed by the courts; nor do individual guards have any control over crowding and other systemic circumstances. All that can be expected is that guards act responsibly under the circumstances that confront them. See McGill v. Duckworth, 944 F.2d 344 (7th Cir.1991). A guard may be responsible without being credulous. Some prisoners are manipulative and cry "wolf" in an effort to have a cell to themselves or choose a favored cellmate. Other prisoners perceive specters in every shadow, even though their fears are unsupported. (There is, for example, no reason to think that the Latin Kings ever had it in for Riccardo. He did not belong to a rival gang, and there is no history of violent or overtly hostile encounters between Riccardo and any gang member.) Guards therefore must discriminate between serious risks of harm and feigned or imagined ones, which is not an easy task given the brief time and scant information available to make each of the many decisions that fill every day's work.

The eighth amendment does not demand that guards perform this task flawlessly. It does not even hold them to the negligence standard. Liability is possible, instead, only when a guard is deliberately indifferent to a substantial risk of serious harm. See Helling v. McKinney, 509 U.S. 25, 113 S. Ct. 2475, 125 L. Ed. 2d 22 (1993); Wilson v. Seiter, 501 U.S. 294, 111 S. Ct. 2321, 115 L. Ed. 2d 271 (1991); Estelle v. Gamble, 429 U.S. 97, 97 S. Ct. 285, 50 L. Ed. 2d 251 (1976). "Deliberate indifference" means subjective awareness. See Farmer v. Brennan, 511 U.S. 825, 114 S. Ct. 1970, 128 L. Ed. 2d 811 (1994). It is not enough, the Court held in Farmer, that the guard ought to have recognized the risk. Instead, "the official must both be aware of facts from which the inference could be drawn that a substantial risk of serious harm exists, and he must also draw the inference." Id. at 837, 114 S. Ct. 1970.

Rausch contends that Riccardo did not face a "substantial risk of serious harm" from Garcia on the evening of May 28, 1997; that, if such a risk was present, Rausch did not appreciate its existence; and that at all events he is entitled to qualified immunity because reasonable officers would not necessarily have understood that the law clearly required Riccardo and Garcia to be in separate cells. The first two matters (the objective and subjective components of the eighth amendment) are for the jury in the first instance, with appellate review limited to the question whether any reasonable juror could have found that the requisite level of risk existed, and that Rausch knew it. Immunity, however, is a matter of law for the court, to be decided without deference to the jury's resolution — and preferably before the case goes to the jury. See Saucier v. Katz, 533 U.S. 194, 202, 121 S. Ct. 2151, 150 L. Ed. 2d 272 (2001); Anderson v. Creighton, 483 U.S. 635, 107 S. Ct. 3034, 97 L. Ed. 2d 523 (1987). The district court brushed aside Rausch's invocation of immunity, writing that a guard cannot benefit from immunity if the action taken was not a reasonable response to a risk actually foreseen. That approach, which merges immunity and the merits, is incompatible with Saucier and its predecessors. See 533 U.S. at 203-06, 121 S. Ct. 2151. Immunity protects officials who act at the "hazy border" (id. at 206, 121 S. Ct. 2151) between the lawful and the forbidden. That Rausch may have overstepped the line does not mean that every reasonable officer would have been bound to know that Rausch acted improperly. We need not pursue the immunity defense, however, because Saucier calls on appellate courts to address the merits first, see id. at 201, 121 S. Ct. 2151, and Rausch is entitled to prevail outright: no reasonable juror could have concluded, on this record, that Rausch actually recognized that placing Garcia and Riccardo together exposed Riccardo to substantial risk.

Now it might seem that Rausch had to appreciate the risk, because (a) Riccardo claimed to fear for his life if celled with any member of the Latin Kings, and (b) Garcia in fact harmed Riccardo. One problem with relying on how things turned out to show knowledge of risk beforehand is that Garcia did not act for the Latin Kings; he told Riccardo that he was fulfilling a personal fantasy, and Riccardo believed this explanation. Thus the risk that Riccardo professed to fear (a "hit") did not come to pass. Even under the law of negligence, this is an important distinction. If a school district entrusts a bus to a driver with a bad drinking record, and the tipsy driver runs the bus off the road while speeding, the school district is liable; but if instead there is an accident for which the driver is not at fault (a tree falls on bus), or the driver collapses of a heart attack while on the road, the district is not liable, because that was not the type of risk created or increased by the negligent conduct even though hiring this particular driver was in the causal chain. See, e.g., Berry v. Sugar Notch Borough, 191 Pa. 345, 43 A. 240 (1899). The risk from which Riccardo sought protection was not realized; for all this record shows, the (objectively evaluated) risk to Riccardo of sharing a cell with Garcia was no greater than the risk of sharing a cell with any other prisoner.

As for Rausch's subjective assessment: though Riccardo initially asserted mortal fear, when later asked whether he "had a problem" with Garcia he shook his head to give a negative answer. Rausch then had to decide which statement to believe. Riccardo argues, and the jury evidently concluded, that Rausch should have believed the first statement, communicated in private, rather than the second, communicated in Garcia's presence. A rational jury could have thought that guards should give priority to statements made in private. (Rausch testified that, if Riccardo had claimed to "have a problem" with Garcia, they would have been separated; but Riccardo might have feared the consequences in a later encounter in the prison's general population.) Still, what Rausch should have believed is not the right question; we need to know what he did believe. No reasonable jury could have found, in light of Riccardo's denial of "a problem" with Garcia and Rausch's decision to act accordingly, that Rausch subjectively appreciated that his action would expose Riccardo to a substantial risk of serious harm.

As we have already explained, prisoners may object to potential cellmates in an effort to manipulate assignments, or out of ignorance; thus although a protest may demonstrate risk it does not necessarily do so. The Constitution does not oblige guards to believe whatever inmates say. How does a reasonable guard separate fact from fiction? Rausch knew when making the assignment at least two things beyond Riccardo's contradictory assertions. First, Rausch knew that Garcia was himself in segregation for protection from the Latin Kings (or at least a subset of them). Perhaps Garcia was manipulating the system himself, falsely asserting to fear the Latin Kings so that he could serve as their assassin; but at least at first cut Garcia could not be deemed a gang enforcer (and, as we learned ex post, his attack on Riccardo was neither a "hit" nor gang related). Second, Rausch knew that Garcia had a clean record in prison. He had not been disciplined for acts of violence (let alone for sexual assault). That makes it reasonable for Rausch to have deemed Riccardo's initial protestation unjustified. It is not as if Rausch housed Riccardo with a known sexual predator.

Lest this observation be thought to leave prisoners without protection from assaults by others with "clean" records, we add that there may be other ways to show both an objectively serious risk and the guards' knowledge of that risk. For example, Riccardo might have attempted to demonstrate that there is a strong correlation between prisoners' professions of fear and actual violence. How many murders (or homosexual assaults) occur in Centralia (or the Illinois prison system) per hundred inmate-years of custody? How many violent events were preceded by requests for protection? How many requests for protection were dishonored, yet nothing untoward happened? Data along these lines would have enabled a jury (and the court) to evaluate actual risks even though Riccardo was unable to show that Rausch should have deemed Garcia to present an especial risk. If violence is common at Centralia, and inmates have good track records in identifying potential aggressors, then guards who do not have their heads in the sand must actually (that is, subjectively) understand the risk an inmate faces when a protest is disregarded. But if violence is rare, or if there is poor correlation between inmates' alarums and subsequent violence, then Riccardo's initial protest would not have provided Rausch with actual knowledge of an impending assault. The record does not contain any evidence along these lines. That leaves only Riccardo's say-so, and for reasons we have already given a prisoner's bare assertion is not enough to make the guard subjectively aware of a risk, if the objective indicators do not substantiate the inmate's assertion.

Rausch also was entitled to believe that his assignment of Garcia and Riccardo to share a cell would last for one night only. During the next day shift the placement office, armed with better information, was supposed to make a fresh evaluation and, if appropriate, a new assignment. Apparently that did not happen; the record does not show why. (Maybe it did happen and the staff approved Rausch's action.) If Rausch knew that the staff charged with this responsibility routinely failed to carry it out, then he might have been obliged to take additional precautions (such as separate interviews of Garcia and Riccardo to probe these issues more deeply) before making an assignment. Rausch himself testified that separate interviews would have been better practice, but the Constitution does not enforce all "better practices"; this is one respect in which the eighth amendment standard differs from the negligence standard. But Riccardo does not contend, and the record does not demonstrate, that disregard of the classification system at Centralia was so common that Rausch was bound to know that his assignment would last indefinitely. Nor was Rausch bound to foresee that, if Riccardo was in genuine fear, he would neglect to complain the next day, when he readily could have done so. (Recall that Riccardo filed two grievances on May 29 about other subjects.) Riccardo testified that he was too terrified to protest and was put off by Rausch's assertion that prisoners in segregation can't refuse assignments; yet grievances are confidential (so Garcia would not have known), and prisoners often appeal over the head of a guard who has told them that something can't or won't be done. Riccardo had already objected to, and obtained the removal of, at least one cellmate assigned to him in segregation. At all events, the question on the table is what Rausch knew (or deliberately avoided learning) on May 28; and there is no evidence that Rausch subjectively believed that Riccardo would fail to use his opportunity to seek further review the next day.

Illinois is free, if it wishes, to give prisoners veto power over the identity of their cellmates. But the eighth amendment does not do so of its own force, and prisoners cannot use the Constitution to achieve this control indirectly by making unsubstantiated assertions. The constitutional question is not what Riccardo (initially) said, but what Rausch actually believed. Some prisoners are manipulative, some are mistaken, and some are not only honest but also accurate. The Constitution does not oblige guards to assume (on pain of absolute liability if an assault later occurs) that all prisoners are in the third category; Farmer articulates a different, and more demanding, standard, preserving room for both independent judgment and honest mistake on the guards' part. This record does not permit a reasonable jury to find that Rausch knew or deliberately disregarded the fact that his actions subjected Riccardo to a substantial risk off serious harm, so the judgment is

REVERSED.

WILLIAMS, Circuit Judge, dissenting.

While I agree that Anthony Riccardo did in fact exhaust his administrative remedies, I disagree with the majority's decision to overturn the judgment in this action as a reasonable jury had ample evidence to find that Lieutenant Larry Rausch was deliberately indifferent to the substantial risk of harm Riccardo faced by being celled with Juan Garcia. Therefore, I respectfully dissent.

On May 30, 1997, while celled with Juan Garcia, a known member of the Latin Kings, Riccardo's head was forcibly shaven by Garcia such that Riccardo was "bleeding pretty bad." Tr. I at 81-82. Garcia then attempted to sodomize Riccardo; however, he was able to resist. Tr. II at 50. After Riccardo resisted, Garcia ejaculated on Riccardo's feet. Tr. I at 83-84. Riccardo was then forced to perform oral sex on Garcia for 15 to 20 minutes. Id. The assault ended when an officer walked by the cell. As the majority notes, the events which give rise to Lt. Rausch's liability are limited to the happenings on the evening of May 28, 1997, when Lt. Rausch replaced Lt. Alemond1 as the lieutenant in charge of the segregation and receiving units at Centralia prison.2

On appeal, Lt. Rausch contends, and the majority agrees, that the evidence introduced at trial was legally insufficient to support a finding of liability under the Eighth Amendment. To sustain overturning a jury verdict, the record must demonstrate no "legally sufficient evidentiary basis for a reasonable jury to find for the non-moving party." Payne v. Milwaukee County, 146 F.3d 430, 432 (7th Cir.1998). While undertaking this assessment, we analyze the "the totality of the evidence," Sheehan v. Donlen Corp., 173 F.3d 1039, 1043 (7th Cir.1999), and are obliged to leave the judgment undisturbed unless the moving party can show that "no rational jury could have brought in a verdict against him." EEOC v. G-K-G, Inc., 39 F.3d 740, 745 (7th Cir.1994). It is not within the province of the appellate courts to "reweigh the evidence." Knox v. State of Indiana, 93 F.3d 1327, 1332 (7th Cir. 1996). Lastly, and most importantly, all reasonable inferences must be analyzed in the light most favorable to Riccardo as the non-moving party. Sheehan, 173 F.3d at 1044.

In Farmer v. Brennan, 511 U.S. 825, 834, 114 S. Ct. 1970, 128 L. Ed. 2d 811 (1994) the Court bifurcated the standard for Eighth Amendment liability into an objective element and a subjective element. First, the potential harm to the inmate must be objectively serious. Id. (citing Wilson v. Seiter, 501 U.S. 294, 298, 111 S. Ct. 2321, 115 L. Ed. 2d 271 (1991)). Second, under the subjective prong, the prison official must "deliberately disregard" this potential harm by being "aware of facts from which the inference could be drawn that a substantial risk of serious harm exists, and [also] draw[ing] the inference." Id. at 838, 114 S. Ct. 1970.

The second inquiry is a question of fact, sustainable through circumstantial evidence, id. at 842, 114 S. Ct. 1970, mandating an "inquiry into a prison official's state of mind." Id. at 837, 114 S. Ct. 1970 (quoting Wilson, 501 U.S. at 299, 111 S. Ct. 2321). "A prisoner normally proves actual knowledge of impending harm by showing that he complained to prison officials about a specific threat to his safety." McGill v. Duckworth, 944 F.2d 344, 349 (7th Cir. 1991) (reasoning that the scienter requirement is satisfied when a prison guard, "[s]uspect[s] something is true but shut[s][his] eyes for fear of what [he] will learn" or "[goes] out of [his] way to avoid acquiring unwelcomed knowledge").

The Supreme Court also cautioned that an "Eighth Amendment claimant need not show that a prison official acted or failed to act believing that harm actually would befall an inmate; it is enough that the official acted or failed to act despite his knowledge of a substantial risk of serious harm." Farmer, 511 U.S. at 842, 114 S. Ct. 1970. Likewise, a claimant need not prove that a prison official was aware of the specific type of harm which befell the prisoner, only that the prison official was aware that a substantial risk of some type of danger existed. See Haley v. Gross, 86 F.3d 630, 643 n. 33 (7th Cir.1996) (applying Farmer, 511 U.S. at 843, 114 S. Ct. 1970) (upholding jury verdict for $1.65 million based on a finding that prison guards were deliberately indifferent by failing to respond to a prisoner's repeated request to be removed from cell when his cellmate set fire to cell causing the cellmate's death and plaintiff's severe burns). Thus, it was Riccardo's burden to show that Lt. Rausch actually knew that there was a substantial risk that Garcia would harm Riccardo. Id. However, Lt. Rausch would be shielded from liability if no objectively serious risk existed, he was unaware of the impending risk, McGill, 944 F.2d at 349, or he took reasonable steps to abate it, whether successful or not, Farmer, 511 U.S. at 844, 114 S. Ct. 1970.

Admittedly, there is evidence in the record to support a finding that Garcia did not objectively pose a substantial threat to Riccardo — namely, the fact that Garcia was also placed in segregation for "enemy protection," allegedly from the Latin Kings, and that Garcia had no history of sexual assault. However, that is not the standard by which this case should be reviewed. The standard is whether there exists a legally sufficient evidentiary basis for a reasonable jury to find in favor of Riccardo. Payne, 146 F.3d at 432.

A jury could have reasonably believed that Lt. Rausch was deliberately indifferent to the substantial risk of harm to Riccardo. It is undisputed that Garcia was a member of the Latin Kings. In their first interaction, Riccardo privately pulled Lt. Rausch aside and expressed his fear of being celled with Garcia. Thus, Riccardo has presented sufficient evidence to support the finding that Lt. Rausch was made aware of the potential harm. See McGill, 944 F.2d at 349. The jury could have further found that Lt. Rausch's decision to question Riccardo in front of Garcia was not a reasonable way to abate the potential danger to Riccardo. Moreover, Lt. Rausch admitted that "[i]f [Riccardo] would have told me he feared for his life, if he refused housing or thought there was a threat to his safety[,] he would not have been placed — they would not have been placed together." Tr. III at 73. Thus, a jury could have determined that Lt. Rausch's admission, coupled with his prior statements to Riccardo on May 28 that Riccardo could not turn down a housing assignment in segregation, and that Riccardo could not be moved to another cell due to a lack of space, amounted to Lt. Rausch "deliberately" avoiding learning that Riccardo was in danger.3 Such an analysis avoids the use of ex post occurrences, such as the fact that Garcia did in fact assault Riccardo, to sustain a finding of deliberate indifference. Moreover, the fact that Lt. Rausch is able to point to evidence in the record to support his position does not mandate reversal of the jury's verdict. Our sole duty as an appellate court is to analyze whether the record supports the jury's determination. It is not our function to reweigh the evidence. Knox, 93 F.3d at 1332.

I am further troubled by the majority's reliance on Riccardo's second statement to Lt. Rausch (made in front of Garcia) to sustain overturning the jury's verdict. As it stands, the deliberate indifference inquiry is an inherently factual determination, Farmer, 511 U.S. at 842, 114 S. Ct. 1970, which requires an "inquiry into a prison official's state of mind," id. at 837, 114 S. Ct. 1970.4 Thus, whether this second statement is used to support the assertion that there was no "objective" risk to Riccardo or that Lt. Rausch was stripped of his "subjective" knowledge of the harm, it is clear that Lt. Rausch's credibility and sincerity are integral components to the usefulness of this interaction. In essence, the majority accepts Lt. Rausch's assertion that his second discussion with Riccardo in front of Garcia was a sincere investigation of the potential risk to Riccardo. However, the jury found otherwise. Further, when asked to review the defendant's Rule 50 motion, the district court aptly stated,

[T]here is ample evidence from which to conclude that Rausch's attempt to ascertain the seriousness of the threat was mere pretense, and that because he did not want to go to the extra effort to find different accommodations for Garcia, he recklessly disregarded what he knew to be a dangerous situation. That decision to essentially disregard the threat is where liability lies. A jury could have reasonably inferred that Rausch crossed the line from gross negligence to deliberate indifference based on the ludicrousness of "asking" each inmate if he had a problem with the other. Credibility had to have been the key to the jury's analysis, thus the Court cannot interject its own credibility determinations; and if it could, having observed both parties' testimony, it may well have reached the same conclusion as the jury. [...]

[A] prison official will only be freed from liability if he responded reasonably to the risk. As mentioned above, Rausch's method of questioning could be perceived as deliberately forcing plaintiff to make a Hobson's choice.

Riccardo v. Rausch, No. 99-CV-372-CJP, at 15 (S.D.Ill. Mar. 7, 2002) (order denying F.R.C.P. Rule 50(b) motion) (citations omitted) (emphasis in original). By taking Lt. Rausch at his word, the majority's decision has the effect of immunizing prison officials from liability based on potentially unreasonable or contrived actions, and sanctions Lt. Rausch's admittedly unreasonable behavior.

In an attempt to break the causal link between Lt. Rausch's actions and the harm to Riccardo, the majority asserts that "the risk that Riccardo professed to fear (a `hit') did not come to pass." Majority opinion at 526. The inquiry, however, is not whether "a hit" was actually put out on Riccardo as that would constitute the sort of impermissible ex post determination eschewed by the majority. Rather, the focus is solely on whether Rausch was made aware that a substantial risk of some type of danger existed prior to the actual event. See Haley, 86 F.3d at 643 n. 33. Under the majority's analysis are we to presume that Riccardo's pleas for protection would only guard against murder or physical beating? For the purposes of an Eighth Amendment inquiry, in the prison context, I find no real distinction between "a hit" and a sexual assault.

The majority also frees Lt. Rausch of liability based on the assertion that Lt. Rausch was under no duty to foresee that Riccardo would not complain to other prison officials between May 28, the date of Lt. Rausch and Riccardo's interaction, and May 30, when the assault actually took place. The record reveals, however, that Garcia closely watched Riccardo's actions impeding Riccardo's ability to have a private conversation with prison guards outside of Garcia's presence. Tr. I at 76-82; Tr. II at 42-43. Further, when Riccardo attempted to alert prison officials, Garcia responded with escalating violence. Tr. II at 46-47. In light of Riccardo's reasonable belief that he could not refuse his housing assignment and that there was no other available cell, see note 3, supra, Riccardo did not realistically have the ability to complain to other guards without alerting Garcia and incurring his wrath.

The majority's decision to question the adequacy of Riccardo's pleas for protection by requiring evidence concerning the overall number of sexual assaults at Centralia prison is also curious. See Lewis, 107 F.3d at 556 (Flaum, C.J., concurring) ("[T]he majority's emphasis upon the adequacy of Lewis's pleas for protection strikes me as inappropriate.") The prison recognizes that "some prisoners are manipulative and cry `wolf'," majority opinion at 525, and has created a procedure to deal with this recurring possibility. A prisoner's request for a cell transfer is always honored if the prisoner alleges a fear for his personal safety, Tr. I at 44, but the prison deals with potential frivolity by treating every request as a potential disciplinary violation. Id.5 Thus, the prison has created a procedure whereby a prisoner is moved first, and questions concerning the sincerity of the request are asked later. In addition, the inquiry under the Eighth Amendment is an individualized one, i.e., Riccardo was required to prove, based on the individual facts of his case, that he was subjected to an objectively serious harm and that Lt. Rausch was "aware of facts from which the inference could be drawn that a substantial risk of serious harm exist[ed]." Farmer, 511 U.S. at 838, 114 S. Ct. 1970.

Finally, though unpursued by the majority, defendant argues that should a constitutional violation be found, he is nevertheless entitled to qualified immunity. According to Rausch, previous case law must show that "no reasonable prison official would have believed it was constitutional either to cell an inmate with someone who gave conflicting answers when asked (in the other inmate's presence) if he had a problem with that inmate, or to rely on a policy requiring review of all placement decisions for suitability within hours." Appellant's Brief at 33. However, the Supreme Court in Hope v. Pelzer, expressly rejected the notion that in order for a right to be "clearly established" previous case law must contain facts which are "materially similar" to the facts contained in the underlying action. 536 U.S. 730, 739, 122 S. Ct. 2508, 153 L. Ed. 2d 666 (2002); see also Burgess v. Lowery, 201 F.3d 942, 944-45 (7th Cir.2000). Rather, the Court focused on whether the prior case law would place officers on notice that their conduct is unlawful. Id. It is clear that Farmer put prison guards on notice that they have a duty, under the Eighth Amendment, to protect inmates from being gratuitously beaten or raped by other inmates. See id. at 833, 114 S. Ct. 1970; see also Haley, 86 F.3d at 646 (rejecting qualified immunity defense in light of Farmer decision which further elucidated "deliberate indifference" standard).

In light of the aforementioned, I must agree with the trial court that a reasonable jury had ample evidence to sustain this verdict and thus I respectfully dissent.

Before FLAUM, Chief Judge, and POSNER, EASTERBROOK, RIPPLE, MANION, KANNE, ROVNER, DIANE P. WOOD, EVANS, and WILLIAMS, Circuit Judges.^†

Plaintiff-appellee filed a petition for rehearing and rehearing en banc on March 12, 2004. In response to this petition, the panel has amended its opinion; the amendments are reflected in the immediately preceding revised opinion. A majority of the judges on the panel voted to deny rehearing. A judge called for a vote on the petition for rehearing en banc, but a majority of the active judges did not favor rehearing en banc. Accordingly, the petition is denied.

RIPPLE, Circuit Judge, with whom ROVNER, DIANE P. WOOD and WILLIAMS, Circuit Judges, join, dissenting from the denial of rehearing en banc.

Today, the court allows to stand the decision of a panel majority that imposes on prison inmates a new and impossibly high standard of proof for establishing deliberate indifference in prison condition cases. The panel majority strongly suggests that, in order to show deliberate indifference, a prisoner must not only identify with particularity the harm he fears but also bolster his own account with a special showing containing material such as statistical evidence of a "strong correlation between prisoners' professions and actual violence." Riccardo v. Rausch, at 527. Neither of these requirements find support in circuit or Supreme Court precedent. Consequently, the panel majority's imposition of these requirements not only does violence to our Eighth Amendment jurisprudence, but it also effectively amends, without any legislative mandate, the standard for judgment as a matter of law under Federal Rule of Civil Procedure 50 for cases brought by prisoners.

Mr. Riccardo told Lt. Rausch that the Latin Kings had put out a "hit" on him and that placement in a cell with Garcia, a known Latin King, would endanger his life. Two days later, Mr. Riccardo was forced to perform sexual acts with Garcia. The panel majority nevertheless concludes that this evidence is not sufficient to establish that Lt. Rausch should have appreciated a serious risk to Mr. Riccardo because the "risk from which Riccardo sought protection was not realized." Id. at 526. In short, in the panel majority's view, because Mr. Riccardo was raped, but not killed, Lt. Rausch cannot be held liable for his failure to respond to Mr. Riccardo's fears.

The Supreme Court in Farmer v. Brennan, 511 U.S. 825, 114 S. Ct. 1970, 128 L. Ed. 2d 811 (1994), eschewed this sort of distinction. The Court made clear that a prison official cannot escape

liability for deliberate indifference by showing that, while he was aware of an obvious, substantial risk to inmate safety, he did not know that the complainant was especially likely to be assaulted by the specific prisoner who eventually committed the assault. The question under the Eighth Amendment is whether prison officials, acting with deliberate indifference, exposed a prisoner to a sufficiently substantial "risk of serious damage to his future health," Helling [v. McKinney, 509 U.S. 25, 35, 113 S. Ct. 2475, 125 L. Ed. 2d 22 (1993),] and it does not matter whether the risk comes from a single source or multiple sources, any more than it matters whether a prisoner faces an excessive risk of attack for reasons personal to him or because all prisoners in his situation face such a risk.

Id. at 843, 114 S. Ct. 1970 (emphasis added). Just as Mr. Riccardo did not need to identify the member of the Latin Kings who would mete out the retaliatory action that he feared, he did not need to identify the specific sort of bodily harm that would be inflicted upon him. Indeed, our case law makes crystal clear that such precision with respect to identifying risks is not required:

While there must be some link between the risk of which the official was aware and the harm that actually occurred — as it would be unfair to hold officials liable for risks they could not have anticipated simply because they ignored other unrelated risks — prison officials need not be specifically aware of the precise risk that unfolds.

Haley v. Gross, 86 F.3d 630, 643 n. 33 (7th Cir.1996). Mr. Riccardo feared reprisal by the Latin Kings. Garcia, a member of that gang, committed an act of physical violence upon Mr. Riccardo. The fact that Mr. Riccardo did not suffer a worse fate does not negate the seriousness or the validity of the threat to Mr. Riccardo that was communicated to Lt. Rausch.

The panel majority further opines that Mr. Riccardo might have overcome the infirmity in his case by showing

that there is a strong correlation between prisoners' professions of fear and actual violence. How many murders (or homosexual assaults) occur in Centralia (or the Illinois prison system) per hundred inmate-years of custody? How many violent events were preceded by requests for protection? How many requests for protection were dishonored, yet nothing untoward happened? Data along these lines would have enabled a jury (and the court) to evaluate actual risks....

Riccardo, at 527. Neither this court, other courts of appeals, nor the Supreme Court ever has required a showing that expressed fears and violence were related in some statistically significant way. It always has been sufficient that the prisoner articulated his fear, that the prison official believed the inmate, but that the official failed to take reasonable actions to protect the prisoner from harm. We often have stated that

[i]f "the circumstances suggest that the defendant-official being sued had been exposed to information concerning the risk and thus `must have known' about it, then such evidence could be sufficient to permit a trier of fact to find that the defendant-official had actual knowledge of the risk."

Sanville v. McCaughtry, 266 F.3d 724, 737 (7th Cir.2001) (quoting Farmer, 511 U.S. at 842-43, 114 S. Ct. 1970; emphasis added). The question of whether the prison official had the "requisite knowledge is a question of fact" reserved for the jury. Id. Requiring the sort of statistical evidence suggested by the panel majority in order to establish the existence of such a risk imposes on inmates a practical burden that is impossible for prisoners to meet. Prisoners certainly should not be relieved of meeting their burden of proof, but, like other litigants, they ought to be permitted to come forward with any evidence, direct or circumstantial, that is probative of the knowledge of prison officials concerning threats reported to them. The judicial inquiry ought not be whether there was statistical evidence from which a jury could conclude that the threat was "real" or should be taken seriously, but whether the record contains any evidence from which the jury reasonably could conclude that the prison official knew that the prisoner was subject to a risk of serious harm. Moreover, there indeed may be times and circumstances when the representation of the prisoner will be a sufficient basis to require action on the part of the prison official. I see no reason for a categorical rule that, under no circumstances, can such a representation be sufficient.

There is no question that Mr. Riccardo articulated to Lt. Rausch his fear of physical violence if celled with a member of the Latin Kings. In violation of Centralia's own policy — one presumably founded on a recognized correlation between complaints and violence — Lt. Rausch did not move Mr. Riccardo to another cell; instead, Lt. Rausch forced Mr. Riccardo to articulate any concerns in the presence of the very source of Mr. Riccardo's fears, Garcia. As the district court determined in rejecting the defendant's Rule 50 motion, there was

ample evidence from which to conclude that Rausch's attempt to ascertain the seriousness of the threat was a mere pretense, and that because he did not want to go to the extra effort to find different accommodations for Garcia, he recklessly disregarded what he knew to be a dangerous situation. That decision to essentially disregard the threat is where liability lies. A jury could have reasonably inferred that Rausch crossed the line from gross negligence to deliberate indifference based on the ludicrousness of "asking" each inmate if he had a problem with the other. Credibility had to have been the key to the jury's analysis, this Court cannot interject its own credibility determinations; and if it could, having observed both parties' testimony, it may very well have reached the same conclusion as the jury.

R.64 at 15-16.

The panel majority makes clear its disagreement with the jury's view of the evidence. The evidence in favor of Mr. Riccardo may not be overwhelming, but it is certainly not legally insufficient. When the evidence is viewed in the light most favorable to Mr. Riccardo, as it must be, a jury could reasonably conclude that Lt. Rausch knew of a serious risk of harm to Mr. Riccardo but failed to take reasonable steps to prevent it. The panel majority may have come to a different conclusion if it had sat as the trier of fact, but its disagreement in that respect is certainly not an appropriate occasion for revising Eighth Amendment and Rule 50 standards.

The panel majority opinion not only deprives Mr. Riccardo of his right to a jury trial, it also takes a very significant step in depriving individuals incarcerated in this circuit of a realistic opportunity to meet their burden of proof in these cases. I therefore respectfully dissent from the denial of rehearing en banc.

WILLIAMS, Circuit Judge, with whom RIPPLE, ROVNER, and DIANE P. WOOD, Circuit Judges, join in dissenting from the denial of rehearing en banc.

The majority's opinion has incorrectly resolved and unjustifiably reframed both the Eighth Amendment standard for deliberate indifference as well as the Rule 50 standard to set aside a jury verdict.

With regard to the Eighth Amendment inquiry, the majority opinion highlights a major gap in the case law. Particularly, what is required to prove the objective prong of the deliberate indifference standard? The Court in Farmer v. Brennan, 511 U.S. 825, 834 n. 3, 114 S. Ct. 1970, 128 L. Ed. 2d 811 (1994) did not address this issue. On page 526-27, the majority asks a plethora of questions which go to this issue, i.e., the overall threat of violence at the prison. Perhaps the dearth of case law on this point is due in part to the general knowledge that prisons are dangerous places where rape and assault occur frequently and therefore threats, such as the one in this case, must be handled with more caution than exhibited by Rausch. See Prison Rape Elimination Act of 2003, 42 U.S.C. § 15601-02 (2004) ("The purpose of this chapter is to: (1) establish a zero-tolerance standard for the incidence of prison rape in prisons in the United States. (6) increase the accountability of prison officials who fail to detect, prevent, reduce, and punish prison rape.") ("Congress makes the following findings: (2) Insufficient research has been conducted and insufficient data reported on the extent of prison rape.... Many inmates have suffered repeated assaults. (13) The high incidence of sexual assault within prisons involves actual and potential violations of the United States Constitution."). Centralia Prison's policy, to move prisoners first and discipline them later for baseless requests, is further proof of the objective danger. Again, evidence of the prison's policy does not hold Rausch liable for mere negligence; rather, it reflects the understood environment in the prison, as accepted by prison officials, guards, and administrators, that threats of violence have a high probability of leading to attacks. Riccardo, therefore, proved that the asserted danger was objectively serious.

Under the more specific inquiry of Farmer's subjective prong, the prison official must "deliberately disregard" a potential harm by being "aware of facts from which the inference could be drawn that a substantial risk of serious harm exists, and [also] draw[ing] the inference." Id. at 838, 114 S. Ct. 1970. This seems to me to create two questions: (1) what facts were presented to Rausch, and (2) did he accept them as true or did his actions evince his intent to purposefully ignore those apparent facts? Once again, the inquiry is a particularized one, focused on the officer's knowledge. Without repeating what I have already set out in my dissent, the jury was presented with evidence that Riccardo told Rausch he did not want to be celled with Garcia because he was in fear of an attack. "A prisoner normally proves actual knowledge of impending harm by showing that he complained to prison officials about a specific threat to his safety." McGill v. Duckworth, 944 F.2d 344, 349 (7th Cir.1991). Therefore, contrary to the majority's assertion, the first part of this inquiry is satisfied.

Accepting the majority's point that Riccardo's mere "say-so" is not enough to establish Eighth Amendment liability, there were plenty of "objective indicators" to support the jury's finding that Rausch's actions amounted to deliberate indifference. The jury heard evidence that Rausch put the two men in front of each other to determine whether a problem existed. It is this act, which the majority uses to exonerate Rausch from liability. However, it is this very act which evinces Rausch's deliberate indifference as found by the jury and reiterated by the district court. What is more, this court has already defined such action as unacceptable under the Eighth Amendment. Id. at 349 (reasoning that the scienter requirement is satisfied when a prison guard, "[s]uspect[s] something is true but shut[s][his] eyes for fear of what [he] will learn" or "[goes] out of [his] way to avoid acquiring unwelcomed knowledge"). The jury found that no reasonable guard would think that asking Riccardo to admit fear of Garcia with Garcia present, would illicit an honest response. Again, it is this act that crosses the line. Recognizing that these specific actions were inconsistent with the Eighth Amendment would not create per se liability for prison officials. Therefore, I respectfully dissent from the court's decision not to rehear this case en banc.