Crawford-El v. Britton

2. Requirement of clear and convincing evidence. There still remains the question whether the defendant's entitlement to summary judgment on qualified immunity before plaintiff's discovery achieves an adequate balance in light of Harlow's purposes. Conventional summary judgment principles supply


some protection to defendants. Plaintiff must do better than "show that there is some metaphysical doubt as to the material facts." Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 586 (1986). "The mere existence of a scintilla of evidence in support of the plaintiff's position will be insufficient [to block summary judgment for defendant]." Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 252 (1986). Here defendants argue that that is not enough. They propose a special standard, which they frame as a requirement of "strong evidence." The United States as amicus proposes a similar heightened standard; framing the proposal in terms of pleading, it suggests that plaintiff be required to "plead specific facts giving rise to a strong inference of the alleged improper motive before any discovery will be permitted." [5]

Two factors make us believe that the standard protection of summary judgment (coupled with the limit on discovery stated above) leave an exposure to both liability and litigation that is impossible to square with Harlow. First, unconstitutional motivation is, as is often said of civil fraud, easy to allege and hard to disprove. Bower v. Jones, 978 F.2d 1004, 1012 (7th Cir. 1992) (citing Hollymatic Corp. v. Holly Systems, Inc., 620 F. Supp. 1366, 1369 (N.D. Ill. 1985) ("[F]raud, focusing as it does on a subjective state of mind, can be very easy to allege and very difficult to prove or disprove.")); see also Ross v. Bolton, 904 F.2d 819, 823 (2d Cir. 1990) (rationale behind heightened pleading requirement for fraud in Rule 9(b) is preventing improvident charges of wrongdoing and strike suits); Charles A. Wright & Arthur R. Miller, 5 Federal Practice and Procedure 1296 (1990) (same). Even cut off from the fruit of depositions and other discovery against the defendant and her colleagues, plaintiff will often be able to depict a selective pattern of decisions that, without

[5 Judge Edwards is correct that neither the Solicitor General nor the government defendants advocated the "clear and convincing" standard, see Edwards Op. at 11, but the difference between that and what the Solicitor General did advocate appears to be mainly that his proposed standard is formulated in language that has much less experience and tradition behind it.]


evidence of a more complete set of comparable ones, and extensive explanation by one or more decision-makers, will look fishy enough that a jury could reasonably find illicit motive by a preponderance.

Second, Harlow plainly views the costs of error in the grant or denial of relief in such cases as asymmetrical. The decision expressed a strong concern about the social costs of damages litigation against officials - namely (to repeat), the conventional costs of litigation, the diversion of the officials - time, deterrence of able persons from even accepting public office, and the chilling of officials' readiness to exercise discretion in the public good. Because of those costs the Court adopted a rule categorically denying recovery where, if the truth could be fully known, there was a malicious perpetration of a constitutional violation (but not a violation of a right so clearly established that a reasonable person would have known he was crossing the line). This can only mean that the Court regarded at least some kinds of officer liability (those turning on subjective intent) as ones where, everything else being equal, the social costs of erroneously denying recovery in some cases were exceeded by the combined social costs of (1) litigating and (2) erroneously affording recovery in other cases.

A standard solution to such a difference in costs between two types of error is to adjust the standard of proof. Criminal law is the best known example, where it is seen as better to allow quite a few actually guilty defendants "perhaps many, in fact" to go free than for one innocent one to be convicted; ergo, the reasonable doubt standard. See, e.g., In re Winship, 397 U.S. 358, 372 (1970) (Harlan, J., concurring) ("[I]t is far worse to convict an innocent man than to let a guilty man go free."); cf. 4 William Blackstone, Commentaries *358 (explaining two-witness rule in perjury cases). But civil law contains frequent applications of a more modest tilt, a requirement that the party seeking to mobilize the state to alter the status quo prove his case by clear and convincing evidence. Courts have set that hurdle in deportation proceedings, Woodby v. Immigration and Naturalization Service, 385 U.S. 276, 285 (1966); denaturalization proceedings,


Schneiderman v. United States, 320 U.S. 118, 123 (1943); civil commitment proceedings, Addington v. Texas, 441 U.S. 418, 423 (1979); cases involving termination of parental rights, Santosky v. Kramer, 455 U.S. 745, 756 (1982); defamation suits against public figures, New York Times Co. v. Sullivan, 376 U.S. 254, 285-86 (1964); and a variety of other civil cases such as civil fraud, lost wills, and oral contracts to make bequests, see Woodby, 385 U.S. at 285 n.18 (citing 9 Wigmore on Evidence 2498 (3d ed. 1940)). Although we understand the specific standards urged by defendants and the United States ("strong evidence" and "strong inference") to be aimed at similar concerns, we do not pursue them because of their uncertainty compared to the familiar clear and convincing standard. [6]

We pause to note a relationship between (1) the costs of litigation regardless of outcome and (2) a different societal valuation of the two types of error. Where the social costs of litigation itself are exceptionally high, assuming no difference at all in societal valuation of the two different types of error, that alone could be a ground for a tilt against the party seeking to alter the status quo. Because a reduction in the probability of success reduces the incentives to bring suit (everything else being equal), such a tilt will automatically reduce the aggregate costs of the affected class of lawsuits - at some cost in increasing the number of good claims that go uncompensated. [7] Accordingly, imposition of a clear and convincing standard may imply (1) simply a perception that the type of litigation involves unusually high costs (so that a tilt against its initiators will decrease its incidence, the court

[6 The "strong inference" standard is used by the Second Circuit in securities fraud cases under Fed. R. Civ. P. 9(b). See, e.g., Acito v. Imcera Group, 47 F.3d 47, 52 (2d Cir. 1995); Shields v. Citytrust Bancorp, 25 F.3d 1124, 1127-28 (2d Cir. 1994). ]

[7 Of course, many plaintiffs in civil rights actions against public officials know that their chances of success on the merits are minimal and may be motivated by purposes other than achieving that success. The tilt makes it easier for district judges to end such cases quickly, thereby reducing the burdens on the defendant and the court that concerned the Court in Harlow. ]


regarding the increase in denials of recovery as an acceptable cost), or (2) a conclusion that errors in defendants' favor are independently to be preferred to errors in plaintiffs' favor, or (3) some combination of the two. If the holding of Harlow represented nothing else, it surely manifested either the first or third of those possibilities; after all, in one stroke it destroyed an entire group of claims for what was, by hypothesis, unconstitutional behavior.

The cases applying a clear and convincing evidence standard frequently allude to the second of these rationales (which of course is encompassed in the third). As the Court observed in Addington, a standard of proof both "indicate[s] the relative importance attached to the ultimate decision" and also "serves to allocate the risk of error between the litigants." 441 U.S. at 423; see also Santosky, 455 U.S. at 755 (citing Addington); Cruzan v. Director, Missouri Dep't of Health, 497 U.S. 261, 283 (1990) (same). The Court illustrated this rationale in New York Times Co. v. Sullivan, quoting a Kansas Supreme Court case to support its actual malice standard: " "[O]ccasional injury to the reputations of individuals must yield to the public welfare, although at times such injury may be great.' " 376 U.S. at 281 (quoting Coleman v. MacLennan, 78 Kan. 711, 724 (1908)). The Supreme Court has used such terms in discussing special gradations of proof. Woodby, 385 U.S. at 284-85; Addington, 441 U.S. at 423-25; Santosky, 455 U.S. at 755.

In developing the New York Times rule of clear and convincing evidence, the Court explicitly drew on the reasoning of Barr v. Matteo, 360 U.S. 564, 571, 575 (1959), in which it had extended and explicated absolute officer immunity for certain types of official acts. 376 U.S. at 282. It recited Barr's entire litany of social costs of officer liability - essentially those later invoked in Harlow - as a parallel justifying its adoption of the New York Times rule. Id. If a heightened standard of proof - clear and convincing evidence - was a sound remedy in the area of public figure defamation, we


think it is equally so in the cognate area of officer damage liability for constitutional torts based on improper motive.

Heightened standards of proof of course apply equivalently at summary judgment and at trial, as a seamless web. In Anderson v. Liberty Lobby, Inc. the Court made clear that just as the reasonable doubt standard for criminal trials implies its use in judicial evaluation of motions for acquittal, the clear and convincing standard for trial of malice for purposes of public figure defamation must imply "a corresponding effect" for motions for a directed verdict and for summary judgment. 477 U.S. at 252-54. [8]

What of the pleadings? The label "heightened pleading" for special requirements for constitutional torts involving improper motive was always a misnomer. A plaintiff is not required to anticipate the defense of qualified immunity in his complaint, Gomez v. Toledo, 446 U.S. 635, 640 (1980), and under the Federal Rules of Civil Procedure is required to file a reply to the defendant's answer only if the district court exercises its authority under Rule 7(a) to order one. At stake has always been the ability of the plaintiff to inflict on the defendant officer liability and the serious burdens of litigation itself "discovery and trial. Although we understand the arguments of the court in Schultea v. Wood, 47 F.3d 1427, 1432-34 (5th Cir. 1995), supporting a rule that where qualified immunity is raised in a case involving illicit motive the district court's discretion not to order a reply "is narrow indeed," we do not see why the limit on discovery and the standard of proof discussed above would not adequately fulfill the implications of Harlow. Of course court-ordered replies and motions

[8 Once the plaintiff has come forward with evidence that a jury could regard as clear and convincing proof of the defendant's unconstitutional motive, his access to discovery on all issues (including motive) would be, in the view of the judges in the plurality, a matter for the district court to determine as in ordinary civil litigation. In other words, although the plaintiff would get no discovery unless he had in hand evidence that would support a jury finding in his favor on the motive issue, if he did have that evidence he could use discovery to obtain additional evidence that might help him win the battle of persuasion at trial. ]


for a more definite statement under Rule 12(e) may simplify and speed the process, but we do not see that protection of substantive rights requires any special rules.

We note briefly the argument of the American Civil Liberties Union as amicus, drawing on the recent decision in Johnson v. Jones, 115 S. Ct. 2151 (1995). In Mitchell v. Forsyth the Supreme Court applied the "collateral order" doctrine of Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541 (1949), to hold that immediate appeal was available for "denial of a defendant's motion for dismissal or summary judgment on the ground of qualified immunity." 472 U.S. at 527. In Johnson the Court expressly limited Mitchell to pure issues of law, id. at 2156, such as the determination that a set of given facts constituted a violation of clearly established law, id. at 2159. This made clear that appeals from denials of summary judgment were not available for questions of evidentiary sufficiency. Id. at 2156; see also Behrens v. Pelletier, 116 S. Ct. 834, 842 (1996) (explicating Johnson). The Court was especially concerned that allowing interlocutory appeals of factual questions about intent "may require reading a vast pretrial record, with numerous conflicting affidavits, depositions and other discovery materials" and would result in unjustifiable delay for the plaintiff. Johnson, 115 S. Ct. at 2158.

The ACLU argues that Johnson concluded that where a dispute about material facts exists in a constitutional tort case, "the goal of shielding defendants from discovery or trial yields to the usual goals of resolving cases on their merits through normal procedures." But Johnson is not remotely so sweeping. As the Court observed in Behrens, "Every denial of summary judgment ultimately rests upon a determination that there are controverted issues of material fact." 116 S. Ct. at 842. The question for purposes of immediate appealability is whether the point at issue is mere sufficiency of the evidence or "more abstract issues of law." Johnson, 115 S. Ct. at 2158; Behrens, 116 S. Ct. at 842. The Court never addressed (or even hinted at) any adjustment in the summary judgment standards for constitutional torts involving improper motive under Harlow. Indeed, no court of


appeals thus far has abandoned its special standards in constitutional motive cases in light of Johnson. See, e.g., Moore v. Valder, 65 F.3d 189, 195, 196 & n.13 (D.C. Cir. 1995); Morin v. Caire, 77 F.3d 116, 121 (5th Cir. 1996); Veney v. Hogan, 70 F.3d 917, 922 (6th Cir. 1995); Hervey v. Estes, 65 F.3d 784, 788-89 (9th Cir. 1995); Gehl Group v. Koby, 63 F.3d 1528, 1535 (10th Cir. 1995). And, of course, this court recognized the distinction drawn in Johnson before that case was decided, see Crawford-El v. Britton, 951 F.2d at 1317 (no immediate review available for district court's treatment of an "I didn't do it" defense on summary judgment); see also Johnson, 115 S. Ct. at 2154 (listing Crawford-El among the decisions on the side that Johnson found correct), yet nonetheless applied special standards. More generally, so far as we know, most if not all trial court proceedings over claims requiring clear and convincing proof plod along without any application of the collateral order doctrine. Limits on the reach of that doctrine of course mean delay in the correction of trial court error and a resulting increased exposure of officials to some adverse consequences, but we do not see why every fine-tuning that limits the immediacy of appeal should connote some anti-defendant shift in the principles to be applied by the district court. [9]

Application to Crawford-El

As we have seen, the district court dismissed Crawford-El's Fourth Amended Complaint under the "heightened pleading" requirement. If dismissal of the complaint were the sole means available to protect defendants from discovery barred by Harlow, then we would confront the issue of whether Rule 8's minimalist standard ("a short and plain

[9 Judge Edwards accuses the plurality of insufficient "judicial restraint," Edwards Op. at 14, but it is not clear by what standard one resolution of a question unanswered by Harlow is more or less "restrained" than another. Nor is it clear why one should view a book review by a member of the plurality, see id., suggesting that courts take a modest role in monitoring the judgments of the political branches, as contradicting an opinion whose tendency (among the various plausible alternatives) is to do exactly that. ]


statement of the grounds") could be applied to the sort of complaints here at issue without violating 28 U.S.C. 2072(b)'s ban on the exercise of rulemaking power to "abridge, enlarge or modify any substantive right." See Leatherman v. Tarrant Co. Narcotics Intelligence & Coordination Unit, 507 U.S. at 166-67 (leaving open question of whether courts are to apply "heightened pleading" requirement to claims against government officials). But we see no reason why the government officials' insulation from discovery would not be amply protected by the principle we have already described, entitling officials to summary judgment resolution of their qualified immunity claims before discovery. That being so, it is unclear how application of conventional pleading standards could amount to the sort of substantive abridgement forbidden by 2072(b). Accordingly, we think it was not correct for the district court to apply, literally, a heightened pleading standard, quite apart from the invalidity of our now-abandoned direct evidence rule.

Quite obviously, however, the court and the litigants have been caught in a vortex of changing standards. And although the defendants have not moved for summary judgment since the filing of the Fourth Amended Complaint, it seems sure that they will do so. Moreover, plaintiff has been on notice at least since our 1991 decision of the need for "nonconclusory allegations that are sufficiently precise to put defendants on notice of the nature of the claim and enable them to prepare a response and, where appropriate, a summary judgment motion on qualified immunity grounds." Crawford-El, 951 F.2d at 1317 (quotations and citations omitted). Accordingly, it seems overwhelmingly likely that the Fourth Amended Complaint represents at least a very close approximation of what Crawford-El can advance in resistance to the motion for summary judgment. In the unusual context of this case, then, we are hardly giving an advisory opinion when we consider whether affidavits embodying the assertions of the Fourth Amended Complaint could successfully withstand Britton's motion for summary judgment, backed by the affidavit she has already filed.


1. Whether Crawford-El Has Alleged a First Amendment Violation. We first examine whether Crawford-El's allegations could possibly constitute a violation of a clearly established constitutional right. See Siegert, 500 U.S. at 227 (question whether the conduct complained of constitutes violation of clearly established law is at an "analytically earlier stage" than question of heightened pleading standard); see also Kartseva v. Dep't of State, 37 F.3d 1524, 1530 (D.C. Cir. 1994) (same); Moore v. Valder, 65 F.3d at 195 (same). Although the question is close, we hold that withholding Crawford-El's property in retaliation for exercise of his First Amendment speech rights would indeed be a violation of clearly established law.

We must answer two questions here: (1) whether Crawford-El's speech was protected under the First Amendment such that retaliation would be violation of a clearly established right and (2) how great the retaliatory injury must be. We start with the first. The Supreme Court's decision in Turner v. Safley, 482 U.S. 78, 88 (1987), summarized existing precedent - including Procunier v. Martinez, 416 U.S. 396, 413-14 (1974), and Pell v. Procunier, 417 U.S. 817, 822 (1974) - and set out the test controlling here: "[W]hen a prison regulation impinges on inmates' constitutional rights, the regulation is valid if it is reasonably related to legitimate penological interests." Although on its face Turner applies only to regulations, several other courts have applied the test to other prison actions, including those in retaliation cases. Frazier v. Dubois, 922 F.2d 560, 562 (10th Cir. 1990) (applying Turner in a First Amendment retaliation case); Jackson v. Cain, 864 F.2d 1235, 1248 (5th Cir. 1989) (same); cf. Cornell v. Woods, 69 F.3d 1383, 1388 (8th Cir. 1995) (in First Amendment retaliation case, applying Pell v. Procunier, 417 U.S. at 822 ("[A] prison inmate retains those First Amendment rights that are not inconsistent with his status as a prisoner or with the legitimate penological objectives of the corrections system.")). Several cases have held that a prisoner's right to have access to the press may be limited. Pell v. Procunier, 417 U.S. at 835 (upholding prison regulation prohibiting face-to-face media interviews with par-


ticular inmates designated by the press); Kimberlin, 6 F.3d at 791 n.6 (upholding under Turner warden's policy prohibiting prisoner press conferences and limiting prisoners' press access to settings expressly authorized under prison regulations). But no court has held that a total ban on communications to the press passes muster. Cf. Nolan v. Fitzpatrick, 451 F.2d 545, 547 (1st Cir. 1971) (striking down ban on prisoner letters to news media insofar as the letters concerned prison matters; emphasizing that prison conditions are "an important matter of public policy" about which prisoners are "peculiarly knowledgeable"). And in light of Turner and related cases, retaliation against Crawford-El for criticism of the prison administration that was truthful, and not otherwise offensive to some penological interest (so far as appears), would have violated a clearly established right of which a reasonable prison official would have known. Cf. Pickering v. Bd. of Educ., 391 U.S. 563, 568, 571-72 (1968) (holding that First Amendment precludes dismissal of a school teacher who criticized Board of Education's handling of a bond issue; public employees should be able to speak freely on issues of public concern without fear of retaliation).

As to the sort of injury cognizable under the First Amendment, Crawford-El here alleges the costs of replacing underwear, tennis shoes, soft shoes, and other items; shipping charges to get his papers back; and mental and emotional distress. In our earlier opinion in this case, we noted that some non-de minimis showing of injury is necessary in a constitutional tort action, 951 F.2d at 1321, 1322, and cited Ingraham v. Wright, 430 U.S. 651, 674 (1978) ("There is, of course, a de minimis level of imposition with which the Constitution is not concerned."), and Bart v. Telford, 677 F.2d 622, 625 (7th Cir. 1982). Bart stated that "even in the field of constitutional torts de minimis non curat lex." Id. "It would trivialize the First Amendment to hold that harassment for exercising the right of free speech was always actionable no matter how unlikely to deter a person of ordinary firmness from that exercise" - for example, a supervisor frowning at an employee in retaliation would not constitute sufficient injury. Id. Still, the effect on freedom of speech of retaliations "need


not be great in order to be actionable." Id.; cf. Memphis Community Sch. Dist. v. Stachura, 477 U.S. 299, 306-311 (1986) (out-of-pocket and mental distress damages recoverable for violation of Due Process Clause and First Amendment right to academic freedom); Hobson v. Wilson, 737 F.2d at 61-62 (mental distress damages recoverable for violation of First Amendment right of political association); Frazier v. Dubois, 922 F.2d at 561 (transfer of prisoner in retaliation for exercise of First Amendment rights is unconstitutional injury; citing cases).

The district court commendably latched onto our approval of Bart and applied a sensible standard - whether an official's acts "would chill or silence a "person of ordinary firmness' from future First Amendment activities." Mem. op. at 13 (quoting Bart). The court then found that the pecuniary losses Crawford-El sustained in the form of the costs of shipping his boxes and replacing clothing, though small, might well deter a person of ordinary firmness in Crawford-El's position from speaking again. We agree that the acts asserted pass that test.

Williams Opinion Continued
Case Intro

Listing of Cases |
Proposition One | Peace Park