CRAWFORD-EL v. BRITTON

WILLIAMS, Circuit Judge:

We decided to hear this case en banc on our own initiative in order to resolve continuing disputes as to how a government official's assertion of qualified immunity, as a defense to a damage action for a constitutional tort, may affect pleading and summary judgment standards where the unconstitutionality of the official's act turns on his motive. Our inquiry is framed by the competing goals described by the Supreme Court in Harlow v. Fitzgerald, 457 U.S. 800, 816-18 (1982) — vindicating constitutional rights but at the same time protecting officials from exposure to discovery and trial that would unduly chill their readiness to exercise discretion in the public interest. We here discard our former solution — a requirement that the plaintiff allege "direct" evidence of unconstitutional motive. See, e.g., Siegert v. Gilley, 895 F.2d 797 (D.C. Cir. 1990), aff'd on other grounds, 500 U.S. 226 (1991). But we read Harlow as calling for alternative rules to protect officials. First, we think Harlow allows an official to get summary judgment resolution of the qualified immunity issue, including the question of the official's state of mind, before the plaintiff has engaged in discovery on that issue. Second, we believe that unless the plaintiff offers clear and convincing evidence on the state-of-mind issue at summary judgment and trial, judgment or directed verdict (as appropriate) should be granted for the individual defendant.

* * *

Crawford-El is a prisoner in the District of Columbia's correctional system serving a life sentence for murder. He filed the present lawsuit in 1989, claiming that the individual defendant, Patricia Britton, a D.C. correctional official, and the District of Columbia had misdelivered boxes belonging to him containing legal papers, clothes and other personal items, thereby violating his constitutional right of access to the courts. When Britton moved for dismissal and for summary judgment on grounds of qualified immunity, the district court denied the motion and Britton appealed. We reviewed Crawford-El's allegations under a "heightened pleading" requirement, insisting that the plaintiff in such a case advance "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 v. Britton, 951 F.2d 1314, 1317 (D.C. Cir. 1991) (quotations omitted). By this standard we found his claims wanting. Because we thought that our heightened pleading doctrine had become clearer in ways adverse to plaintiff since his pleading, however, we remanded the case to the district court in case that court, in its discretion, should decide to permit repleading. Id. at 1322.

On remand the district court indeed granted permission, and Crawford-El filed his Fourth Amended Complaint. There he repleaded the access-to-courts claim, but without adding material to fill the gap identified in our first opinion. He also pleaded a due process claim. The district court dismissed both claims, and a panel of this court affirmed. Crawford-El v. Britton, No. 94-7203, mem. op. at 1-2 (D.C. Cir. Nov. 28, 1995). In addition, Crawford-El charged that the defendants' alleged misdelivery of his belongings was in retaliation for various feisty communications with the press and thus in violation of the First Amendment. (This claim had initially appeared in his briefing on the first round in this court. See Crawford-El, 951 F.2d at 1316.) The district court granted the defendants' motion to dismiss the First Amendment claim as well, saying that the complaint did not allege "direct" evidence of unconstitutional motivation and citing Siegert v. Gilley, 895 F.2d 797, 800-802 (D.C. Cir. 1990), aff'd on other grounds, 500 U.S. 226, 231 (1991), our court's most emphatic statement of the "direct" evidence requirement. Crawford-El v. Britton, No. 89-3076, mem. op. at 14-15 (D.D.C. Feb. 15, 1994). After affirming dismissal of the first two claims, the panel suggested, and the court en banc agreed, that the dismissal of the First Amendment retaliation claim should be heard by the court en banc.

Our order for rehearing en banc relates only to the qualified immunity raised by plaintiff's action against Britton. But the District of Columbia is, as noted in the text, still in the case. The district court had dismissed it as a defendant, but since in his successive amended complaints Crawford-El repeatedly named the District as a defendant and the District did not object, the district court held that the District had waived a law-of-the-case argument and therefore reinstated it as defendant. Crawford-El v. Britton, No. 89-3076, mem. op. at 1 n. 1 (D.D.C. Feb. 15, 1994). Because Crawford-El's claims against the District do not concern the questions for which we granted rehearing en banc, they are to be resolved by the panel.
Another extant part of the complaint is a pendant District law claim for conversion of Crawford-El's property. The survival of this claim (in the federal courts) turns on whether, after the remand ordered here, there is any federal claim to which it may be appended.

The background law on subjective motivation and qualified immunity.

In Harlow v. Fitzgerald the Court reformulated its test for officials' qualified immunity in constitutional tort actions. For acts to which qualified immunity may apply, it held that the plaintiff can prevail only by showing not just that there was a violation, but that defendant's acts violated "clearly established statutory or constitutional rights of which a reasonable person would have known." 457 U.S. at 818. It thus excluded liability where there was a violation (but not of a right so clearly established that a reasonable person would have known of it) even when the official acted "with the malicious intention to cause a deprivation of constitutional rights or other injury." Id. at 815 (quoting Wood v. Strickland, 420 U.S. 308, 322 (1975)).

The qualified immunity defense is unavailable for ministerial acts, see Harlow, 457 U.S. at 816; see also Davis v. Scherer, 468 U.S. 183, 196 n. 14 (1984), and unnecessary for acts for which the officer enjoys absolute immunity, see Harlow, 457 U.S. at 807.

The Court was quite explicit as to the purpose of its change. It noted that claims against officers necessarily included ones "against the innocent as well as the guilty," and that among the "social costs" of such suits were "the expenses of litigation, the diversion of official energy from pressing public issues, and the deterrence of able citizens from acceptance of public office." Id. at 814. Last but not least, it invoked Judge Hand's opinion in Gregoire v. Biddle, 177 F.2d 579 (2d Cir. 1949), which had argued that the fear of being sued would "dampen the ardor of all but the most resolute, or the most irresponsible [public officials], in the unflinching discharge of their duties." 177 F.2d at 581 (quoted in Harlow, 457 U.S. at 814). It saw the inclusion of liability based on subjective malice as greatly increasing all these costs. Because such liability opened up a wide field of inquiry, often with "no clear end to the relevant evidence" bearing on the official's "experiences, values, and emotions," and typically not susceptible of disposition by summary judgment, its resolution was "peculiarly disruptive of effective government." Id. at 816-17. Most notably for our purposes, the Court underscored the burdensome character of discovery flowing from such liability. See id. at 817 (speaking of the "broad-ranging discovery" that would result from allowing such claims); id. at 818 (speaking of the resulting "broad-reaching discovery"). Moreover, the Court said, such liability would thwart what had been its assumption in its earlier definition of qualified immunity — that "[i]nsubstantial lawsuits" would be quickly terminated." Id. at 814 (quoting Butz v. Economou, 438 U.S. 478, 507-508 (1978)). Accordingly the Court held that qualified immunity could be penetrated only on a showing of objective unreasonableness — the now familiar requirement of "clearly established" rights. Id. at 818. Henceforth, "bare allegations of malice should not suffice to subject government officials either to the costs of trial or to the burdens of broad-reaching discovery." Id. at 817-18. The Court later described Harlow as having "purged qualified immunity doctrine of its subjective components." Mitchell v. Forsyth, 472 U.S. 511, 517 (1985); see also Davis v. Scherer, 468 U.S. 183, 191 (1984).

In fact, under the decisions of every circuit court addressing the matter, Harlow has not accomplished the stated purpose. This circuit and others have understood Harlow to allow inquiry into subjective motivation where an otherwise constitutional act becomes unconstitutional only when performed with some sort of forbidden motive (such as, here, the claim that Britton's decisions routing Crawford-El's parcels were driven by a desire to penalize his exercise of free speech rights). See, e.g., Siegert v. Gilley, 895 F.2d at 800-801; Whitacre v. Davey, 890 F.2d 1168, 1171 (D.C. Cir. 1989); Martin v. D.C. Metropolitan Police Dept., 812 F.2d 1425, 1431 (D.C. Cir. 1987); Gooden v. Howard County, Md., 954 F.2d 960, 969-70 (4th Cir. 1992) (en banc); Pueblo Neighborhood Health Ctrs., Inc. v. Losavio, 847 F.2d 642, 649 (10th Cir. 1988); Elliott v. Thomas, 937 F.2d 338, 344-45 (7th Cir. 1991); Branch v. Tunnell, 14 F.3d 449, 452 (9th Cir. 1994); cf. Halperin v. Kissinger, 807 F.2d 180, 186-87 (D.C. Cir. 1986) (noting this court's and others' decisions to allow unconstitutional motive claims in areas other than national security). Even though it has entailed many of the "social costs" of inquiry into subjective motivation stated in Harlow, courts have concluded that the vindication of constitutional rights calls for damages liability — often the only device available for such vindication. Halperin, 807 F.2d at 186. In Hobson v. Wilson, 737 F.2d 1 (D.C. Cir. 1984), we recognized the problem, noting that a plaintiff's claim of unconstitutional motive could easily lead to discovery and trial, with no hope of success, and the "result would be precisely the burden Harlow sought to prevent." Id. at 29. We decided that for claims of which unconstitutional intent was an essential part, "nonconclusory allegations of evidence of such intent must be present in a complaint for litigants to proceed to discovery on the claim. The allegations on this issue need not be extensive, but they will have to be 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." Id. This did not speak explicitly to the issue of whether a plaintiff must surmount any particular burden in order to secure discovery. But in Martin v. D.C. Metropolitan Police we specifically took the view that the substantive characteristics of cases involving qualified immunity and unconstitutional motive required deviation from garden-variety application of the Federal Rules of Civil Procedure's liberal pleading and discovery rules. We quoted at length and with evident approbation from a Fifth Circuit decision:

What is a federal trial judge to do? One thing he may not do: face it as just another lawsuit in which the notice pleading's liberal policy of F.R. Civ. P. 8 counts on pre-trial discovery to ascertain the factual basis for the claim[.] . . . Allowing pretrial depositions, especially those taken adversely of the government official to ferret all of his actions and the reasons therefor . . . would defeat and frustrate the function and purpose of the . . . immunity[.] . . . [U]se of liberal discovery to establish the basis of a claim is directly at odds with the Court's direction in Harlow that government officials entitled to immunity [are to] be freed from the burdens, the stress, the anxieties and the diversions of pretrial preparations.

Martin, 812 F.2d at 1437 (R.B. Ginsburg, J.) (quoting Elliott v. Perez, 751 F.2d 1472, 1479 (5th Cir. 1985)) (footnotes omitted).

Our holding in Martin both imposed a "direct evidence" requirement and related it to the problem of discovery. To get to trial, we said, a plaintiff must produce "something more than inferential or circumstantial support for his allegation of unconstitutional motive. That is, some direct evidence [of improper motivation] must be produced. . . ." 812 F.2d at 1435 (emphasis added). But we formulated no explicit rule on discovery. While we quoted Elliott's exhortation about protecting officials from "the burdens, the stress, the anxieties and the diversions of pretrial preparations," we also said that a complete ban on plaintiff's discovery of defendant before resolution of qualified immunity issues on summary judgment might turn the prior decisions allowing plaintiffs to raise claims of unconstitutional motive into an "empty gesture," id. at 1437, and that we were "leaving some space for discovery," id. We told district courts to employ "with particular care and sensibility their large authority to exercise control over discovery" in order to balance all the concerns properly. Id. at 1436-37.

Then-Judge Ginsburg later observed that in Martin the court had "cut back allowable discovery severely, permitting only a sharply limited, precisely defined line of inquiry, and even then, only because of special exigencies in the particular case." Bartlett v. Bowen, 824 F.2d 1240, 1245 (D.C. Cir. 1987) (R.B. Ginsburg, J., concurring in denial of rehearing en banc in Martin and several other cases).

In Whitacre v. Davey we read Martin to require allegations of direct evidence of unconstitutional motive to survive a motion to dismiss and get discovery, 890 F.2d at 1171 n. 4, but the point was not necessary to the case because the allegations of circumstantial evidence were inadequate even under the less demanding standard of Title VII, see id. at 1172. Finally, in Siegert v. Gilley, 895 F.2d at 802, we specifically held that "in order to obtain even limited discovery, such [unconstitutional] intent must be pleaded with specific, discernible facts or offers of proof that constitute direct as opposed to merely circumstantial evidence of the intent." The pleading requirement entailed the discovery consequence: if defendant was entitled to dismissal of the case in the absence of specific assertions of direct evidence, there would be no occasion for discovery. Although the Supreme Court granted certiorari on the question whether "a `heightened pleading' standard which precludes limited discovery prior to disposition on a summary judgment motion violates applicable law," Pet. for Cert. i, quoted in Siegert v. Gilley, 500 U.S. 226, 237 (1991) (Marshall, J., dissenting), the Court in fact affirmed on a different, "preliminary" issue, namely its conclusion that plaintiff had failed to allege a constitutional violation at all. Id. at 232-35. In Kimberlin v. Quinlan, 6 F.3d 789, 793-94 (D.C. Cir. 1993), we applied our "direct evidence" requirement, and denied rehearing en banc with a flurry of concurring and dissenting opinions, 17 F.3d 1525 (D.C. Cir. 1994). The Supreme Court granted certiorari, 115 S.Ct. 929 (1995), but then vacated and remanded, 115 S.Ct. 2552 (1995), for consideration in the light of Johnson v. Jones, 115 S.Ct. 2151 (1995), which clarified the circumstances permitting an interlocutory appeal from denial of a summary judgment motion by a defendant invoking qualified immunity; we then dismissed the Kimberlin appeal. No. 91-5315, 1995 WL 759464 (D.C. Cir. Nov. 8, 1995) (order remanding case to district court).

Because the district court here applied the "direct evidence" rule, mem. op. at 5 n. 4, and found Crawford-El's complaint wanting, id. at 15-17, the present case calls on us to decide whether the circuit should continue to apply that rule, foreclosing discovery unless the pleadings assert "direct evidence" of illicit motive. We find that question easy, at least if, as we believe, there are adequate alternative means of reconciling Harlow's twin purposes in the context of constitutional torts dependent on the official's having an improper motive. We first address the drawbacks of the "direct evidence" rule, and then consider alternative extrapolations from the logic of Harlow. Deficiencies of the "direct evidence" requirement.

First, the distinction between direct and circumstantial evidence has no direct correlation with the strength of the plaintiff's case. While a perjured claim of having heard a confession of unconstitutional motive would meet the test, a massive circumstantial case would not. See Siegert v. Gilley, 500 U.S. at 236 (Kennedy, J., concurring) (rejecting D.C. Circuit's direct/circumstantial test on this ground); Elliott v. Thomas, 937 F.2d at 345 (same). Second, the distinction does not appear calibrated in any other way to the trade-offs found determinative by the Court in Harlow and qualified immunity doctrine generally. Although the rule presumably did reduce the incidence of motive-related damage suits against officers, we have no reason to think that it did any better as a screen than, say, a random rejection of nine out of every ten claims. The abandonment of circuit precedent en banc is of course not to be lightly undertaken. Critical Mass Energy Project v. NRC, 975 F.2d 871, 875 (D.C. Cir. 1992) (en banc) (quoting Arizona v. Rumsey, 467 U.S. 203, 212 (1984)). We have noted in contemplating such overrulings that treatment of the issue in other circuits is a factor to be considered. Id. at 876. Here, the only courts to consider our direct evidence rule have rejected it emphatically, see Elliott v. Thomas; Branch v. Tunnell, 937 F.2d 1382, 1386-87 (9th Cir. 1991), as have the four Supreme Court justices who have chosen to speak on the matter. Siegert v. Gilley, 500 U.S. at 235-36 (Kennedy, J., concurring); id. at 245-46 (Marshall, J., with whom Blackmun Stevens, JJ., concurred, dissenting). Under the circumstances, we think it readily justifiable to overrule our precedents establishing the direct/circumstantial distinction, without even addressing the question whether formulation of the rule as a pleading requirement violates the liberal pleading concepts established by the Federal Rules of Civil Procedure. See Leatherman v. Tarrant Co. Narcotics Intelligence Coordination Unit, 507 U.S. 163 (1993) (invalidating heightened pleading requirement invoked by municipal government unit as defense to constitutional tort, as violation of Rules 8 and 9(b), but reserving issue of holding's application to claims against individual government officials). Alternative protections inferred from Harlow.

In Harlow the Supreme Court assumed that it had established principles of officer liability that eliminated the litigation burdens associated with an official's state of mind, or, as it put the point in Mitchell v. Forsyth, that it had "purged qualified immunity doctrine of its subjective components." 472 U.S. at 517. For that proposition to be literally true, it would be necessary to reject any officer liability for constitutional torts in which the officer's intent is an essential element in rendering the conduct unconstitutional. See Elliott v. Thomas, 937 F.2d at 344 (carrying out "the program of Harlow" would require imputing to defendants the best intent they could possibly have); see also Silberman Op., post (reading Harlow to extinguish liability for such torts). As Elliott noted, however, that would eliminate any damage remedy even for "egregious wrongdoing." 937 F.2d at 344; see also Halperin, 807 F.2d at 186. What, then, does Harlow suggest are appropriate devices to balance the interest in providing remedies against the interest in protecting officials from the undue litigation burdens, including, as Harlow emphasized, discovery itself? We think the crux of the answer lies at the summary judgment phase of litigation. It divides into two questions: First, what methods may plaintiff use to secure evidence to resist the defendant's motion for summary judgment? Second, must plaintiff's evidence substantively meet some higher standard than the conventional preponderance test?

1. Methods available to plaintiff for securing evidence for purposes of summary judgment resolution of qualified immunity. The primary burdens of litigation occur in discovery and trial. If the plaintiff can defer summary judgment while he uses discovery to extract evidence as to defendant's state of mind, Harlow's concern about exposing officials to debilitating discovery will generally be defeated in constitutional tort cases dependent on improper motive. After describing its objective test, the Court said, "Until this threshold immunity question is resolved, discovery should not be allowed." 457 U.S. at 818. We can protect the sequence apparently insisted upon by Harlow — no discovery until there has been at least one cut at the qualified immunity issue — by the straightforward rule that plaintiff cannot defeat a summary judgment motion unless, prior to discovery, he offers specific, non-conclusory assertions of evidence, in affidavits or other materials suitable for summary judgment, from which a fact finder could infer the forbidden motive. In his concurring opinion in Siegert, Justice Kennedy adumbrated this approach. Observing that "heightened pleading" was inconsistent with Federal Rules of Civil Procedure 8 and 9(b), he said:

But avoidance of disruptive discovery is one of the very purposes for the official immunity doctrine, and it is no answer to say that the plaintiff has not yet had the opportunity to engage in discovery. The substantive defense of immunity controls.

Upon the assertion of a qualified immunity defense the plaintiff must put forward specific, nonconclusory factual allegations which establish malice, or face dismissal.

500 U.S. at 236 (emphasis added).

In Elliott v. Thomas, 937 F.2d at 344-46, Judge Easterbrook spelled out the point in more detail. "Unless the plaintiff has the kernel of a case in hand [specific, nonconclusory allegations which establish the necessary mental state], the defendant wins on immunity grounds in advance of discovery." Id. at 344-45. Because the substantive law — the law of qualified immunity per Harlow — tells the court what is needed for summary judgment, there is no conflict with Rule 56's provision for summary judgment:

If a rule of law crafted to carry out the promise of Harlow requires the plaintiff to produce some evidence, and the plaintiff fails to do so, then Rule 56(c) allows the court to grant the motion for summary judgment without ado.

937 F.2d at 345 (emphasis added). This is, of course, substantially similar in result to the imposition of a "heightened pleading" standard, in that both prevent serious invasion of the defendant's time unless the plaintiff can, without discovery, offer specifics of his case as to defendant's motivation. See, e.g., Elliott v. Perez; Sawyer v. County of Creek, 908 F.2d 663, 665, 668 (10th Cir. 1990) (noting that because plaintiff conceded inability to amend complaint without discovery, dismissal would be with prejudice).

Although neither Elliott nor Justice Kennedy's concurrence in Siegert expressly addressed Rule 56(f), which authorizes the district judge to defer ruling on summary judgment and to provide for depositions and other discovery, the solution flows from their analysis of Harlow — its articulation of the substantive right of qualified immunity. To allow the plaintiff to engage in discovery, in order to carry his burden of establishing a basis for inferring improper motive, would violate Harlow's determination to protect the official from discovery until the qualified immunity issue has been resolved. Under the Rules Enabling Act, the Federal Rules of Civil Procedure "shall not abridge, enlarge or modify any substantive right," 28 U.S.C. Section(s) 2072(b), so that any reading of the Rules to trump officials' substantive entitlements is impermissible.

We note that the rule preventing discovery concerning illicit motivation does not bar discovery concerning a defendant official's state of mind for other purposes. A claim for damages for an allegedly unreasonable search or seizure will often turn on whether the defendant was in possession of facts that would have led a reasonable officer to suppose he had probable cause or exigent circumstances. See, e.g., Anderson v. Creighton, 483 U.S. 635, 640-41 (1987) (relevant question in that case was "the objective (albeit fact-specific) question whether a reasonable officer could have believed Anderson's warrantless search to be lawful, in light of clearly established law and the information the searching officers possessed") (emphasis added). Although the Anderson Court appeared to discourage discovery even in that context, see id. at 646-47 n. 6, we do not understand its message as remotely approaching an absolute bar. Similarly, in Billman v. Indiana Dep't of Corrections, 56 F.3d 785, 788-89 (7th Cir. 1995), the Seventh Circuit said it would permit discovery to allow a prisoner to identify the proper defendants in an Eighth Amendment case where a defendant would be liable if it were shown that he knew plaintiff's cellmate was HIV-positive and had a tendency to rape cellmates, and was responsible for the assignment. The state-of-mind showings the plaintiffs had to make in Anderson and Billman thus went simply to the defendants' acquisition of particular facts, not the broader inquiry into motivation at stake here. Our case would be equivalent if Crawford-El had simply to show that Britton knew the boxes contained legal papers (or something else of value to plaintiff) and was responsible for their transfer.

Thus, unlike Judge Edwards, see Edwards at 7-8, we do not see any schism in the Seventh Circuit, between Elliott's requirement that plaintiff himself supply evidence of defendant's illicit motivation in order to withstand defendant's summary judgment motion, 937 F.2d at 345, and Billman's allowing plaintiff discovery to develop evidence that defendant was aware of facts that would, if known to defendant, render his conduct violative of the 8th Amendment.

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."

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.

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 Section(s) 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 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 Section(s) 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.

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).

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. 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 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.

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.

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.

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.

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 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.

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.

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 statement of the grounds") could be applied to the sort of complaints here at issue without violating 28 U.S.C. Section(s) 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 Section(s) 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 particular 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. 2. Whether a Jury Could Reasonably Find Clear and Convincing Evidence of Retaliatory Action. The Fourth Amended Complaint alleges a variety of encounters between Crawford-El and Britton from which plaintiff believes it can be inferred that the misdelivery of his goods must have been in retaliation for various activities that are protected by the First Amendment.

Crawford-El sets the stage with allegations that Britton was hostile to him because of his actions on behalf of fellow prisoners even before his contacts with the press. While he was Clerk for the Occoquan Facility Housing and Adjustment Board at Lorton (from about October 1985 to April 1986), he had frequent contact with Britton since she often served on that Board and Crawford-El often went to the nearby block containing Britton's office to photocopy. He claims that Britton, while despising all prisoners, was particularly hostile to him because he had been in charge of the law library when housed at the central facility at Lorton and had helped many prisoners prepare administrative grievances. According to Crawford-El, Britton deemed him "too big for his britches." In April 1986 Crawford-El apparently invited reporters from the Washington Post to visit the prison, correctly noting on the visitor application form submitted to Britton that the proposed visitors' address was 1150 15th Street, NW, Washington, DC 20071, but discreetly omitting that this was the Post's address. Britton approved the application. A reporter came, and on April 20, 1986 the Post published a front-page article under the headline "Jail Crisis Spills Into Occoquan Unit," subheaded "Crowding, Anger Grow as D.C. Inmates are Shifted to Va. Facility." It quoted Crawford-El's account of an alleged irregularity — that on his arrival at Occoquan a correctional officer had obtained trousers for him by searching in other prisoners' lockers for an extra pair. The next day, says Crawford-El, Britton called him into her office and told him he had "tricked" her and that "so long as [Crawford-El] was incarcerated she was going to do everything she had to to make it as hard for him as possible."

Between April 1986 and Crawford-El's next successful use of the press, he brought a variety of lawsuits against the District. In one, for property allegedly lost through prison officials' negligence, he recovered about $500; in three others he complained on behalf of himself and a class about the lack of food compatible with prisoners' Islamic beliefs, alleged interference with his religious beliefs, and sued, curiously, "for legal malpractice." In December 1988, while the suits were pending, he and a group of other prisoners were transferred to the Spokane County Jail. While assembled for the trip in shackles, the prisoners were videotaped. Plaintiff says that he and others protested the videotaping as a violation of their privacy rights, to which Britton responded, "You're a prisoner, you don't have any rights." Shortly after arrival at Spokane, Crawford-El again spoke with a reporter from the Post. On December 18, 1988, another front-page article appeared, "Sudden Move Severs Inmates' Ties to D.C.; Isolation of Spokane County Jail Puts Prisoners `In a Firecracker Mood.'" It credited Crawford-El with the firecracker metaphor and also quoted him as claiming that the prisoners sent to Spokane were "the civil litigants of Lorton who have been put here to get us out of their hair so our lawsuits will be dismissed on procedural grounds." Shortly after the publication of this article, according to Crawford-El, Britton told a Spokane County Jail official that Crawford-El was "a legal troublemaker," meaning, according to the complaint, "a prisoner who asserts her or his legal rights, or seeks administrative or judicial redress of grievances." As we noted before, "even prison officials free of hostility toward Crawford-El might regard `troublemaker' as an apt moniker." Crawford-El, 951 F.2d at 1319.

The alleged retaliatory act — the misdelivery of boxes — occurred in the course of Crawford-El's transfer back from Spokane to Lorton and thence on to a federal prison in Marianna, Florida, a transfer over which Britton had charge. At Spokane, Crawford-El was instructed to give his property to officials there for forwarding to him. Crawford-El alleges Britton was aware of the boxes' importance to him, saying that when he and two other prisoners met Britton on August 18, 1989 at the Western Missouri Correctional Center en route back to Lorton, they told her that their boxes contained legal papers needed for ongoing cases. She allegedly said that she understood Crawford-El's need for the personal property and legal materials and that the boxes would be sent to her office. (In her affidavit Britton contests the claim that she was ever told of the papers: "I do not recall plaintiff telling me that there were legal documents in his personal property, nor did I have knowledge of the contents of the three sealed boxes." She said she had the boxes sent to her office to keep them from being lost.)

On the trip back to Lorton, supervised by Britton, the property Crawford-El was carrying with him (and that of other prisoners as well) was put into storage on the bus and apparently lost. Crawford-El won an uncontested small claims court suit against Britton for $72.50 based on this loss.

In late August, after arriving back at Lorton, Crawford-El allegedly wrote to Britton requesting that his property be sent to him as soon as she received it. Shortly afterward, he noticed that some other prisoners returning from Washington State had got their property. Just before he was transferred, he checked with a Lorton "Property Officer" named Ward, who told him that he could have his property sent to him at his final destination by writing a request to that effect after arrival at that final destination. At still another intermediate stop, the federal prison in Petersburg, Virginia, Crawford-El learned from other D.C. prisoners that Britton had been calling their families asking them to pick up the prisoners' property because otherwise she would throw it away. He called his parents, who told him his brother-in-law Jesse Carter had picked up his boxes. (Crawford-El was "upset" at this, since he believed he would have difficulty getting permission to receive the property once it had left the prison system.) According to Crawford-El's own allegation in the Fourth Amended Complaint, Carter told Crawford-El that Britton had told him that she was concerned about Crawford-El's legal materials and other property and was afraid the boxes would be lost if she sent them to the Lorton Property Officer for mailing to Crawford-El, and that federal prisons would not accept shipments of D.C. prisoner property. That account meshes with Britton's affidavit, which says that she asked Carter to take Crawford-El's property "only to insure its safety and protection from loss, and for no other reason whatsoever." (Britton also stated that "we had been advised by the Federal Bureau of Prisons that they would not accept the personal property of the prisoners.") But Crawford-El also says that Britton told Carter that Crawford-El "should be happy she did not throw [his property] in the trash."

In the course of Crawford-El's attempts to get his property back, his lawyer received a copy of a letter from the Corporation Counsel's office, stating:

As has been our past practice, inmates transferring from DCDOC [the D.C. Department of Corrections] to BOP [the federal Bureau of Prisons] custody are permitted only a small amount of personal property which should be limited to personal care items and legal documents.

The letter also said that there were "significant differences among DCDOC and BOP property policies and differences between individual BOP facilities" and noted that "[i]n special cases, we ask that DCDOC contact individual facility Inmate Systems staff for permission prior to mailing any inmate personal property to a BOP facility." Though Crawford-El's mother forwarded the boxes on to him at the prison at Marianna, Florida, Crawford-El had some difficulty getting them, as he had expected. Crawford-El asserts that this was because they arrived outside prison channels.

The allegations supplying the strongest evidence of Britton's alleged malign intent are her threat to Crawford-El after the 1986 Post article to make things "as hard as possible for him" and her remark to Carter about throwing the boxes in the trash. But those comments — for both of which Crawford-El is the only source mentioned — are suspect as self-serving assertions. The complaint undermines the "trash" comment by affirmatively asserting that Carter said Britton told him she was giving him the property out of concern about its getting lost, an account that Britton's affidavit supports. As for the allegation that Britton told a Spokane County Jail official that Crawford-El was "a legal troublemaker," the complaint itself defines that term in such a way as to make it impossible to deny that the description is apt. The letter by Corporation Counsel on its face suggests some confusion about the federal Bureau of Prisons policy concerning transfer of D.C. inmate's personal property, reducing the likelihood that Britton's handing the property to his brother-in-law was a deliberate scheme to keep it away from Crawford-El. Indeed, in the absence of some reason to believe Britton thought Carter had it in for Crawford-El or was hopelessly incompetent (neither of which is claimed by Crawford-El), or thought that federal prison officials would much more readily allow Crawford-El to receive the property if sent by the D.C. Department of Corrections than if sent from outside the prison system, transfer of the boxes to the brother-in-law makes an awkward fit with any serious purpose to keep them from Crawford-El. In addition, Crawford-El's own complaint states that Britton had telephoned other D.C. prisoners' families to ask them to pick up those prisoners' property at Lorton — behavior further reducing the chance that Britton's treatment of Crawford-El had any retaliatory purpose. In short, a jury could not reasonably find that Crawford's nonconclusory assertions constitute clear and convincing evidence of unconstitutional intent. On remand, Crawford-El may attempt to bolster his evidence — perhaps in part through discovery, if by amplifying his independent assertions he secures district court permission to conduct discovery pursuant to Judge Ginsburg's separate opinion, which is controlling on the issues as the opinion consistent with the disposition on the narrowest grounds, i.e., a "common denominator" of the reasoning of the majority, see King v. Palmer, 950 F.2d 771, 780-81 (D.C. Cir. 1991) (en banc)). If he adds no evidence, the district court should grant any future motion for summary judgment by Britton on the federal claims against her.

* * *

Accordingly we vacate the dismissal of Crawford-El's First Amendment retaliation claim against Britton, and the pendent conversion claim (see supra note 1), and, once the panel has resolved the issues between Crawford-El and the District (see id.), remand the case to the district court for further proceedings.

So ordered.

[48] SILBERMAN, Circuit Judge, concurring:

Crawford-El, a D.C. prisoner serving a life sentence for murder and a chronic litigant whom we have previously described as a "trouble maker," Crawford-El v. Britton, 951 F.2d 1314, 1320 (D.C. Cir. 1991), cert. denied, 506 U.S. 818 (1992), has brought a damage claim (now amended four times) against a prison official who allegedly retaliated against him for the exercise of his constitutional rights to bring innumerable law suits (and talk to the press) by allowing his boxes of "legal material" to be picked up by his brother-in-law (horrors!) when the plaintiff was transferred from one prison to another.

There was a time, not too many years ago, when any American lawyer or judge hearing that such a case was the subject of an en banc hearing in a federal court of appeals, even that it plausibly could be brought as a claim in a federal district court, would have been incredulous. Before I discuss what I believe to be the appropriate resolution of the case — given the state of present law on qualified immunity of government officials — I think it worthwhile to trace the jurisprudential steps that have led us to this situation. Particularly is this so because some justices have expressed legitimate concerns about the degree of judicial "policymaking" implicated in fashioning the substantive and procedural framework of qualified immunity, see Wyatt v. Cole, 504 U.S. 158, 171-72 (1992) (Kennedy, J., concurring, joined by Justice Scalia); see also Chief Judge Edwards' Sep. Op. at 14-15, overlooking the much more fundamental — and troublesome — judicial policymaking involved in creating the causes of action that have given us the problem.

Qualified immunity is not a new innovation and some have expressed concern insofar as it has been extended beyond its common-law boundaries. But at common law, we did not have constitutional torts as such. Moreover, pre-trial discovery in the nineteenth century was not burdensome (in sharp contrast to our current system) due to the severe restrictions placed on it, if it was allowed at all, even in the most permissive of states. See Wolfson, Addressing the Adversarial Dilemma of Civil Discovery, 36 Clev. St. L. Rev. 17, 25-27 (1987).

I.

Federal damage actions that typically raise qualified immunity concerns are those brought against federal officials as Bivens actions or against state officers under Section(s) 1 of the 1871 Civil Rights Act (hereinafter 1983). Section 1983 reads:

Every person who, under color of any statute, ordinance, regulation, custom, or usage of any State or Territory or the District of Columbia, 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, shall be liable to the party injured in an action at law, suit in equity, or other proper proceeding for redress.

42 U.S.C. Section(s) 1983 (1994). Ironically, Section(s) 1983 was the least controversial provision in the 1871 Act, attracting little attention or debate. And for almost 100 years the federal courts read that statute as it was clearly intended, to attack the so-called "Black Codes" passed by Southern states after the civil war, not private torts. See, e.g., Lane v. Wilson, 307 U.S. 268 (1939); Brawner v. Irvin, 169 F. 964, 968 (C.C.N.D. Ga. 1909) (dismissing case that alleged that the police chief had whipped petitioner for striking his relative since it alleged only a private tort). But in 1961 in Monroe v. Pape, 365 U.S. 167, the Supreme Court extended the statute to reach the behavior of Chicago police officers who did not claim their actions were sanctioned under state law. Indeed, there was little doubt that the plaintiffs had a tort remedy under Illinois law. But as is so often true when the Supreme Court hands down a decision that substantially expands federal judicial power, the facts were dramatic: thirteen Chicago police officers broke into the Monroes' apartment, forced the Monroes to stand naked at gunpoint in the middle of their living room, struck their children, and called Mr. Monroe "nigger" and "black boy." Id. at 203 (Frankfurter, J., dissenting in part). The Court overrode what seems to me to be the characteristically impeccable reasoning of Justice Frankfurter (when he was relying on reasoning rather than rhetoric) in dissent, and turned 1983 into a provision that the post-civil war Congress could not possibly have visualized. See Zagrans, "Under Color Of" What Law: A Reconstructed Model of Section 1983 Liability, 71 Va. L. Rev. 499 (1985). The Court's construction effectively read out of the statute the "under color of law" limitation, making it synonymous with the Fourteenth Amendment's state action requirement. Subsequently, the Court discovered a whole series of new constitutional rights and applied the Bill of Rights to the states. As a result, the 296 federal civil rights actions against government officials filed in 1961 have exploded into over 40,000 by 1988, over half of which were filed by prisoners. In just the period between 1975 and 1984, the number of prisoner civil rights cases increased by approximately 200%, from 6,606 to a staggering 18,856. See Eisenberg Schwab, The Reality of Constitutional Tort Litigation, 72 Cornell L. Rev. 641, 667 (1987). In contrast, there were only 21 cases decided under Section(s) 1983 in its first 50 years. See Comment, The Civil Rights Act: Emergence of an Adequate Federal Civil Remedy?, 26 Ind. L.J. 361, 363 (1951).

The Court's interpretation of "under color of law" has not been its only creative interpretation of Section(s) 1983. It has allowed litigants to use Section(s) 1983 to enforce statutes that have no connection to the Fourteenth Amendment or the post-civil war civil rights legislation. See Maine v. Thiboutot, 448 U.S. 1 (1980). The Court was not discomforted that its interpretation would result in the scope of Section(s) 1983 being vastly greater than its jurisdictional counterpart (which was the only conceivable basis for Section(s) 1983 suits until Section(s) 1331 was passed some years later). The dissent in Thiboutot indicated that it is "idiotic" to interpret Section(s) 1983 in this fashion. Id. at 21 n. 9.

In fairness, "incorporation" of the Bill of Rights had begun a long time before. See, e.g., Smyth v. Ames, 169 U.S. 466, 525-26 (1898) (applying the "Takings Clause" to state rate regulation of railroads).

Then, in 1971 the Court, in perhaps an even more stunning exercise of judicial policymaking, fashioned a federal cause of action for damages against federal officials for a "constitutional tort." In Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388 (1971), the facts were again grim; six federal law enforcement officials without a warrant broke into the apartment of the plaintiff to conduct a search. He was arrested in front of his wife and children — who were also threatened with arrest — for a narcotics violation and was subsequently interrogated, searched, and booked. The case against him was ultimately dismissed. Bivens reflected the Court's policy proclivity to "equalize" the obligations of constitutional law imposed on state government to those imposed on federal government. See, e.g., Bolling v. Sharpe, 347 U.S. 497, 500 (1954); Butz v. Economou, 438 U.S. 478, 501-03 (1978). To be sure, prior to 1875 and the passage of the general federal question jurisdiction statute, an injured party could bring a common-law suit in state court against a governmental actor. The governmental official would then raise as a defense that he was acting pursuant to a statute or authority vested in him — a defense which could be defeated by showing that the statute or delegated authority was unconstitutional. For instance, the Fourth Amendment's prohibition against unreasonable searches and seizures was enforced by bringing a common-law trespass action against a governmental official, an action which an official could not defeat by invoking a claim of authority violative of the Fourth Amendment. See Boyd v. United States, 116 U.S. 616, 626-27 (1886). Of course, there was no a priori assurance that there would always be a common-law right guaranteeing a remedy for an official's unconstitutional action (although there normally would be), but this is only a problem if one thinks that there is an a priori reason to believe that every constitutional violation must be remedied. Our historical practice simply does not support the proposition that the Constitution is self-executing. Cf. Webster v. Doe, 486 U.S. 592, 613 (1988) (Scalia, J. dissenting) (explaining that it is "untenable that there must be a judicial remedy for every constitutional violation"). After 1875, the Court started down a different path. It gradually concluded that an implied cause of action under the Constitution existed where the remedy sought was an injunction. The Court by "almost imperceptible steps . . . appears to have come to treat the remedy of injunction as conferred directly by federal law for any abuse of state authority which in the view of federal law ought to be remediable." Hart, The Relations Between State and Federal Law, 54 Colum. L. Rev. 489, 524 (1954). This process culminated in Ex Parte Young, 209 U.S. 123 (1908), in which the Court upheld an injunction of a state official where the alleged wrong was the threat of future prosecutions. Whatever the validity of this reasoning in an era when the courts had license to create general federal common law, see Swift v. Tyson, 41 U.S. (16 Pet.) 1 (1842), there is no question that the Court's finding of an implied right to an injunction against a government official in his official capacity is on far more solid ground than the creation of an implied right to damages against a governmental official as an individual.

In actions for trespass, the defendant would typically seek damages against the trespasser. See, e.g., Huckle v. Money, 95 Eng.Rep. 768 (1763). While one might be tempted to argue that since the framers envisioned the Fourth Amendment being enforced through actions for damages — where the Fourth Amendment negated the government official's defense — the important point is that the underlying cause of action was a creature of state law.

One should keep in mind that even under the most narrow construction of Harlow v. Fitzgerald, 457 U.S. 800 (1982), it is inevitable that some meritorious suits will be barred. This should not be surprising since the very notion of an "immunity" from suit, as opposed to a "defense," entails that valid constitutional claims will be barred.

After all, the Constitution (with a few exceptions such as the Thirteenth Amendment) is concerned with limitations on the power of government. Individuals are implicated only insofar as they act as agents of the government as opposed to private tortfeasors. Moreover, the Court has for the last hundred years consistently followed this line of reasoning in finding an implied right to an injunction; Bivens suits lack such a pedigree. See, e.g., Davis v. Passman, 442 U.S. 228, 241-43 (1979) (explaining this tradition); Chamber of Commerce v. Reich, 74 F.3d 1322, 1327-28 (D.C. Cir. 1996); Collins, "Economic Rights," Implied Constitutional Actions, and the Scope of Section 1983, 77 Geo. L.J. 1493, 1510 (1989). The availability of the historically recognized right to injunctive relief obviates the need for a judicially-created damages remedy. As Justices Frankfurter and Brandeis explained (and as implicitly recognized by Justice Harlan in his Bivens concurrence) "remedies" are independent of "rights." Remedies can vary based on the weighing of numerous policy considerations even while the right being enforced remains the same. See Truax v. Corrigan, 257 U.S. 312, 354-57 (1921) (Brandeis, J., dissenting); F. Frankfurter N. Greene, The Labor Injunction 205-223 (1930). The lack of a damages remedy does not denigrate or change the nature of the underlying right.

The Court has implied a damages remedy in order to enforce the Fifth Amendment's prohibition on the taking of private property for public use without just compensation. See Jacobs v. United States, 290 U.S. 13 (1933). However, the damages remedy was against the government and has explicit textual support in the Amendment's requirement that "just compensation" be paid. See First English Evangelical Lutheran Church of Glendale v. County of Los Angeles, 482 U.S. 304, 316 n. 9 (1987) (citing cases that "make clear that it is the Constitution that dictates the remedy for interference with property rights amounting to a taking").
Some have argued that the Court after 1875, even if rarely, also implied a personal damages remedy. In most of these cases, the Court seems to have conceived of the cause of action, although admittedly sometimes artificially, as based upon the common law. In Wiley v. Sinkler, 179 U.S. 58 (1900), and Swafford v. Templeton, 185 U.S. 487 (1902), the two most cited examples of this implied damages remedy, the Court concluded that the lower federal court had federal question jurisdiction to entertain a suit for damages against state officials for their interference with the plaintiffs' right to vote in federal elections since it involved the construction and application of the Constitution. The Court in these cases was focused on whether the suits raised federal questions, not the legitimacy of the damages remedy, although the two questions admittedly do overlap.

The point to bear in mind then, before turning to the qualified immunity question, is that the causes of action that largely create the problem that qualified immunity addresses were not created by Congress; they were devised by the Supreme Court without any legislative or constitutional (in the sense of positive law) guidance. Justice Harlan candidly admitted in his concurring opinion in Bivens, and as subsequently affirmed by the whole Court, see Bush v. Lucas, 462 U.S. 367, 376-78 (1983), that the Court, in crafting a remedy, feels free to take into account the range of policy considerations "at least as broad as the range of those a legislature would consider." 403 U.S. at 407.

As Justice Rehnquist pointed out in dissent in Carlson v. Green, 446 U.S. 14, 36 (1980), this quasi-Article I legislative function of open-ended balancing of different policy considerations and goals is ill-suited for the judiciary. The best solution to the whole problem would be the flat overruling of both Bivens — as Justice Rehnquist called for in Carlson — and Pape, putting the issue of damage remedies against state or federal officials for constitutional torts where it belongs — with states and Congress. But since the Supreme Court, in accordance with public choice theory, see generally J. Buchanan G. Tulloch, The Calculus of Consent (1962) (arguing that all rational actors, including those in government, pursue power), follows its own version of the Breznev Doctrine — no significant retreat from extensions of federal constitutional power (unless perhaps, if confronted by Congress) — that is a vain hope.

To the extent that the Bivens Court relied on the Court's authority to infer private damages remedies in the face of statutory silence, see Bivens, 403 U.S. 388, 398 (Harlan, J., concurring), this has been undermined by subsequent case law. See Carlson, 446 U.S. at 39 n. 5.

It could be argued that the Supreme Court's withdrawal from Lochner is an exception, but of course substantive due process grew back anew in "politically correct" gardens.

II.

As I have indicated, shocking factual allegations played no small part in the development of the law in Pape and Bivens. (Many journalists and lawyers describe as a virtue a hypothetical Supreme Court justice's disposition to decide in accordance with the facts of a particular case; they mean the justice should decide how the dispute should be resolved using a Solomonic policy-oriented methodology and then the law should be fashioned to accommodate that resolution.) It is hard to imagine a similar outcome in either case if facts akin to Crawford-El's had been presented. In other words, if Pape or Bivens had involved constitutional tort claims that depended on allegations that the actor's motive was proscribed, I am confident that the Supreme Court would not have gone down either path, especially in light of the probative difficulties that motive-based wrongs necessarily involve, because virtually any ostensibly legal action taken by a government official can be thought unconstitutional if prompted by an unconstitutional motive. Viewed in this light, Judge (now Justice) Ginsburg's "heightened pleading" requirement that a plaintiff allege direct evidence to show an unconstitutional motive for actions that would otherwise be perfectly legal might be thought an effort to keep a Bivens claim close to the kinds of facts that moved the Supreme Court to create the cause of action in the first place. See Martin v. D.C. Metropolitan Police Dep't, 812 F.2d 1425 (D.C. Cir. 1987). Theoretically, circumstantial evidence is not inherently weaker than direct evidence, but I think Judge Williams' opinion overstates the matter — by a good deal — when it argues that "we have no reason to think that it did any better as a screen, than, say a random rejection of nine out of every ten claims." Judge Williams' Op. at 10. Since direct evidence of an unconstitutional motive for an ostensible legal act is virtually never available (I do not recall ever seeing such a case since Martin was decided), the Martin heightened pleading requirement effectively kept Bivens unconstitutional motive cases from going to discovery and trial in our circuit for 10 years. That result, no matter how reached, is not only desirable, it is implicitly contemplated, as I explain below, by Harlow v. Fitzgerald, 457 U.S. 800 (1982). Thus, whatever the logical flaws in the direct versus circumstantial evidence distinction or the designation of a "heightened pleading" requirement, I would have been content to hold to Martin as precedent — which under Judge Edwards' reasoning would be the most judicially restrained choice — but as long as the court is determined to reexamine the doctrine, I prefer a somewhat different approach than does Judge Williams.

In actual practice, Judge Williams' clear and convincing test applied at the summary judgment stage may well have the same ultimate impact as the Martin test. Under both standards, it would appear quite difficult for a plaintiff to gain discovery, let alone a trial, if the government official's bad motivation is the key to making out the constitutional tort. Still, the test as set forth by Judge Williams holds out the prospect of confusion in application. I am not sure I understand just what sort of showing a plaintiff must make to meet the "specific, non-conclusory assertions of evidence, in affidavits or other materials suitable for summary judgment" test. Judge Williams' Op. at 13 (emphasis added). Or how that differs from Justice Kennedy's requirement that "the plaintiff must put forward specific, nonconclusory factual allegations which establish malice." Siegert v. Gilley, 500 U.S. 226, 236 (1991) (emphases added). Or how either standard differs from the Seventh Circuit's cryptic phrase "[u]nless the plaintiff has the kernel of a case in hand, the defendant wins on immunity grounds in advance of discovery." Elliott v. Thomas, 937 F.2d 338, 345 (1991). For example, would Judge Williams' disposition differ if the plaintiff produced an affidavit asserting that in a private conversation with him the defendant unequivocally stated that she intended to punish him, for his vexing litigation, by giving his "legal papers" to his brother-in-law? Would Justice Kennedy's or Judge Easterbrook's? I fear Judge Williams' approach, while certainly preferable to Judge Edwards', will promise a good deal of further litigation with very little return in providing relief in supposedly meritorious cases.

I quite agree with Judge Williams' discussion of Leatherman v. Tarrant Co. Narcotics Intelligence Coordination Unit, 507 U.S. 163 (1993), and also agree that a plaintiff is entitled to discovery for certain other purposes. Judge Williams' Op. at 13-14.

Judge Ginsburg's approach promises even more confusion — and ungoverned variance among the practice of district judges. By permitting discovery upon a showing based on "specific evidence within the plaintiff's command, that such discovery will uncover evidence sufficient to sustain a jury finding in the plaintiff's favor," Judge Ginsburg asks each judge to use his or her crystal ball rather than a rule of decision. The ex ante impact on potential defendants' behavior would not under this formulation differ meaningfully from Judge Edwards' position. In my view it will induce more paralysis than discouragement of wicked actions. It is perhaps one of the simplest axioms of law and economics that overdeterrence as well as underdeterrence yields inefficient results. See P. Schuck, Suing Government 68-75 (1983).

I think the more straightforward solution, following Harlow's reasoning, is to hold that when the defendant asserts a legitimate motive for his or her action, only an objective inquiry into the pretextuality of the assertion is allowed. If the facts establish that the purported motivation would have been reasonable, the defendant is entitled to qualified immunity. Cf. Halperin v. Kissinger, 807 F.2d 180, 188 (D.C. Cir. 1986). Although Harlow dealt specifically with a different subjective aspect of an official's motivation — his knowledge or appreciation of governing constitutional law — as Judge Williams notes, the Court in Mitchell v. Forsyth, 472 U.S. 511, 517 (1985), read Harlow as having "purged qualified immunity doctrine of its subjective components." See also Anderson v. Creighton, 483 U.S. 635, 645 (1987) (explaining that the Harlow Court "completely reformulated qualified immunity along principles not at all embodied in the common law, replacing the inquiry into subjective malice so frequently required at common law with an objective inquiry into the legal reasonableness of the official action"). More important, Harlow itself unequivocally states that "[u]ntil this threshold immunity question is resolved, discovery should not be allowed." 457 U.S. at 818. That thought certainly strongly suggests that a factual dispute over whether the defendant's otherwise legal action is rendered illegal because of an unconstitutional motive cannot defeat a qualified immunity defense. The Harlow Court manifested a clear awareness of the peculiar difficulties that litigation over any kind of motivational disputes entail:

There are special costs to "subjective" inquiries. . . . In contrast with the thought processes accompanying "ministerial" tasks, the judgments surrounding discretionary action almost inevitably are influenced by the decision maker's experiences, values, and emotions. These variables explain in part why questions of subjective intent so rarely can be decided by summary judgment . . .

Id. at 816. The Court specifically noted that "petitioners advance persuasive arguments that the dismissal of insubstantial lawsuits without trial — a factor presupposed in the balance of competing interests struck by our prior cases — requires an adjustment of the `good faith' standard established by our decisions." Id. at 814-15. The gravamen of the petitioners' argument was that the qualified immunity available under Butz was undermined by district courts which "routinely denied motions for summary judgment on the ground that the claim of malice or bad faith automatically raised a triable issue of fact as to the defendant's state of mind." It would be odd if the Court found this concern persuasive and yet reformulated the qualified immunity inquiry in a way that was not responsive to the difficulty of defeating at summary judgment intent — based constitutional suits. Nor is it at all clear that allowing a government official, as Judge Williams puts it, Judge Williams' Op. at 16, to maliciously perpetrate a constitutional violation (so long as the constitutional right was not so clearly established that a "merely reasonable person" would not have known it) is less "egregious," Judge Williams' Op. at 11, than allowing the same official to take an objectively reasonable action that would be blameless if the defendant's motives were benign. The very logic that leads my colleagues to reject the distinction between direct and circumstantial evidence, it seems to me, could lead to a similar rejection of the distinction between two subjective elements (knowledge of the law and actual motivation) of the constitutional tort/qualified immunity analysis.

Yet, as Judge Williams correctly notes, the circuit courts have shrunk from that interpretation of Harlow. They have done so, it appears, because of a concern that has driven much of American jurisprudence in the latter half of the twentieth century; the prospect of a racially discriminatory act. See, e.g., Kennedy, The State, Criminal Law, and Racial Discrimination: A Comment, 107 Harv. L. Rev. 1255 (1994) (discussing the impact of race on the evolution of criminal law). Thus in Elliott, the Seventh Circuit recognized that:

[c]arrying out the program of Harlow seems to imply attributing to the defendants the best intent they (objectively) could have under the circumstances, and asking whether the law at the time clearly establishes that persons with such an intent violate the Constitution. Yet that would be the functional equivalent of eliminating all recoveries when a mental state is part of the definition of the wrong — as it is in cases of racial discrimination, excessive punishment, and many other constitutional torts.

937 F.2d at 344 (emphasis added). Similarly, in Halperin, where we actually so applied Harlow, at least to "national security cases," see 807 F.2d at 187-88, we revealingly suggested that to conclude that Harlow meant to preclude inquiry into all intent would permit a defendant "to discriminate on the basis of race." Id. at 186. Giving Harlow its logical extension does not, in my view, present any special problems of encouraging racial discrimination, because, as I will discuss shortly, there are other restraints on discriminatory official action. Therefore, I would extend to all unconstitutional motive actions the principle adopted in Halperin, where we held that if the government defendants' actions (wiretaps) in a Bivens case would be "validated" by a legitimate national security motive, the defendants are entitled to immunity if they purport to act for national security reasons, unless a jury could conclude that it was objectively unreasonable for the defendants to so act. A simple hypothetical illustrates its ease of application. Suppose a plaintiff claims that a defendant (perhaps a judicial official not covered by Civil Service or Title VII legislation), see Whitacre v. Davey, 890 F.2d 1168 (D.C. Cir. 1989), cert. denied, 497 U.S. 1038 (1990), impermissibly fired her because of her race. The defendant claims that the plaintiff was discharged because of budget constraints. If the defendant's rationale would have been objectively reasonable under the circumstances, the defendant wins on summary judgment. In contrast, if a reasonable trier of fact could find that budget constraints were objectively unreasonable under the circumstances (if, for instance, the official's division recently received a windfall of funds or hired a number of additional workers), the case would proceed to trial. Cf. Halperin, 807 F.3d at 189 (noting that the defendants win on summary judgment if they "adduce sufficient facts that no reasonable jury . . . could conclude that it was objectively unreasonable for the defendants to be acting for national security reasons").To be sure, as I have noted, in Halperin we limited our holding to national security cases, perceiving a particular need to protect the executive branch from probing into motivations that touch such sensitive issues. There the government's wiretap was thought to be unconstitutional unless it was motivated by national security concerns — so it appeared as if it was the government that put motivation at issue. But, I think that formulation is deceptive. Any action, the discharge of a government employee say, could be phrased the same way; as either illegal if motivated by unconstitutional discrimination or constitutional if not. Perhaps all the Halperin panel meant by the notion of a "validating" intent is that, as a matter of substantive law, the burden was on the government to prove its motivation was driven by national security concerns. But suppose a government agency discharged an employee for alleged national security reasons and the employee claimed it was for racially discriminatory grounds. Judge Williams does not explain whether under such circumstances Halperin or his clear and convincing test governs. (Indeed neither Judge Williams, Judge Ginsburg, nor Judge Edwards discusses the relevance of Halperin to their respective tests — making it a silent orphan — so it is wholly indeterminate whether it is affected by this en banc proceeding.) Halperin's reasoning avoids this analytical difficulty: if the challenged defendants' actions, without regard to their actual intent, are consistent with an objectively reasonably intent, the defendants are entitled to immunity. And even if the defendants are not able to meet this burden, they are still entitled to immunity if they are able to prove that their actual motivation was legitimate.

Of course, if the government official (or the government) does not deny that the defendant acted with an unconstitutional motive, that is another matter.

Harlow allows the use of evidence concerning subjective motivation if it benefits the government. "[I]f the official pleading the defense claims extraordinary circumstances and can prove that he neither knew nor should have known of the relevant legal standard, the defense should be sustained." 457 U.S. at 819. The Court then notes, somewhat cryptically, that "[b]ut again, the defense would turn primarily on objective factors."

Judge Ginsburg (and to a lesser extent Judge Williams), although assiduously avoiding a reference to Halperin, criticizes my approach as creating inevitable incentives to unconstitutional behavior. But, of course, the same criticism can be made against either of their positions insofar as they strengthen a defendant's hand even fractionally over Judge Edwards' position. There is simply no escape from a judgment, without any empirical data, as to where along the spectrum to draw the line between the interests of discouraging unconstitutional behavior and avoiding the peculiar difficulties that the threat of personal damage suits against public officials entail.

In any event, I do not think the matter is quite as simple or self-evident as Judge Ginsburg's downward sloping demand curve. We should bear in mind that in these cases, which often arise in an employment context, the defendant, even if he or she acts in part with a proscribed motive, that motive typically is only a contributing factor to a decision. This has led to terribly complicated jurisprudential efforts to develop techniques to measure the relative importance of the proscribed motive. Cf. NLRB v. Wright Line, 662 F.2d 899 (1st Cir. 1981), cert. denied, 455 U.S. 989 (1982); Price Waterhouse v. Hopkins, 490 U.S. 228 (1989), as modified by the Civil Rights Act of 1991, Pub.L. No. 102-166, 105 Stat. 1071 (1991) (codified and amended in scattered sections of 42 U.S.C. (Supp. V 1993). Take the present case. How would we really distinguish between the defendant's hard feelings (if they could be established) toward the plaintiff because he is a self-evident pest as opposed to the more grandly phrased "because of his exercise of his First Amendment rights"?

The truth of the matter — as most practitioners in the labor and EEO field well know — is that a determination as to the existence and relative importance of an illegal motive is difficult, often artificially relying on certain presumptions. And the behavior of the potential defendant ex ante is typically directed at avoiding those indicia of the proscribed motive that will tend to be relied upon in that substantive area of the law. (Can one imagine an employer deciding whether to discharge a employee for theft attempting, perhaps through yoga, to cleanse his mind of any hostility because of the employee's union status?)

Still, it is difficult to deny that, at least theoretically, Judge Williams' view and even more Judge Ginsburg's position creates a greater disincentive to government officials taking action with an unconstitutional motive than does mine. But, personal damage suits are decidedly not the only disincentive. We should bear in mind of what my colleagues fail to take sufficient account — that there are restraints against such behavior other than Section(s) 1983 or Bivens damage suits. When officials violate citizens' rights, they expose themselves to disciplinary sanctions, harm to their professional reputations, and reduced opportunities for advancement. See, e.g., Schuck, supra, at 69. Unlike normal tort law, federal and state officials are sworn to uphold the Constitution; violating one's oath may mean a reputation for deceit and unreliability. Certainly a rational actor would avoid this result, if only to avoid a decrease in his or her value as an employee. Cf. Epstein, In Defense of the Contract at Will, 51 U. Chi. L. Rev. 947, 967 (1984); R. Posner, Overcoming Law 109- 44 (1995). To the extent an individual fears moral retribution, the oath will further induce proper behavior. I hope I will be forgiven for assuming that such an oath, like a monetary disincentive, can affect the behavior of government officials. See Edward J. DeBartolo Corp. v. Florida Gulf Coast Bldg. Constr. Trades Council, 485 U.S. 568, 575 (1988); Webster, 486 U.S. at 613 (Scalia, J., dissenting). Individuals who fear divine punishment also face a downward-sloping demand curve: as the level of sin rises, the punishment increases.

Moreover, a number of federal statutes are aimed at governmental unconstitutional conduct. Even in the absence of suits for money damages, government officials will be deterred by the threat of criminal prosecution. Government officials possess no general immunity from such actions. See, e.g., Imbler v. Pachtman, 424 U.S. 409, 429 (1976) (noting that the Court has "never suggested that the policy considerations which compel civil immunity for certain government officials also place them beyond the reach of the criminal law. Even judges, who have long been cloaked with absolute immunity from damages, could be punished criminally for willful deprivations of constitutional rights on the strength of 18 U.S.C. Section(s) 242, the criminal analog of 1983").

In addition to Section(s) 1983, plaintiffs can sue officials for monetary relief under 42 U.S.C. Section(s) 1981 (1994) (civil action for denying persons the "full and equal benefit of all laws and proceedings" guaranteeing security of persons and property); Section(s) 1982 (civil action for interference with citizens' property rights on the basis of race); 1985 (civil action for conspiracy to deprive persons of equal protection of the laws); and Section(s) 1986 (civil action for failure to prevent a conspiracy to interfere with Section(s) 1985 rights).
Admittedly, as Judge Ginsburg notes, qualified immunity may apply to these actions for money damages as well.

See, e.g., 18 U.S.C. Section(s) 241 (1994) (criminal action for conspiracy to "injure, oppress, threaten, or intimidate" a person in the exercise of his constitutional rights); Section(s) 242 (criminal action for deprivation of a person's constitutional rights on account of a person being an alien or by reason of his race).

Federal statutes providing causes of action against the government itself — particularly those targeted at discrimination — provide additional deterrence. The government undoubtedly looks askance to official misconduct that subjects it to liability. See, e.g., Laura Oren, Immunity and Accountability in Civil Rights Litigation: Who Should Pay?, 50 Pitt. L. Rev. 935, 1003 (1989) ("Deterrence . . . is most effective at the level where control lies. It is the government and not the individual employee, which has the ability to change policy, discipline misconduct, and require a different kind of training."). And with respect to the actions of state or D.C. officials, there are, as Justice Frankfurter noted, state causes of action for damages. Insofar as this panoply of remedies contains lacunae, I would leave it to Congress to fill them. The gaps tolerated by the Supreme Court and this circuit undermine Judge Ginsburg's argument that without resort to Section(s) 1983 and Bivens' suits, individuals like Crawford-El may not have redress. In Schweiker, for example, the Court acknowledged that "[t]he trauma to respondents, and thousands of others like them, must surely have gone beyond what anyone of normal sensibilities would wish to see imposed on innocent disabled citizens." Schweiker, 487 U.S. at 428-29. Nonetheless, the Court deferred to Congress' decision whether to leave a gap. Similarly, in Spagnola, this Circuit denied the appellants' argument that, because "no remedy whatsoever"existed for individuals aggrieved by minor personnel actions under the Civil Service Reform Act, the court was obliged to create a Bivens' remedy. This deference makes sense as a constitutional and practical matter; given their greater resources and access to information, legislators are more likely than district court judges to reach the most socially beneficial result.

See, e.g., Federal Tort Claims Act, 28 U.S.C. Section(s) 2674 (1994) (providing for a cause of action for some federal governmental activity that constitutes a tort under state law); Title VII of the Civil Rights Act of 1964, 42 U.S.C. Section(s) 2000e et seq. (1988 Supp. V 1993); Civil Service Reform Act of 1978, Pub.L. No. 95-454, 92 Stat. 111 (codified as amended in scattered sections of 5 U.S.C. (1994)) (establishing the Office of Special Counsel to investigate and prosecute allegations of supervisory abuse within the civil service disciplinary structure). Age Discrimination and Employment Act, 29 U.S.C. Section(s) 621-634 (1994) (civil action for employment discrimination based on age); Rehabilitation Act, 29 U.S.C. Section(s) 794 (1988) (civil action for discrimination on the basis of disability).

See, e.g., Bush v. Lucas, 462 U.S. 367, 390 (1983) (declining to extend Bivens' action to civil service employees, even while assuming that existing remedies do not provide complete relief for plaintiffs, "because we are convinced that Congress is in a better position to decide whether or not the public interest would be served by creating it"); Schweiker v. Chilicky, 487 U.S. 412, 429 (1988) (refusing to create a Bivens' remedy in light of an elaborate scheme devised by Congress and noting "[w]hether or not we believe that its response was the best response, Congress is the body charged with making the inevitable compromises required in the design of a massive and complex . . . program"); Spagnola v. Mathis, 859 F.2d 223, 228 (D.C. Cir. 1988) (en banc) (holding that "courts must withhold their power to fashion damages remedies when Congress has put in place a comprehensive system to administer public rights, has "not inadvertently" omitted damages remedies for certain claimants and has not plainly expressed an intention that the courts preserve Bivens' remedies" (citation omitted) (emphasis added)).

See, e.g., Bush, 462 U.S. at 389 ("Not only has Congress developed considerable familiarity with balancing governmental efficiency and the rights of employees, but it also may inform itself through factfinding procedures such as hearings that are not available to the courts."); United States v. Gilman, 347 U.S. 507, 511-513 (1954) ("The selection of that policy which is most advantageous to the whole involves a host of considerations that must be weighed and appraised. That function is more appropriately for those who write the laws, rather than for those who interpret them.").

In any event, that there are real gaps is doubtful: by 1985 only 30 Bivens suits out of more than 12,000 resulted in a monetary judgment for the plaintiff at the trial level with only four judgments actually having been paid. See Written Statement of John J. Farley, III, Director, Torts Branch, Civil Division, U.S. Department of Justice, to the Litigation Section of the Bar of the District of Columbia (May 1985) at 1. Obviously, the vast majority of these suits are meritless. See Fallon, Meltzer Shapiro, Hart and Wechsler's The Federal Courts and the Federal System 1122 (4th ed. 1996) ("The view that constitutional tort actions are less likely to prove meritorious than civil litigation has been confirmed as to both prisoner and nonprisoner actions . . ., although it is in the former class that the general lack of substance is most striking."). Prisoner suits serve less as a necessary deterrent to unconstitutional conduct (to put it mildly) than as a diversion from the monotony of prison life to plaintiffs such as Crawford-El, whose injury is the inconvenience of having some boxes being turned over to his brother-in-law. Perhaps all sides in this dispute would have been better off if the prison officials had agreed to provide an alternative form of entertainment to Crawford-El, maybe free cable, in return for not having to go through the expense and hassle of this lawsuit.

Congress has already taken steps to limit prisoner suits. See Prison Litigation Reform Act of 1995, Pub.L. No. 104-134, Section(s) 801 et al., 110 Stat. 1321 (1996).

* * * *

Although my reading of Harlow will reduce the costs to government officials — and the public — caused by Bivens actions and the impact of Pape on Section(s) 1983, much the better would be for Congress to legislate on the whole subject as it has on certain aspects of prisoner suits. The Supreme Court has recognized that when and if it does, the federal judiciary should beat a hasty retreat. See Bush, 462 U.S. at 368, 390. Ginsburg, Circuit Judge, concurring: I agree with the clear majority of my colleagues who conclude that the direct-evidence rule of Martin v. D.C. Metropolitan Police Dep't, 812 F.2d 1425, 1431 (D.C. Cir. 1987), should be abandoned. I also concur in Judge Williams' opinion insofar as it requires that a Section(s) 1983 or Bivens plaintiff who seeks damages from a government official for a constitutional tort must prove the defendant's unconstitutional motive (where that is an element of the tort) by clear and convincing evidence. As Judge Williams details, a plaintiff will feel the weight of this burden not only at trial but also in opposing a motion for summary judgment; in both contexts the plaintiff will have to present evidence that a jury could consider clear and convincing proof of the defendant's unconstitutional motive.

I cannot concur, however, in Judge Williams' attempt to place an even greater burden upon the plaintiff at the summary judgment stage. He would require the district court to grant summary judgment prior to discovery unless the plaintiff already has in hand evidence of the defendant's motive that a reasonable jury could find "clear and convincing." That seems a rather bold intrusion into the district court's management of the fact-finding process, an area in which we generally defer to the trial judge. The consequences would be twofold. First, Judge Williams' proposal would put compensation beyond the reach of even the plaintiffs with the most meritorious claims — a consequence arguably consistent with Harlow v. Fitzgerald, 457 U.S. 800, 814 (1982), in which the Supreme Court accepted that some deserving plaintiffs would be denied compensation in order to reduce the social costs of litigation against government officials. Second, Judge Williams' approach would invite an increase in the number of constitutional torts that are committed — a consequence more difficult to square with Harlow.

I. General Principles

In relating this case to Harlow, we must consider not only the compensatory role of constitutional tort liability but also its deterrent purpose. The rule announced in Harlow probably did not increase the frequency with which public officials knowingly violate someone's constitutional rights. An official who knows that the action he is contemplating would violate an individual's constitutional rights can hardly be confident that a court will later disagree — more precisely, that the court will conclude that the official's action was objectively reasonable under the law as clearly established at the time. Harlow is cold comfort, ex ante, to that official. This is why the Court could say in Harlow that the rule announced there would "provide no license to lawless conduct." 457 U.S. at 819.

We cannot make the same statement about the requirement that the plaintiff prove his case by clear and convincing evidence; as sure as we are that demand curves slope downward and that there will be more of a behavior when the price (or penalty) goes down, we can be confident that raising the plaintiff's burden of persuasion will embolden some additional Government officials to take actions that they know are unconstitutional. Although we cannot know the magnitude of that effect (i.e. the slope of the demand curve for tortious conduct), I agree with Judge Williams that we are justified in taking this step to contain the social cost of litigating constitutional torts that turn upon the defendant's motive.

My colleague, however, would take not only this but a second step beyond Harlow; he would not only raise the plaintiff's burden of persuasion but also require the plaintiff to obtain evidence without the ability to compel its production from those most likely to have it. No matter whether the plaintiff can demonstrate that he has a reasonable chance — or for that matter a virtual certainty — of obtaining such evidence from the defendant or even a third party, such as one of the defendant's coworkers, Judge Williams would deny him any discovery. This would further reduce the deterrent effect of constitutional tort liability, perhaps to a point below what is justified.

Judge Williams overlooks the point; Judge Silberman faces up to it but reminds us that "personal damage suits are decidedly not the only disincentive" to unconstitutional conduct. The federal statutes that he cites, however, do not justify the balance that he or Judge Williams would strike between the interests of injured plaintiffs and the public interest in avoiding unfounded litigation against government officials. First, those statutes do not reach all the motive-based constitutional torts for which a plaintiff can seek redress under Bivens or Section(s) 1983. Second, a plaintiff who seeks damages against a public official under any of the cited statutes has no greater access to discovery than does a plaintiff who sues for the same remedy under Bivens or Section(s) 1983; qualified immunity shields the public official from personal damage liability regardless of the particular type of action brought against him. See Todd v. Hawk, 72 F.3d 443, 445 n. 7 (5th Cir. 1995) ("Racial discrimination claims brought under Section(s) 1981 are subject to the defense of qualified immunity"); Hobson v. Wilson, 737 F.2d 1, 19 (D.C. Cir. 1994) ("section 1985(3) encompasses actions against federal officers, subject, of course, to considerations of qualified immunity").

Effective deterrence of unconstitutional conduct depends unavoidably upon exposing public officials to some risks that might also chill them in the proper exercise of their discretion. In order to obtain any other remedy or impose any sanction, the plaintiff or prosecutor respectively will have to show that some public official acted with a prohibited motive — racial, religious, or gender discrimination, retaliation for protected speech, or what have you. Although the public official will be shielded from personal liability and, perhaps, from the cost of retaining counsel, he will not be shielded from the demands upon his time, the risk of injury to his reputation, the emotional distress likely to attend an adversarial inquiry into whether his actions were basely motivated, or the possibility of unpleasant consequences apart from the litigation (such as losing his job) if the inquiry shows that his actions were improperly motivated. Therefore, that a particular rule, such as the one Judge Silberman proposes, would leave in place some deterrent effect because some types of cases might still be brought tells us little about whether the rule strikes an appropriate balance between our interest in deterring constitutional torts generally and our interest in reducing the social costs of litigation against public officials. Judge Silberman suggests, based upon the low success-rate of Bivens and Section(s) 1983 actions, that there is not much out there to deter. He does not consider, however, that the low-success rate is, in part, a result of the qualified immunity doctrine and other legal rules. We cannot know how much additional unconstitutional mischief the rules proposed by Judges Silberman and Williams would elicit, but that seems reason enough to proceed with more caution than either of them displays. A more prudent and discriminating approach — one that may preserve the desired deterrent while still lessening the burden now placed upon defendant public officials — would be to provide more guidance than we have heretofore given to district judges faced with the task of balancing, case by case, the competing values accommodated by the institution of qualified immunity. We could then rely upon them, as we normally do, to manage the fact-finding process that my colleagues would truncate with clear but Draconian rules.

When a defendant files a motion for summary judgment and the plaintiff argues that he needs discovery in order to withstand the motion, Rule 56(f) invests the district court with discretion to (1) deny the motion for summary judgment, (2) continue the motion pending discovery, or (3) "make such other order as is just." In a case involving qualified immunity, the district court abuses this discretion if it fails duly to consider not only the competing interests of the parties — as in any civil litigation — but also the social costs associated with discovery had against a government official.

Hence, while this court has acknowledged that "in the mine-run of cases" summary judgment is generally inappropriate until all discovery has been completed, Martin, 812 F.2d at 1436, we have also recognized that "creditable pleas of official immunity remove cases from the mine-run category," id. at 1436-37. Although we now reject then-Judge Ruth Bader Ginsburg's elevation of direct over circumstantial evidence, see id. at 1435, we ought not forget her description of our task in a case such as this — to "leav[e] some space for discovery" while "minimiz[ing] the burdens imposed upon government officials." Id. at 1437. In Martin we required the plaintiff to make factual allegations sufficiently precise to enable the district court to "employ with particular care and sensibility [its] large authority to exercise control over discovery." Id. at 1437. We expected that district courts would protect government officials from "unnecessary involvement in litigation" by "permit[ting] particularized interrogation of the defendants for the circumscribed purpose of ascertaining whether there is any substance" to the plaintiff's specific factual allegations. Id. at 1438.

Rather than looking further back, as Chief Judge Edwards does, to the concern expressed in Hobson, 737 F.2d at 30-31, that "in some circumstances plaintiffs are able to paint only with a very broad and speculative brush at the pre-discovery stage," we should go forward along the path to which Justice Ginsburg pointed us in Martin. Consideration of the social costs associated with litigation against public officials (which, as Harlow teaches, weighs heavily against discovery) should constrain to this extent the district court's discretion to continue a summary judgment motion pending discovery: If, when the defendant moves for summary judgment, the plaintiff cannot present evidence that would support a jury in finding that the defendant acted with an unconstitutional motive, then the district court should grant the motion for summary judgment unless the plaintiff can establish, based upon such evidence as he may have without the benefit of discovery and any facts to which he can credibly attest, a reasonable likelihood that he would discover evidence sufficient to support his specific factual allegations regarding the defendant's motive.

Chief Judge Edwards too speaks of requiring a "reasonable likelihood that additional discovery will uncover evidence to buttress the claim," but that is not the same as requiring a reasonable likelihood, based upon specific evidence within the plaintiff's command, that discovery will uncover evidence sufficient to sustain a jury finding in the plaintiff's favor. Moreover, the Chief Judge's emphasis upon some plaintiffs' ability to "paint only with a broad and speculative brush," and upon the district court's almost unfettered discretion (in the mine-run of cases, that is) to continue a summary judgment motion pending discovery, suggests a substantial difference in our expectations of the district court.

Permitting a plaintiff to pursue limited discovery only upon showing that he has a reasonable likelihood of turning up evidence that a jury could consider clear and convincing proof of the defendant's unconstitutional motive would leave more space for discovery than would Judge Williams or Judge Silberman, would still protect the public from the costs of pointless discovery against Government officials, and would not usurp the district court's authority over the course of the litigation. Moreover, I see no reason to doubt the district court's willingness or ability to strike anew in each case the balance that underlies the doctrine of qualified immunity. Indeed, a district judge, whose experience with the management of discovery is far more extensive than ours, whose familiarity with the case and with the litigants is more immediate, and whose tools for controlling the course of litigation are more subtle and precise, is eminently qualified for this task.

II. Application to this Case

I agree with my colleagues who conclude that Crawford-El adequately alleged a violation of clearly established constitutional law and that we must therefore remand this case to the district court. But if on remand Britton moves for summary judgment prior to discovery and Crawford-El cannot substantially supplement the record now before us, then it would be an abuse of discretion for the district court to deny the motion or to continue it pending discovery.

A. The Summary Judgment Standard

In Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 254-56 (1986), the Supreme Court explained how a district court should determine whether a plaintiff has submitted evidence sufficient to withstand a summary judgment motion when the plaintiff must prove an element of his claim — in that libel case it was actual malice — by clear and convincing evidence:

[T]here is no genuine issue if the evidence presented in the opposing affidavits is of insufficient caliber or quantity to allow a rational finder of fact to find actual malice by clear and convincing evidence.

Thus, in ruling on a motion for summary judgment, the judge must view the evidence presented through the prism of the substantive evidentiary burden. . . . It makes no sense to say that a jury could reasonably find for either party without some benchmark as to what standards govern its deliberations and within what boundaries its ultimate decision must fall, and these standards and boundaries are in fact provided by the applicable evidentiary standards.

Our holding that the clear — and — convincing standard of proof should be taken into account in ruling on summary judgment motions does not denigrate the role of the jury. It by no means authorizes trial on affidavits. Credibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge, whether he is ruling on a motion for summary judgment or for a directed verdict. The evidence of the non-movant is to be believed, and all justifiable inferences are to be drawn in his favor. . . .

In sum, we conclude that the determination of whether a given factual dispute requires submission to a jury must be guided by the substantive evidentiary standards that apply to the case. . . . Thus, where the factual dispute concerns actual malice, clearly a material issue in a New York Times case, the appropriate summary judgment question will be whether the evidence in the record could support a reasonable jury finding either that the plaintiff has shown actual malice by clear and convincing evidence or that the plaintiff has not.

Thus, although the plaintiff is entitled to have all rational inferences drawn in his favor on intermediate facts — such as hostility, in this case — those facts must add up to clear and convincing evidence of the ultimate facts that he must prove — here, that Britton (1) in order to retaliate against Crawford-El for exercising his constitutional rights (2) knowingly gave Crawford-El's legal papers to his brother-in-law.

B. Crawford-El's Complaint

Let us now look at Crawford-El's sworn declarations to see whether they are sufficient to withstand Britton's no doubt imminent motion for summary judgment. In paragraph 6 of his fourth amended complaint, Crawford-El declares that

[1] Ms. Britton persistently displayed toward prisoners a cavalier attitude — manifesting a view that prisoners were beneath her, disentitled to dignity, and unworthy of civil treatment. [2] Ms. Britton was hostile to plaintiff, in particular, because she knew plaintiff . . . had been in charge of the law library [and] had helped many prisoners prepare . . . grievance forms or appeals of disciplinary actions, and had a reputation for asserting legal rights and knowing the administrative procedures for doing so. [3] Ms. Britton deemed plaintiff "too big for his britches."

The first sentence establishes merely that Britton did not like prisoners generally; it says nothing specific about her alleged unconstitutional motive. The second sentence states a fact about Britton's state of mind, to which Crawford-El may not testify without laying a foundation. See Fed.R.Civ.Pro. 56(e) ("affidavits shall be made on personal knowledge"); and Fed.R.Evid. 602 (accord) and 701 ("testimony in the form of opinions or inferences is limited to those opinions or inferences which are . . . rationally based on the perception of the witness"). The third sentence, provided without context, does not tell us why Britton said that Crawford-El was "too big for his britches" or even whether the statement manifests hostility.

In paragraph 9 of the complaint, Crawford-El declares that Ms. Britton was among those who were hostile to the Inmate Grievance Committee and to plaintiff's efforts to seek redress of prisoner grievances. On one occasion when plaintiff was typing [Housing and Adjustment] Board papers in the Q Block office, Ms. Britton came in and said to Cpt. (then Lt.) Brummell in a caustic manner that she (Cpt. Brummell) should watch out for plaintiff and make sure he wasn't using the typewriter to write up [grievance forms] or lawsuits. As Ms. Britton said this she stood over plaintiff to see what he was typing.

Britton's concern, even if caustically expressed, that Crawford-El not conduct his jailhouse law practice when he was supposed to be performing administrative work is not evidence of hostility to Crawford-El's efforts to seek redress of prisoner grievances.

In paragraph 12 of the complaint, Crawford-El declares that The day after the [first Washington Post] article was published [April 21, 1986], defendant Britton ordered plaintiff into her office.

Corporal Barrett, then Officer in Charge of Dorm K2, escorted plaintiff there. Ms. Britton was visibly upset. After ignoring plaintiff for a considerable period, she asked him if he had arranged the visit by the reporter. When plaintiff said that he had, she asked him how he had done it. Plaintiff showed her the visitor application naming the reporters invited and their address and pointed out that Ms. Britton had approved the application. [7] Ms. Britton became enraged and accused plaintiff of tricking her. Plaintiff denied tricking her. Ms. Britton said plaintiff had embarrassed her before her coworkers by having the reporter come. Ms. Britton made a telephone call trying to get plaintiff placed in restrictive confinement in Q Block. [11] When this effort failed she said that so long as plaintiff was incarcerated she was going to do everything she had to do to make it as hard for him as possible. A few days later Ms. Britton had plaintiff transferred to the Department's Central Facility.

Crawford-El's statement (in the 7th sentence) that Britton "became enraged" when she thought she had been duped by Crawford-El does not help his case. On the contrary, that she was angered at being tricked — Crawford-El has no constitutional right to trick his keeper — provides a qualifying context for Crawford-El's most significant declarations: that Britton said that she was embarrassed by the article and that she would make life hard for Crawford-El.

Judge Williams brushes the allegations aside as "self-serving." Self-serving opinions, inferences, and conclusions without a basis in perceptible fact may not be sufficient to withstand a summary judgment motion even under the mere preponderance standard; but neither is summary judgment "a procedure for resolving a swearing contest" over concrete facts, see Jackson v. Duckworth, 955 F.2d 21, 22 (7th Cir. 1992) (Posner, J.) (Section(s) 1983 action against prison officials), should such a contest arise — Britton has not contradicted Crawford-El's declaration with her own sworn statement. Recall Anderson, in which the Supreme Court instructed, again on a summary judgment motion in a case where the plaintiff must prove an element by clear and convincing evidence, that "[c]redibility determinations . . . are jury functions, not those of a judge," and that "[t]he evidence of the non-movant is to be believed." Suppose Britton (or Corporal Barrett) were to corroborate the alleged threat, however; without more it would not clearly and convincingly indicate that Britton's decision to deliver Crawford-El's property to his brother-in-law was unconstitutionally motivated. Britton allegedly made the threat in a moment of anger in April 1986; she delivered Crawford-El's property to his brother-in-law in September 1989, at the same time (according to Crawford-El's own declaration) that she was calling the families of other prisoners, who, like Crawford-El, were being sent to the federal prison in Petersburg, Virginia and threatening to discard the prisoner's property if a family member did not come to collect it.

In paragraph 15 of the complaint, Crawford-El declares that [During a transfer to the Spokane County Jail in Washington State] Correctional Officer Ballard, with Ms. Britton's knowledge, made a videotape of prisoners [including Crawford-El] while they were handcuffed, leg-shackled, and chained about their waists. Plaintiff and several others protested to Ms. Britton that the videotaping violated their privacy rights. Plaintiff said to her that the videotaping could not be done without the prisoners' written authorization. Ms. Britton responded, "You're a prisoner, you don't have any rights."

What does this show? That Britton was generally insensitive to the constitutional rights of prisoners? Maybe. More likely it shows simply that she did not believe that a prisoner has a right not to be videotaped. In either event, it is not very probative on the question whether (nine months later) she retaliated against Crawford-El for exercising his first amendment rights.

In paragraph 17 of his complaint, Crawford-El alleges that shortly after publication of a second Washington Post article (December 1988) in which he was quoted on the topic of jailhouse lawyers, Britton told one Captain Manning of the Spokane County Jail (to which Crawford-El had been transferred) that Crawford-El was a "legal troublemaker." As Judge Williams observes, Britton's describing Crawford-El as a "legal troublemaker" is scant evidence of hostility. Indeed, viewed as an expression of hostility it is too mild to support the inference that she bore a grudge against Crawford-El nine months later when she gave his legal papers to his brother-in-law.

Finally, Crawford-El alleges that on August 18, 1989, when he and other prisoners told Britton that property left in her possession included important legal material, she "smirked and spoke in a cavalier manner," but "informed [Crawford-El] that she understood his need both for his personal property and his legal material and that she would personally see to it that [he] would get them." Crawford-El alleges also that upon arriving at the federal prison in Petersburg, Virginia several other D.C. prisoners informed him that Britton had asked their families to pick up their property or she would throw it away. Crawford-El offers no evidence indicating that Britton bore an unconstitutional animus toward any of these other prisoners; on the contrary, that she apparently treated the property of several prisoners in the same manner jibes with her sworn declaration that she was motivated by what she understood to be the policy of the Federal Bureau of Prisons.

In sum, even if Crawford-El were by discovery to get corroboration of every sworn declaration in his fourth amended complaint, he would not have evidence that would clearly and convincingly indicate to a reasonable jury what he must prove. At best, his evidence would establish that in a moment of anger in April 1986, Britton threatened to retaliate against him for embarrassing her by making statements to a Washington Post reporter, and that as recently as December 1988 she resented his jailhouse lawyering. Crawford-El points to no evidence that Britton did anything to make good on the 1986 threat before she delivered his legal papers to his brother-in-law in September 1989, nor to anything suggesting that he could discover evidence of such rabid hostility toward him that it would constitute clear — and — convincing circumstantial evidence that Britton was retaliating against Crawford-El by treating him as she treated other similarly situated prisoners. Under the clear — and — convincing evidence standard, no reasonable jury could find on these facts that Britton acted with an unconstitutional motive in 1989 and Crawford-El has not offered a reason to believe that more evidence can be discovered. If on remand he has nothing more significant to offer, then the plaintiff should be denied discovery and the defendant's motion for summary judgment should be granted.

[106] KAREN LECRAFT HENDERSON, Circuit Judge, concurring:

It is high time that we scuttle the awkward direct/circumstantial evidence distinction and I fully endorse the clear and convincing standard the plurality adopts in its stead. I am at a loss to understand, however, why my colleagues chose this case to do so. Despite repeated opportunities to replead below, both pro se and through appointed counsel, the plaintiff has failed, as he has so many times before, to allege facts demonstrating the deprivation of any constitutional right (clearly established or not). In short, his constitutional claims are frivolous and the district court would have done well to dismiss the complaint sua sponte under the in forma pauperis statute, either before or after our first remand. See 28 U.S.C. Section(s) 1915(d) (authorizing district court to dismiss in forma pauperis suit "if satisfied that the action is frivolous or malicious"). Nevertheless, my colleagues choose yet again to ignore the hopeless infirmity of the plaintiff's claims and insist on maintaining life support. On remand, the district court will no doubt at long last lay the plaintiff's meritless claims to rest. I would have pulled the plug long ago.

See, e.g., Best v. District of Columbia, No. 92-7196 (D.C. Cir. 1995) (summarily affirming district court's dismissal of claim of wrongful videotaping of prisoners); Crawford-El v. Meese, No. 88-8034 (D.C. Cir. 1990) (summarily affirming dismissal of challenge to prison diet); Crawford-El v. District of Columbia Dep't of Corrections, No. 91-2413 (D.D.C. 1992) (dismissing claim for damages resulting from snakebite allegedly caused by guards' negligence); Crawford-El v. Barry, No. 88-0715, (D.D.C. 1989) (sua sponte dismissing claims of wrongful deprivation of visitation privileges and of denial of prison religious classes); Crawford-El v. Shapiro, No 88-2339 (D.D.C. 1988) (dismissing malpractice claim).

See Siegert v. Gilley, 500 U.S. 226, 233 (1991) (finding qualified immunity where plaintiff "failed not only to allege the violation of a constitutional right that was clearly established at the time of [the defendant's] actions, but also to establish the violation of any constitutional right at all").

The gist of the plaintiff's retaliation claim is this: In September 1989 defendant Britton handed the plaintiff's belongings over to his brother-in-law rather than sending them directly to his new penal home, intending thereby to wreak vengeance upon the plaintiff for speaking to the press in 1986 and 1988. This is absurd. The only allegation that even suggests a retaliatory motive is that more than three years earlier, on April 21, 1986, the day after the first article was published, Britton accused the plaintiff of tricking her into signing the reporter's visitor's pass and made a general threat that she would "do everything she had to to make it as hard for him as possible." Whatever probative force the alleged threat might otherwise have is undercut by the length of time that elapsed before the "diversion" of the plaintiff's property. The plaintiff's own factual allegations, on the other hand, reveal an innocent, even beneficent, motive for Britton's handling of the plaintiff's property. According to the fourth amended complaint, the plaintiff's brother-in-law, who was employed at the Department of Corrections, "informed plaintiff that he had been called by Ms. Britton, that she had told him she was concerned about his legal material and other property, that she was afraid that the property might get lost were she to send it from her office to the Lorton Property Officer for mailing to plaintiff." Appellant's App. 24-25. Thus, it appears Britton simply wanted to ensure that the property reached the plaintiff promptly and intact. And it would have had not the plaintiff himself prevented its delivery. In any event, intent aside, what Britton did had the effect of providing the plaintiff with exactly what he claims he wanted: prompt access to his property, if not in the precise manner he would have chosen (or at the taxpayer's expense). Thus, the complaint's claim of unconstitutional retaliation is "nonsensical on its face." See Adams v. Rice, 40 F.3d 72, 75 (4th Cir. 1994) (so characterizing inmate's complaint claiming unconstitutional reprisal by prison officials who, after he requested "protective custody," placed him in segregation, which, as the court noted, gave him "the protective custody he requested or its approximate equivalent").

The complaint does contain several allegations which, if true, may indicate Britton's general hostility toward the plaintiff and growing impatience with his complaints and litigiousness. While such evidence might support the plaintiff's now defunct claim of interference with his first amendment right to petition the court, it does not demonstrate intent to retaliate for the press interviews.

Even assuming, against common sense, that Britton's handling of the plaintiff's property amounted to some sort of punishment, he has no claim under 42 U.S.C. Section(s) 1983. As the panel noted in the plaintiff's first appeal, there is a "general principle that some showing of injury is a prerequisite to a constitutional tort action." Crawford-El v. Britton, 951 F.2d 1314, 1322 (D.C. Cir. 1991) (citing Butz v. Economou, 438 U.S. 478, 504 (1978)) (Crawford-El I). In addition, the injury must be of constitutional dimension: "There is, of course, a de minimis level of imposition with which the Constitution is not concerned." Ingraham v. Wright, 430 U.S. 651, 674 (1977). The plaintiff's retaliation claim is below the de minimis level. The only alleged injuries attributable to Britton are the costs of mailing three boxes of belongings to Florida — incurred when the plaintiff finally allowed his mother to send them — and, perhaps, a brief delay in receiving them and the consequent cost of temporarily replacing a few items, as well as the emotional distress flowing therefrom. Such slight harm does not cross the constitutional threshold. Cf. Buthy v. Commissioner of Office of Mental Health, 818 F.2d 1046, 1050 (2d Cir. 1987) (holding that state mental institution rule requiring forensic unit patients to remain awake for fixed 16-hour period is "a de minimis imposition on individual liberty" that cannot support due process claim); Walsh v. Louisiana High Sch. Athletic Ass'n, 616 F.2d 152, 158 (5th Cir. 1980) (rejecting student's challenge to "student transfer rule," making student attending high school outside his home district ineligible to participate in interscholastic athletics for one year, because of "the de minimis nature of the burden placed on the plaintiffs' free exercise of religion"). It is therefore redressable, if at all, through a local conversion suit, not in federal court under section 1983. See Crawford-El I, 951 F.2d at 1318 ("At worst, the act might constitute a common law conversion. . . ."); Paul v. Davis, 424 U.S. 693, 699-701 (1976) (state law tort does not a constitutional deprivation make).

Any other damages resulted not from Britton's decision but from the plaintiff's own intransigence. It is even doubtful that he would have suffered delay or replacement costs if he had allowed his mother to forward his belongings promptly.

It is true that an ordinarily permissible act may become a constitutional deprivation if performed in retaliation for the exercise of a first amendment right. See, e.g., Perry v. Sindermann, 408 U.S. 593 (1972) (decision not to renew untenured professor's contract); Cornell v. Woods, 69 F.3d 1383, 1387-88 (8th Cir. 1995) (transfer of inmate to different prison); Meriwether v. Coughlin, 879 F.2d 1037 (2d Cir. 1989) (change in inmate's work assignment); Jackson v. Cain, 864 F.2d 1235 (5th Cir. 1989) (filing disciplinary charges). The threshold injury requirement nevertheless remains. A retaliation claim is actionable precisely "because retaliatory actions may tend to chill individuals' exercise of constitutional rights." American Civil Liberties Union of Md., Inc. v. Wicomico County, 999 F.2d 780, 785 (4th Cir. 1993) (citing Perry v. Sindermann, 408 U.S. at 597). Thus, the "test" for whether one exists "is whether the adverse action taken by the defendants is likely to chill the exercise of constitutionally protected speech." McGill v. Board of Educ., 602 F.2d 774, 780 (7th Cir. 1979) (citing Pickering v. Board of Educ., 391 U.S. 563 (1968)); see also DiMeglio v. Haines, 45 F.3d 790, 806 (4th Cir. 1994) ("Not every restriction is sufficient to chill the exercise of First Amendment rights, nor is every restriction actionable, even if retaliatory."); Bart v. Telford, 677 F.2d 622, 625 (7th Cir. 1982) ("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."). The plaintiff's claim flunks the test. It is difficult to imagine that the minimal adverse effect (if any) of Britton's actions was likely to chill or deter him (or any reasonable person) from exercising his first amendment rights. Thus, even if retaliatory, Britton's conduct cannot give rise to a constitutional cause of action. See DiMeglio v. Haines, 45 F.3d at 806-07 (stating that claim of retaliatory reassignment of zoning investigator "to a geographic subset of the very region from which he formerly had derived his zoning assignments" "likely would not be sufficiently adverse to implicate the First Amendment"); Raymon v. Alvord Indep. Sch. Dist., 639 F.2d 257 (5th Cir. March Unit A 1981) (holding that student's claim of retaliatory lowering of algebra grade, resulting in "insignificant decrease in her overall grade point average" that did not affect her class rank, was "patently insubstantial"). In sum, the plaintiff's meritless claims should have been long since booted and, in any event, should never have been dignified with en banc review. Nevertheless, the issues have been joined and I concur in the plurality's disposition of them.

[111] EDWARDS, Chief Judge, with whom WALD, RANDOLPH, ROGERS, and TATEL, Circuit Judges, concur, concurring in the judgment to remand:

Justice Felix Frankfurter once wrote:

[T]he only sure safeguard against crossing the line between adjudication and legislation is an alert recognition of the necessity not to cross it and instinctive, as well as trained, reluctance to do so.

Felix Frankfurter, Some Reflections on the Reading of Statutes, 47 Colum. L. Rev. 527, 535 (1947). This admonition has been thoroughly lost on my colleagues who have a different view of this case. Without any directive from Congress or mandate from the Supreme Court, my colleagues run roughshod over the Federal Rules of Civil Procedure and invent new evidentiary standards that would make it all but certain that an entire category of constitutional tort claims against government officials — whether or not meritorious — would never be able to survive a defendant's assertion of qualified immunity. This result is both unfathomable and astonishing.

Fortunately, a clear majority of the court agrees that plaintiffs who file constitutional tort claims alleging that governmental officials acted with unconstitutional intent are not obligated to meet any form of heightened pleading standard in their initial complaint. Rather, it is clear that plaintiffs need only adhere to the basic notice pleading requirements of Federal Rule of Civil Procedure 8(c), and need not anticipate an affirmative defense of qualified immunity. Further, the opinions of the court make it clear that we reject any heightened pleading rule that would require plaintiffs to plead direct, rather than circumstantial, evidence.

However, I strongly disagree with Judge Williams's and Judge Henderson's opinions suggesting that, in the face of a defendant's claim for qualified immunity, a plaintiff faces dismissal (even without any discovery) unless he or she can put forward specific, nonconclusory factual allegations establishing the defendant's unconstitutional intent by "clear and convincing" evidence. I similarly reject Judge Silberman's opinion that would go even further and completely rewrite the law to say that a motive-based claim can never survive a motion to dismiss so long as the defendant's behavior can be seen as consistent with any possible legal motivation, i.e., without regard to whether it can be demonstrated that the presumed legal motivation is not what actually prompted the actions that are at issue.

It is not surprising that these opinions (along with the separate opinion of Judge Ginsburg) can find no safe path to common ground. These opinions offer judgments that are in complete defiance of the Federal Rules of Civil Procedure, inventing evidentiary standards out of whole cloth and overlaying them onto the established procedures for adjudicating lawsuits in our federal courts. Because there is no principled basis for these judgments, the opinions flounder in their rationales and command no majority position. There are some telling similarities in the opinions, for each suffers from the same glaring infirmities: the opinions are completely unmoored to any legislative enactment or Supreme Court precedent, and they are contrary to the law of every other court of appeals in the nation. The net result is judicial activism at its most extreme. Because I believe that this court has no authority to amend the Federal Rules and to ignore established precedent, I reject the positions offered by my colleagues.

Although all members of the court appear to agree that this case must be remanded for further proceedings, the court is sharply divided over the basis for remand. Judge Williams suggests that Judge Ginsburg's opinion (pursuant to which Crawford-El might get discovery) provides a "common denominator" of the reasoning of a majority, as the opinion consistent with the disposition on the narrowest grounds. Whether or not Judge Ginsburg's opinion controls, it is clear that a majority of the court agrees that the trial judge must have discretion to consider the appropriate circumstances under which discovery should be allowed.

* * * *

A. This Circuit's Jurisprudence

The issue raised by this case is not a new one. Ever since the Supreme Court's opinion in Harlow v. Fitzgerald, 457 U.S. 800 (1982), holding that government officials generally can be held liable for civil damages only if they "violate clearly established statutory or constitutional rights," id. at 818, federal appeals courts have been forced to apply the principles of Harlow to cases in which plaintiffs allege that defendants took action against them with unconstitutional motivation. The difficulty with these cases is that, in some instances, plaintiffs might allege facts demonstrating that defendants have acted lawfully, append a claim that they did so with an unconstitutional motive, and as a consequence usher defendants into discovery, and perhaps trial, with no hope of success on the merits. The result would be precisely the burden Harlow sought to prevent.

Hobson v. Wilson, 737 F.2d 1, 29 (D.C. Cir. 1984), cert. denied, 470 U.S. 1084 (1985).

In order to prevent frivolous claims from reaching such an advanced stage in the proceedings, the Hobson court required that these motive-based complaints provide "nonconclusory allegations of evidence of such intent" in order to survive a motion to dismiss and proceed to discovery. Id. According to the court, "[t]he allegations on this issue need not be extensive, but they will have to be sufficiently precise to put defendants on notice of the nature of the claim and enable them to prepare a response." Id. Unlike the rule proposed by Judge Williams, this test did not create a new judge — made evidentiary standard, but was simply a "firm application of the Federal Rules of Civil Procedure," as called for by the Supreme Court in Harlow, 457 U.S. at 819-20 n. 35 (internal quotation omitted). Moreover, in Hobson, we noted that, "in some circumstances plaintiffs are able to paint only with a very broad and speculative brush at the pre-discovery stage, and that overly rigid application of the rule . . . could lead to dismissal of meritorious claims;" we therefore warned district court judges to act cautiously and dismiss only those claims that were "devoid of factual support." Hobson, 737 F.2d at 30-31.

In subsequent cases, this court, while purporting to remain faithful to Hobson, appeared to invent a new requirement, that plaintiffs plead only direct, as opposed to circumstantial, evidence of defendants' unconstitutional motivation. Given that the Supreme Court has stated that the probative value of circumstantial evidence "is intrinsically no different from testimonial evidence," Holland v. United States, 348 U.S. 121, 140 (1954), and that such evidence can in some cases be "more certain, satisfying and persuasive than direct evidence," Michalic v. Cleveland Tankers, Inc., 364 U.S. 325, 330 (1960), the so-called direct-evidence rule never has made any sense. I therefore join my colleagues in emphatically rejecting such an illogical and unjustified requirement. I also agree with my colleagues that any evidentiary standard to be applied on a motion for summary judgment must be applied consistently at every subsequent stage of the proceedings before the trial court. Otherwise, as with the direct evidence rule, plaintiffs would face a higher burden to survive a pre-trial motion than they would face in order to prevail at trial.

Having corrected our wrong turn towards a direct-evidence rule, we should not now reach out and invent yet another arbitrary and unjust standard. The correct decision in this case is to return to the sound principles set forth by the court in Hobson.

B. The Hobson Standard

Although this court has sometimes referred to the rule enunciated in Hobson as a "heightened pleading standard," see, e.g., Siegert v. Gilley, 895 F.2d 797, 801 (D.C. Cir. 1990), aff'd on other grounds, 500 U.S. 226 (1991); Smith v. Nixon, 807 F.2d 197, 200 (D.C. Cir. 1986), that label is misleading because application of the Hobson principles does not necessarily affect what the plaintiff must put in the complaint. Indeed, the Supreme Court made it clear in Gomez v. Toledo, 446 U.S. 635, 640 (1980), that, as a matter of substantive law, "two — and only two — allegations are required in order to state a cause of action" under 42 U.S.C. Section(s) 1983 (1994). A plaintiff must allege only that the defendant "has deprived him of a federal right" and has "acted under color of state or territorial law." Gomez, 446 U.S. at 640. A defendant's qualified immunity is an affirmative defense, and, therefore, "the burden of pleading it rests with the defendant" under the Federal Rules, which provide that the defendant must plead any "`matter constituting an avoidance or affirmative defense.'" Id. (quoting Fed.R.Civ.P. 8(c)). Thus, pursuant to Gomez, a plaintiff has no obligation to anticipate or respond to a potential qualified immunity defense in the initial complaint.

Once the defendant actually asserts the qualified immunity defense, however, the court must then determine whether the plaintiff can offer a sufficient factual basis to support the allegations of unconstitutional animus and therefore overcome qualified immunity. Under the Federal Rules, there are a number of appropriate mechanisms available by which the plaintiff can provide this additional factual support. For example, pursuant to Rule 7(a), the plaintiff may file a reply that sets out the plaintiff's evidence relevant to immunity and the material that the plaintiff claims is reasonably likely to lead to pertinent additional evidence. See Schultea v. Wood, 47 F.3d 1427, 1432-33 (5th Cir. 1995) (en banc). Alternatively, the plaintiff may file an amended complaint or a more definite statement, or the court can use its discretionary power over discovery under Rule 26(b) to limit initial discovery to a brief interrogatory concerning the plaintiff's evidence relevant to immunity. In any of these scenarios, the trial court is able to evaluate the nature of the plaintiff's allegations concerning unconstitutional intent, prior to ruling on a defense motion for summary judgment.

Federal Rule of Civil Procedure 7(a) states that "the court may order a reply to an answer or a third-party answer."

Federal Rule of Civil Procedure 15(a) permits a party to amend its complaint at any time "by leave of court."

Federal Rule of Civil Procedure 12(e) permits a Motion for More Definite Statement if a pleading "is so vague or ambiguous that a party cannot reasonably be required to frame a responsive pleading."

Federal Rule of Civil Procedure 26(b)(2)(iii) permits the court to alter the limits on discovery if the trial judge determines that "the burden or expense of the proposed discovery outweighs its likely benefit, taking into account the needs of the case, the amount in controversy, the parties' resources, the importance of the issues at stake in the litigation, and the importance of the proposed discovery in resolving the issues."

Thus, rather than refer to the Hobson test as a heightened pleading requirement, I agree with Judge Easterbrook that we should "speak instead of the minimum quantum of proof required to defeat the initial motion for summary judgment." Elliott v. Thomas, 937 F.2d 338, 345 (7th Cir. 1991), cert. denied, 502 U.S. 1121 (1992). Under the principles enunciated in Hobson, plaintiffs can survive an initial motion for summary judgment, prior to discovery, by providing "nonconclusory allegations of evidence" of the defendant's unconstitutional intent. Hobson also rightly recognized that the courts should be cautious in applying this standard, lest meritorious claims be dismissed. For example, it will sometimes be the case that the relevant evidence is in the possession of the defendant and is therefore unavailable to the plaintiff without further discovery. Thus, if the plaintiff can show a reasonable likelihood that additional discovery will uncover evidence to buttress the claim, the trial judge may invoke Rule 56(f) and deny the summary judgment motion.

Federal Rule of Civil Procedure 56(f) expressly grants the trial judge broad discretion to order discovery prior to ruling on a summary judgment motion, where the party opposing the motion cannot "present by affidavit facts essential to justify the party's opposition." This court has explicitly held that the decision whether or not to stay discovery pursuant to Rule 56(f) is committed to the sound discretion of the District Court. White v. Fraternal Order of Police, 909 F.2d 512, 517 (D.C. Cir. 1990). Yet, the new evidentiary standard proposed by Judge Williams would effectively strip the trial judge of this discretion, by denying any discovery to plaintiffs unless they can provide "clear and convincing" evidence prior to discovery.

These procedures are part of the standard apparatus provided by the Federal Rules to enable trial judges in civil suits to differentiate meritorious claims from frivolous ones, and the Supreme Court has never suggested that this same apparatus is somehow inadequate when it comes to the particular immunity concerns expressed in Harlow. Indeed, as Justice Kennedy has pointed out, the objective standard for qualified immunity articulated in Harlow was based on the fact that the standards for summary judgment at the time "made it difficult for a defendant to secure summary judgment regarding a factual question such as subjective intent." Wyatt v. Cole, 504 U.S. 158, 171 (1992) (Kennedy, J., concurring). Now, however, "subsequent clarifications to summary-judgment law have alleviated that problem, by allowing summary judgment to be entered against a nonmoving party `who fails to make a showing sufficient to establish the existence of an element essential to that party's case, and on which that party will bear the burden of proof at trial.'" Id. (quoting Celotex Corp. v. Catrett, 477 U.S. 317, 322 (1986)). As a result, Rule 56 is now more than adequate to dispose of unmeritorious claims without appellate judges taking it upon themselves to invent new evidentiary standards designed to address particular categories of cases.

C. The Proposed Standard

Judge Williams's opinion argues that its Draconian rule requiring "clear and convincing" evidence is necessary to vindicate the substantive right of qualified immunity. I am inclined to agree with the view of the Seventh Circuit that "it is hard to depict a `right not to be tried'" as a substantive, rather than procedural, right. Elliott, 937 F.2d at 345. However, even were I to assume that qualified immunity is a substantive right, there is no valid justification for requiring plaintiffs to satisfy a "clear and convincing" evidence test in the cases here at issue. Indeed, there is great irony in the judgment offered by those judges who subscribe to Judge Williams's opinion, for the new rule that they propose would have a devastating impact on potential plaintiffs who already face substantial burdens in attempting to pursue civil rights claims. Recognizing these burdens, Chief Judge Posner has argued that there is a "peculiar perversity" in imposing a heightened standard in cases involving prison inmates because "it is far more difficult for a prisoner to write a detailed complaint than for a free person to do so" due to the fact that prisoners have no power to investigate their claims and gather evidence prior to obtaining discovery. Billman v. Indiana Dep't of Corrections, 56 F.3d 785, 789-90 (7th Cir. 1995); see also Kit Kinports, Qualified Immunity in Section 1983 Cases: The Unanswered Questions, 23 Ga. L. Rev. 597, 647 (1989) (The author argues that, in a case where the motive underlying a defendant's actions is a fact solely within the knowledge of the defendant, a court could not fairly grant defendant's motion for summary judgment before plaintiff has been given an opportunity to conduct discovery on this issue.). Thus, with regard to such cases, the standard proposed by Judge Williams, while purporting to permit some intent-based qualified immunity claims, would, as a practical matter, make it virtually impossible for these claims ever to survive a motion to dismiss. See David Rudovsky, The Qualified Immunity Doctrine in the Supreme Court: Judicial Activism and the Restriction of Constitutional Rights, 138 U. Pa. L. Rev. 23, 63 (1989) ("Where the plaintiff must establish the culpability element as part of the constitutional claim, denial of discovery on this issue would make it impossible to prove certain cases.").

In reading the opinions by Judge Williams and Judge Henderson, one is left with the impression that a "clear and convincing" standard is deemed necessary because, without it, some plaintiffs in section 1983 cases might actually prevail on their claims. Yet, it is overwhelmingly clear that the Court in Harlow never for a moment intended to insulate government officials from liability in all cases where the official's state of mind is a necessary element of the constitutional violation alleged. In fact, in Harlow itself the plaintiff alleged that the defendants had violated his First Amendment rights by dismissing him in retaliation for testifying before a congressional committee. As then-Judge Ruth Bader Ginsburg has pointed out,

[h]ad the Court intended its formulation of the qualified immunity defense to foreclose all inquiry into the defendants' state of mind, the Court might have instructed the entry of judgment for defendants . . . on the constitutional claim without further ado. In fact, the Court returned the case to the district court in an open-ended remand, a disposition hardly consistent with a firm intent to delete the state of mind inquiry from every constitutional tort calculus.

Martin v. District of Columbia Metro. Police Dep't, 812 F.2d 1425, 1432 (D.C. Cir.), vacated in part, 817 F.2d 144 (D.C. Cir.), reinstated, 824 F.2d 1240 (D.C. Cir. 1987).

Moreover, if a "clear and convincing" evidence standard were truly necessary to vindicate defendants' alleged substantive right not to be tried, as some of my colleagues seem to believe, one wonders why no other circuit has seen fit to embrace such a rule. Indeed, although nearly every other federal appeals court in the nation has addressed the precise issue that we face today, not one has adopted a standard even approaching the positions offered by my colleagues who view this case differently. Instead, all ten circuits that have addressed the issue have adopted formulations that are essentially identical to the one laid out in Hobson and echoed in Justice Kennedy's concurrence in Siegert v. Gilley, 500 U.S. 226, 236 (1991) (Kennedy, J., concurring) ("Upon the assertion of a qualified immunity defense the plaintiff must put forward specific, nonconclusory factual allegations which establish malice, or face dismissal.").

Although some of the circuit courts have actually adopted some form of so-called "heightened pleading" requirement and have chosen to test the plaintiff's claims at the complaint stage (an approach that I believe runs counter to Gomez), the more important point is that, regardless of when they apply the test, the courts have been quite consistent in articulating the appropriate evidentiary burden. See Blue v. Koren, 72 F.3d 1075, 1084 (2d Cir. 1995) ("[T]he plaintiff must proffer particularized evidence of direct or circumstantial facts . . . supporting the claim of an improper motive in order to avoid summary judgment."); Colburn v. Upper Darby Township, 838 F.2d 663, 666 (3d Cir. 1988) ("The dual policy concerns of protecting state officials from a deluge of frivolous claims and providing state officials with sufficient notice of the claims asserted to enable preparation of responsive pleadings have led us to impose on section 1983 claims the additional pleading requirement that the complaint contain a modicum of factual specificity, identifying the particular conduct of defendants that is alleged to have harmed the plaintiffs." (internal quotation omitted)), cert. denied, 489 U.S. 1065 (1989); Gooden v. Howard County, Md., 954 F.2d 960, 969-70 (4th Cir. 1992) (en banc) ("To avoid evisceration of the purposes of qualified immunity . . . plaintiffs alleging unlawful intent . . . [must] plead specific facts in a nonconclusory fashion to survive a motion to dismiss."); Schultea v. Wood, 47 F.3d 1427, 1434 (5th Cir. 1995) (en banc) ("The district court need not allow any discovery unless it finds that plaintiff has supported his claim with sufficient precision and factual specificity to raise a genuine issue as to the illegality of defendant's conduct at the time of the alleged acts."); Veney v. Hogan, 70 F.3d 917, 922 (6th Cir. 1995) (Plaintiff must respond to an assertion of qualified immunity with "specific, non-conclusory allegations of fact that will enable the district court to determine that those facts, if proved, will overcome the defense of qualified immunity."); Elliott v. Thomas, 937 F.2d 338, 344-45 (7th Cir. 1991) ("[T]he plaintiff [is required] to produce specific, nonconclusory factual allegations which establish the necessary mental state, or face dismissal." (internal quotation and alteration omitted)), cert. denied, 502 U.S. 1121 (1992); Edgington v. Missouri Dep't of Corrections, 52 F.3d 777, 779 (8th Cir. 1995) ("Complaints seeking damages against governmental officials . . . are subject to a heightened standard of pleading with sufficient specificity to put defendants on notice of the nature of the claim."); Branch v. Tunnell, 937 F.2d 1382, 1387 (9th Cir. 1991) ("We believe a requirement that a plaintiff must put forward nonconclusory allegations of subjective motivation . . . satisfies Harlow's directive that government officials should be shielded from `insubstantial' lawsuits, while at the same time preserving the opportunity for plaintiffs to pursue meritorious claims."); Walter v. Morton, 33 F.3d 1240, 1243 (10th Cir. 1994) ("To survive a summary judgment motion, a plaintiff must point to specific evidence showing the official's actions were improperly motivated."); Oladeinde v. City of Birmingham, 963 F.2d 1481, 1485 (11th Cir. 1992) ("In pleading a section 1983 action, some factual detail is necessary."), cert. denied, 507 U.S. 987 (1993). The First Circuit has not adopted a particular formulation of the standard, but instead has made it clear that intent-based claims can be sufficient to overcome qualified immunity, see Feliciano-Angulo v. Rivera-Cruz, 858 F.2d 40, 46 (1st Cir. 1988), and has indicated that motions for summary judgment in qualified immunity cases will be handled under the Federal Rules just like any other case, see Alexis v. McDonald's Restaurants of Mass., 67 F.3d 341, 348-49 n. 7 (1st Cir. 1995). Thus, no other federal jurisdiction operates under a standard even approaching the harshness of the positions endorsed by the judges who view this case differently.

Further, we have been presented with no evidence to indicate that, under these formulations, government officials around the country are being subjected to intolerable litigation burdens from intent-based civil rights suits or that district court judges are routinely permitting frivolous claims to go forward. Indeed, it is worth noting that neither the Solicitor General nor the government defendants themselves even advocated a "clear and convincing" evidence standard in their submissions to this court.

To the contrary, at least one empirical study of constitutional tort litigation concludes that "the image of a civil rights litigation explosion is overstated and borders on myth." Theodore Eisenberg Stewart Schwab, The Reality of Constitutional Tort Litigation, 72 Cornell L. Rev. 641, 643 (1987).

Although the brief of J. Michael Quinlan and Loye W. Miller, Jr., as amici curiae, does suggest a "clear and convincing" standard as a possible alternative to the direct evidence rule, amici offer no legal precedent requiring or supporting such a standard, arguing instead that, as a policy matter, the proposed standard would be appropriate given "the venerable principle that government officials are presumed to act in good faith." Brief of J. Michael Quinlan and Loye W. Miller, Jr. as Amici Curiae at 25.

A rule requiring plaintiffs to meet a higher evidentiary standard in qualified immunity cases has never been endorsed by the Supreme Court, and (contrary to the suggestion in Judge Williams's opinion) Harlow itself gives no indication that the Court contemplated such an onerous requirement. Indeed, Judge Williams's opinion completely ignores the fact that, although the Court in Harlow stated that "insubstantial suits against high public officials should not be allowed to proceed to trial," the decision relies on the "firm application of the Federal Rules of Civil Procedure" to achieve this objective. Harlow, 457 U.S. at 819-20 n. 35 (internal quotations omitted). Thus, nothing in Harlow gives appellate courts free-reign to perform their own cost-benefit analysis or to select new evidentiary standards out of thin air. Furthermore, the recent case of Leatherman v. Tarrant County Narcotics Intelligence and Coordination Unit, 507 U.S. 163 (1993), broadly repudiates the use of heightened, judge-made standards to fulfill policy-related goals such as those advanced by the judges who view this case differently. Although Leatherman addressed only claims against municipalities, it is significant that the Court explicitly rejected the justifications for a heightened standard that had been offered by the defendants, and instead insisted that the Federal Rules remain the sole touchstone for determining the sufficiency of the plaintiff's case. As the Court stated, additional requirements can be imposed only "by the process of amending the Federal Rules, and not by judicial interpretation." Id. at 168.

Finally, my colleagues' attempt to justify a clear and convincing evidence standard by reference to New York Times Co. v. Sullivan, 376 U.S. 254 (1964), is in vain. In that case, nothing less than the First Amendment's guarantee of freedom of the press was at stake, and the Court concluded that this vital interest, enshrined in the Bill of Rights, justified a heightened evidentiary burden. See, e.g., id., at 270 ("[W]e consider this case against the background of a profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open. . . ."). Given that there is no analogous constitutional right protecting public officials from lawsuits, this case cannot possibly qualify as a "cognate" area of law. Despite the complete lack of judicial precedent or evidence that an alternative judicial remedy is either appropriate or necessary, the judges who view this case differently have reached out and attempted to devise new rules of law that would have devastating consequences in many civil rights lawsuits. Thus, at precisely the moment that we have finally dispensed with our absurd and anomalous direct-evidence rule, some members of the court would once again concoct an arbitrary and unfair evidentiary standard that diverges from the settled law of every other court in the land.

D. Application to This Case

Although it is far from clear whether the plaintiff in this case can prevail on the merits of his claims, his complaint is sufficient to survive motions to dismiss or for summary judgment. Prison inmate Leonard Rollon Crawford-El alleges that prison official Patricia Britton unconstitutionally retaliated against him for exercising his First Amendment rights by deliberately retaining and then misdelivering his personal items during a series of prison transfers. Far from simply appending a claim of unconstitutional motivation to an otherwise questionable claim, however, Crawford-El's complaint, as recounted by the District Court, includes several specific factual allegations:

— Crawford-El alleges that Britton treated him worse than other prisoners because she knew that when he had been in charge of the law library at the Central Facility, he had helped other prisoners prepare their Administrative Remedy Procedure grievance forms or their appeals of disciplinary decisions. Crawford-El had "a reputation for asserting legal rights and knowing the administrative procedures for doing so," and that made Britton hostile towards him. (Fourth Amended Complaint, at Para(s) 6.)

— On April 20, 1986, The Washington Post published a front-page article about jail overcrowding based on interviews with Crawford-El. The next day, Britton chastised Crawford-El for tricking her and for embarrassing her before her co-workers. She threatened to make life hard for him in jail any way she could. (Fourth Amended Complaint, at Para(s) 12.)

— Britton stated on another occasion that prisoners like Crawford-El "don't have any rights." (Fourth Amended Complaint, at Para(s) 15.)

— After the publication of a second Washington Post article, which reported inmates' suspicions that "they were handpicked for transfer [from the District of Columbia to the State of Washington] because they were "jailhouse lawyers' — troublemaking `writ-writers' who tied up the courts with occasionally successful lawsuits against the prison system" and quoted Crawford-El to that effect, Britton told another prison official that Crawford-El was a "legal troublemaker." (Fourth Amended Complaint, at WW 16-17.)

Crawford-El v. Britton, 844 F. Supp. 795, 802-03 (D.D.C. 1994). Based on these and other allegations, the trial judge concluded that a jury "might reasonably infer . . . that Britton diverted and withheld Crawford-El's property out of an unconstitutional desire to retaliate against a `legal troublemaker.'" Id. at 803. Thus, an experienced member of our District Court found that, under the Federal Rules of Civil Procedure, there is a legitimate case to go to the jury. Yet, under the evidentiary standard proposed by several of my colleagues, Crawford-El's allegations would not be sufficient even to proceed to discovery. Nothing in the Supreme Court's qualified immunity jurisprudence justifies such a result; indeed, it would appear that Crawford-El's complaint would survive a motion for summary judgment under the rules adopted by every other court of appeals in the nation.

Conclusion

Given the lack of any Supreme Court decision indicating that a "clear and convincing" standard can or should be invented by judges and overlaid onto the Federal Rules of Civil Procedure, the result proposed by the judges who view this case differently suggests an extraordinary use of judicial authority. One would have thought that the outcome they propose would be anathema to judges who advocate a philosophy of judicial restraint, particularly when the more prudent course is to insist on a firm application of the Federal Rules until such a time as the Supreme Court commands us to do otherwise, or an amendment is made either to the Federal Rules or to section 1983 itself. "[J]ust as masons building a cathedral should not supplant the architect, even though both are creating a work of art, a judge should not supplant the politician or administrator though all are seeking sound governance." Stephen F. Williams, The Roots of Deference, 100 Yale L.J. 1103, 1111 (1991); see also, e.g., Laurence H. Silberman, Chevron — The Intersection of Law Policy, 58 Geo. Wash. L. Rev. 821, 821-22 (1990) ("decr[ying] the extraordinary expansion of judicial power in the latter half of this century," and observing that the one concept that most distinguishes those who advocate "judicial restraint" is "avoidance of judicial policy making").

The simple truth here is that Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388 (1971), and its progeny (not to mention the Civil Rights Act of 1871), remain the law of the land and control the actions of this court. And, as Justice Kennedy recently pointed out, courts must be cautious about "devising limitations to a remedial statute, enacted by Congress, which `on its face does not provide for any immunities.'" Wyatt, 504 U.S. at 171 (Kennedy, J., concurring) (quoting Malley v. Briggs, 475 U.S. 335, 342 (1986)). I therefore find it incredible that some members of this court seek to create new rules that would effectively render impossible all Bivens-type civil rights actions that turn on the intent of government officials. Until the Supreme Court finally resolves the question once and for all, it appears that this circuit might sit alone among all the federal courts of appeal in its approach to this issue. Citizens of the United States who legitimately use the legal system to render representatives of their government accountable for unconstitutional action should not find the courthouse door in our nation's capital slammed shut. I hope that will not be the consequence of today's decision.