with whom
Justice Marshall joins, and with whom Justice Stevens joins as to Part III, concurring in part and dissenting in part.In experiments designed to test the effects of lysergic acid diethylamide (LSD), the Government of the United States treated thousands of its citizens as though they were laboratory animals, dosing them with this dangerous drug without their consent. One of the victims, James B. Stanley, seeks compensation from the Government officials who injured him. The Court holds that the Constitution provides him with no remedy, solely because his injuries were inflicted while he performed his duties in the Nation’s Armed Forces. If our Constitution required this result, the Court’s decision, though legally necessary, would expose a tragic flaw in that document. But in reality, the Court disregards the commands of our Constitution, and bows instead to the purported requirements of a different master, military discipline, declining to provide Stanley with a remedy because it finds “special factors counselling hesitation.” Bivens v. Six Unknown Fed. Narcotics Agents, 403 U. S. 388, 396 (1971). This is abdication, not hesitation. I dissent.1
*687I
Before addressing the legal questions presented, it is important to place the Government’s conduct in historical context. The medical trials at Nuremberg in 1947 deeply impressed upon the world that experimentation with unknowing human subjects is morally and legally unacceptable. The United States Military Tribunal established the Nuremberg Code as a standard against which to judge German scientists who experimented with human subjects. Its first principle was:
“1. The voluntary consent of the human subject is absolutely essential.
“The duty and responsibility for ascertaining the quality of the consent rests upon each individual who initiates, directs or engages in the experiment. It is a personal duty and responsibility which may not be delegated to another with impunity.” United States v. Brandt (The Medical Case), 2 Trials of War Criminals Before the Nuremberg Military Tribunals Under Control Council Law No. 10, pp. 181-182 (1949) (emphasis added).
The United States military developed the Code, which ap- ' plies to all citizens — soldiers as well as civilians.2
*688In the 1950’s, in defiance of this principle, military intelligence agencies and the Central Intelligence Agency (CIA) began surreptitiously testing chemical and biological materials, including LSD. These programs, which were “designed to determine the potential effects of chemical or biological agents when used operationally against individuals unaware that they had received a drug,” included drug testing on “unwitting, nonvolunteer” Americans. S. Rep. No. 94-755, Book I, p. 385 (1976) (S. Rep.).3 James B. Stanley, a master sergeant in the Army, alleges that he was one of 1,000 soldiers covertly administered LSD by Army Intelligence between 1955 and 1958. See id., at 392.4
The Army recognized the moral and legal implications of its conduct. In a 1959 Staff Study, the United States Army Intelligence Corps (USAINTC) discussed its covert administration of LSD to soldiers:
“Tt was always a tenet of Army Intelligence that the basic American principle of dignity and welfare of the individual will not be violated. ... In intelligence, the stakes involved and the interests of national security may permit a more tolerant interpretation of moral-ethical values, but not legal limits, through necessity. . . . Any claim against the US Government for alleged injury due *689to EA 1729 [LSD] must be legally shown to have been due to the material. Proper security and appropriate operational techniques can protect the fact of employment of EA 1729.”’ Id., at 416-417 (quoting USAINTC Staff Study, Material Testing Program EA 1729, p. 26 (Oct. 15, 1959)).
That is, legal liability could be avoided by covering up the LSD experiments.
When the experiments were uncovered, the Senate agreed with the Army’s conclusion that its experiments were of questionable legality, and issued a strong condemnation:
“[I]n the Army’s tests, as with those of the CIA, individual rights were . . . subordinated to national security considerations; informed consent and follow-up examinations of subjects were neglected in efforts to maintain the secrecy of the tests. Finally, the command and control problems which were apparent in the CIA’s programs are paralleled by a lack of clear authorization and supervision in the Army’s programs.” S. Rep., at 411.5
Having invoked national security to conceal its actions, the Government now argues that the preservation of military discipline requires that Government officials remain free to violate the constitutional rights of soldiers without fear of money damages. What this case and others like it demonstrate, however, is that Government officials (military or civilian) must not be left with such freedom. See, e. g., Jaffee v. United States, 663 F. 2d 1226 (CA3 1981) (en banc) (exposure of soldiers to nuclear radiation during atomic weapons testing); Schnurman v. United States, 490 F. Supp. 429 (ED *690Va. 1980) (exposure of unknowing soldier to mustard gas); Thornwell v. United States, 471 F. Supp. 344 (DC 1979) (soldiers used to test the effects of LSD without their knowledge); cf. Barrett v. United States, No. 76 Civ. 381 (SDNY, May 5, 1987) (death of mental hospital patient used as the unconsenting subject of an Army experiment to test mescaline derivative).6
II
Serious violations of the constitutional rights of soldiers must be exposed and punished. Of course, experimentation with unconsenting soldiers, like any constitutional violation, may be enjoined if and when discovered. An injunction, however, comes too late for those already injured; for these victims, “it is damages or nothing.” Bivens, 403 U. S., at 410 (Harlan, J., concurring). The solution for Stanley and *691other soldiers, as for any citizen, lies in a Bivens action — an action for damages brought directly under the Constitution for the violation of constitutional rights by federal officials. But the Court today holds that no Bivens remedy is available for service-connected injuries, because “special factors counsel] hesitation. ” Id., at 396. The practical result of this decision is absolute immunity from liability for money damages for all federal officials who intentionally violate the constitutional rights of those serving in the military.
First, I will demonstrate that the Court has reached this result only by ignoring governing precedent. The Court confers absolute immunity from money damages on federal officials (military and civilian alike) without consideration of longstanding case law establishing the general rule that such officials are liable for damages caused by their intentional violations of well-established constitutional rights. If applied here, that rule would require a different result. Then I will show that the Court denies Stanley’s Bivens action solely on the basis of an unwarranted extension of the narrow exception to this rule created in Chappell v. Wallace, 462 U. S. 296 (1983). The Court’s reading of Chappell tears it from its analytical moorings, ignores the considerations decisive in our immunity cases, and leads to an unjust and illogical result.
A
The Court acknowledges that Stanley may bring a Bivens action for damages under the Constitution unless there are “special factors counselling hesitation in the absence of affirmative action by Congress.” Bivens, supra, at 396. Ascertaining the propriety of a damages award is the purpose of both the Bivens “special factors” analysis and the inquiry into whether these federal officials are entitled to absolute immunity from money damages.7 As a practical *692matter, the immunity inquiry and the “special factors” inquiry are the same; the policy considerations that inform them are identical, and a court can examine these considerations only once.8
In Davis v. Passman, 442 U. S. 228 (1979), the Court explicitly acknowledged that the immunity question and the “special factors” question are intertwined. The Court recognized that “a suit against a Congressman for putatively unconstitutional actions taken in the course of his official conduct does raise special concerns counseling hesitation” under Bivens, but held that “these concerns are coextensive with the protections afforded by the Speech or Debate Clause,” id., at 246, which “shields federal legislators with absolute immunity,” id., at 236, n. 11.9 Absent immunity, the Court said, legislators ought to be liable in damages, as are ordinary persons. See id., at 246. The same analysis applies to federal officials making decisions in military matters. Absent immunity, they are liable for damages, as are all citizens.
*693As the Court notes, I do not dispute that the question whether a Bivens action exists is “analytically distinct from the question of official immunity from Bivens liability.” Ante, at 684. I contend only that the “special factors” analysis of Bivens and the functional analysis of immunity are based on identical judicial concerns which, when correctly applied, should not and do not (as either a logical or practical matter) produce different outcomes. Justice Stevens explained it well:
“The practical consequences of a holding that no remedy has been authorized against a public official are essentially the same as those flowing from a conclusion that the official has absolute immunity. Moreover, similar factors are evaluated in deciding whether to recognize an implied cause of action or a claim of immunity. In both situations, when Congress is silent, the Court makes an effort to ascertain its probable intent.” Mitchell v. Forsyth, 472 U. S. 511, 538-539 (1985) (concurring opinion).
Thus, the redundance which so troubles the Court in equation of the “special factors” analysis and the immunity analysis strikes me as evidence only that the analyses are being properly performed. And Davis cannot be characterized, as the Court asserts, as a unique case in which the “special factors” of Bivens were coextensive with the immunity granted.10
*694When performing the Bivens analysis here, therefore, the Court should examine our cases discussing immunity for federal officials.11
B
The Court historically has conferred absolute immunity on officials who intentionally violate the constitutional rights of citizens only in extraordinary circumstances. Qualified immunity (that is, immunity for acts that an official did not know, or could not have known, violated clearly established constitutional law) “represents the norm.” See Harlow v. Fitzgerald, 457 U. S. 800, 807 (1982) (Presidential aides); Mitchell, supra (United States Attorney General); Butz v. Economou, 438 U. S. 478 (1978) (Cabinet officers).12
In Butz, we balanced “the need to protect officials who are required to exercise their discretion and the related public interest in encouraging the vigorous exercise of official authority,” id., at 506, against the crucial importance of a damages remedy in deterring federal officials from committing con*695stitutional wrongs and vindicating the rights of citizens, id., at 504-505.13 After full consideration of potential adverse consequences, we decided that the extension of absolute immunity to federal officials would “seriously erode the protection provided by basic constitutional guarantees,” id., at 505, and undermine the basic assumption of our jurisprudence: “that all individuals, whatever their position in government, are subject to federal law.” Id., at 506 (emphasis added). Thus, we concluded that it is “not unfair to hold liable the official who knows or should know he is acting outside the law,” and that “insisting on awareness of clearly established constitutional limits will not unduly interfere with the exercise of official judgment.” Id., at 506-507.
In Butz we acknowledged that federal officials may receive absolute immunity in the exercise of certain functions, but emphasized that the burden is on the official to demonstrate that an “exceptional situatio[n]” exists, in which “absolute immunity is essential for the conduct of the public business.” See Butz, supra, at 507; Harlow, 457 U. S., at 812. The official seeking immunity “first must show that the responsibilities of his office embraced a function so sensitive as to require a total shield from liability,” and “then must demonstrate that he was discharging the protected function when performing the act for which liability is asserted.” Id., at 813.
Even when, as here, national security is invoked,14 federal officials bear the burden of demonstrating that the usual rule *696of qualified immunity should be abrogated. In Mitchell v. Forsyth, 472 U. S. 511 (1985), the Court found “no . . . historical or common-law basis for an absolute immunity for officers carrying out tasks essential to national security.” Id., at 521. In language applicable here, the Court pointed out: “National security tasks . . . are carried out in secret .... Under such circumstances, it is far more likely that actual abuses will go uncovered than that fancied abuses will give rise to unfounded and burdensome litigation.” Id., at 522.15 The Court highlighted the “danger that high federal officials will disregard constitutional rights in their zeal to protect the national security,” and deemed it “sufficiently real to counsel against affording such officials an absolute immunity.” Id., at 523.
This analysis of official immunity in the national security context applies equally to officials giving orders to the military. In Scheuer v. Rhodes, 416 U. S. 232 (1974), the Governor, the Adjutant General of the Ohio National Guard, and other National Guard officers were sued under 42 U. S. C. § 1983 for damages arising from injuries suffered when the Guard was deployed and ordered to fire its guns during a civil disturbance. The Court awarded only qualified immunity to the highest military officer of the State — the Governor (who commanded the State National Guard) — and to executive and military officers involved in the decision to take military *697action.16 Scheuer demonstrates that executive officials may receive only qualified immunity even when the function they perform is military decisionmaking.17
Whoever the officials in this case are (and we do not know), and whatever their functions, it is likely that under the Court’s usual analysis, they, like most Government officials, are not entitled to absolute immunity. The record does not reveal what offices the individual petitioners held, let alone what functions they normally performed, or what functions they were performing at the time they (somehow) participated in the decision to administer LSD to Stanley (and 1,000 other soldiers). The Court has no idea whether those officials can carry “the burden of showing that public policy requires [absolute immunity]” for effective performance of those functions. Butz, 438 U. S., at 506. Yet the Court grants them absolute immunity, so long as they intentionally inflict only service-connected injuries, doing violence to the principle that “extension of absolute immunity from damages liabil*698ity to all federal executive officials would seriously erode the protection provided by basic constitutional guarantees. ” Id., at 505. The case should be remanded and petitioners required to demonstrate that absolute immunity was necessary to the effective performance of their functions.
C
It is well accepted that when determining whether and what kind of immunity is required for Government officials, the Court’s decision is informed by the common law. See Nixon v. Fitzgerald, 457 U. S. 731, 747 (1982); Mitchell, supra, at 521; Butz, supra, at 508. My conclusion that qualified, rather than absolute, immunity is the norm for Government officials, even in cases involving military matters, is buttressed by the common law.
At common law, even military superiors received no exemption from the general rule that officials may be held accountable for their actions in damages in a civil court of law.18 “[T]he English judiciary refused to adopt absolute immunity as an essential protection of [intramilitary] discipline,”19 and “[t]he original American decisions in intramilitary cases [also] *699adopted a qualified immunity in intentional tort cases.” Zillman, Intramilitary Tort Law: Incidence to Service Meets Constitutional Tort, 60 N. C. L. Rev. 489, 498, 499 (1982).20 The best-known American case is Wilkes v. Dinsman, 7 How. 89 (1849), after remand, Dinsman v. Wilkes, 12 How. 390 (1852). In that case, this Court permitted a Navy seaman to bring a claim against his superior officer for injuries resulting from willful torts. Although the Court suggested that the commander was entitled to a jury charge providing some immunity, it refused to confer absolute immunity from liability for intentional torts:
“It must not be lost sight of . . . that, while the chief agent of the government, in so important a trust, when conducting with skill, fidelity, and energy, is to be protected under mere errors of judgment in the discharge of his duties, yet he is not to be shielded from responsibility if he acts out of his authority or jurisdiction, or inflicts private injury either from malice, cruelty, or any species of oppression, founded on considerations independent of public ends.
“The humblest seaman or marine is to be sheltered under the aegis of the law from any real wrong, as well as the highest in office.” 7 How., at 123.21
As noted above, the Court subsequently used Wilkes as an example of the usual practice of affording only qualified im*700munity to government officials. See Butz, 438 U. S., at 491. In addition, in Chappell v. Wallace, 462 U. S., at 305, n. 2, the Court distinguished Wilkes, plainly indicating that Chappell did not hold that soldiers could never sue for service-connected injury inflicted by an intentional tort. Indeed, by preserving Wilkes, the Court suggested that even military officials would not always be absolutely immune from liability for such conduct.
Although Chappell reveals that we have moved away from the common-law rule in cases involving the command relationship between soldiers and their superiors, our immunity cases and a close analysis of Chappell, see infra this page and 701-707, reveal that there is no justification for straying further.
Ill
A
In Chappell the Court created a narrow exception to the usual rule of qualified immunity for federal officials. Repeatedly referring to the “ ‘peculiar and special relationship of the soldier to his superiors,’” and to the need for “immediate compliance with military procedures and orders,” the Court held that “enlisted military personnel may not maintain a suit to recover damages from a superior officer for alleged constitutional violations.” 462 U. S., at 300, 305 (quoting United States v. Brown, 348 U. S. 110, 112 (1954)).22 Although the Court concedes this central focus of Chappell, it gives short shrift to the obvious and important distinction be*701tween Chappell and the present case, namely, that the defendants are not alleged to be Stanley’s superior officers. Instead the Court seizes upon the statement in Chappell that our analysis in that case was guided by the concerns underlying the Feres doctrine, and dramatically expands the carefully limited holding in Chappell, extending its reasoning beyond logic and its meaning beyond recognition.
The Court reasons as follows: In Chappell we stated that the concern for “military discipline” underlying the Feres doctrine would guide our analysis of the soldiers’ Bivens claims against their superior officers. 462 U. S., at 299. In United States v. Johnson, 481 U. S. 681 (1987), we held that the concerns underlying the Feres doctrine precluded a soldier’s FTCA claim for service-connected injury, even against civilian federal officials. Thus, the Court concludes, the concerns underlying the Feres doctrine preclude Stanley’s Bivens action for service-connected injury against civilian federal officials.
This argument has a number of flaws. First, in Chappell we said with good reason that our analysis would be “guided,” not governed, by concerns underlying Feres. The Bivens context differs significantly from the FTCA context; Bivens involves not negligent acts, but intentional constitutional violations that must be deterred and punished. Because Chappell involved the relationship at the heart of the Feres doctrine — the relationship between soldier and superior — the Court found Feres considerations relevant, and provided direct military superiors with absolute immunity from damages actions filed by their subordinates. Here, however, the defendants are federal officials who perform unknown functions and bear an unknown relationship to Stanley. Thus, we must assure ourselves that concerns underlying the Feres doctrine actually do require absolute immunity from money damages before we take the drastic step of insulating officials from liability for intentional constitutional violations. This the Court utterly fails to do.
*702Second, two of the three Feres. rationales that decided Johnson, supra, are entirely inapplicable here.23 Thus, the Court relies solely upon the third Feres rationale — a solicitude for military discipline. The Feres’ concern for military discipline itself has three components. The first, the concern for the instinctive obedience of soldiers to orders, is of central importance in the Feres doctrine.24 That rationale profoundly and exclusively concerned the Court in Chappell which involved the relationship between a superior officer and those in his or her command.25 This concern for instinc*703tive obedience is not at all implicated where a soldier sues civilian officials.26
As for the other components of the concern for military discipline, their application is entirely different in the Bivens context. The Court fears that military affairs might be disrupted by factual inquiries necessitated by Bivens actions. The judiciary is already involved, however, in cases that implicate military judgments and decisions, as when a soldier sues for nonservice-connected injury, when a soldier sues civilian contractors with the Government for service-connected injury, and when a civilian is injured and sues a civilian contractor with the military or a military tortfeasor. See John*704son, 481 U. S., at 700 (Scalia, J., dissenting).27 Although the desire to limit the number of such cases might justify the decision not to allow soldiers’ FTCA suits arising from negligent conduct by civilian Government employees, see United States v. Johnson, supra, it is insufficient to preclude suits against civilians for intentional violations of constitutional rights. Unless the command relationship (or some other consideration requiring absolute immunity) is involved, these violations should receive moral condemnation and legal redress without limitation to that accorded negligent acts.
Finally, the Court fears that the vigor of military decision-making will be sapped if damages can be awarded for an incorrect (albeit intentionally incorrect) choice. Of course, this case involves civilian decisionmakers, but because the injury was service connected, we must assume that these civilian judgments are somehow intertwined with conduct of the military mission. See Johnson, supra, at 691. The significant difference between the Feres (FTCA) and Bivens (constitutional claim) contexts, however, is that, in the latter, the vigorous-decisionmaking concern has already been taken into account in our determination that qualified immunity is the general rule for federal officials, who should be required “on occasion ... to pause to consider whether a proposed course of action can be squared with the Constitution.” Mitchell, 472 U. S., at 524. The special requirements of com*705mand that concerned us in Chappell are not implicated in this case, and neither the Government nor the Court offers any plausible reason to extend absolute immunity to these civilian officials for their intentional constitutional violations.
In Chappell, the Court did not create an inflexible rule, requiring a blind application of Feres in soldiers’ cases raising constitutional claims. Given the significant interests protected by Bivens actions, the Court must consider a constitutional claim in light of the concerns underlying Feres. If those concerns are not implicated by a soldier’s constitutional claim, Feres should not thoughtlessly be imposed to prevent redress of an intentional constitutional violation.28
The Court decides that here (as indeed in any case) one might select a higher level of generality for the Chappell holding, and concludes that any Bivens action arising from a service-connected injury is foreclosed by “special factors counselling hesitation.” Bivens, 403 U. S., at 396. The Court concedes that “[t]his is essentially a policy judgment,” which depends upon “how much occasional, unintended impairment of military discipline one is willing to tolerate.” Ante, at 681. But the Court need not make a policy judgment; in our immunity cases we have an established legal framework within which to consider whether absolute immunity from money damages is required in any particular situa*706tion. Were I to concede that military discipline is somehow implicated by the award of damages for intentional torts against civilian officials (which I do not, see supra, at 702-703), I would nonetheless conclude, in accord with our usual immunity analysis, that the decisionmaking of federal officials deliberately choosing to violate the constitutional rights of soldiers should be impaired. I cannot comprehend a policy judgment that frees all federal officials from any doubt that they may intentionally, and in bad faith, violate the constitutional rights of those serving in the Armed Forces. The principles of accountability embodied in Bivens — that no official is above the law, and that no violation of right should be without a remedy — apply.
B
The second “special factor” in Chappell — congressional activity “providing] for the review and remedy of complaints and grievances such as those presented by” the injured soldier—is not present here. Chappell, 462 U. S., at 302.29 The Veterans’ Benefits Act is irrelevant where, as here, the injuries alleged stem (in large part) from pain and suffering in forms not covered by the Act. The UCMJ assists only when the soldier is on active duty and the tortfeasor is another military member. Here, in contrast to the situation in Chappell, no intramilitary system “provides for the . . . remedy” of Stanley’s complaint. 462 U. S., at 302. See also Bush v. Lucas, 462 U. S. 367, 386, 388, 378, n. 14 (1983) (special factors counseling hesitation found because claims were “fully cognizable” within an “elaborate remedial system,” *707providing “comprehensive,” “meaningful,” and “constitutionally adequate” remedies).
Nonetheless, the Court finds Congress’ activity (and inactivity) of particular significance here, because we are confronted with a constitutional authorization for Congress to “‘make Rules for the Government and Regulation of the land and naval Forces.’” Ante, at 679 (quoting U. S. Const. Art. I, § 8, cl. 14). First, the existence of a constitutional provision authorizing Congress to make intramilitary rules does not answer the question whether civilian federal officials are immune from damages in actions arising from service-connected injury. Second, any time Congress acts, it does so pursuant to either an express or implied grant of power in the Constitution. If a Bivens action were precluded any time Congress possessed a constitutional grant of authority to act in a given area, there would be no Bivens. In fact, many administrative agencies exist and function entirely at the pleasure of Congress, yet the Court has not hesitated to infer Bivens actions against these agencies’ officials. This is so no matter how explicitly or frequently the Constitution authorizes Congress to act in a given area. Even when considering matters most clearly within Congress’ constitutional authority, we have found that a Bivens action will lie. See Davis v. Passman, 442 U. S. 228 (1979).
In Chappell the Court found that both the imperatives of military discipline and the congressional creation of constitutionally adequate remedies for the alleged violations constituted “special factors counselling hesitation,” and refused to infer a Bivens action. In this case, the invocation of “military discipline” is hollow, and congressional activity nonexistent; a Bivens action must lie.
> h-H
The soldier s case is instructive: Subject to most unilateral discipline, forced to risk mutiliation and death, conscripted without, perhaps against, his will — he is still *708conscripted with his capacities to act, to hold his own or fail in situations, to meet real challenges for real stakes. Though a mere ‘number’ to the High Command, he is not a token and not a thing. (Imagine what he would say if it turned out that the war was a game staged to sample observations on his endurance, courage, or cowardice.)” H. Jonas, Philosophical Reflections on Experimenting with Human Subjects, in Experimentation with Human Subjects 3 (P. Freund ed. 1969).
The subject of experimentation who has not volunteered is treated as an object, a sample. James Stanley will receive no compensation for this indignity. A test providing absolute immunity for intentional constitutional torts only when such immunity was essential to maintenance of military discipline would “take into account the special importance of defending our Nation without completely abandoning the freedoms that make it worth defending.” Goldman v. Weinberger, 475 U. S. 503, 530-531 (1986) (O’Connor, J., dissenting). But absent a showing that military discipline is concretely (not abstractly) implicated by Stanley’s action, its talismanic invocation does not counsel hesitation in the face of an intentional constitutional tort, such as the Government’s experimentation on an unknowing human subject. Soldiers ought not be asked to defend a Constitution indifferent to their essential human dignity. I dissent.
I agree with the Court that Stanley’s cause of action under the Federal Tort Claims Act (FTCA) should not have been reinstated by the Court of Appeals. Thus, I join in Part I of the Court’s opinion.
See, e. g., Mulford, Experimentation on Human Beings, 20 Stan. L. Rev. 99, 105, n. 34 (1967) (Military personnel cannot be compelled to submit to nontherapeutic procedures) (citing Johnson, Civil Rights of Military Personnel Regarding Medical Care and Experimental Procedures, 117 Science 212-215 (1953)). Indeed, the application of such principles to all citizens, including soldiers, is essential in a society governed by law:
“[Human experimentation authorized by the state] dramatizes the notion that the state is free to treat its nationals in the manner it chooses because it perceives itself as the source of all rights, and therefore as beyond the reach of law, rather than regarding rights as inalienable, that is, not subject to arbitrary cancellation by the State.” Bassiouni, Baffes, & Evrard, An Appraisal of Human Experimentation in International Law and Practice: The Need for International Regulation of Human Experimentation, 72 J. of Crim. L. & C. 1597, 1607 (1981).
This massive Senate Report is the product of a select Committee which “conducted] an investigation and study of governmental operations with respect to intelligence activities and of the extent, if any, to which illegal, improper, or unethical activities were engaged in by any agency of the Federal Government.” S. Rep., at 2. The Committee’s function was “to illustrate the problems before Congress and the'country.” Id., at 5. Significantly, the Report added that “[t]he Justice Department and the courts in turn have their proper roles to play.” Ibid.
The intelligence community believed that it was necessary “to conceal these activities from the American public in general,” because public knowledge of the “unethical and illicit, activities would have serious repercussions in political and diplomatic circles and would be detrimental to the accomplishment of its mission.” Id., at 394 (quoting CIA Inspector General’s Survey of the Technical Services Division, p. 217 (1957)).
See also S. Rep., at 403:
“Though it was known that the testing was dangerous, the lives of subjects were placed in jeopardy and their rights were ignored during the ten years of testing which followed Dr. Olsen’s death. [Dr. Olsen, a civilian employee of the Army, committed suicide after being administered LSD without his knowledge.] Although it was clear that the laws of the United States were being violated, the testing continued.”
In Jaffee v. United States, 663 F. 2d 1226 (CA3 1981), a former enlisted member of the Army sought damages arising from injuries received in 1953 at Camp Desert Rock, Nevada, where his commanding officers ordered him and thousands of other soldiers to stand unprotected from nuclear radiation while an atomic bomb was exploded nearby. Jaffee developed inoperable cancer in 1977 and alleged that the radiation exposure was the cause.
Between 1945 and 1963, an estimated 250,000 military personnel were exposed to large doses of radiation while engaged in maneuvers designed to determine the effectiveness of combat troops in nuclear battlefield conditions. Veterans’. Claims for Disabilities from Nuclear Weapons Testing: Hearing before the Senate Committee on Veterans Affairs, 96th Cong., 1st Sess., 2 (1979). Soldiers were typically positioned one to three miles from nuclear detonation. They were issued no protective clothing (although Atomic Energy Commission personnel were) and were not warned as to the possible dangers of radiation. They were instructed to cover their eyes at detonation; “soldiers with their eyes shut could see the bones in their forearms at the moment of the explosion.” Schwartz, Making Intramilitary Tort Law More Civil: A Proposed Reform of the Feres Doctrine, 95 Yale L. J. 992, 994, n. 16 (1986) (discussing firsthand accounts in T. Saffer & O. Kelly, Countdown Zero 43, 75, 152 (1982)). The exposed servicemembers have been disproportionately likely to be afflicted with inoperable cancer and leukemia, as well as a number of nonmalignant disorders.
The Court made clear in Davis v. Passman, 442 U. S. 228, 244 (1979), that the question whether a plaintiff has a cause of action under the Constitution is different from the question whether that plaintiff is entitled to *692damages if he or she prevails on the merits. The latter is the relevant inquiry when a Bivens claim is made. Of course, if the plaintiff fails either to plead a cause of action or to demonstrate the damages are appropriate as a matter of law, the complaint is dismissed under Federal Rule of Civil Procedure 12(b)(6). In the first instance, the complaint is dismissed for “failure to state a claim,” while in the latter instance, the complaint is dismissed because it is not one “upon which relief can be granted.”
The Court has acknowledged that the damages remedy made available in Bivens would be “drained of meaning if federal officials were entitled to absolute immunity for their constitutional transgressions,” because “a suit under the Constitution could provide no redress to the injured citizen, nor would it in any degree deter federal officials from committing constitutional wrongs.” Butz v. Economou, 438 U. S. 478, 501, 505 (1978) (internal quotation omitted).
The Court in Davis, supra, did not decide whether Passman was absolutely immune from damages, but instead remanded the action to the Court of Appeals for a determination of that question. Analytically, the Court therefore postponed decision on the propriety of damages until a lower court could ascertain whether immunity, a “special factor,” shielded Passman from damages.
The Court does not provide an example of a situation in which the Bivens inquiry and the immunity inquiry might reach contrary conclusions. Of course, I cannot produce “any reason for creating” an equivalency between the two analyses as to this particular case. Ante, at 685. Neither I nor the Court has any idea what functions were performed by the petitioner officials, so I cannot argue that the considerations militating in favor of qualified immunity here also militate in favor of permitting a cause of action.
The Court’s use of the doctrine of Feres v. United States, 340 U. S. 135 (1950), in its analysis of soldiers’ Bivens actions reveals the connection between the “special factors” inquiry and the absolute immunity inquiry. In Feres, the Court decided that, in the FTCA, Congress had not waived sovereign immunity from damages for claims arising out of negligent acts of federal officials causing service-connected injury. When, as here, the Court decides whether a Bivens action exists, it necessarily decides whether the policies underlying Feres alter the usual rule of qualified immunity for federal officials. In both cases the question is how policies underpinning Feres affect immunity from money damages.
The President, see Nixon v. Fitzgerald, 457 U. S. 731 (1982), prosecutors, Imbler v. Packtman, 424 U. S. 409 (1976), and federal officials with prosecutorial and adjudicative functions, see Butz, supra, at 508-517, possess absolute immunity from damages actions arising from the violation of clearly established constitutional rights. But most public servants receive qualified, rather than absolute, immunity. See Procunier v. Navarette, 434 U. S. 555 (1978) (prison officials); O’Connor v. Donaldson, 422 U. S. 563 (1975) (state hospital administrators); Scheuer v. Rhodes, 416 U. S. 232 (1974) (state executive officers); Pierson v. Ray, 386 U. S. 547 (1967) (police).
Qualified immunity for executive officials is the result of the balancing of “fundamentally antagonistic social policies.” Barr v. Mateo, 360 U. S. 564, 576 (1959) (plurality opinion). Civil damages compensate victims of wrongdoing and deter tortious conduct, while immunity encourages participation in government, allows courageous action in public service, and provides officials with the freedom to concentrate on their public responsibilities.
The Government suggests that federal officers and agents gave LSD to Stanley and 1,000 other soldiers “for the purpose of ‘ascertain[ing] the effects of the drug on their ability to function as soldiers’ and ‘to evaluate *696the validity of the traditional security training ... in the face of unconventional, drug enhanced interrogations.’ ” Brief for United States 3, n. 1 (quoting S. Rep. 411-412).
Again in analysis equally applicable here, the Court observed that most officials who receive absolute immunity from suits for damages with regard to certain functions are subject to other checks “that help to prevent abuses of authority from going unredressed.” 472 U. S., at 522 (legislators are accountable to their constituents, and the judicial process is theoretically self-correcting by appellate review). But “[s]imilar built-in restraints on the Attorney General’s activities in the name of national security ... do not exist.” Id., at 523.
See also Butz, 438 U. S., at 491-492 (In finding qualified immunity for federal officials, the Court relied in part upon “a case involving military discipline, [in which] the Court issued a similar ruling [authorizing immunity absent willful or malicious conduct]”); Burgess, Official Immunity and Civil Liability for Constitutional Torts Committed by Military Commanders After Butz v. Economou, 89 Mil. L. Rev. 25, 46-47 (1980) (reading Butz to militate against intramilitary immunity in suits alleging constitutional violations).
Chappell v. Wallace, 462 U. S. 296 (1983), which I discuss in detail infra, at 700-707, is not to the contrary. There the Court found that a “special factor” — the need for effective performance of active duty command-entitled military officers to absolute immunity from damages for injuries inflicted upon direct subordinates. Relying on the “special nature of military life — the need for unhesitating and decisive action by military officers and equally disciplined responses by enlisted personnel” — the Court decided that the military command function “would be undermined by a judicially created remedy exposing officers to personal liability at the hands of those they are charged to command.” Id., at 304. The unique requirements of intramilitary authority drove the Court in Chappell; those concerns do not govern here, where we address the immunity of officials whose relationship with Stanley is unknown.
See W. Winthrop, Military Law and Precedents 880-885 (2d ed. 1920) (collecting decisions in which servicemembers sued their superiors for the intentional torts of libel, malicious prosecution, false imprisonment, and other service-related injuries). The Winthrop treatise reveals that military officers had only a defense of an absence of malice respecting actions within the scope of their authority, a defense closely resembling qualified immunity.
Some English cases do suggest absolute immunity for intramilitary torts. See Sutton v. Johnstone, 1 T. R. 492, 99 Eng. Rep. 1215, 1246 (K. B. 1786) (dictum); Dawkins v. Lord Rokeby, 4 F. & F. 806, 841, 176 Eng. Rep. 800, 815 (N. P. 1866); Dawkins v. Lord Paulet, L. R. 5 Q. B. 94, 115 (1869). But there is strong authority on the other side, and, before the question became the subject of statutory law, see Crown Proceedings Act, 1947, 10 & 11 Geo. 6, ch. 44, the English courts considered the matter unresolved. See Fraser v. Balfour, 87 L. J. K. B. 1116, 1118 (1918) (court observed that the question of immunity in intramilitary torts was “still open”).
See, e. g., Wilson v. MacKenzie, 7 Hill 96 (N. Y. Sup. Ct. 1845) (citing cases) (naval officer was sued for the beating and imprisonment of an enlisted man; court rejected defense of absolute immunity, stating that English courts had allowed suits for acts done under the rubric of military discipline); Maurice v. Worden, 54 Md. 233 (1880) (professor at the United States Naval Academy sued his superior officers for libel; state court rejected defense of absolute immunity).
See also Dinsman v. Wilkes, 12 How. 390, 403 (1852) (although discipline may be endangered by civil damages suits, the Nation will be dishonored if a servicemember can “be oppressed and injured by his commanding officer, from malice or ill-will, or the wantoness of power, without giving him redress in the courts of justice”).
The Court concedes that “Stanley and the lower courts may well be correct that Chappell implicates military chain-of-command concerns more directly than do the facts alleged here,” and that in Chappell we “explicitly focus[ed] on the officer-subordinate relationship that existed in [that] case,” using language that “would not be applicable here.” Ante, at 680. For example, we highlighted the need for “special regulations in relation to military discipline” and the “hierarchical structure of discipline and obedience to command, unique in its application to the military establishment and wholly different from civilian patterns,” Chappell, 462 U. S., at 300 (emphasis added).
First, in Feres the Court feared that allowing FTCA recovery, which varies from State to State, would impinge upon the military’s need for uniformity. In contrast, Bivens actions are governed by uniform federal law. Second, the “swift,” “efficient,” and “generous statutory disability and death benefits” of the Veterans’ Benefits Act (VBA), 72 Stat. 1118, as amended, 38 U. S. C. § 301 et seq., constitute “an independent reason why the Feres doctrine bars suit for service-related injuries.” United States v. Johnson, 481 U. S. 681, 689 (1987). But the VBA fails to address the violation of constitutional rights unaccompanied by personal injury that is not defined as disabling. Those whose constitutional rights are infringed, resulting in humiliation or “in mere pain and suffering, but no lasting permanent physical injury, would not be compensated at all.” Donaldson, Constitutional Torts and Military Effectiveness: A Proposed Alternative to the Feres Doctrine, 23 A. F. L. Rev. 171, 198-199 (1982-1983).
In Johnson, supra, when the Court extended the application of Feres to preclude suits for service-connected injuries against civilian officials, the Court did not refer to, or rely upon, Feres’ concern with obedience to orders. Of course, this aspect of military discipline would not be implicated in Johnson, or in any cases involving tortious conduct by a civilian official. But in Johnson, two of the three major rationales underlying Fern — the concern for uniformity and the congressional provision of thoroughgoing compensation — were relevant. Neither of these rationales applies here. See n. 22, supra.
Stanley points out that he was administered LSD without his knowledge so that he could not have disobeyed any order given him. Had his military superior surreptitiously administered the LSD to him, this fact alone might distinguish a suit for damages against that official from the suit in Chappell. Here, however, the fact that the LSD was given Stanley without his knowledge simply removes the case one step further from *703the concern for obedience to orders that the Court chose to protect in Chappell.
I do not mean to imply that Chappell suggests that Bivens actions against military officials other than direct superiors are precluded. Criticisms of the blanket application of Feres in the Bivens context have equal force in the context of intentional intramilitary torts that do not involve the direct chain of command. “The policy argument for absolute immunity . . . rests on the dubious proposition that a serviceman is more likely to respect authority when he has no recourse for the intentional or malicious deprivation of his constitutional rights. The contrary argument — that safeguarding rights compatible with military needs will engender respect for authority and promote discipline — is more appealing.” Note, Intramilitary Immunity and Constitutional Torts, 80 Mich. L. Rev. 312, 328 (1981). Cf. Johnson, supra, at 700 (Scalia, J., dissenting) (“Or perhaps — most fascinating of all to contemplate — Congress thought that barring recovery by servicemen might adversely affect military discipline”).
Nor does the military view the authority intentionally to violate the constitutional rights of soldiers as essential to its mission. See Uniform Code of Military Justice (UCMJ), 10 U. S. C. §§ 938, 939, discussed infra, at n. 27. Moreover, the military does not require instinctive or reflexive obedience of the soldier in all contexts (combat being the obvious counterexample). Soldiers are subject to criminal sanctions if they obey certain orders. See United States v. Galley, 22 U. S. C. M. A. 534, 48 C. M. R. 19 (1973) (obedience to orders no defense where defendant should have known that order to kill civilians was illegal); United States v. Kinder, 14 C. M. R. 742 (USAF Ct. Mil. Rev. 1954) (obedience to orders no defense for soldier who executed order to shoot subdued prisoner at South Korean air base).
In addition, judicial involvement occurs when courts review court-martial proceedings (through federal habeas corpus jurisdiction), see Burns v. Wilson, 346 U. S. 137, 142 (1953), when the Court of Claims reviews cases involving interference with military career advantage, see 28 U. S. C. § 1491, and when soldiers bring claims for injunctive and declaratory relief from statutory and constitutional violations. See also UCMJ, 10 U. S. C. § 938 (providing complaint procedure for “[a]ny member of the armed forces who believes himself wronged by his commanding officer”); § 939 (providing procedure for damages arising from willful damage to property of any soldier by another member of the Armed Services); Colson v. Bradley, 477 F. 2d 639 (CA8 1973) (judicial review of § 938 claim); Cortright v. Resor, 447 F. 2d 245 (CA2 1971) (same).
The Court states that, in the FTCA context, it is theoretically “freer to compromise military concerns . . . since we were confronted with an explicit congressional authorization for judicial involvement that was, on its face, unqualified,” while in the Bivens context, we “rely upon inference for our own authority to allow money damages.” Ante, at 681-682. One could approach the question from an entirely different angle. The usual rule with regard to suing the United States is sovereign immunity, so the FTCA creates an exception to that rule which must be narrowly interpreted. The usual rule is individual accountability for injury done, and qualified immunity of federal officials represents a judge-made exception to that rule. Our decision to find “special factors” in a Bivens case and grant absolute immunity to federal officials with regard to a certain class of injuries represents a further and indefensible enlargement of a special status.
In Chappell, the Court makes plain that, standing alone, the “special nature of [the] military” would not have sufficed to confer absolute immunity upon military superiors for wrongs inflicted upon those in their command. It was the “unique disciplinary structure of the Military Establishment and Congress’ activity in the field” that “[tjaken together” constituted those special factors precluding any damages award. 462 U. S., at 304 (emphasis added).