This is not an equal protection case.1 An empirical judgment that most persons formally accused of criminal conduct are probably guilty would provide a rational basis for a set of rules that treat them like convicts until they establish their innocence. No matter how rational such an approach might be — no matter how acceptable in a community where equality of status is the dominant goal — it is obnoxious to the concept of individual freedom protected by the Due Process Clause. If ever accepted in this country, it would work a fundamental change in the character of our free society.
Nor is this an Eighth Amendment case.2 That provision of the Constitution protects individuals convicted of crimes from punishment that is cruel and unusual. The pretrial detainees whose rights are at stake in this case, however, are innocent men and women who have been convicted of no crimes. Their claim is not that they have been subjected to cruel and unusual punishment in violation of the Eighth Amendment, but that to subject them to any form of punishment at all is an unconstitutional deprivation of their liberty.
*580This is a due process case.3 The most significant — and I venture to suggest the most enduring' — part of the Court’s opinion today is its recognition of this initial constitutional premise. The Court squarely holds that “under the Due Process Clause, a detainee may not be punished prior to an adjudication of guilt in accordance with due process of law.”4 Ante, at 535.
This right to be free of punishment is not expressly embodied in any provision in the Bill of Rights. Nor is the source of this right found in any statute. The source of this fundamental freedom is the word “liberty” itself as used in the Due Process Clause, and as informed by “history, reason, the past course of decisions,” and the judgment and experience of “those whom the Constitution entrusted” with interpreting that word. Anti-Fascist Committee v. McGrath, 341 U. S. 123, 162-163 (Frankfurter, J., concurring). See Leis v. Flynt, 439 U. S. 438, 457 (Stevens, J., dissenting).
In my opinion, this latter proposition is obvious and indisputable.5 Nonetheless, it is worthy of emphasis because the Court has now accepted it in principle. Ante, at 535. In recent years, the Court has mistakenly implied that the concept of liberty encompasses only those rights that are either created by statute or regulation or are protected by an express provision of the Bill of Rights.6 Today, however, without the help of any statute, regulation, or express provision of the Constitution, the Court has derived the innocent person’s right not to be punished from the Due Process Clause itself. It has accordingly abandoned its parsimonious definition of *581the “liberty” protected by the majestic words of the Clause. I concur in that abandonment. It is with regard to the scope of this fundamental right that we part company.
I
Some of the individuals housed in the Metropolitan Correction Center (MCC) are convicted criminals.7 As to them, detention may legitimately serve a punitive goal, and there is strong reason, even apart from the rules challenged here, to suggest that it does.8 But the same is not true of the detainees who are also housed there and whose rights we are called upon to address. Notwithstanding the impression created by the Court’s opinion, see, e. g., ante, at 562, these people are not “prisoners”: 9 they have not been convicted of any crimes, and their detention may serve only a more limited, regulatory purpose.10 See Houchins v. KQED, Inc., 438 U. S. 1, 37-38 (Stevens, J., dissenting).
*582Prior to conviction every individual is entitled to the benefit of a presumption both that he is innocent of prior criminal conduct and that he has no present intention to commit any offense in the immediate future.11 That presumption does *583not imply that he may not be detained or otherwise subjected to restraints on the basis of an individual showing of probable cause that he poses relevant risks to the community. For our system of justice has always and quite properly functioned on the assumption that probable cause to believe (1) that a person has committed a crime, and (2) that absent the posting of bail he poses at least some risk of flight,12 justifies pretrial detention to ensure his presence at trial.13
The fact that an individual may be unable to pay for a bail bond, however, is an insufficient reason for subjecting him to indignities that would be appropriate punishment for convicted felons. Nor can he be subject on that basis to onerous restraints that might properly be considered regulatory with respect to particularly obstreperous or dangerous arrestees. An innocent man who has no propensity toward immediate violence, escape, or subversion may not be dumped into a pool of second-class citizens and subjected to restraints designed to regulate others who have. For him, such treatment *584amounts to punishment. And because the due process guarantee is individual and personal, it mandates that an innocent person be treated as an individual human being and be free of treatment which, as to him, is punishment.14
It is not always easy to determine whether a particular restraint serves the legitimate, regulatory goal of ensuring a detainee’s presence at trial and his safety and security in the meantime, or the unlawful end of punishment. But the courts have performed that task in the past, and can and should continue to perform it in the future. Having recognized the constitutional right to be free of punishment, the Court may not point to the difficulty of the task as a justification for confining the scope of the punishment concept so narrowly that it effectively abdicates to correction officials the judicial responsibility to enforce the guarantees of due process.
In addressing the constitutionality of the rules at issue in this case, the Court seems to say that as long as the correction officers are not motivated by “an expressed intent to punish” their wards, ante, at 538, and as long as their rules are not “arbitrary or purposeless,” ante, at 539, these rules are an acceptable form of regulation and not punishment. Lest that test be too exacting, the Court abjectly defers to the prison administrator unless his conclusions are “ 'conclusively shown to be wrong.’ ” Ante, at 555, quoting Jones v. North Carolina Prisoners’ Labor Union, 433 U. S. 119, 132.15
*585Applying this test, the Court concludes that enforcement of the challenged restrictions does not constitute punishment because there is no showing of a subjective intent to punish and there is a rational basis for each of the challenged rules. In my view, the Court has reached an untenable conclusion because its test for punishment is unduly permissive.
The requirement that restraints have a rational basis provides an individual with virtually no protection against punishment. Any restriction that may reduce the cost of the facility’s warehousing function could not be characterized as “arbitrary or purposeless” and could not be “conclusively shown” to have no reasonable relation to the Government’s mission.16 This is true even of a restraint so severe that it might be cruel and unusual.
Nor does the Court’s intent test ensure the individual the protection that the Constitution guarantees. For the Court seems to use the term “intent” to mean the subjective intent of the jail administrator. This emphasis can only “encourage hypocrisy and unconscious self-deception.”17 While a *586subjective intent may provide a sufficient reason for finding that punishment has been inflicted, such an intent is clearly not a necessary nor even the most common element of a punitive sanction.
In short, a careful reading of the Court’s opinion reveals that it has attenuated the detainee’s constitutional protection against punishment into nothing more than a prohibition against irrational classifications or barbaric treatment. Having recognized in theory that the source of that protection is the Due Process Clause, the Court has in practice defined its scope in the far more permissive terms of equal protection and Eighth Amendment analysis.
Prior to today, our cases have unequivocally adopted a less obeisant and more objective approach to punishment than the one the Court applies here. In my judgment, those decisions provide the framework for the correct analysis of the punishment issue in this case.
The leading case is Kennedy v. Mendoza-Martinez, 372 U. S. 144. The Court’s conclusion that the statute in question was punitive was expressly based on “the objective manifestations of congressional purpose.” Id., at 169.18 The Court also recognized that in many cases such manifestations as it relied upon — the wording and construction of predecessor *587provisions as well as the congressional Reports on the provision itself, id., at 169-184 — would be unavailable19 or untrustworthy.20 In such cases, which surely include those in which the actions of an administrator rather than an Act of Congress are at issue, the Court stated that certain other “criteria” must be applied “to the face” of the official action to determine if it is punitive. Ibid. Illustrative of these objective “criteria” were several listed by the Court:
“Whether the sanction involves an affirmative disability or restraint, whether it has historically been regarded as a punishment, whether it comes into play only on a finding of scienter, whether its operation will promote the traditional aims of punishment — retribution and deterrence, whether the behavior to which it applies is already a crime, whether an alternative purpose to which it may rationally be connected is assignable for it, and whether it appears excessive in relation to the alternative purpose assigned .. . .” Id., at 168-169.
Today the Court does not expressly disavow the objective criteria identified in Mendoza-Martinez. In fact, in a footnote, see ante, at 539 n. 20, it relies on one of those criteria in order to answer an otherwise obvious criticism of the test the Court actually applies in this case. Under the test as the Court explains it today, prison guards could make regular use of dungeons, chains, and shackles, since such practices would make it possible to maintain security with a smaller number of guards. Commendably, however, the Court expressly rejects this application of its test by stating that the avail*588ability of less harsh alternatives would give rise to an inference that the practice was motivated by an intent to punish.
Although it is not easy to reconcile the footnote rejection of chains and shackles with the rest of the Court’s analysis, this footnote confirms my view that a workable standard must allow a court to infer that punishment has been inflicted by evaluating objective criteria such as those delineated in Mendoza-Martinez. When sanctions involve “affirmative disabilities]” and when they have “historically been regarded as a punishment,” Kennedy v. Mendoza-Martinez, 372 U. S., at 168-169, courts must be sensitive to the possibility that those sanctions are punitive. So, too, when the rules governing detention fail to draw any - distinction among those who are detained — suggesting that all may be subject to rules designed for the most dangerous few — careful scrutiny must be applied. Finally, and perhaps most important, when there is a significant and unnecessary disparity between the severity of the harm to the individual and the demonstrated importance of the regulatory objective, see ibid., courts must be justified in drawing an inference of punishment.
II
When measured against an objective standard, it is clear that the four rules discussed in Part III of the Court’s opinion are punitive in character. All of these rules were designed to forestall the potential harm that might result from smuggling money, drugs, or weapons into the institution. Such items, it is feared, might be secreted in hard-cover books, packages of food or clothing, or body cavities. That fear provides the basis for a total prohibition on the receipt of hard-cover books (except from publishers, book clubs, or bookstores) or packages of food, for a visual search of body cavities after every visit, and for excluding the detainee from his cell while his personal belongings are searched by a guard.
There is no question that jail administrators have a legitimate interest in preventing smuggling. But it is equally *589clear that that interest is being served here in a way that punishes many if not all of the detainees.
The challenged practices concededly deprive detainees of fundamental rights and privileges of citizenship beyond simply the right to leave. The Court recognizes this premise, but it dismisses its significance by asserting that detainees may be subjected to the “ ‘withdrawal or limitation' ” of fundamental rights. Ante, at 546, quoting Price v. Johnston, 334 U. S. 266, 285.21 I disagree. The withdrawal of rights is *590itself among the most basic punishments that society can exact, for such a withdrawal qualifies the subject’s citizenship and violates his dignity.22 Without question that kind of harm is an “affirmative disability” that “has historically been regarded as a punishment.” 23
This withdrawal of fundamental rights is not limited to those for whom punishment is proper, or to those detainees *591posing special security risks. The MCC houses convicted persons along with pretrial detainees. The former may constitutionally be punished, so long as that punishment is not cruel and unusual. And the fact of their long-term confinement may provide greater justification for concerns with ongoing smuggling operations, violence, or escape.24 Moreover, there may certainly be among the pretrial detainees, who cannot be punished, some whose background or history suggests a special danger that they will attempt to smuggle contraband into the jail. The rules at issue here, however, are not limited to those who may be constitutionally punished, or to those particularly dangerous detainees for whom onerous restraint is an appropriate regulation. Rather, the rules apply indiscriminately to all.
It is possible, of course, that the MCC officials have determined not to punish the convicted criminals who are confined there, but merely to regulate or detain them. It is possible, too, that as to the detainees, the rules that have been adopted and that are at issue here serve to impose only those restraints *592needed to regulate the least dangerous of the group. But the Government does not even suggest that the convicted criminals are not being punished during the confinement at MCC.25 And common sense suggests that if one set of rules is applied indiscriminately to detainees, those rules will serve to regulate the most dangerous — not the least — of the group. Indeed, prison security might well be in jeopardy were it otherwise. If that is true, and if the restraints are as substantial and fundamental as those here, then the conclusion that at least some, if not all, of the detainees are being punished is virtually inescapable.
That this is indeed the case here is confirmed by the excessive disparity between the harm to the individuals occasioned by these rules and the importance of their regulatory objective. The substantiality of the harm to the detainees cannot be doubted. The rights involved are among those that are specifically protected by the Constitution. That fact alone underscores our societal evaluation of their importance. The enforcement of these rules in the MCC, moreover, is a clear affront to the dignity of the detainee as a human being.26 *593To prohibit detainees from receiving books or packages communicates to the detainee that he, his friends, and his family cannot be trusted. And in the process, it eliminates one of his few remaining contacts with the outside world. The practice of searching the detainee’s private possessions in his absence, frequently without care, United States ex rel. Wolfish v. Levi, 439 F. Supp. 114, 149 (SDNY 1977), offends not only his privacy interest, but also his interest in “minimal dignity,” ibid. Finally, the search of private body cavities has been found to engender “deep degradation” and “terror” in the inmates, id., at 147: the price of such searches is so high as to lead detainees to forgo visits with friends and family altogether. Id., at 148.
In contrast to these severe harms to the individual, the interests served by these rules appear insubstantial. As to the room searches, nothing more than the convenience of the corrections staff supports the refusal to allow detainees to observe at a reasonable distance. While petitioners have raised the fear that inmates may become violent during such searches and may distract the guards, the District Court specifically found that they had made no showing of any pattern of violence or disruption to support these purported fears. Id., at 149. And absent such a showing, there is no more reason to ban all detainees from observing the searches of their rooms than there would be to ban them from every area in the MCC where guards or other inmates are present.
The prohibitions on receiving books and packages fare no better. The District Court found no record of “untoward experience” with respect to the book rule, United States ex rel. Wolfish v. United States, 428 F. Supp. 333, 340 (SDNY 1977), and no support in the evidence for the petitioners’ “dire predictions” as to packages, 439 F. Supp., at 152. The simple *594fact is, and the record and the case law make clear, that in many prisons housing criminals convicted of serious crimes— where the inmates as a class may well be more dangerous, where smuggling is likely to be a far more serious problem, and where punishment is appropriate — packages of various sorts are routinely admitted subject to inspection. Ibid. The administrators here have hardly established that the corrections staff at MCC is incapable of performing similar inspections with respect to an inmate population which has a far greater entitlement to them. And the unsupported claim that food or goods may be used for barter or may introduce sanitation problems ignores not only the possibility of reasonable regulation, but also the fact that similar goods are sold in the MCC commissary, id., at 152-153, and are no more immune from barter or spoilage.
The body-cavity search- — clearly the greatest personal indignity — may be the least justifiable measure of all. After every contact visit a body-cavity search is mandated by the rule. The District Court’s finding that these searches have failed in practice to produce any demonstrable improvement in security, id., at 147, is hardly surprising.27 Detainees and their visitors are in full view during all visits, and are fully clad. To insert contraband in one’s private body cavities during such a visit would indeed be “an imposing challenge to nerves and agility.” Ibid. There is no reason to expect, and the petitioners have established none, that many pretrial detainees would attempt, let alone succeed, in surmounting this challenge absent the challenged rule. Moreover, as the District Court explicitly found, less severe alternatives are available to ensure that contraband is not transferred during visits. Id., at 147-148. Weapons and other dangerous instruments, the items of greatest legitimate concern, may be *595discovered by the use of metal detecting devices or other equipment commonly used for airline security. In addition, inmates are required, even apart from the body-cavity searches, to disrobe, to have their clothing inspected, and to present open hands and arms to reveal the absence of any concealed objects. These alternative procedures, the District Court found, “amply satisf[y]” the demands of security. Id., at 148. In my judgment, there is no basis in this record to disagree.
It may well be, as the Court finds, that the rules at issue here were not adopted by administrators eager to punish those detained at MCC. The rules can all be explained as the easiest way for administrators to ensure security in the jail. But the easiest course for jail officials is not always one that our Constitution allows them to take. If fundamental rights are withdrawn and severe harms are indiscriminately inflicted on detainees merely to secure minimal savings in time and effort for administrators, the guarantee of due process is violated.
In my judgment, each of the rules at issue here is unconstitutional. The four rules do indiscriminately inflict harm on all pretrial detainees in MCC. They are all either unnecessary or excessively harmful, particularly when judged against our historic respect for the dignity of the free citizen. I think it is unquestionably a form of punishment to deny an innocent person the right to read a book loaned to him by a friend or relative while he is temporarily confined, to deny him the right to receive gifts or packages, to search his private possessions out of his presence, or to compel him to exhibit his private body cavities to the visual inspection of a guard. Absent probable cause to believe that a specific individual detainee poses a special security risk, none of these practices would be considered necessary, or even arguably reasonable, if the pretrial detainees were confined in a facility separate and apart from convicted prisoners. If reasons of *596convenience justify intermingling the two groups, it is not too much to require the facility’s administrator to accept the additional inspection burdens that would result from denying them the right to subject citizens to these humiliating indignities. I would affirm the judgment of the Court of Appeals as to all four of these rules.28
Ill
The so-called “double-bunking” issue was resolved by the District Court on cross-motions for summary judgment. The record was compiled and the issue decided on the basis of a legal test that all of us now agree was erroneous.29 If the record is incomplete, or if it discloses any material question of fact concerning the punitive character of the housing conditions at MCC, a remand for trial is required. Three basic facts dictate that result.
First, as earlier emphasized, MCC houses convicted prisoners along with pretrial detainees. Both classes of inmates are subjected to the same conditions. It may be that the Government — despite representations to the contrary, see 439 F. Supp., at 153 — conceives of the confinement of convicts in the facility as a vacation for them from the punitive rigors of prison life. But the opposite conclusion — that the detainees are instead being subjected to some of those rigors — is at least an equally justifiable inference from the facts revealed by the record, particularly in view of the other rules applicable to both classes.
Second, the Government acknowledges that MCC has been used to house twice as many inmates as it was designed to *597accommodate.30 The design capacity of a building is one crucial indication of its purpose. So is the later abandonment of that design in favor of a substantially more crowded and *598oppressive one. Certainly, the inference that what the architect designed to detain, the jailer has used to punish, is permissible, even if it may not be compelled or even probable.
Finally, MCC officials experienced little difficulty in complying with the preliminary order of the District Court to return the facility to its design capacity. The Court dismisses this fact as not conclusive on the question of purpose and reasonableness. Ante, at 542-543, n. 25. But the fact that the Government’s lawful regulatory purpose could so easily be served by less severe conditions is certainly some evidence of a punitive purpose and of excessiveness. If the lawful purpose may be equally served by those new conditions at no greater cost, the record provides a basis for arguing that there is no legitimate reason for the extra degree of severity that has characterized the overcrowded conditions in the past.31
While I by no means suggest that any of these facts demonstrates that the detention conditions are punitive,32 taken *599together they raise an issue of fact that should not be resolved by this Court, or even by the District Court, on a motion for summary judgment.
It is admittedly easier to conclude that the Due Process Clause prohibits preconviction punishment than it is to articulate a standard for determining if such punishment has occurred. But if the standard is to afford any meaningful protection for the citizen’s liberty, it must require something more than either an explicit statement by the administrator that his rule is designed to inflict punishment, or a sanction that is so arbitrary that it would be invalid even if it were not punitive. However the test is phrased, it must at least be satisfied by an unexplained and significant disparity between the severity of the harm to the individual and the demonstrated importance of the nonpunitive objective served by it. I therefore respectfully dissent from the conclusion that the demeaning and unnecessary practices described in Part III of the Court’s opinion do not constitute punishment, and also from the conclusion that the overcrowded housing conditions discussed in Part II do not even give rise to an inference that they have punitive qualities.
No State shall . . . deny to any person within its jurisdiction the equal protection of the laws.” U. S. Const., Arndt. 14, § 1.
“Excessive bail shall not be required, nor excessive fines imposed, nor cruel and unusual punishment inflicted.” U. S. Const., Arndt. 8.
Because this is a federal facility, it is, of course, the Fifth Amendment that applies. It provides, in relevant part: “No person shall be . . . deprived of life, liberty, or property, without due process of law . . . .”
Because Mr. Justice Marshall does not accept this basis for analysis, see ante, at 568-569,1 have added this separate dissent even though I agree with much of his analysis and most of his criticism of the Court.
See Meachum v. Fano, 427 U. S. 215, 230 (Stevens, J., dissenting).
See Leis v. Flynt, 439 U. S. 438, 443; Paul v. Davis, 424 U. S. 693.
The facility is used to house convicted persons who are temporarily in New York for court appearances and the like, as well as some who are confined there for the duration of short sentences.
There is neither time, staff, nor opportunity to offer convicted inmates at MCC the kind of training or treatment that is sometimes available in a prison environment.
See Webster’s Third International Dictionary 1804 (1961) (As “often" used, a “prison” is “an institution for the imprisonment of persons convicted of major crimes or felonies: a penitentiary as distinguished from a reformatory, local jail, or detention home”).
Long-term incarceration and other postconviction sanctions have significant backward-looking, personal, and normative components. Because they are primarily designed to inflict pain or to “correct” the individual because of some past misdeed, the sanctions are considered punitive. See E. Pincoffs, The Rationale of Legal Punishment 51-57 (1966). See also Gregg v. Georgia, 428 U. S. 153, 184, and n. 30 (opinion of Stewart, Powell, and SteveNS, JJ.); H. Hart, Punishment and Responsibility 4-5 (1968); id., at 158-173; F. Dostoevskii, Crime and Punishment (Coulson transí. 1964); I. Kant, The Philosophy of Law 195-198 (W. Hastie transí. 1887).
By contrast, pretrial detention is acceptable as a means of assuring the *582detainee’s presence at trial and of maintaining his and his fellows’ safety in the meantime. Its focus is therefore essentially forward looking, general, and nonnormative. Because this type of government sanction is primarily designed for the future benefit of the public at large and implies no moral judgment about the person affected, it is properly classified as regulatory. See H. Packer, The Limits of the Criminal Sanction 5 (1968).
The Court’s bill of attainder cases have recognized the distinction between regulation and punishment in analyzing the concept of “legislative punishment.” Thus, on the one hand, post bellum statutes excluding persons who had been sympathetic to the Confederacy from certain professions were found unconstitutional because of the backward-looking focus on the acts of specific individuals. Ex parte Garland, 4 Wall. 333; Cummings v. Missouri, 4 Wall. 277. However, later statutes requiring persons to take loyalty oaths before getting the benefits of certain labor legislation and before being employed in a public job were found constitutional because of their future orientation and more general purpose. American Communications Assn. v. Douds, 339 U. S. 382, 413-415; Garner v. Board of Public Works, 341 U. S. 716, 722-725.
On at least two occasions, this Court has relied upon this presumption as a justification for shielding a person awaiting trial from potentially oppressive governmental actions. McGinnis v. Royster, 410 U. S. 263, 273 (“[I]t would hardly be appropriate for the State to undertake in the pretrial detention period programs to rehabilitate a man still clothed with a presumption of innocence”); Stack v. Boyle, 342 U. S. 1, 4 (“Unless [the] right to bail before trial is preserved, the presumption of innocence, secured only after centuries of struggle, would lose its meaning”). These cases demonstrate that the presumption — or, as it was called last Term, the “assumption” — of innocence that is indulged until evidence has convinced a jury to the contrary beyond a reasonable doubt, see Taylor v. Kentucky, 436 U. S. 478, 484 n. 12, colors all of the government’s actions toward persons not yet convicted. In sum, although there may be some question as to what it means to treat a person as if he were guilty, there can be no dispute that the government may never do so at any point in advance of conviction.
Relying on nothing more than the force of assertion, and without even mentioning McGinnis and Stack, the Court states that the presumption of innocence “has no application to a determination of the rights of a pretrial *583detainee during confinement before his trial has even begun.” Ante, at 533. But having so recently reiterated that the presumption is “fundamental,” see Taylor v. Kentucky, supra, at 483, I cannot believe the Court means what it seems to be saying.
In many instances, detention will occur although the risk of flight is exceedingly low. This is because there is “a large class of persons for whom any bail at all is 'excessive bail.’ They are the people loosely referred to as ‘indigents.’ Studies of the operation of the bail system have demonstrated that even at the very lowest levels of bail — say $500, where the bail bond premium may be only $25 or $50 — there is a very substantial percentage of persons who do not succeed in making bail and are therefore held in custody pending trial.” Packer, supra n. 10, at 216.
American jurisdictions have traditionally relied on a pretrial system of “bail or jail” to assure that arrestees appear at trial. Id., at 211. As to the bail aspect of the system, the Eighth Amendment is explicit that whatever steps the Government takes must not be excessive in relation to that purpose. Stack v. Boyle, supra, at 5. See 18 U. S. C. §3146 (a). Although not expressed in the Constitution, a like restraint on the other half of the pretrial system is a logical corollary to the “No Excess Bail” Clause.
Indeed, this Court has recognized on previous occasions that individualization is sometimes necessary to prevent clearly punitive sanctions from being administered in a cruel and unusual manner. Woodson v. North Carolina, 428 U. S. 280, 304; Trop v. Dulles, 356 U. S. 86, 100.
Even if the Court were to apply this aspect of its test in a meaningful way, it would add little to the concept of punishment that is impermissible under the Due Process Clause. The Court states this test as follows: “[I]f a restriction or condition is not reasonably related to a legitimate goal — if it is arbitrary or purposeless — a court permissibly may infer that the purpose of the governmental action is punishment that may not constitution*585ally be inflicted upon detainees qua detainees.” Ante, at 539. It is readily apparent that this standard is nothing more than the “rational basis” requirement that even presumptively valid economic and social regulations must satisfy to pass muster under the Due Process Clause. Accordingly, if a court followed the path proposed in the quotation above, it would take unnecessary steps. For governmental activity that affects even minor interests and is “arbitrary or purposeless” is unconstitutional whether or not it is punishment. See, e. g., Rinaldi v. Yeager, 384 U. S. 305; Illinois Elections Board v. Socialist Workers Party, 440 U. S. 173.
Beyond excluding expressly intended punishment, the Court puts no restrictions on the goals that it recognizes as legitimate; under its test the Government need only show some rational nexus to security, order, or the apparently open-ended class of “operational concerns” facing the jail administrator, ante, at 540, and the restriction will be upheld,
“[The subjective approach] focuses on what an interested party intends rather than on what a detached observer thinks, thereby depriving the distinction [between punishment and other types of government activity] of any pretense to objectivity. If a prison warden thinks that his *586inmates are better off in his custody than they would be in the world outside, then by [the subjective] definition what he is administering is Treatment rather than Punishment. If the legislature that passes a compulsory commitment statute for narcotics addicts is motivated by hostility toward addicts, commitment is Punishment; if it is motivated by compassion, commitment is Treatment. And if it is motivated by both hostility and compassion? Other objections aside, what use can possibly be made of such a definition?
“Other objections cannot be left aside, because they demonstrate that [the subjective] definition not only is unintelligible but leads to quite dangerous consequences. . . . [For] [t]o allow the characterization to turn on the intention of the administrator is to encourage hypocrisy and unconscious self-deception.” Packer, supra n. 10, at 32-33.
Accord, United States v. Lovett, 328 U. S. 303, 311.
Some state courts have had to resort to such criteria even when analyzing the punitive content of legislation because many state assemblies publish no record of their deliberations. E. g., Starkweather v. Blair, 245 Minn. 371, 71 N. W. 2d 869 (1955).
“[E] ven a clear legislative classification of a statute as ‘non-penal’ would not alter the fundamental nature of a plainly penal statute.” Trop v. Dulles, 356 U. S., at 95 (plurality opinion).
Although the Court's discussion of this point is laced with citations of prison eases such as Price, ante, at 545-547, it fails to mention a single precedent dealing with pretrial detainees. Cf. Houchins v. KQED, Inc., 438 U. S. 1, 37-38 (Stevens, J., dissenting); O’Brien v. Skinner, 414 U. S. 524; Goosby v. Osser, 409 U. S. 512.
Having concluded that detainees’ rights are “limited,” the Court is reduced, for example, to analyzing restrictions on First Amendment rights in the deferential language of “minimum rationality” — language traditionally applied to restrictions on economic activities such as selling hot dogs or eyeglasses. New Orleans v. Dukes, 427 U. S. 297; Williamson v. Lee Optical Co., 348 U. S. 483.
The First Amendment is not the only victim of the Court’s analysis. It also devalues the Fourth Amendment as it applies to pretrial detainees. This is particularly evident with respect to the Court’s discussion of body-cavity searches. Although it recognizes the detainee’s constitutionally protected interest in privacy, the Court immediately demeans that interest by affording it “diminished scope.” The reason for the diminution is the detainee's limited expectation of privacy. Ante, at 557, 558. At first blush, the Court’s rationale appears to be that once the detainee is told that he will not be permitted to carry on any of his activities in private, he cannot “reasonably” expect otherwise. But “reasonable expectations of privacy” cannot have this purely subjective connotation lest we wake up one day to headlines announcing that henceforth the Government will not recognize the sanctity of the home but will instead enter residences at will. The reasonableness of the expectation must include an objective component that refers to those aspects of human activity that the “reasonable person” typically expects will be protected from unchecked Government observation. Cf. Katz v. United States, 389 U. S. 347, 361 (Harlan, J., concurring) . Hence, the question must be whether the Government may, without violating the Fourth Amendment, tell the detainee by words or by action that he has no or virtually no right to privacy. In my view, the *590answer to this question must be negative: despite the fact of his confinement and the impossibility of retreat to the privacy of his home, the detainee must have the right to privacy that we all retain when we venture out into public places. And surely the scope of that privacy is not so diminished that it does not include an expectation that body cavities will not be exposed to view. Absent probable cause, therefore, I would hold that such searches of pretrial detainees may not occur.
The classic example of the coincidence of punishment and the total deprivation of rights is voting. Thus, in Richardson v. Ramirez, 418 U. S. 24, the Court, although recognizing the importance of the right to vote, id., at 54, see Reynolds v. Sims, 377 U. S. 533, 561, found support in §2 of the Fourteenth Amendment for denying convicted felons the right to vote. Cf. O’Brien v. Skinner, supra (finding certain restrictions on absentee voting by pretrial detainees unconstitutional under the Equal Protection Clause). See also Goosby v. Osser, supra.
This is certainly not to say that the fact of conviction justifies the total deprivation of all constitutionally protected rights. Having abandoned the concept of the prisoner as a slave of the state, e. g., Morrissey v. Brewer, 408 U. S. 471, the Court has also rejected any ironclad exclusion of such persons from the protection of the Constitution. E. g., Wolff v. McDonnell, 418 U. S. 539, 555-556; Pell v. Procunier, 417 U. S. 817, 822; Cruz v. Beto, 405 U. S. 319; Lee v. Washington, 390 U. S. 333. Nonetheless, it also recognizes “that a prison inmate retains [only those] rights that are not inconsistent . . . with the legitimate penological objectives of the corrections system.” Pell v. Procunier, supra, at 822. Cf. Lanza v. New York, 370 U. S. 139.
E. g., Wolff v. McDonnell, supra, at 555; Richardson v. Ramirez, supra, at 43-53. The Court has probably relied upon historical analysis more often than on any of the other objective factors discussed in Kennedy v. Mendoza-Martinez, in determining whether some government sanction is punitive. E. g., Cummings v. Missouri, 4 Wall. 277; Ex parte Wilson, 114 U. S. 417, 426-429; Mackin v. United States, 117 U. S. 348, 350-352; Wong Wing v. United States, 163 U. S. 228, 237-238.
The prospect of long-term incarceration facing an inmate increases his incentive to use illicit means to obtain luxuries that his imprisonment would otherwise deny him. Moreover, the fact of long-term incarceration of a large number of persons is conducive to the development of an institutional subeconomy and even subgovemment that often thrives on contraband and is inconsistent with the orderly operation of the facility. See, e. g., H. Mattick, The Prosaic Sources of Prison Violence, Occasional Papers of the University of Chicago Law School, No. 3, Mar. 15, 1972.
As the foregoing indicates, I believe the analysis of the four rules as applied to convicted prisoners is different from that as applied to pretrial detainees. Not only do the due process and other rights of the two have different scope, but the Government's security interests also differ. In my view, the courts below, in erroneously applying the same standards to both' sets of inmates and in focusing on detainees, did not adequately develop the record with respect to convicts. Accordingly, I would remand the question of the validity of the four rules in the context of convicted prisoners for further proceedings. Cf. United States ex rel. Miller v. Twomey, 479 F. 2d 701, 719 (CA7 1973).
In fact, the Government admitted below that the “restrictions on the possession of personal property” at MCC “serve the legitimate purpose of punishment” with respect to convicted inmates as well as the security purposes relied on in the present context of pretrial detainees. United States ex rel. Wolfish v. Levi, 439 F. Supp. 114, 153 (SDNY 1977).
This affront may itself constitute punishment because of its retributive character. Mendoza-Martinez makes clear that a sanction is punitive if it "will promote [a] traditional ai[m] of punishment — retribution.” 372 U. S., at 168-169. In its retributive aspect, “ '[pjunishment is the way in which society expresses its denunciation for wrong doing.’ ” Gregg v. Georgia, 428 U. S., at 184, and n. 30 (opinion of Stewart, Powell, and Stevens, JJ.), quoting Lord Justice Denning’s testimony before the Royal Commission on Capital Punishment. See also letter from Judge Learned Hand to the editors of the University of Chicago Law Review (undated), reprinted in 22 U. Chi. L. Rev. 319 (1965); sources cited in the first paragraph of n. 10, supra. A focus of this “denunciatory” approach is the right of society, in significant respects, to deny the civic and human dignity *593of persons who have been convicted, of doing wrong. Cf. Gregg v. Georgia, supra, at 173, 182 (fundamental violations of “human dignity” may constitute cruel and unusual punishment).
Indeed, the District Court found the searches entirely ineffective in some of their most offensive manifestations (e. g., anal searches). 439 F. Supp., at 147.
The District Court reserved decision on all of these practices save the restriction on receipt of hardback books until a full trial on the merits. It is accordingly appropriate to resolve these issues now without a remand.
I do not understand how the Court, having quite thoroughly demonstrated that the District Court applied an erroneous legal test, ante, at 530, 532-535, can nonetheless rely on that court’s conclusion that no disputed issues of material fact prevented it from applying its erroneous test to the housing issue. Ante, at 541 n. 24.
“The decisive reality, however, not seriously open to debate, is that the rooms were designed and built to hold a single person, not more. The conclusion is compelled by an array of undisputed facts. To begin with, petitioners invoke the high authority of the architect who designed the MCC and who, in sworn testimony recorded in this court, has described a room like the ones he drew, housing one inmate, as a 'very basic planning principle.’ Contrasting dormitories with rooms, he went on to say:
“ 'Dormitories are a much more flexible kind of a thing, you see. That is the only real area in that particular facility. One of the reasons why there’s been a tendency to go to single rooms is because it's a very clear and apparent violation of capacity when you try to put two people in a room. You can’t put one and a third persons in a room. You can always up the population of a space, in which you put people in, and you can through more imaginative planning get better utilization of the space but there is an absoluteness of a room which is designed for one person, and to try to convert it into a two-person room, it’s a clear violation of the capability of that space. There is no question there. There is more than enough, you know, objections to double-celling.’
“It is not necessary by any means to rely solely on what the architect said; the plain visual evidence of what he did demonstrates that the rooms he designed were for one inmate, not two or more. There is no place for each of two people, assigned by others to this unwanted intimacy, to walk or eat or write a letter or be quiet or be outside another’s toilet. There is one shelf for toiletries and one for other things, neither adequate for two people. In the larger group of 100 double-celled rooms there is no place to hang a garment. The double-decker bunks by which these rooms have been changed from singles are so constructed that air from a vent, cold during our winter visit, blows out onto the upper bed a foot or so above body level. Many of the prisoners have blocked the vents to cope with this architecturally unintended unpleasantness. And, as a result the rooms are musty and unpleasant smelling. The single beds originally designed for these rooms each had two drawers built under them, mounted on casters for reasonably convenient use. In the reconstruction to house two inmates, it was found necessary to dismantle these- caster arrangements; now each 'double’ room has one of the old drawers lying loose under the lower bed or none at all for the two assigned occupants.” United States ex rel. Wolfish v. United States, 428 F. Supp. 333, 336-337 (SDNY 1977) (footnote omitted; emphasis in original).
To these facts may be added some of the findings of the District Court: (1) Even at design capacity, “movement is more restricted at the MCC than in most other federal facilities,” including those that exclusively house convicts, 439 F. Supp., at 125; (2) the doubling of the design capacity of individual cells leaves “no place for each of two people, assigned by others to this unwanted intimacy, to walk or eat or write a letter or be quiet or be outside another’s toilet,” places the person in the newly added upper bunk directly under the cold air vent, renders some of the furniture designed for the rooms unusable, and in general subjects the inmate to “foul odors, social stigma, humiliation, and denials of minimal privacy,” 428 F. Supp., at 337, 339; (3) overall, the “living conditions [are] grossly short of minimal decency, and [have] no semblance of justification except [for] the general defense that the facilities of the Bureau of Prisons are in toto insufficient to house all the people consigned to them,” 439 F. Supp., at 135. Without so stating expressly, the Court has rejected these findings. Ante, at 542-543. Because that rejection is not permissible absent a determination of clear error, and because no such determination has been made, its treatment of the District Court’s findings is inexplicable. See Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U. S. 100, 123.
The ameliorative factors discussed by the Court, ante, at 542-543, might well convince the factfinder that the housing conditions are not punitive.