CHIEF JUSTICE ROBERTS, with whom JUSTICE SCALIA joins, dissenting.

The Court creates constitutional law by surmising what is typical when a social guest encounters an entirely atypical situation. The rule the majority fashions does not implement the high office of the Fourth Amendment to protect privacy, but instead provides protection on a random and happenstance basis, protecting, for example, a co-occupant who happens to be at the front door when the other occupant consents to a search, but not one napping or watching television in the next room. And the cost of affording such random protection is great, as demonstrated by the recurring cases in which abused spouses seek to authorize police entry into a home they share with a nonconsenting abuser.

The correct approach to the question presented is clearly mapped out in our precedents: The Fourth Amendment protects privacy. If an individual shares information, papers, or places with another, he assumes the risk that the other person will in turn share access to that information or those papers or places with the government. And just as an individual who has shared illegal plans or incriminating documents with another cannot interpose an objection when that other person turns the information over to the government, just because the individual happens to be present at the time, so too someone who shares a place with another cannot inter-pose an objection when that person decides to grant access to the police, simply because the objecting individual happens to be present.

A warrantless search is reasonable if police obtain the voluntary consent of a person authorized to give it. Co-occupants have "assumed the risk that one of their number might permit [a] common area to be searched." United States v. Matlock, 415 U.S. 164, 171, n. 7 (1974). Just as Mrs. Randolph could walk upstairs, come down, and turn her husband's cocaine straw over to the police, she can consent to police entry and search of what is, after all, her home, too.

I

In Illinois v. Rodriguez, 497 U.S. 177 (1990), this Court stated that "[w]hat [a person] is assured by the Fourth Amendment . . . is not that no government search of his house will occur unless he consents; but that no such search will occur that is 'unreasonable.'" Id., at 183. One element that can make a warrantless government search of a home "'reasonable'" is voluntary consent. Id., at 184; Schneckloth v. Bustamonte, 412 U.S. 218, 219 (1973). Proof of voluntary consent "is not limited to proof that consent was given by the defendant," but the government "may show that permission to search was obtained from a third party who possessed common authority over or other sufficient relationship to the premises." Matlock, supra, at 171. Today's opinion creates an exception to this otherwise clear rule: A third-party consent search is unreasonable, and therefore constitutionally impermissible, if the co-occupant against whom evidence is obtained was present and objected to the entry and search.

This exception is based on what the majority describes as "widely shared social expectations" that "when people living together disagree over the use of their common quarters, a resolution must come through voluntary accommodation." Ante, at 111, 113-114. But this fundamental predicate to the majority's analysis gets us nowhere: Does the objecting co-tenant accede to the consenting co-tenant's wishes, or the other way around? The majority's assumption about voluntary accommodation simply leads to the common stalemate of two gentlemen insisting that the other enter a room first.

Nevertheless, the majority is confident in assuming — confident enough to incorporate its assumption into the Constitution — that an invited social guest who arrives at the door of a shared residence, and is greeted by a disagreeable co-occupant shouting "'stay out,'" would simply go away. Ante, at 113. The Court observes that "no sensible person would go inside under those conditions," Ibid., and concludes from this that the inviting co-occupant has no "authority" to insist on getting her way over the wishes of her co-occupant, ante, at 114. But it seems equally accurate to say — based on the majority's conclusion that one does not have a right to prevail over the express wishes of his co-occupant — that the objector has no "authority" to insist on getting his way over his co-occupant's wish that her guest be admitted.

The fact is that a wide variety of differing social situations can readily be imagined, giving rise to quite different social expectations. A relative or good friend of one of two feuding roommates might well enter the apartment over the objection of the other roommate. The reason the invitee appeared at the door also affects expectations: A guest who came to celebrate an occupant's birthday, or one who had traveled some distance for a particular reason, might not readily turn away simply because of a roommate's objection. The nature of the place itself is also pertinent: Invitees may react one way if the feuding roommates share one room, differently if there are common areas from which the objecting roommate could readily be expected to absent himself. Altering the numbers might well change the social expectations: Invitees might enter if two of three co-occupants encourage them to do so, over one dissenter.

The possible scenarios are limitless, and slight variations in the fact pattern yield vastly different expectations about whether the invitee might be expected to enter or to go away. Such shifting expectations are not a promising foundation on which to ground a constitutional rule, particularly because the majority has no support for its basic assumption — that an invited guest encountering two disagreeing co-occupants would flee — beyond a hunch about how people would typically act in an atypical situation.

And in fact the Court has not looked to such expectations to decide questions of consent under the Fourth Amendment, but only to determine when a search has occurred and whether a particular person has standing to object to a search. For these latter inquiries, we ask whether a person has a subjective expectation of privacy in a particular place, and whether "the expectation [is] one that society is prepared to recognize as 'reasonable.'" Katz v. United States, 389 U.S. 347, 361 (1967) (Harlan, J., concurring); see Minnesota v. Olson, 495 U.S. 91, 95-96, 100 (1990) (extending Katz test to standing inquiry). But the social expectations concept has not been applied to all questions arising under the Fourth Amendment, least of all issues of consent. A criminal might have a strong expectation that his longtime confidant will not allow the government to listen to their private conversations, but however profound his shock might be upon betrayal, government monitoring with the confidant's consent is reasonable under the Fourth Amendment. See United States v. White, 401 U.S. 745, 752 (1971) (plurality opinion).

The majority suggests that "widely shared social expectations" are a "constant element in assessing Fourth Amendment reasonableness," ante, at 111 (citing Rakas v. Illinois, 439 U.S. 128, 144, n. 12 (1978)), but that is not the case; the Fourth Amendment precedents the majority cites refer instead to a "legitimate expectation of privacy," id., at 143, n. 12 (emphasis added; internal quotation marks omitted). Whatever social expectation the majority seeks to protect, it is not one of privacy. The very predicate giving rise to the question in cases of shared information, papers, containers, or places is that privacy has been shared with another. Our common social expectations may well be that the other person will not, in turn, share what we have shared with them with another — including the police — but that is the risk we take in sharing. If two friends share a locker and one keeps contraband inside, he might trust that his friend will not let others look inside. But by sharing private space, privacy has "already been frustrated" with respect to the locker-mate. United States v. Jacobsen, 466 U.S. 109, 117 (1984). If two roommates share a computer and one keeps pirated software on a shared drive, he might assume that his room-mate will not inform the government. But that person has given up his privacy with respect to his roommate by saving the software on their shared computer.

A wide variety of often subtle social conventions may shape expectations about how we act when another shares with us what is otherwise private, and those conventions go by a variety of labels — courtesy, good manners, custom, protocol, even honor among thieves. The Constitution, however, protects not these but privacy, and once privacy has been shared, the shared information, documents, or places remain private only at the discretion of the confidant.

II

Our cases reflect this understanding. In United States v. White, we held that one party to a conversation can consent to government eavesdropping, and statements made by the other party will be admissible at trial. 401 U.S., at 752. This rule is based on privacy: "Inescapably, one contemplating illegal activities must realize and risk that his companions may be reporting to the police [I]f he has no doubts, or allays them, or risks what doubt he has, the risk is his." Ibid.

The Court has applied this same analysis to objects and places as well. In Frazier v. Cupp, 394 U.S. 731 (1969), a duffel bag "was being used jointly" by two cousins. Id., at 740. The Court held that the consent of one was effective to result in the seizure of evidence used against both: "[I]n allowing [his cousin] to use the bag and in leaving it in his house, [the defendant] must be taken to have assumed the risk that [his cousin] would allow someone else to look inside." Ibid.

As the Court explained in United States v. Jacobsen, supra:

"It is well settled that when an individual reveals private information to another, he assumes the risk that his confidant will reveal that information to the authorities, and if that occurs the Fourth Amendment does not prohibit governmental use of that information. Once frustration of the original expectation of privacy occurs, the Fourth Amendment does not prohibit governmental use of the now non-private information: 'This Court has held repeatedly that the Fourth Amendment does not prohibit the obtaining of information revealed to a third party and conveyed by him to Government authorities, even if the information is revealed on the assumption that it will be used only for a limited purpose and the confidence placed in a third party will not be betrayed.'" Id., at 117 (quoting United States v. Miller, 425 U.S. 435, 443 (1976)).

The same analysis applies to the question whether our privacy can be compromised by those with whom we share common living space. If a person keeps contraband in common areas of his home, he runs the risk that his co-occupants will deliver the contraband to the police. In Coolidge v. New Hampshire, 403 U.S. 443 (1971), Mrs. Coolidge retrieved four of her husband's guns and the clothes he was wearing the previous night and handed them over to police. We held that these items were properly admitted at trial because "when Mrs. Coolidge of her own accord produced the guns and clothes for inspection, . . . it was not incumbent on the police to stop her or avert their eyes." Id., at 489.

Even in our most private relationships, our observable actions and possessions are private at the discretion of those around us. A husband can request that his wife not tell a jury about contraband that she observed in their home or illegal activity to which she bore witness, but it is she who decides whether to invoke the testimonial marital privilege. Trammel v. United States, 445 U.S. 40,53 (1980). In Trammel, we noted that the former rule prohibiting a wife from testifying about her husband's observable wrongdoing at his say-so "goes far beyond making 'every man's house his castle/and permits a person to convert his house into 'a den of thieves.'" Id., at 51-52 (quoting 5 J. Bentham, Rationale of Judicial Evidence 340 (1827)).

There is no basis for evaluating physical searches of shared space in a manner different from how we evaluated the privacy interests in the foregoing cases, and in fact the Court has proceeded along the same lines in considering such searches. In Matlock, police arrested the defendant in the front yard of a house and placed him in a squad car, and then obtained permission from Mrs. Graff to search a shared bedroom for evidence of Matlock's bank robbery. 415 U.S., at 166. Police certainly could have assumed that Matlock would have objected were he consulted as he sat handcuffed in the squad car outside. And in Rodriguez, where Miss Fischer offered to facilitate the arrest of her sleeping boy-friend by admitting police into an apartment she apparently shared with him, 497 U.S., at 179, police might have noted that this entry was undoubtedly contrary to Rodriguez's social expectations. Yet both of these searches were reasonable under the Fourth Amendment because Mrs. Graff had authority, and Miss Fischer apparent authority, to admit others into areas over which they exercised control, despite the almost certain wishes of their present co-occupants.

The common thread in our decisions upholding searches conducted pursuant to third-party consent is an understanding that a person "assume[s] the risk" that those who have access to and control over his shared property might consent to a search. Matlock, 415 U.S., at 171, n. 7. In Matlock, we explained that this assumption of risk is derived from a third party's "joint access or control for most purposes" of shared property. Ibid. And we concluded that shared use of property makes it "reasonable to recognize that any of the co-inhabitants has the right to permit the inspection in his own right." Ibid.

In this sense, the risk assumed by a joint occupant is comparable to the risk assumed by one who reveals private information to another. If a person has incriminating information, he can keep it private in the face of a request from police to share it, because he has that right under the Fifth Amendment. If a person occupies a house with incriminating information in it, he can keep that information private in the face of a request from police to search the house, because he has that right under the Fourth Amendment. But if he shares the information — or the house — with another, that other can grant access to the police in each instance.

To the extent a person wants to ensure that his possessions will be subject to a consent search only due to his own consent, he is free to place these items in an area over which others do not share access and control, be it a private room or a locked suitcase under a bed. Mr. Randolph acknowledged this distinction in his motion to suppress, where he differentiated his law office from the rest of the Randolph house by describing it as an area that "was solely in his control and dominion." App. 3. As to a "common area," however, co-occupants with "joint access or control" may consent to an entry and search. Matlock, supra, at 171, n. 7.

By emphasizing the objector's presence and noting an occupant's understanding that obnoxious guests might "be admitted in [one's] absence," ante, at 111, the majority appears to resurrect an agency theory of consent suggested in our early cases. See Stoner v. California, 376 U.S. 483, 489 (1964) (stating that a hotel clerk could not consent to a search of a guest's room because the guest had not waived his rights "by word or deed, either directly or through an agent"); Chapman v. United States, 365 U.S. 610, 616-617 (1961). This agency theory is belied by the facts of Matlock and Rodriguez — both defendants were present but simply not asked for consent — and the Court made clear in those cases that a co-occupant's authority to consent rested not on an absent occupant's delegation of choice to an agent, but on the consenting co-occupant's "joint access or control" of the property. Matlock, supra, at 171, n. 7; see Rodriguez, supra, at 181; United States v. McAlpine, 919 F. 2d 1461, 1464, n. 2 (CA10 1990) ("[A]gency analysis [was] put to rest by the Supreme Court's reasoning in Matlock").

The law acknowledges that although we might not expect our friends and family to admit the government into common areas, sharing space entails risk. A person assumes the risk that his co-occupants — just as they might report his illegal activity or deliver his contraband to the government — might consent to a search of areas over which they have access and control. See United States v. Karo, 468 U.S. 705, 726 (1984) (O'Connor, J., concurring in part and concurring in judgment) (finding it a "relatively easy case . . . when two persons share identical, overlapping privacy interests in a particular place, container, or conversation. Here both share the power to surrender each other's privacy to a third party").

III

The majority states its rule as follows: "[A] warrantless search of a shared dwelling for evidence over the express refusal of consent by a physically present resident cannot be justified as reasonable as to him on the basis of consent given to the police by another resident." Ante, at 120.

Just as the source of the majority's rule is not privacy, so too the interest it protects cannot reasonably be described as such. That interest is not protected if a co-owner happens to be absent when the police arrive, in the backyard gardening, asleep in the next room, or listening to music through earphones so that only his co-occupant hears the knock on the door. That the rule is so random in its application confirms that it bears no real relation to the privacy protected by the Fourth Amendment. What the majority's rule protects is not so much privacy as the good luck of a co-owner who just happens to be present at the door when the police arrive. Usually when the development of Fourth Amendment jurisprudence leads to such arbitrary lines, we take it as a signal that the rules need to be rethought. See California v. Acevedo, 500 U.S. 565, 574, 580 (1991). We should not embrace a rule at the outset that its sponsors appreciate will result in drawing fine, formalistic lines. See ante, at 121.

Rather than draw such random and happenstance lines — and pretend that the Constitution decreed them — the more reasonable approach is to adopt a rule acknowledging that shared living space entails a limited yielding of privacy to others, and that; the law historically permits those to whom we have yielded our privacy to in turn cooperate with the government. Such a rule flows more naturally from our cases concerning Fourth Amendment reasonableness and is logically grounded in the concept of privacy underlying that Amendment.

The scope of the majority's rule is not only arbitrary but obscure as well. The majority repeats several times that a present co-occupant's refusal to permit entry renders the search unreasonable and invalid "as to him." Ante, at 106, 120, 122. This implies entry and search would be reasonable "as to" someone else, presumably the consenting co-occupant and any other absent co-occupants. The normal Fourth Amendment rule is that items discovered in plain view are admissible if the officers were legitimately on the premises; if the entry and search were reasonable "as to" Mrs. Randolph, based on her consent, it is not clear why the cocaine straw should not be admissible "as to" Mr. Randolph, as discovered in plain view during a legitimate search "as to" Mrs. Randolph. The majority's differentiation between entry focused on discovering whether domestic violence has occurred (and the consequent authority to seize items in plain view), and entry focused on searching for evidence of other crime, is equally puzzling. See ante, at 118-119. This Court has rejected subjective motivations of police officers in assessing Fourth Amendment questions, see Whren v. United States, 517 U.S. 806, 812-813 (1996), with good reason: The police do not need a particular reason to ask for consent to search, whether for signs of domestic violence or evidence of drug possession.

While the majority's rule protects something random, its consequences are particularly severe. The question presented often arises when innocent co-tenants seek to disassociate or protect themselves from ongoing criminal activity. See, e. g., United States v. Hendrix, 595 F. 2d 883, 884 (CADC 1979) ( per curiam) (wife asked police "'to get her baby and take [a] sawed-off shotgun out of her house'"); People v. Cosme, 48 N. Y. 2d 286, 288-289, 293, 397 N.E.2d 1319, 1320, 1323 (1979) (woman asked police to remove cocaine and a gun from a shared closet); United States v. Botsch, 364 F. 2d 542, 547 (CA2 1966). Under the majority's rule, there will be many cases in which a consenting co-occupant's wish to have the police enter is overridden by an objection from another present co-occupant. What does the majority imagine will happen, in a case in which the consenting co-occupant is concerned about the other's criminal activity, once the door clicks shut? The objecting co-occupant may pause briefly to decide whether to destroy any evidence of wrongdoing or to inflict retribution on the consenting co-occupant first, but there can be little doubt that he will attend to both in short order. It is no answer to say that the consenting co-occupant can depart with the police; remember that it is her home, too, and the other co-occupant's very presence, which allowed him to object, may also prevent the consenting co-occupant from doing more than urging the police to enter.

Perhaps the most serious consequence of the majority's rule is its operation in domestic abuse situations, a context in which the present question often arises. See Rodriguez, 497 U.S., at 179; United States v. Donlin, 982 F. 2d 31 (CA1 1992); Hendrix, supra; People v. Sanders, 904 P. 2d 1311 (Colo. 1995) (en banc); Brandon v. State, 778 P. 2d 221 (Alaska App. 1989). While people living together might typically be accommodating to the wishes of their co-tenants, requests for police assistance may well come from co-inhabitants who are having a disagreement. The Court concludes that because "no sensible person would go inside" in the face of disputed consent, ante, at 113, and the consenting co-tenant thus has "no recognized authority" to insist on the guest's admission, ante, at 114, a "police officer [has] no better claim to reasonableness in entering than the officer would have in the absence of any consent at all," ibid. But the police officer's superior claim to enter is obvious: Mrs. Randolph did not invite the police to join her for dessert and coffee; the officer's precise purpose in knocking on the door was to assist with a dispute between the Randolphs — one in which Mrs. Randolph felt the need for the protective presence of the police. The majority's rule apparently forbids police from entering to assist with a domestic dispute if the abuser whose behavior prompted the request for police assistance objects.

The majority acknowledges these concerns, but dismisses them on the ground that its rule can be expected to give rise to exigent situations, and police can then rely on an exigent circumstances exception to justify entry. Ante, at 116-117, n. 6. This is a strange way to justify a rule, and the fact that alternative justifications for entry might arise does not show that entry pursuant to consent is unreasonable. In addition, it is far from clear that an exception for emergency entries suffices to protect the safety of occupants in domestic disputes. See, e. g., United States v. Davis, 290 F. 3d 1239, 1240-1241 (CA10 2002) (finding no exigent circumstances justifying entry when police responded to a report of domestic abuse, officers heard no noise upon arrival, defendant told officers that his wife was out of town, and wife then appeared at the door seemingly unharmed but resisted husband's efforts to close the door).

Rather than give effect to a consenting spouse's authority to permit entry into her house to avoid such situations, the majority again alters established Fourth Amendment rules to defend giving veto power to the objecting spouse. In response to the concern that police might be turned away under its rule before entry can be justified based on exigency, the majority creates a new rule: A "good reason" to enter, coupled with one occupant's consent, will ensure that a police officer is "lawfully in the premises." Ante, at 118. As support for this "consent plus a good reason" rule, the majority cites a treatise, which itself refers only to emergency entries. Ibid, (citing 4 W. LaFave, Search and Seizure § 8.3(d), p. 161 (4th ed. 2004)). For the sake of defending what it concedes are fine, formalistic lines, the majority spins out an entirely new framework for analyzing exigent circumstances. Police may now enter with a "good reason" to believe that "violence (or threat of violence) has just occurred or is about to (or soon will) occur." Ante, at 118. And apparently a key factor allowing entry with a "good reason" short of exigency is the very consent of one co-occupant the majority finds so inadequate in the first place.

The majority's analysis alters a great deal of established Fourth Amendment law. The majority imports the concept of "social expectations," previously used only to determine when a search has occurred and whether a particular person has standing to object to a search, into questions of consent. Ante, at 111, 113., To determine whether entry and search are reasonable, the majority considers a police officer's subjective motive in asking for consent, which we have otherwise refrained from doing in assessing Fourth Amendment questions. Ante, at 118. And the majority creates a new exception to the warrant requirement to justify warrantless entry short of agency in potential domestic abuse situations. Ibid.

Considering the majority's rule is solely concerned with protecting a person who happens to be present at the door when a police officer asks his co-occupant for consent to search, but not one who is asleep in the next room or in the backyard gardening, the majority has taken a great deal of pain in altering Fourth Amendment doctrine, for precious little (if any) gain in privacy. Perhaps one day, as the consequences of the majority's analytic approach become clearer, today's opinion will be treated the same way the majority treats our opinions in Matlock and Rodriguez — as a "loose end" to be tied up. Ante, at 121.

One of the concurring opinions states that if it had to choose between a rule that a co-tenant's consent was valid or a rule that it was not, it would choose the former. Ante, at 125 (opinion of BREYER, J.). The concurrence advises, however, that "no single set of legal rules can capture the ever-changing complexity of human life," ibid., and joins what becomes the majority opinion, "[g]iven the case-specific nature of the Court's holding," ante, at 127. What the majority establishes, in its own terms, is "the rule that a physically present inhabitant's express refusal of consent to a police search is dispositive as to him, regardless of the consent of a fellow occupant." Ante, at 122-123 (emphasis added). The concurrence joins with the apparent "understandin[g]" that the majority's "rule" is not a rule at all, but simply a "case-specific" holding. Ante, at 127 (opinion of BREYER, J.). The end result is a complete lack of practical guidance for the police in the field, let alone for the lower courts.

* * *

Our third-party consent cases have recognized that a person who shares common areas with others "assume[s] the risk that one of their number might permit the common area to be searched." Matlock, 415 U.S., at 171, n. 7. The majority reminds us, in high tones, that a man's home is his castle, ante, at 115, but even under the majority's rule, it is not his castle if he happens to be absent, asleep in the keep, or otherwise engaged when the constable arrives at the gate. Then it is his co-owner's castle. And, of course, it is not his castle if he wants to consent to entry, but his co-owner objects. Rather than constitutionalize such an arbitrary rule, we should acknowledge that a decision to share a private place, like a decision to share a secret or a confidential document, necessarily entails the risk that those with whom we share may in turn choose to share — for their own protection or for other reasons — with the police.

I respectfully dissent.

JUSTICE SCALIA, dissenting.

I join the dissent of THE CHIEF JUSTICE, but add these few words in response to JUSTICE STEVENS' concurrence.

It is not as clear to me as it is to JUSTICE STEVENS that, at the time the Fourth Amendment was adopted, a police officer could enter a married woman's home over her objection, and could not enter with only her consent. Nor is it clear to me that the answers to these questions depended solely on who owned the house. It is entirely clear, however, that if the matter did depend solely on property rights, a latter-day alteration of property rights would also produce a latter-day alteration of the Fourth Amendment outcome — without altering the Fourth Amendment itself.

JUSTICE STEVENS' attempted critique of originalism confuses the original import of the Fourth Amendment with the background sources of law to which the Amendment, on its original meaning, referred. From the date of its ratification until well into the 20th century, violation of the Amendment was tied to common-law trespass. See Kyllo v. United States, 533 U.S. 27, 31-32 (2001); see also California v. Acevedo, 500 U.S. 565, 581, 583 (1991) (SCALIA, J., concurring in judgment). On the basis of that connection, someone who had power to license the search of a house by a private party could authorize a police search. See 1 Restatement of Torts § 167, and Comment 6 (1934); see also Williams v. Howard, 110 S. C. 82, 96 S. E. 251 (1918); Fennemore v. Armstrong, 29 Del. 35, 96 A. 204 (Super.Ct. 1915). The issue of who could give such consent generally depended, in turn, on "historical and legal refinements" of property law. United States v. Matlock, 415 U.S. 164, 171, n. 7 (1974). As property law developed, individuals who previously could not authorize a search might become able to do so, and those who once could grant such consent might no longer have that power. But changes in the law of property to which the Fourth Amendment referred would not alter the Amendment's meaning: that anyone capable of authorizing a search by a private party could consent to a warrantless search by the police.

There is nothing new or surprising in the proposition that our unchanging Constitution refers to other bodies of law that might themselves change. The Fifth Amendment provides, for instance, that "private property" shall not "be taken for public use, without just compensation"; but it does not purport to define property rights. We have consistently held that "the existence of a property interest is determined by reference to 'existing rules or understandings that stem from an independent source such as state law.'" Phillips v. Washington Legal Foundation, 524 U.S. 156, 164 (1998) (quoting Board of Regents of State Colleges v. Roth, 408 U.S. 564, 577 (1972)). The same is true of the Fourteenth Amendment Due Process Clause's protection of "property." See Castle Rock v. Gonzales, 545 U.S. 748, 756 (2005). This reference to changeable law presents no problem for the originalist. No one supposes that the meaning of the Constitution changes as States expand and contract property rights. If it is indeed true, therefore, that a wife in 1791 could not authorize the search of her husband's house, the fact that current property law provides otherwise is no more troublesome for the originalist than the well-established fact that a State must compensate its takings of even those property rights that did not exist at the time of the founding.

In any event, JUSTICE STEVENS' panegyric to the equal rights of women under modern property law does not support his conclusion that "[a]ssuming . . . both spouses are competent, neither one is a master possessing the power to override the other's constitutional right to deny entry to their castle." Ante, at 125. The issue at hand is what to do when there is a conflict between two equals. Now that women have authority to consent, as JUSTICE STEVENS claims men alone once did, it does not follow that the spouse who refuses consent should be the winner of the contest. JUSTICE STEVENS could just as well have followed the same historical developments to the opposite conclusion: Now that "the male and the female are equal partners," Ibid., and women can consent to a search of their property, men can no longer obstruct their wishes. Men and women are no more "equal" in the majority's regime, where both sexes can veto each other's consent, than on the dissent's view, where both sexes cannot.

Finally, I must express grave doubt that today's decision deserves JUSTICE STEVENS' celebration as part of the forward march of women's equality. Given the usual patterns of domestic violence, how often can police be expected to encounter the situation in which a man urges them to enter the home while a woman simultaneously demands that they stay out? The most common practical effect of today's decision, insofar as the contest between the sexes is concerned, is to give men the power to stop women from allowing police into their homes — which is, curiously enough, precisely the power that JUSTICE STEVENS disapprovingly presumes men had in 1791.

JUSTICE THOMAS, dissenting.

The Court has long recognized that "[i]t is an act of responsible citizenship for individuals to give whatever information they may have to aid in law enforcement." Miranda v. Arizona, 384 U.S. 436, 477-478 (1966). Consistent with this principle, the Court held in Coolidge v. New Hampshire, 403 U.S. 443 (1971), that no Fourth Amendment search occurs where, as here, the spouse of an accused voluntarily leads the police to potential evidence of wrongdoing by the accused. Id., at 486-490. Because Coolidge squarely controls this case, the Court need not address whether police could permissibly have conducted a general search of the Randolph home, based on Mrs. Randolph's consent. I respectfully dissent.

In the instant case, Mrs. Randolph told police responding to a domestic dispute that respondent was using a substantial quantity of cocaine. Upon police request, she consented to a general search of her residence to investigate her statements. However, as the Court's recitation of the facts demonstrates, ante, at 107, the record is clear that no such general search occurred. Instead, Sergeant Brett Murray asked Mrs. Randolph where the cocaine was located, and she showed him to an upstairs bedroom, where he saw the "piece of cut straw" on a dresser. Corrected Tr. of Motion to Suppression Hearing in Case No. 2001R-699 (Super.Ct. Sumter Cty., Ga., Oct. 3, 2002), pp. 8-9. Upon closer examination, Sergeant Murray observed white residue on the straw, and concluded the straw had been used for ingesting cocaine. Id., at 8. He then collected the straw and the residue as evidence. Id., at 9.

Sergeant Murray's entry into the Randolphs' home at the invitation of Mrs. Randolph to be shown evidence of respondent's cocaine use does not constitute a Fourth Amendment search. Under this Court's precedents, only the action of an agent of the government can constitute a search within the meaning of the Fourth Amendment, because that Amendment "was intended as a restraint upon the activities of sovereign authority, and was not intended to be a limitation upon other than governmental agencies." Burdeau v. McDowell, 256 U.S. 465, 475 (1921) (emphasis added). See also Coolidge, 403 U.S., at 487. Applying this principle in Coolidge, the Court held that when a citizen leads police officers into a home shared with her spouse to show them evidence relevant to their investigation into a crime, that citizen is not acting as an agent of the police, and thus no Fourth Amendment search has occurred. Id., at 488-498.

Review of the facts in Coolidge clearly demonstrates that it governs this case. While the police interrogated Coolidge as part of their investigation into a murder, two other officers were sent to his house to speak with his wife. Id., at 485. During the course of questioning Mrs. Coolidge, the police asked whether her husband owned any guns. Id., at 486. Mrs. Coolidge replied in the affirmative, and offered to retrieve the weapons for the police, apparently operating under the assumption that doing so would help to exonerate her husband. Ibid. The police accompanied Mrs. Coolidge to the bedroom to collect the guns, as well as clothing that Mrs. Coolidge told them her husband had been wearing the night of the murder. Ibid.

Before this Court, Coolidge argued that the evidence of the guns and clothing should be suppressed as the product of an unlawful search because Mrs. Coolidge was acting as an "'instrument,'" or agent, of the police by complying with a "'demand'" made by them. Id., at 487. The Court recognized that, had Mrs. Coolidge sought out the guns to give to police wholly on her own initiative, "there can be no doubt under existing law that the articles would later have been admissible in evidence." Ibid. That she did so in cooperation with police pursuant to their request did not transform her into their agent; after all, "it is no part of the policy underlying the Fourth and Fourteenth Amendments to discourage citizens from aiding to the utmost of their ability in the apprehension of criminals." Id., at 488. Because the police were "acting normally and properly" when they asked about any guns, and questioning Mrs. Coolidge about the clothing was "logical and in no way coercive," the Fourth Amendment did not require police to "avert their eyes" when Mrs. Coolidge produced the guns and clothes for inspection. Id., at 488-489.

This case is indistinguishable from Coolidge, compelling the conclusion that Mrs. Randolph was not acting as an agent of the police when she admitted Sergeant Murray into her home and led him to the incriminating evidence. Just as Mrs. Coolidge could, of her own accord, have offered her husband's weapons and clothing to the police without implicating the Fourth Amendment, so too could Mrs. Randolph have simply retrieved the straw from the house and given it to Sergeant Murray. Indeed, the majority appears to concede as much. Ante, at 116 ("The co-tenant acting on his own initiative may be able to deliver evidence to the police, Coolidge, supra, at 487-489 . . ., and can tell the police what he knows, for use before a magistrate in getting a warrant"). Drawing a constitutionally significant distinction between what occurred here and Mrs. Randolph's independent production of the relevant evidence is both inconsistent with Coolidge and unduly formalistic.

Accordingly, the trial court appropriately denied respondent's motion to suppress the evidence Mrs. Randolph provided to the police and the evidence obtained as a result of the consequent search warrant. I would therefore reverse the judgment of the Supreme Court of Georgia.