announced the judgment of the Court and delivered an opinion, in which The Chief Justice and Justice Scalia join.
In this case we must decide whether a failure to give a suspect the warnings prescribed by Miranda v. Arizona, *634384 U. S. 436 (1966), requires suppression of the physical fruits of the suspect’s unwarned but voluntary statements. The Court has previously addressed this question but has not reached a definitive conclusion. See Massachusetts v. White, 439 U. S. 280 (1978) (per curiam) (dividing evenly on the question); see also Patterson v. United States, 485 U. S. 922 (1988) (White, J., dissenting from denial of.certiorari). Although we believe that the Court’s decisions in Oregon v. Elstad, 470 U. S. 298 (1985), and Michigan v. Tucker, 417 U. S. 433 (1974), are instructive, the Courts of Appeals have split on the question after our decision in Dickerson v. United States, 530 U. S. 428 (2000). See, e. g., United States v. Villalba-Alvarado, 345 F. 3d 1007 (CA8 2003) (holding admissible the physical fruits of a Miranda violation); United States v. Sterling, 283 F. 3d 216 (CA4 2002) (same); United States v. DeSumma, 272 F. 3d 176 (CA3 2001) (same); United States v. Faulkingham, 295 F. 3d 85 (CA1 2002) (holding admissible the physical fruits of a negligent Miranda violation). Because the Miranda rule protects against violations of the Self-Incrimination Clause, which, in turn, is not implicated by the introduction at trial of physical evidence resulting from voluntary statements, we answer the question presented in the negative.
I
In June 2001, respondent, Samuel Francis Patane, was arrested for harassing his ex-girlfriend, Linda O’Donnell. He was released on bond, subject to a temporary restraining order that prohibited him from contacting O’Donnell. Respondent apparently violated the restraining order by attempting to telephone O’Donnell. On June 6, 2001, Officer Tracy Fox of the Colorado Springs Police Department began to investigate the matter. On the same day, a county probation officer informed an agent of the Bureau of Alcohol, Tobacco and Firearms (ATF), that respondent, a convicted felon, illegally possessed a .40 Glock pistol. The ATF relayed this information to Detective Josh Benner, who worked *635closely with the ATF. Together, Detective Benner and Officer Fox proceeded to respondent’s residence.
After reaching the residence and inquiring into respondent’s attempts to contact O’Donnell, Officer Fox arrested respondent for violating the restraining order. Detective Benner attempted to advise respondent of his Miranda rights but got no further than the right to remain silent. At that point, respondent interrupted, asserting that he knew his rights, and neither officer attempted to complete the warning.1 App. 40.
Detective Benner then asked respondent about the Glock. Respondent was initially reluctant to discuss the matter, stating: ‘T am not sure I should tell you anything about the Glock because I don’t want you to take it away from me.” Id., at 41. Detective Benner persisted, and respondent told him that the pistol was in his bedroom. Respondent then gave Detective Benner permission to retrieve the pistol. Detective Benner found the pistol and seized it.
A grand jury indicted respondent for possession of a firearm by a convicted felon, in violation of 18 U. S. C. § 922(g)(1). The District Court granted respondent’s motion to suppress the firearm, reasoning that the officers lacked probable cause to arrest respondent for violating the restraining order. It therefore declined to rule on respondent’s alternative argument that the gun should be suppressed as the fruit of an unwarned statement.
The Court of Appeals reversed the District Court’s ruling with respect to probable cause but affirmed the suppression order on respondent’s alternative theory. The court rejected the Government’s argument that this Court’s decisions in Elstad, supra, and Tucker, supra, foreclosed application of the fruit of the poisonous tree doctrine of Wong Sun *636v. United States, 371 U. S. 471 (1963), to the present context. 304 F. 3d 1013, 1019 (CA10 2002). These holdings were, the Court of Appeals reasoned, based on the view that Miranda announced a prophylactic rule, a position that it found to be incompatible with this Court’s decision in Dickerson, swpra, at 444 (“Miranda announced a constitutional rule that Congress may not supersede legislatively”).2 The Court of Appeals thus equated Dickerson’s announcement that Miranda is a constitutional rule with the proposition that a failure to warn pursuant to Miranda is itself a violation of the Constitution (and, more particularly, of the suspect’s Fifth Amendment rights). Based on its understanding of Dickerson, the Court of Appeals rejected the post-Dickerson views of the Third and Fourth Circuits that the fruits doctrine does not apply to Miranda violations. 304 F. 3d, at 1023-1027 (discussing United States v. Sterling, 283 F. 3d 216 (CA4 2002), and United States v. DeSumma, 272 F. 3d 176 (CA3 2001)). It also disagreed with the First Circuit’s conclusion that suppression is not generally required in the case of negligent failures to warn, 304 F. 3d, at 1027-1029 (discussing United States v. Faulkingham, 295 F. 3d 85 (CAI 2002)), explaining that “[djeterrence is necessary not merely to deter intentional wrongdoing, but also to ensure that officers diligently (non-negligently) protect — and properly are trained to protect — the constitutional rights of citizens,” 304 F. 3d, at 1028-1029. We granted certiorari. 538 U. S. 976 (2003).
As we explain below, the Miranda rule is a prophylactic employed .to protect against violations of the Self-Incrimination Clause. The Self-Incrimination Clause, however, is not implicated by the admission into evidence of the physical fruit of a- voluntary statement. Accordingly, there is no justification for extending the Miranda rule to this con*637text. And just as the Self-Incrimination Clause primarily focuses on the criminal trial, so too does the Miranda rule. The Miranda rule is not a code of police conduct, and police do not violate the Constitution (or even the Miranda rule, for that matter) by mere failures to warn. For this reason, the exclusionary rule articulated in cases such as Wong Sun does not apply. Accordingly, we reverse the judgment of the Court of Appeals and remand the case for further proceedings.
II
The Self-Incrimination Clause provides: “No person . . . shall be compelled in any criminal case to be a witness against himself.” U. S. Const., Amdt. 5. We need not decide here the precise boundaries of the Clause’s protection. For present purposes, it suffices to note that the core protection afforded by the Self-Incrimination Clause is a prohibition on compelling a criminal defendant to testify against himself at trial. See, e. g., Chavez v. Martinez, 538 U. S. 760, 764-768 (2003) (plurality opinion); id., at 777-779 (Souter, J., concurring in judgment); 8 J. Wigmore, Evidence § 2263, p. 378 (J. McNaughton rev. ed. 1961) (explaining that the Clause “was directed at the employment of legal process to extract from the person's own lips an admission of guilt, which would thus take the place of other evidence”); see also United States v. Hubbell, 530 U. S. 27, 49-56 (2000) (Thomas, J., concurring) (explaining that the privilege might extend to bar the compelled production of any incriminating evidence, testimonial or otherwise). The Clause cannot be violated by the introduction of nontestimonial evidence obtained as a result of voluntary statements. See, e. g., id., at 34 (noting that the word “ ‘witness’ ” in the Self-Incrimination Clause “limits the relevant category of compelled incriminating communications to those that are ‘testimonial’ in character”); id., at 35 (discussing why compelled blood samples do not violate the Clause; cataloging other examples and citing cases); Elstad, 470 U. S., at 304 (“The Fifth Amendment, of *638course, is not concerned with nontestimonial evidence”); id., at 306-307 (“The Fifth Amendment prohibits use by the prosecution in its case in chief only of compelled testimony”); Withrow v. Williams, 507 U. S. 680, 705 (1993) (O’CONNOR, J., concurring in part and dissenting in part) (describing “true Fifth Amendment claims [as] the extraction and use of compelled testimony”); New York v. Quarles, 467 U. S. 649, 665-672, and n. 4 (1984) (O’CONNOR, J., concurring in judgment in part and dissenting in part) (explaining that the physical fruit of a Miranda violation need not be suppressed for these reasons).
To be sure, the Court has recognized and applied several prophylactic rules designed to protect the core privilege against self-incrimination. See, e. g., Chavez, supra, at 770-772 (plurality opinion). For example, although the text of the Self-Incrimination Clause at least suggests that “its coverage [is limited to] compelled testimony that is used against the defendant in the trial itself,” Hubbell, supra, at 37, potential suspects may, at times, assert the privilege in proceedings in which answers might be used to incriminate them in a subsequent criminal case. See, e. g., United States v. Balsys, 524 U. S. 666, 671-672 (1998); Minnesota v. Murphy, 465 U. S. 420, 426 (1984); cf. Kastigar v. United States, 406 U. S. 441 (1972) (holding that the Government may compel grand jury testimony from witnesses over Fifth Amendment objections if the witnesses receive “use and derivative use immunity”); Uniformed Sanitation Men Assn., Inc. v. Commissioner of Sanitation of City of New York, 392 U. S. 280, 284 (1968) (allowing the Government to use economic compulsion to secure statements but only if the Government grants appropriate immunity). We have explained that “[t]he natural concern which underlies [these] decisions is that an inability to protect the right at one stage of a proceeding may make its invocation useless at a later stage.” Tucker, 417 U. S., at 440-441.
*639Similarly, in Miranda, the Court concluded that the possibility of coercion inherent in custodial interrogations unacceptably raises the risk that a suspect’s privilege against self-incrimination might be violated. See Dickerson, 530 U. S., at 434-435; Miranda, 384 U. S., at 467. To protect against this danger, the Miranda rule creates a presumption of coercion, in the absence of specific warnings, that is generally irrebuttable for purposes of the prosecution’s case in chief.
But because these prophylactic rules (including the Miranda rule) necessarily sweep beyond the actual protections of the Self-Incrimination Clause, see, e. g,, Withrow, supra, at 690-691; Elstad, supra, at 306, any further extension of these rules must be justified by its necessity for the protection of the actual right against compelled self-incrimination, Chavez, supra, at 778 (Souter, J., concurring in judgment) (requiring a “ ‘powerful showing’ ” before “expanding] . . . the privilege against compelled self-incrimination”). Indeed, at times the Court has declined to extend Miranda even where it has perceived a need to protect the privilege against self-incrimination. See, e. g., Quarles, supra, at 657 (concluding “that the need for answers to questions in a situation posing a threat to the public safety outweighs the need for the prophylactic rule protecting the Fifth Amendment’s privilege against self-incrimination”).
It is for these reasons that statements taken without Miranda warnings (though not actually compelled) can be used to impeach a defendant’s testimony at trial, see Elstad, supra, at 307-308; Harris v. New York, 401 U. S. 222 (1971), though the fruits of actually compelled testimony cannot, see New Jersey v. Portash, 440 U. S. 450, 458-459 (1979). More generally, the Miranda rule “does not require that the statements [taken without complying with the rule] and their fruits be discarded as inherently tainted,” Elstad, 470 U. S., at 307. Such a blanket suppression rule could not be justi*640fied by reference to the “Fifth Amendment goal of assuring trustworthy evidence” or by any deterrence rationale, id., at 308; see Tucker, supra, at 446-449; Harris, supra, at 225-226, and n. 2, and would therefore fail our close-fit requirement.
Furthermore, the Self-Incrimination Clause contains its own exclusionary rule. It provides that “[n]o person . . . shall be compelled in any criminal case to be a witness against himself.” Arndt. 5. Unlike the Fourth Amendment’s bar on unreasonable searches, the Self-Incrimination Clause is self-executing. We have repeatedly explained “that those subjected to coercive police interrogations have an automatic protection from the use of their involuntary statements (or evidence derived from their statements) in any subsequent criminal trial.” Chavez, 538 U. S., at 769 (plurality opinion) (citing, for example, Elstad, supra, at 307-308). This explicit textual protection supports a strong presumption against expanding the Miranda rule any further. Cf. Graham v. Connor, 490 U. S. 386 (1989).
Finally, nothing in Dickerson, including its characterization of Miranda as announcing a constitutional rule, 530 U. S., at 444, changes any of these observations. Indeed, in Dickerson, the Court specifically noted that the Court’s “subsequent cases have reduced the impact of the Miranda rule on legitimate law enforcement while reaffirming [Miranda]’s core ruling that unwarned statements may not be used as evidence in the prosecution’s case in chief.” Id., at 443-444. This description of Miranda, especially the emphasis on the use of “unwarned statements ... in the prosecution’s case in chief,” makes clear our continued focus on the protections of the Self-Incrimination Clause. The Court’s reliance on our Miranda precedents, including both Tucker and Elstad, see, e. g., Dickerson, supra, at 438, 441, further demonstrates the continuing validity of those decisions. In short, nothing in Dickerson calls into question our continued *641insistence that the closest possible fit be maintained between the Self-Incrimination Clause and any rule designed to protect it.
Ill
Our cases also make clear the related point that a mere failure to give Miranda warnings does not, by itself, violate a suspect’s constitutional rights or even the Miranda rule. So much was evident in many of our pre-Dickerson cases, and we have adhered to this view since Dickerson. See Chavez, 538 U. S., at 772-773 (plurality opinion) (holding that a failure to read Miranda warnings did not violate the respondent’s constitutional rights); 538 U. S., at 789 (Kennedy, J., concurring in part and dissenting in part) (agreeing “that failure to give a Miranda warning does not, without more, establish a completed violation when the unwarned interrogation ensues”); Elstad, supra, at 308; Quarles, 467 U. S., at 654; cf. Chavez, supra, at 777-779 (Souter, J., concurring in judgment). This, of course, follows from the nature of the right protected by the Self-Incrimination Clause, which the Miranda rule, in turn, protects. It is “ ‘a fundamental trial right.’” Withrow, 507 U.S., at 691 (quoting United States v. Verdugo-Urquidez, 494 U. S. 259, 264 (1990)). See also Chavez, 538 U. S., at 766-768 (plurality opinion); id., at 790 (Kennedy, J., concurring in part and dissenting in part) (“The identification of a Miranda violation and its consequences, then, ought to be determined at trial”).
It follows that police do not violate a suspect’s constitutional rights (or the Miranda rule) by negligent or even deliberate failures to provide the suspect with the full panoply of warnings prescribed by Miranda. Potential violations occur, if at all, only upon the admission of unwarned statements into evidence at trial. And, at that point, “[t]he exclusion of unwarned statements ... is a complete and sufficient *642remedy” for any perceived Miranda violation. Chavez, supra, at 790.3
Thus, unlike unreasonable searches under the Fourth Amendment or actual violations of the Due Process Clause or the Self-Incrimination Clause, there is, with respect to mere failures to warn, nothing to deter. There is therefore no reason to apply the “fruit of the poisonous tree” doctrine of Wong Sun, 371 U. S., at 488.4 See also Nix v. Williams, 467 U. S. 431, 441 (1984) (discussing the exclusionary rule in the Sixth Amendment context and noting that it applies to “illegally obtained evidence [and] other incriminating evidence derived from [it]” (emphasis added)). It is. not for this Court to impose its preferred police practices on either federal law enforcement officials or their state counterparts.
IV
In the present case, the Court of Appeals, relying on Dickerson, wholly adopted the position that the taking of unwarned statements violates a suspect’s constitutional rights. 304 F. 3d, at 1028-1029.5 And, of course, if this were so, a *643strong deterrence-based argument could be made for suppression of the fruits. See, e.g., Nix, supra, at 441-444; Wong Sun, supra, at 484-486; cf. Nardone v. United States, 308 U. S. 338, 341 (1939).
But Dickerson’s characterization of Miranda as a constitutional rule does not lessen the need to maintain the closest possible fit between the Self-Incrimination Clause and any judge-made rule designed to protect it. And there is no such fit here. Introduction of the nontestimonial fruit of a voluntary statement, such as respondent’s Glock, does not' implicate the Self-Incrimination Clause. The admission of such fruit presents no risk that a defendant’s coerced statements (however defined) will be used against him at a criminal trial. In any case, “[t]he exclusion of unwarned statements ... is a complete and sufficient remedy” for any perceived Miranda violation. Chavez, supra, at 790 (Kennedy, J., concurring in part and dissenting in part). See also H. Friendly, Benchmarks 280-281 (1967). There is simply no need to extend (and therefore no justification for extending) the prophylactic rule of Miranda to this context.
Similarly, because police cannot violate the Self-Incrimination Clause by taking unwarned though voluntary statements, an exclusionary rule cannot be justified by reference to a deterrence effect on law enforcement, as the Court of Appeals believed, 304 F. 3d, at 1028-1029. Our decision not to apply Wong Sun to mere failures to give Miranda warnings was sound at the time Tucker and Elstad were decided, and we decline to apply Wong Sun to such failures now.
The Court of Appeals ascribed significance to the fact that, in this case, there might be “little [practical] difference between [respondent’s] confessional statement” and the actual physical evidence. 304 F. 3d, at 1027. The distinction, the court said, “appears to make little sense as a matter of policy.” Ibid. But, putting policy aside, we have held that “[t]he word ‘witness’ in the constitutional text limits the” *644scope of the Self-Incrimination Clause to testimonial evidence. Hubbell, 530 U. S., at 34-35. The Constitution itself makes the distinction.6 And although it is true that the Court requires the exclusion of the physical fruit of actually coerced statements, it must be remembered that statements taken without sufficient Miranda warnings are presumed to have been coerced only for certain purposes and then only when necessary to protect the privilege against self-incrimination. See Part II, supra. For the reasons discussed above, we decline to extend that presumption further.7
Accordingly, we reverse the judgment of the Court of Ap: peals and remand the case for further-proceedings.
It is so ordered.
The Government concedes that respondent’s answers to subsequent on-the-scene questioning are inadmissible at trial under Miranda v. Arizona, 384 U. S. 436 (1966), despite the partial warning and respondent’s assertions that he knew his rights.
The Court of Appeals also distinguished Oregon v. Elstad, 470 U. S. 298 (1985), on the ground that the second (and warned) confession at issue there was the product of the defendant’s volition. 304 F. 3d, at 1019, 1021. For the reasons discussed below, we do not find this distinction relevant.
We acknowledge that there is language in some of the Court’s post-Miranda decisions that might suggest that the Miranda rule operates as a direct constraint on police. See, e. g., Stansbury v. California, 511 U. S. 318, 322 (1994) (per curiam); Moran v. Burbine, 475 U. S. 412, 420 (1986) (stating that “Miranda imposed on the police an obligation to follow certain procedures”); cf. Edwards v. Arizona, 451 U. S. 477, 485 (1981). But Miranda itself made clear that its focus was the admissibility of statements, see, e. g,, 384 U. S., at 439, 467, a view the Court reaffirmed in Dickerson v. United States, 530 U. S. 428, 443-444 (2000) (equating the Miranda rule with the proposition that “unwarned statements may not be used as evidence in the prosecution’s case in chief” (emphasis added)).
We reject respondent’s invitation to apply the balancing test of Nardone v. United States, 308 U. S. 338 (1939). Brief for Respondent 15-33. At issue in Nardone was the violation of a federal wiretap statute, and the Court employed an exclusionary rule to deter those violations. But, once again, there are no violations (statutory or constitutional) to deter here.
It is worth mentioning that the Court of Appeals did not have' the benefit of our decision in Chavez v. Martinez, 538 U. S. 760 (2003).
While Fourth Amendment protections extend to “persons, houses, papers, and effects,” the Self-Incrimination Clause prohibits only compelling a defendant to be “a witness against himself,” Arndt. 5.
It is not clear whether the Government could have used legal processes actually to compel respondent to produce the Glock, though there is a reasonable argument that it could have. See, e. g., United States v. Hubbell, 530 U. S. 27, 42-45 (2000); Baltimore City Dept. of Social Servs. v. Bouknight, 493 U. S. 549, 554-556 (1990); Fisher v. United States, 425 U. S. 391 (1976); Warden, Md. Penitentiary v. Hayden, 387 U. S. 294, 302-303 (1967); Schmerber v. California, 384 U. S. 757, 761 (1966). But see Commonwealth v. Hughes, 380 Mass. 583, 404 N. E. 2d 1239 (1980); Goldsmith v. Superior Court, 152 Cal. App. 3d 76, 199 Cal. Rptr. 366 (1984). In light-of this, it would be especially odd to exclude the Glock here.