HUDSON v. MICHIGAN

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
586 OCTOBER TERM, 2005
Syllabus
HUDSON v. MICHIGAN
certiorari to the court of appeals of michigan
No. 04–1360. Argued January 9, 2006—Reargued May 18, 2006—Decided
June 15, 2006
Detroit police executing a search warrant for narcotics and weapons en
tered petitioner Hudson’s home in violation of the Fourth Amendment’s
“knock-and-announce” rule. The trial court granted Hudson’s motion
to suppress the evidence seized, but the Michigan Court of Appeals re
versed on interlocutory appeal. Hudson was convicted of drug posses
sion. Affirming, the State Court of Appeals rejected Hudson’s renewed
Fourth Amendment claim.
Held: The judgment is affirmed.
Affirmed.
Justice Scalia delivered the opinion of the Court with respect to
Parts I, II, and III, concluding that violation of the “knock-and
announce” rule does not require suppression of evidence found in a
search. Pp. 589–599.
(a) Because Michigan has conceded that the entry here was a knock
and-announce violation, the only issue is whether the exclusionary rule
is appropriate for such a violation. Pp. 589–590.
(b) This Court has rejected “[i]ndiscriminate application” of the ex
clusionary rule, United States v. Leon, 468 U. S. 897, 908, holding it
applicable only “where its deterrence benefits outweigh its ‘substantial
social costs,’ ” Pennsylvania Bd. of Probation and Parole v. Scott, 524
U. S. 357, 363. Exclusion may not be premised on the mere fact that a
constitutional violation was a “but-for” cause of obtaining the evidence.
The illegal entry here was not the but-for cause, but even if it were,
but-for causation can be too attenuated to justify exclusion. Attenua
tion can occur not only when the causal connection is remote, but also
when suppression would not serve the interest protected by the consti
tutional guarantee violated. The interests protected by the knock
and-announce rule include human life and limb (because an unannounced
entry may provoke violence from a surprised resident), property (be
cause citizens presumably would open the door upon an announcement,
whereas a forcible entry may destroy it), and privacy and dignity of the
sort that can be offended by a sudden entrance. But the rule has never
protected one’s interest in preventing the government from seeing or
taking evidence described in a warrant. Since the interests violated

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
587 Cite as: 547 U. S. 586 (2006)
Syllabus
here have nothing to do with the seizure of the evidence, the exclusion
ary rule is inapplicable. Pp. 590–594.
(c) The social costs to be weighed against deterrence are considerable
here. In addition to the grave adverse consequence that excluding rele
vant incriminating evidence always entails—the risk of releasing dan
gerous criminals—imposing such a massive remedy would generate a
constant flood of alleged failures to observe the rule, and claims that
any asserted justification for a no-knock entry had inadequate support.
Another consequence would be police officers’ refraining from timely
entry after knocking and announcing, producing preventable violence
against the officers in some cases, and the destruction of evidence in
others. Next to these social costs are the deterrence benefits. The
value of deterrence depends on the strength of the incentive to commit
the forbidden act. That incentive is minimal here, where ignoring
knock-and-announce can realistically be expected to achieve nothing but
the prevention of evidence destruction and avoidance of life-threatening
resistance, dangers which suspend the requirement when there is “rea
sonable suspicion” that they exist, Richards v. Wisconsin, 520 U. S. 385,
394. Massive deterrence is hardly necessary. Contrary to Hudson’s
argument that without suppression there will be no deterrence, many
forms of police misconduct are deterred by civil-rights suits, and by the
consequences of increasing professionalism of police forces, including a
new emphasis on internal police discipline. Pp. 594–599.
Justice Scalia, joined by The Chief Justice, Justice Thomas,
and Justice Alito, concluded in Part IV that Segura v. United States,
468 U. S. 796, New York v. Harris, 495 U. S. 14, and United States v.
Ramirez, 523 U. S. 65, confirm the conclusion that suppression is unwar
ranted in this case. Pp. 599–602.
Scalia, J., delivered the opinion of the Court with respect to Parts I,
II, and III, in which Roberts, C. J., and Kennedy, Thomas, and Alito,
JJ., joined, and an opinion with respect to Part IV, in which Roberts,
C. J., and Thomas and Alito, JJ., joined. Kennedy, J., filed an opinion
concurring in part and concurring in the judgment, post, p. 602. Breyer,
J., filed a dissenting opinion, in which Stevens, Souter, and Ginsburg,
JJ., joined, post, p. 604.
David A. Moran argued and reargued the cause for peti
tioner. With him on the briefs were Timothy O’Toole, Ste
ven R. Shapiro, Michael J. Steinberg, Kary L. Moss, and
Richard D. Korn.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
588 HUDSON v. MICHIGAN
Opinion of the Court
Timothy A. Baughman argued and reargued the cause
and filed a brief for respondent.
David B. Salmons argued and reargued the cause for the
United States as amicus curiae urging affirmance. With
him on the brief were Solicitor General Clement, Assistant
Attorney General Fisher, Deputy Solicitor General Dreeben,
and Deborah Watson.*
Justice Scalia delivered the opinion of the Court, except
as to Part IV.
We decide whether violation of the “knock-and-announce”
rule requires the suppression of all evidence found in the
search.
I
Police obtained a warrant authorizing a search for drugs
and firearms at the home of petitioner Booker Hudson.
They discovered both. Large quantities of drugs were
found, including cocaine rocks in Hudson’s pocket. A loaded
gun was lodged between the cushion and armrest of the chair
in which he was sitting. Hudson was charged under Michi
gan law with unlawful drug and firearm possession.
This case is before us only because of the method of entry
into the house. When the police arrived to execute the war
rant, they announced their presence, but waited only a short
time—perhaps “three to five seconds,” App. 15—before turn
ing the knob of the unlocked front door and entering Hud
son’s home. Hudson moved to suppress all the inculpatory
evidence, arguing that the premature entry violated his
Fourth Amendment rights.
The Michigan trial court granted his motion. On interloc
utory review, the Michigan Court of Appeals reversed, re
*Tracey Maclin, Timothy Lynch, and Joshua L. Dratel filed a brief for
the Cato Institute et al. as amici curiae urging reversal.
Kent S. Scheidegger and Charles L. Hobson filed a brief for the Criminal
Justice Legal Foundation as amicus curiae urging affirmance.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
589 Cite as: 547 U. S. 586 (2006)
Opinion of the Court
lying on Michigan Supreme Court cases holding that sup
pression is inappropriate when entry is made pursuant to
warrant but without proper “ ‘knock and announce.’ ” App.
to Pet. for Cert. 4 (citing People v. Vasquez, 461 Mich. 235,
602 N. W. 2d 376 (1999) (per curiam); People v. Stevens, 460
Mich. 626, 597 N. W. 2d 53 (1999)). The Michigan Supreme
Court denied leave to appeal. 465 Mich. 932, 639 N. W. 2d
255 (2001). Hudson was convicted of drug possession. He
renewed his Fourth Amendment claim on appeal, but the
Court of Appeals rejected it and affirmed the conviction.
App. to Pet. for Cert. 1–2. The Michigan Supreme Court
again declined review. 472 Mich. 862, 692 N. W. 2d 385
(2005). We granted certiorari. 545 U. S. 1138 (2005).
II
The common-law principle that law enforcement officers
must announce their presence and provide residents an op
portunity to open the door is an ancient one. See Wilson v.
Arkansas, 514 U. S. 927, 931–932 (1995). Since 1917, when
Congress passed the Espionage Act, this traditional protec
tion has been part of federal statutory law, see 40 Stat. 229,
and is currently codified at 18 U. S. C. § 3109. We applied
that statute in Miller v. United States, 357 U. S. 301 (1958),
and again in Sabbath v. United States, 391 U. S. 585 (1968).
Finally, in Wilson, we were asked whether the rule was also
a command of the Fourth Amendment. Tracing its origins
in our English legal heritage, 514 U. S., at 931–936, we con
cluded that it was.
We recognized that the new constitutional rule we had an
nounced is not easily applied. Wilson and cases following it
have noted the many situations in which it is not necessary
to knock and announce. It is not necessary when “circum
stances presen[t] a threat of physical violence,” or if there is
“reason to believe that evidence would likely be destroyed if
advance notice were given,” id., at 936, or if knocking and

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
590 HUDSON v. MICHIGAN
Opinion of the Court
announcing would be “futile,” Richards v. Wisconsin, 520
U. S. 385, 394 (1997). We require only that police “have a
reasonable suspicion . . . under the particular circumstances”
that one of these grounds for failing to knock and announce
exists, and we have acknowledged that “[t]his showing is not
high.” Ibid.
When the knock-and-announce rule does apply, it is not
easy to determine precisely what officers must do. How
many seconds’ wait are too few? Our “reasonable wait
time” standard, see United States v. Banks, 540 U. S. 31, 41
(2003), is necessarily vague. Banks (a drug case, like this
one) held that the proper measure was not how long it would
take the resident to reach the door, but how long it would
take to dispose of the suspected drugs—but that such a time
(15 to 20 seconds in that case) would necessarily be extended
when, for instance, the suspected contraband was not easily
concealed. Id., at 40–41. If our ex post evaluation is sub
ject to such calculations, it is unsurprising that, ex ante,
police officers about to encounter someone who may try to
harm them will be uncertain how long to wait.
Happily, these issues do not confront us here. From the
trial level onward, Michigan has conceded that the entry was
a knock-and-announce violation. The issue here is remedy.
Wilson specifically declined to decide whether the exclu
sionary rule is appropriate for violation of the knock-and
announce requirement. 514 U. S., at 937, n. 4. That ques
tion is squarely before us now.
III
A
In Weeks v. United States, 232 U. S. 383 (1914), we adopted
the federal exclusionary rule for evidence that was unlaw
fully seized from a home without a warrant in violation of
the Fourth Amendment. We began applying the same rule
to the States, through the Fourteenth Amendment, in Mapp
v. Ohio, 367 U. S. 643 (1961).

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
591 Cite as: 547 U. S. 586 (2006)
Opinion of the Court
Suppression of evidence, however, has always been our
last resort, not our first impulse. The exclusionary rule gen
erates “substantial social costs,” United States v. Leon, 468
U. S. 897, 907 (1984), which sometimes include setting the
guilty free and the dangerous at large. We have therefore
been “cautio[us] against expanding” it, Colorado v. Connelly,
479 U. S. 157, 166 (1986), and “have repeatedly emphasized
that the rule’s ‘costly toll’ upon truth-seeking and law en
forcement objectives presents a high obstacle for those urg
ing [its] application,” Pennsylvania Bd. of Probation and
Parole v. Scott, 524 U. S. 357, 364–365 (1998). We have re
jected “[i]ndiscriminate application” of the rule, Leon, supra,
at 908, and have held it to be applicable only “where its re
medial objectives are thought most efficaciously served,”
United States v. Calandra, 414 U. S. 338, 348 (1974)—that is,
“where its deterrence benefits outweigh its ‘substantial so
cial costs,’ ” Scott, supra, at 363 (quoting Leon, supra, at
907).
We did not always speak so guardedly. Expansive dicta
in Mapp, for example, suggested wide scope for the exclu
sionary rule. See, e. g., 367 U. S., at 655 (“[A]ll evidence ob
tained by searches and seizures in violation of the Consti
tution is, by that same authority, inadmissible in a state
court”). Whiteley v. Warden, Wyo. State Penitentiary, 401
U. S. 560, 568–569 (1971), was to the same effect. But we
have long since rejected that approach. As explained in Ar
izona v. Evans, 514 U. S. 1, 13 (1995): “In Whiteley, the
Court treated identification of a Fourth Amendment viola
tion as synonymous with application of the exclusionary rule
to evidence secured incident to that violation. Subsequent
case law has rejected this reflexive application of the exclu
sionary rule.” (Citation omitted.) We had said as much in
Leon, a decade earlier, when we explained that “[w]hether
the exclusionary sanction is appropriately imposed in a par
ticular case . . . is ‘an issue separate from the question
whether the Fourth Amendment rights of the party seeking

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
592 HUDSON v. MICHIGAN
Opinion of the Court
to invoke the rule were violated by police conduct.’ ” 468
U. S., at 906 (quoting Illinois v. Gates, 462 U. S. 213, 223
(1983)).
In other words, exclusion may not be premised on the
mere fact that a constitutional violation was a “but-for”
cause of obtaining evidence. Our cases show that but-for
causality is only a necessary, not a sufficient, condition for
suppression. In this case, of course, the constitutional viola
tion of an illegal manner of entry was not a but-for cause of
obtaining the evidence. Whether that preliminary misstep
had occurred or not, the police would have executed the war
rant they had obtained, and would have discovered the gun
and drugs inside the house. But even if the illegal entry
here could be characterized as a but-for cause of discovering
what was inside, we have “never held that evidence is ‘fruit
of the poisonous tree’ simply because ‘it would not have come
to light but for the illegal actions of the police.’ ” Segura v.
United States, 468 U. S. 796, 815 (1984). See also id., at 829
(Stevens, J., dissenting) (“We have not . . . mechanically
applied the [exclusionary] rule to every item of evidence that
has a causal connection with police misconduct”). Rather,
but-for cause, or “causation in the logical sense alone,”
United States v. Ceccolini, 435 U. S. 268, 274 (1978), can be
too attenuated to justify exclusion, id., at 274–275. Even in
the early days of the exclusionary rule, we declined to
“hold that all evidence is ‘fruit of the poisonous tree’
simply because it would not have come to light but for
the illegal actions of the police. Rather, the more apt
question in such a case is ‘whether, granting establish
ment of the primary illegality, the evidence to which in
stant objection is made has been come at by exploitation
of that illegality or instead by means sufficiently distin
guishable to be purged of the primary taint.’ ” Wong
Sun v. United States, 371 U. S. 471, 487–488 (1963) (quot
ing J. Maguire, Evidence of Guilt 221 (1959); emphasis
added).

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
593 Cite as: 547 U. S. 586 (2006)
Opinion of the Court
Attenuation can occur, of course, when the causal connec
tion is remote. See, e. g., Nardone v. United States, 308
U. S. 338, 341 (1939). Attenuation also occurs when, even
given a direct causal connection, the interest protected by
the constitutional guarantee that has been violated would
not be served by suppression of the evidence obtained.
“The penalties visited upon the Government, and in turn
upon the public, because its officers have violated the law
must bear some relation to the purposes which the law is
to serve.” Ceccolini, supra, at 279. Thus, in New York v.
Harris, 495 U. S. 14 (1990), where an illegal warrantless ar
rest was made in Harris’s house, we held:
“[S]uppressing [Harris’s] statement taken outside the
house would not serve the purpose of the rule that made
Harris’ in-house arrest illegal. The warrant require
ment for an arrest in the home is imposed to protect the
home, and anything incriminating the police gathered
from arresting Harris in his home, rather than else
where, has been excluded, as it should have been; the
purpose of the rule has thereby been vindicated.” Id.,
at 20.
For this reason, cases excluding the fruits of unlawful war
rantless searches, see, e. g., Boyd v. United States, 116 U. S.
616 (1886); Weeks, 232 U. S. 383; Silverthorne Lumber Co. v.
United States, 251 U. S. 385 (1920); Mapp, supra, say nothing
about the appropriateness of exclusion to vindicate the in
terests protected by the knock-and-announce requirement.
Until a valid warrant has issued, citizens are entitled to
shield “their persons, houses, papers, and effects,” U. S.
Const., Amdt. 4, from the government’s scrutiny. Exclusion
of the evidence obtained by a warrantless search vindicates
that entitlement. The interests protected by the knock
and-announce requirement are quite different—and do not
include the shielding of potential evidence from the govern
ment’s eyes.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
594 HUDSON v. MICHIGAN
Opinion of the Court
One of those interests is the protection of human life and
limb, because an unannounced entry may provoke violence in
supposed self-defense by the surprised resident. See, e. g.,
McDonald v. United States, 335 U. S. 451, 460–461 (1948)
(Jackson, J., concurring). See also Sabbath, 391 U. S., at 589;
Miller, 357 U. S., at 313, n. 12. Another interest is the pro
tection of property. Breaking a house (as the old cases typi
cally put it) absent an announcement would penalize some
one who “ ‘did not know of the process, of which, if he had
notice, it is to be presumed that he would obey it . . . .’ ”
Wilson, 514 U. S., at 931–932 (quoting Semayne’s Case, 5 Co.
Rep. 91a, 91b, 77 Eng. Rep. 194, 195–196 (K. B. 1603)). The
knock-and-announce rule gives individuals “the opportunity
to comply with the law and to avoid the destruction of prop
erty occasioned by a forcible entry.” Richards, 520 U. S., at
393, n. 5. See also Banks, 540 U. S., at 41. And thirdly, the
knock-and-announce rule protects those elements of privacy
and dignity that can be destroyed by a sudden entrance. It
gives residents the “opportunity to prepare themselves for”
the entry of the police. Richards, 520 U. S., at 393, n. 5.
“The brief interlude between announcement and entry with
a warrant may be the opportunity that an individual has to
pull on clothes or get out of bed.” Ibid. In other words, it
assures the opportunity to collect oneself before answering
the door.
What the knock-and-announce rule has never protected,
however, is one’s interest in preventing the government from
seeing or taking evidence described in a warrant. Since the
interests that were violated in this case have nothing to do
with the seizure of the evidence, the exclusionary rule is
inapplicable.
B
Quite apart from the requirement of unattenuated causa
tion, the exclusionary rule has never been applied except
“where its deterrence benefits outweigh its ‘substantial so
cial costs,’ ” Scott, 524 U. S., at 363 (quoting Leon, 468 U. S.,

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
595 Cite as: 547 U. S. 586 (2006)
Opinion of the Court
at 907). The costs here are considerable. In addition to the
grave adverse consequence that exclusion of relevant incrim
inating evidence always entails (viz., the risk of releasing
dangerous criminals into society), imposing that massive
remedy for a knock-and-announce violation would gener
ate a constant flood of alleged failures to observe the rule,
and claims that any asserted Richards justification for a no
knock entry, see 520 U. S., at 394, had inadequate support.
Cf. United States v. Singleton, 441 F. 3d 290, 293–294 (CA4
2006). The cost of entering this lottery would be small, but
the jackpot enormous: suppression of all evidence, amounting
in many cases to a get-out-of-jail-free card. Courts would
experience as never before the reality that “[t]he exclusion
ary rule frequently requires extensive litigation to deter
mine whether particular evidence must be excluded.” Scott,
supra, at 366. Unlike the warrant or Miranda require
ments, compliance with which is readily determined (either
there was or was not a warrant; either the Miranda warning
was given, or it was not), what constituted a “reasonable wait
time” in a particular case, Banks, supra, at 41 (or, for that
matter, how many seconds the police in fact waited), or
whether there was “reasonable suspicion” of the sort that
would invoke the Richards exceptions, is difficult for the
trial court to determine and even more difficult for an appel
late court to review.
Another consequence of the incongruent remedy Hudson
proposes would be police officers’ refraining from timely
entry after knocking and announcing. As we have ob
served, see supra, at 590, the amount of time they must wait
is necessarily uncertain. If the consequences of running
afoul of the rule were so massive, officers would be inclined
to wait longer than the law requires—producing preventable
violence against officers in some cases, and the destruction
of evidence in many others. See Gates, 462 U. S., at 258
(White, J., concurring in judgment). We deemed these con
sequences severe enough to produce our unanimous agree

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
596 HUDSON v. MICHIGAN
Opinion of the Court
ment that a mere “reasonable suspicion” that knocking and
announcing “under the particular circumstances, would be
dangerous or futile, or that it would inhibit the effective in
vestigation of the crime,” will cause the requirement to
yield. Richards, supra, at 394.
Next to these “substantial social costs” we must consider
the deterrence benefits, existence of which is a necessary
condition for exclusion. (It is not, of course, a sufficient con
dition: “[I]t does not follow that the Fourth Amendment re
quires adoption of every proposal that might deter police
misconduct.” Calandra, 414 U. S., at 350; see also Leon,
supra, at 910.) To begin with, the value of deterrence de
pends upon the strength of the incentive to commit the for
bidden act. Viewed from this perspective, deterrence of
knock-and-announce violations is not worth a lot. Violation
of the warrant requirement sometimes produces incriminat
ing evidence that could not otherwise be obtained. But ig
noring knock-and-announce can realistically be expected to
achieve absolutely nothing except the prevention of destruc
tion of evidence and the avoidance of life-threatening resist
ance by occupants of the premises—dangers which, if there
is even “reasonable suspicion” of their existence, suspend the
knock-and-announce requirement anyway. Massive deter
rence is hardly required.
It seems to us not even true, as Hudson contends, that
without suppression there will be no deterrence of knock
and-announce violations at all. Of course even if this asser
tion were accurate, it would not necessarily justify suppres
sion. Assuming (as the assertion must) that civil suit is not
an effective deterrent, one can think of many forms of police
misconduct that are similarly “undeterred.” When, for ex
ample, a confessed suspect in the killing of a police officer,
arrested (along with incriminating evidence) in a lawful war
ranted search, is subjected to physical abuse at the station
house, would it seriously be suggested that the evidence
must be excluded, since that is the only “effective deter

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
597 Cite as: 547 U. S. 586 (2006)
Opinion of the Court
rent”? And what, other than civil suit, is the “effective de
terrent” of police violation of an already-confessed suspect’s
Sixth Amendment rights by denying him prompt access to
counsel? Many would regard these violated rights as more
significant than the right not to be intruded upon in one’s
nightclothes—and yet nothing but “ineffective” civil suit is
available as a deterrent. And the police incentive for those
violations is arguably greater than the incentive for disre
garding the knock-and-announce rule.
We cannot assume that exclusion in this context is neces
sary deterrence simply because we found that it was neces
sary deterrence in different contexts and long ago. That
would be forcing the public today to pay for the sins and
inadequacies of a legal regime that existed almost half a cen
tury ago. Dollree Mapp could not turn to Rev. Stat. § 1979,
42 U. S. C. § 1983, for meaningful relief; Monroe v. Pape, 365
U. S. 167 (1961), which began the slow but steady expansion
of that remedy, was decided the same Term as Mapp. It
would be another 17 years before the § 1983 remedy was ex
tended to reach the deep pocket of municipalities, Monell v.
New York City Dept. of Social Servs., 436 U. S. 658 (1978).
Citizens whose Fourth Amendment rights were violated by
federal officers could not bring suit until 10 years after
Mapp, with this Court’s decision in Bivens v. Six Unknown
Fed. Narcotics Agents, 403 U. S. 388 (1971).
Hudson complains that “it would be very hard to find a
lawyer to take a case such as this,” Tr. of Oral Arg. 7, but 42
U. S. C. § 1988(b) answers this objection. Since some civil
rights violations would yield damages too small to justify the
expense of litigation, Congress has authorized attorney’s fees
for civil-rights plaintiffs. This remedy was unavailable in
the heydays of our exclusionary-rule jurisprudence, because
it is tied to the availability of a cause of action. For years
after Mapp, “very few lawyers would even consider repre
sentation of persons who had civil rights claims against the
police,” but now “much has changed. Citizens and lawyers

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
598 HUDSON v. MICHIGAN
Opinion of the Court
are much more willing to seek relief in the courts for police
misconduct.” M. Avery, D. Rudovsky, & K. Blum, Police
Misconduct: Law and Litigation, p. v (3d ed. 2005); see gener
ally N. Aron, Liberty and Justice for All: Public Interest
Law in the 1980s and Beyond (1989) (describing the growth
of public-interest law). The number of public-interest law
firms and lawyers who specialize in civil-rights grievances
has greatly expanded.
Hudson points out that few published decisions to date
announce huge awards for knock-and-announce violations.
But this is an unhelpful statistic. Even if we thought that
only large damages would deter police misconduct (and that
police somehow are deterred by “damages” but indifferent
to the prospect of large § 1988 attorney’s fees), we do not
know how many claims have been settled, or indeed how
many violations have occurred that produced anything more
than nominal injury. It is clear, at least, that the lower
courts are allowing colorable knock-and-announce suits to go
forward, unimpeded by assertions of qualified immunity.
See, e. g., Green v. Butler, 420 F. 3d 689, 700–701 (CA7 2005)
(denying qualified immunity in a knock-and-announce civil
suit); Holland ex rel. Overdorff v. Harrington, 268 F. 3d
1179, 1193–1196 (CA10 2001) (same); Mena v. Simi Valley,
226 F. 3d 1031, 1041–1042 (CA9 2000) (same); Gould v. Davis,
165 F. 3d 265, 270–271 (CA4 1998) (same). As far as we
know, civil liability is an effective deterrent here, as we have
assumed it is in other contexts. See, e. g., Correctional
Services Corp. v. Malesko, 534 U. S. 61, 70 (2001) (“[T]he
threat of litigation and liability will adequately deter federal
officers for Bivens purposes no matter that they may enjoy
qualified immunity” (as violators of knock-and-announce do
not)); see also Nix v. Williams, 467 U. S. 431, 446 (1984).
Another development over the past half-century that de
ters civil-rights violations is the increasing professionalism
of police forces, including a new emphasis on internal police
discipline. Even as long ago as 1980 we felt it proper to

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
599 Cite as: 547 U. S. 586 (2006)
Opinion of Scalia, J.
“assume” that unlawful police behavior would “be dealt with
appropriately” by the authorities, United States v. Payner,
447 U. S. 727, 733–734, n. 5 (1980), but we now have increas
ing evidence that police forces across the United States take
the constitutional rights of citizens seriously. There have
been “wide-ranging reforms in the education, training, and
supervision of police officers.” S. Walker, Taming the Sys
tem: The Control of Discretion in Criminal Justice 1950–
1990, p. 51 (1993). Numerous sources are now available to
teach officers and their supervisors what is required of them
under this Court’s cases, how to respect constitutional guar
antees in various situations, and how to craft an effective
regime for internal discipline. See, e. g., D. Waksman & D.
Goodman, The Search and Seizure Handbook (2d ed. 2006);
A. Stone & S. DeLuca, Police Administration: An Introduc
tion (2d ed. 1994); E. Thibault, L. Lynch, & R. McBride, Pro
active Police Management (4th ed. 1998). Failure to teach
and enforce constitutional requirements exposes municipali
ties to financial liability. See Canton v. Harris, 489 U. S.
378, 388 (1989). Moreover, modern police forces are staffed
with professionals; it is not credible to assert that internal
discipline, which can limit successful careers, will not have a
deterrent effect. There is also evidence that the increasing
use of various forms of citizen review can enhance police
accountability.
In sum, the social costs of applying the exclusionary rule
to knock-and-announce violations are considerable; the incen
tive to such violations is minimal to begin with, and the ex
tant deterrences against them are substantial—incompara
bly greater than the factors deterring warrantless entries
when Mapp was decided. Resort to the massive remedy of
suppressing evidence of guilt is unjustified.
IV
A trio of cases—Segura v. United States, 468 U. S. 796
(1984); New York v. Harris, 495 U. S. 14 (1990); and United

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
600 HUDSON v. MICHIGAN
Opinion of Scalia, J.
States v. Ramirez, 523 U. S. 65 (1998)—confirms our conclu
sion that suppression is unwarranted in this case.
Like today’s case, Segura involved a concededly illegal
entry. Police conducting a drug crime investigation waited
for Segura outside an apartment building; when he arrived,
he denied living there. The police arrested him and brought
him to the apartment where they suspected illegal activity.
An officer knocked. When someone inside opened the door,
the police entered, taking Segura with them. They had nei
ther a warrant nor consent to enter, and they did not an
nounce themselves as police—an entry as illegal as can be.
Officers then stayed in the apartment for 19 hours awaiting
a search warrant. 468 U. S., at 800–801; id., at 818–819
(Stevens, J., dissenting). Once alerted that the search war
rant had been obtained, the police—still inside, having se
cured the premises so that no evidence could be removed—
conducted a search. Id., at 801. We refused to exclude the
resulting evidence. We recognized that only the evidence
gained from the particular violation could be excluded, see
id., at 799, 804–805, and therefore distinguished the effects
of the illegal entry from the effects of the legal search: “None
of the information on which the warrant was secured was
derived from or related in any way to the initial entry into
petitioners’ apartment . . . ,” id., at 814. It was therefore
“beyond dispute that the information possessed by the
agents before they entered the apartment constituted an in
dependent source for the discovery and seizure of the evi
dence now challenged.” Ibid.
If the search in Segura could be “wholly unrelated to the
prior entry,” ibid., when the only entry was warrantless, it
would be bizarre to treat more harshly the actions in this
case, where the only entry was with a warrant. If the prob
able cause backing a warrant that was issued later in time
could be an “independent source” for a search that proceeded
after the officers illegally entered and waited, a search war

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
601 Cite as: 547 U. S. 586 (2006)
Opinion of Scalia, J.
rant obtained before going in must have at least this much
effect.1
In the second case, Harris, the police violated the defend
ant’s Fourth Amendment rights by arresting him at home
without a warrant, contrary to Payton v. New York, 445 U. S.
573 (1980). Once taken to the station house, he gave an in
criminating statement. See 495 U. S., at 15–16. We re
fused to exclude it. Like the illegal entry which led to dis
covery of the evidence in today’s case, the illegal arrest in
Harris began a process that culminated in acquisition of the
evidence sought to be excluded. While Harris’s statement
was “the product of an arrest and being in custody,” it “was
not the fruit of the fact that the arrest was made in the house
rather than someplace else.” Id., at 20. Likewise here:
While acquisition of the gun and drugs was the product of
a search pursuant to warrant, it was not the fruit of the
fact that the entry was not preceded by knock-and-announce.2
1 Justice Breyer’s insistence that the warrant in Segura was “obtained
independently without use of any information found during the illegal
entry,” post, at 617 (dissenting opinion), entirely fails to distinguish it from
the warrant in the present case. Similarly inapposite is his appeal to
Justice Frankfurter’s statement in Wolf v. Colorado, 338 U. S. 25, 28
(1949), that the “ ‘knock at the door, . . . as a prelude to a search, without
authority of law . . . [is] inconsistent with the conception of human rights
enshrined in [our] history,’ ” see post, at 620. “How much the more offen
sive,” Justice Breyer asserts, “when the search takes place without any
knock at all,” ibid. But a no-knock entry “without authority of law” (i. e.,
without a search warrant) describes not this case, but Segura—where the
evidence was admitted anyway.
Justice Breyer’s assertion that Segura, unlike our decision in the
present case, had no effect on deterrence, see post, at 625–626, does not
comport with the views of the Segura dissent. See, e. g., 468 U. S., at 817
(Stevens, J., dissenting) (“The Court’s disposition, I fear, will provide
government agents with an affirmative incentive to engage in unconstitu
tional violations of the privacy of the home”).
2 Harris undermines two key points of the dissent. First, the claim
that “whether the interests underlying the knock-and-announce rule are
implicated in any given case is, in a sense, beside the point,” post, at

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
602 HUDSON v. MICHIGAN
Opinion of Kennedy, J.
United States v. Ramirez, supra, involved a claim that po
lice entry violated the Fourth Amendment because it was
effected by breaking a window. We ultimately concluded
that the property destruction was, under all the circum
stances, reasonable, but in the course of our discussion we
unanimously said the following: “[D]estruction of property in
the course of a search may violate the Fourth Amendment,
even though the entry itself is lawful and the fruits of the
search are not subject to suppression.” Id., at 71. Had the
breaking of the window been unreasonable, the Court said,
it would have been necessary to determine whether there
had been a “sufficient causal relationship between the break
ing of the window and the discovery of the guns to warrant
suppression of the evidence.” Id., at 72, n. 3. What clearer
expression could there be of the proposition that an imper
missible manner of entry does not necessarily trigger the
exclusionary rule?
* * *
For the foregoing reasons we affirm the judgment of the
Michigan Court of Appeals.
It is so ordered.
Justice Kennedy, concurring in part and concurring in
the judgment.
Two points should be underscored with respect to today’s
decision. First, the knock-and-announce requirement pro
tects rights and expectations linked to ancient principles in
our constitutional order. See Wilson v. Arkansas, 514 U. S.
927, 934 (1995). The Court’s decision should not be inter
preted as suggesting that violations of the requirement are
621. This is flatly refuted by Harris’s plain statement that the reason for
a rule must govern the sanctions for the rule’s violation. 495 U. S., at 17,
20; see also supra, at 593. Second, the dissent’s attempt to turn Harris
into a vindication of the sanctity of the home, see post, at 626–628. The
whole point of the case was that a confession that police obtained by
illegally removing a man from the sanctity of his home was admissible
against him.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
603 Cite as: 547 U. S. 586 (2006)
Opinion of Kennedy, J.
trivial or beyond the law’s concern. Second, the continued
operation of the exclusionary rule, as settled and defined by
our precedents, is not in doubt. Today’s decision determines
only that in the specific context of the knock-and-announce
requirement, a violation is not sufficiently related to the later
discovery of evidence to justify suppression.
As to the basic right in question, privacy and security in
the home are central to the Fourth Amendment’s guaran
tees as explained in our decisions and as understood since
the beginnings of the Republic. This common understand
ing ensures respect for the law and allegiance to our institu
tions, and it is an instrument for transmitting our Constitu
tion to later generations undiminished in meaning and force.
It bears repeating that it is a serious matter if law enforce
ment officers violate the sanctity of the home by ignoring
the requisites of lawful entry. Security must not be subject
to erosion by indifference or contempt.
Our system, as the Court explains, has developed proce
dures for training police officers and imposing discipline for
failures to act competently and lawfully. If those measures
prove ineffective, they can be fortified with more detailed
regulations or legislation. Supplementing these safeguards
are civil remedies, such as those available under Rev. Stat.
§ 1979, 42 U. S. C. § 1983, that provide restitution for discrete
harms. These remedies apply to all violations, including, of
course, exceptional cases in which unannounced entries cause
severe fright and humiliation.
Suppression is another matter. Under our precedents the
causal link between a violation of the knock-and-announce
requirement and a later search is too attenuated to allow
suppression. Cf. United States v. Ramirez, 523 U. S. 65, 72,
n. 3 (1998) (application of the exclusionary rule depends on
the existence of a “sufficient causal relationship” between the
unlawful conduct and the discovery of evidence). When, for
example, a violation results from want of a 20-second pause
but an ensuing, lawful search lasting five hours discloses evi

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
604 HUDSON v. MICHIGAN
Breyer, J., dissenting
dence of criminality, the failure to wait at the door cannot
properly be described as having caused the discovery of
evidence.
Today’s decision does not address any demonstrated pat
tern of knock-and-announce violations. If a widespread pat
tern of violations were shown, and particularly if those viola
tions were committed against persons who lacked the means
or voice to mount an effective protest, there would be reason
for grave concern. Even then, however, the Court would
have to acknowledge that extending the remedy of exclusion
to all the evidence seized following a knock-and-announce vi
olation would mean revising the requirement of causation
that limits our discretion in applying the exclusionary rule.
That type of extension also would have significant practical
implications, adding to the list of issues requiring resolution
at the criminal trial questions such as whether police officers
entered a home after waiting 10 seconds or 20.
In this case the relevant evidence was discovered not be
cause of a failure to knock and announce, but because of a
subsequent search pursuant to a lawful warrant. The Court
in my view is correct to hold that suppression was not re
quired. While I am not convinced that Segura v. United
States, 468 U. S. 796 (1984), and New York v. Harris, 495 U. S.
14 (1990), have as much relevance here as Justice Scalia
appears to conclude, the Court’s holding is fully supported
by Parts I through III of its opinion. I accordingly join
those Parts and concur in the judgment.
Justice Breyer, with whom Justice Stevens, Justice
Souter, and Justice Ginsburg join, dissenting.
In Wilson v. Arkansas, 514 U. S. 927 (1995), a unanimous
Court held that the Fourth Amendment normally requires
law enforcement officers to knock and announce their pres
ence before entering a dwelling. Today’s opinion holds that
evidence seized from a home following a violation of this re
quirement need not be suppressed.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
Cite as: 547 U. S. 586 (2006) 605
Breyer, J., dissenting
As a result, the Court destroys the strongest legal incen
tive to comply with the Constitution’s knock-and-announce
requirement. And the Court does so without significant
support in precedent. At least I can find no such support in
the many Fourth Amendment cases the Court has decided in
the near century since it first set forth the exclusionary prin
ciple in Weeks v. United States, 232 U. S. 383 (1914). See
Appendix, infra.
Today’s opinion is thus doubly troubling. It represents a
significant departure from the Court’s precedents. And it
weakens, perhaps destroys, much of the practical value of
the Constitution’s knock-and-announce protection.
I
This Court has set forth the legal principles that ought to
have determined the outcome of this case in two sets of basic
Fourth Amendment cases. I shall begin by describing that
underlying case law.
A
The first set of cases describes the constitutional knock
and-announce requirement, a requirement that this Court
initially set forth only 11 years ago in Wilson, supra.
Cf. Sabbath v. United States, 391 U. S. 585 (1968) (suppress
ing evidence seized in violation of federal statutory knock
and-announce requirement); Miller v. United States, 357
U. S. 301 (1958) (same). In Wilson, tracing the lineage of
the knock-and-announce rule back to the 13th century, 514
U. S., at 932, and n. 2, we wrote that
“[a]n examination of the common law of search and sei
zure leaves no doubt that the reasonableness of a search
of a dwelling may depend in part on whether law en
forcement officers announced their presence and author
ity prior to entering.” Id., at 931.
We noted that this “basic principle” was agreed upon by
“[s]everal prominent founding-era commentators,” id., at 932,

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
606 HUDSON v. MICHIGAN
Breyer, J., dissenting
and “was woven quickly into the fabric of early American
law” via state constitutions and statutes, id., at 933. We
further concluded that there was
“little doubt that the Framers of the Fourth Amendment
thought that the method of an officer’s entry into a
dwelling was among the factors to be considered in
assessing the reasonableness of a search or seizure.”
Id., at 934.
And we held that the “common-law ‘knock and announce’
principle forms a part of the reasonableness inquiry under
the Fourth Amendment.” Id., at 929. Thus, “a search or
seizure of a dwelling might be constitutionally defective if
police officers enter without prior announcement.” Id., at
936; see United States v. Banks, 540 U. S. 31, 36 (2003);
United States v. Ramirez, 523 U. S. 65, 70 (1998); Richards
v. Wisconsin, 520 U. S. 385, 387 (1997).
B
The second set of cases sets forth certain well-established
principles that are relevant here. They include:
Boyd v. United States, 116 U. S. 616 (1886). In this semi
nal Fourth Amendment case, decided 120 years ago, the
Court wrote, in frequently quoted language, that the Fourth
Amendment’s prohibitions apply
“to all invasions on the part of the government and its
employe´ s of the sanctity of a man’s home and the priva
cies of life. It is not the breaking of his doors, and the
rummaging of his drawers, that constitutes the essence
of the offence; but it is the invasion of his indefeasible
right of personal security, personal liberty and private
property.” Id., at 630.
Weeks, supra. This case, decided 28 years after Boyd,
originated the exclusionary rule. The Court held that the
Federal Government could not retain evidence seized uncon

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
607 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
stitutionally and use that evidence in a federal criminal trial.
The Court pointed out that “[i]f letters and private docu
ments” could be unlawfully seized from a home “and used
in evidence against a citizen accused of an offense, the pro
tection of the Fourth Amendment declaring his right to
be secure against such searches and seizures is of no value,
and . . . might as well be stricken from the Constitution.”
232 U. S., at 393.
Silverthorne Lumber Co. v. United States, 251 U. S. 385
(1920). This case created an exception to (or a qualification
of) Weeks’ exclusionary rule. The Court held that the Gov
ernment could not use information obtained during an illegal
search to subpoena documents that they illegally viewed dur
ing that search. Writing for the Court, Justice Holmes
noted that the exclusionary rule “does not mean that the
facts [unlawfully] obtained become sacred and inaccessible.
If knowledge of them is gained from an independent source
they may be proved like any others . . . .” 251 U. S., at
392. Silverthorne thus stands for the proposition that the
exclusionary rule does not apply if the evidence in question
(or the “fruits” of that evidence) was obtained through a
process unconnected with, and untainted by, the illegal
search. Cf. Nix v. Williams, 467 U. S. 431, 444 (1984) (de
scribing related “inevitable discovery” exception).
Wolf v. Colorado, 338 U. S. 25 (1949), and Mapp v. Ohio,
367 U. S. 643 (1961). Both of these cases considered
whether Weeks’ exclusionary rule applies to the States. In
Wolf, the Court held that it did not. It said that “[t]he secu
rity of one’s privacy against arbitrary intrusion by the police
. . . is . . . implicit in ‘the concept of ordered liberty’ and as
such enforceable against the States through the Due Process
Clause.” 338 U. S., at 27–28. But the Court held that the
exclusionary rule is not enforceable against the States as “an
essential ingredient of the right.” Id., at 29. In Mapp, the
Court overruled Wolf. Experience, it said, showed that al
ternative methods of enforcing the Fourth Amendment’s re

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
608 HUDSON v. MICHIGAN
Breyer, J., dissenting
quirements had failed. See 367 U. S., at 651–653; see, e. g.,
People v. Cahan, 44 Cal. 2d 434, 447, 282 P. 2d 905, 913 (1955)
(majority opinion of Traynor, J.) (“Experience [in California]
has demonstrated, however, that neither administrative,
criminal nor civil remedies are effective in suppressing law
less searches and seizures”). The Court consequently held
that “all evidence obtained by searches and seizures in viola
tion of the Constitution is, by that same authority, inadmissi
ble in a state court.” Mapp, 367 U. S., at 655. “To hold
otherwise,” the Court added, would be “to grant the right
but in reality to withhold its privilege and enjoyment.”
Id., at 656.
II
Reading our knock-and-announce cases, Part I–A, supra,
in light of this foundational Fourth Amendment case law,
Part I–B, supra, it is clear that the exclusionary rule should
apply. For one thing, elementary logic leads to that conclu
sion. We have held that a court must “conside[r]” whether
officers complied with the knock-and-announce requirement
“in assessing the reasonableness of a search or seizure.”
Wilson, 514 U. S., at 934; see Banks, supra, at 36. The
Fourth Amendment insists that an unreasonable search or
seizure is, constitutionally speaking, an illegal search or sei
zure. And ever since Weeks (in respect to federal prosecu
tions) and Mapp (in respect to state prosecutions), “the use
of evidence secured through an illegal search and seizure” is
“barred” in criminal trials. Wolf, supra, at 28 (citing Weeks,
232 U. S. 383); see Mapp, supra, at 655.
For another thing, the driving legal purpose underlying
the exclusionary rule, namely, the deterrence of unlawful
government behavior, argues strongly for suppression. See
Elkins v. United States, 364 U. S. 206, 217 (1960) (purpose of
the exclusionary rule is “to deter—to compel respect for the
constitutional guaranty . . . by removing the incentive to dis
regard it”). In Weeks, Silverthorne, and Mapp, the Court
based its holdings requiring suppression of unlawfully ob

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
609 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
tained evidence upon the recognition that admission of that
evidence would seriously undermine the Fourth Amend
ment’s promise. All three cases recognized that failure to
apply the exclusionary rule would make that promise a hol
low one, see Mapp, supra, at 657, reducing it to “a form of
words,” Silverthorne, supra, at 392, “of no value” to those
whom it seeks to protect, Weeks, supra, at 393. Indeed, this
Court in Mapp held that the exclusionary rule applies to the
States in large part due to its belief that alternative state
mechanisms for enforcing the Fourth Amendment’s guaran
tees had proved “worthless and futile.” 367 U. S., at 652.
Why is application of the exclusionary rule any the less
necessary here? Without such a rule, as in Mapp, police
know that they can ignore the Constitution’s requirements
without risking suppression of evidence discovered after an
unreasonable entry. As in Mapp, some government officers
will find it easier, or believe it less risky, to proceed with
what they consider a necessary search immediately and with
out the requisite constitutional (say, warrant or knock-and
announce) compliance. Cf. Mericli, The Apprehension of
Peril Exception to the Knock and Announce Rule—Part I,
16 Search and Seizure L. Rep. 129, 130 (1989) (hereinafter
Mericli) (noting that some “[d]rug enforcement authorities
believe that safety for the police lies in a swift, surprising
entry with overwhelming force—not in announcing their of
ficial authority”).
Of course, the State or the Federal Government may pro
vide alternative remedies for knock-and-announce violations.
But that circumstance was true of Mapp as well. What rea
son is there to believe that those remedies (such as private
damages actions under Rev. Stat. § 1979, 42 U. S. C. § 1983),
which the Court found inadequate in Mapp, can adequately
deter unconstitutional police behavior here? See Kamisar,
In Defense of the Search and Seizure Exclusionary Rule, 26
Harv. J. L. & Pub. Pol’y 119, 126–129 (2003) (arguing that
“five decades of post-Weeks ‘freedom’ from the inhibiting ef

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
610 HUDSON v. MICHIGAN
Breyer, J., dissenting
fect of the federal exclusionary rule failed to produce any
meaningful alternative to the exclusionary rule in any juris
diction” and that there is no evidence that “times have
changed” post-Mapp).
The cases reporting knock-and-announce violations are le
gion. See, e. g., 34 Geo. L. J. Ann. Rev. Crim. Proc. 31–35
(2005) (collecting Courts of Appeals cases); Bremer, 85
A. L. R. 5th 1 (2001) (collecting state-court cases); Brief for
Petitioner 16–17 (collecting federal and state cases). In
deed, these cases of reported violations seem sufficiently fre
quent and serious as to indicate “a widespread pattern.”
Ante, at 604 (Kennedy, J., concurring in part and concurring
in judgment). Yet the majority, like Michigan and the
United States, has failed to cite a single reported case in
which a plaintiff has collected more than nominal damages
solely as a result of a knock-and-announce violation. Even
Michigan concedes that, “in cases like the present one . . . ,
damages may be virtually nonexistent.” Brief for Respond
ent 35, n. 66. And Michigan’s amici further concede that
civil immunities prevent tort law from being an effective sub
stitute for the exclusionary rule at this time. Brief for
Criminal Justice Legal Foundation 10; see also Hope v.
Pelzer, 536 U. S. 730, 739 (2002) (difficulties of overcoming
qualified immunity defenses).
As Justice Stewart, the author of a number of significant
Fourth Amendment opinions, explained, the deterrent effect
of damages actions “can hardly be said to be great,” as such
actions are “expensive, time-consuming, not readily avail
able, and rarely successful.” The Road to Mapp v. Ohio and
Beyond: The Origins, Development and Future of the Ex
clusionary Rule in Search-and-Seizure Cases, 83 Colum. L.
Rev. 1365, 1388 (1983). The upshot is that the need for de
terrence—the critical factor driving this Court’s Fourth
Amendment cases for close to a century—argues with at
least comparable strength for evidentiary exclusion here.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
611 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
To argue, as the majority does, that new remedies, such
as 42 U. S. C. § 1983 actions or better trained police, make
suppression unnecessary is to argue that Wolf, not Mapp, is
now the law. (The Court recently rejected a similar argu
ment in Dickerson v. United States, 530 U. S. 428, 441–442
(2000).) To argue that there may be few civil suits because
violations may produce nothing “more than nominal injury”
is to confirm, not to deny, the inability of civil suits to deter
violations. See ante, at 598. And to argue without evi
dence (and despite myriad reported cases of violations, no
reported case of civil damages, and Michigan’s concession of
their nonexistence) that civil suits may provide deterrence
because claims may “have been settled” is, perhaps, to
search in desperation for an argument. See ibid. Rather,
the majority, as it candidly admits, has simply “assumed”
that, “[a]s far as [it] know[s], civil liability is an effective
deterrent,” ibid., a support-free assumption that Mapp and
subsequent cases make clear does not embody the Court’s
normal approach to difficult questions of Fourth Amend
ment law.
It is not surprising, then, that after looking at virtually
every pertinent Supreme Court case decided since Weeks,
I can find no precedent that might offer the majority support
for its contrary conclusion. The Court has, of course, recog
nized that not every Fourth Amendment violation necessar
ily triggers the exclusionary rule. Ante, at 590–592; cf.
Illinois v. Gates, 462 U. S. 213, 223 (1983) (application of
the exclusionary rule is a separate question from whether
the Fourth Amendment has been violated). But the class of
Fourth Amendment violations that do not result in suppres
sion of the evidence seized, however, is limited.
The Court has declined to apply the exclusionary rule only:
(1) where there is a specific reason to believe that appli
cation of the rule would “not result in appreciable deter
rence,” United States v. Janis, 428 U. S. 433, 454 (1976);
see, e. g., United States v. Leon, 468 U. S. 897, 919–920

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
612 HUDSON v. MICHIGAN
Breyer, J., dissenting
(1984) (exception where searching officer executes defec
tive search warrant in “good faith”); Arizona v. Evans,
514 U. S. 1, 14 (1995) (exception for clerical errors by
court employees); Walder v. United States, 347 U. S. 62
(1954) (exception for impeachment purposes), or
(2) where admissibility in proceedings other than crimi
nal trials was at issue, see, e. g., Pennsylvania Bd. of
Probation and Parole v. Scott, 524 U. S. 357, 364 (1998)
(exception for parole revocation proceedings); INS v.
Lopez-Mendoza, 468 U. S. 1032, 1050 (1984) (plurality
opinion) (exception for deportation proceedings); Janis,
supra, at 458 (exception for civil tax proceedings);
United States v. Calandra, 414 U. S. 338, 348–350 (1974)
(exception for grand jury proceedings); Stone v. Powell,
428 U. S. 465, 493–494 (1976) (exception for federal ha
beas proceedings).
Neither of these two exceptions applies here. The second
does not apply because this case is an ordinary criminal trial.
The first does not apply because (1) officers who violate the
rule are not acting “as a reasonable officer would and should
act in similar circumstances,” Leon, supra, at 920, (2) this
case does not involve government employees other than po
lice, Evans, supra, and (3), most importantly, the key ration
ale for any exception, “lack of deterrence,” is missing, see
Pennsylvania Bd. of Probation, supra, at 364 (noting that
the rationale for not applying the rule in noncriminal cases
has been that the deterrence achieved by having the rule
apply in those contexts is “minimal” because “application of
the rule in the criminal trial context already provides sig
nificant deterrence of unconstitutional searches”); Michigan
v. Tucker, 417 U. S. 433, 447 (1974) (noting that deterrence
rationale would not be served if rule applied to police officers
acting in good faith, as the “deterrent purpose of the exclu
sionary rule necessarily assumes that the police have en
gaged in willful, or at the very least negligent, conduct”).
That critical latter rationale, which underlies every excep

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
613 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
tion, does not apply here, as there is no reason to think that,
in the case of knock-and-announce violations by the police,
“the exclusion of evidence at trial would not sufficiently
deter future errors,” Evans, supra, at 14, or “ ‘further the
ends of the exclusionary rule in any appreciable way,’ ” Leon,
supra, at 919–920.
I am aware of no other basis for an exception. The Court
has decided more than 300 Fourth Amendment cases since
Weeks. The Court has found constitutional violations in
nearly a third of them. See W. Greenhalgh, The Fourth
Amendment Handbook: A Chronological Survey of Supreme
Court Decisions 27–130 (2d ed. 2003) (collecting and summa
rizing 332 post-Weeks cases decided between 1914 and 2002).
The nature of the constitutional violation varies. In most
instances officers lacked a warrant; in others, officers pos
sessed a warrant based on false affidavits; in still others, the
officers executed the search in an unconstitutional manner.
But in every case involving evidence seized during an illegal
search of a home (federally since Weeks, nationally since
Mapp), the Court, with the exceptions mentioned, has either
explicitly or implicitly upheld (or required) the suppression
of the evidence at trial. See Appendix, infra. In not one
of those cases did the Court “questio[n], in the absence of a
more efficacious sanction, the continued application of the
[exclusionary] rule to suppress evidence from the State’s
case” in a criminal trial. Franks v. Delaware, 438 U. S. 154,
171 (1978).
I can find nothing persuasive in the majority’s opinion that
could justify its refusal to apply the rule. It certainly is not
a justification for an exception here (as the majority finds) to
find odd instances in other areas of law that do not auto
matically demand suppression. Ante, at 596–597 (suspect
confesses, police beat him up afterwards; suspect confesses,
then police apparently arrest him, take him to station, and
refuse to tell him of his right to counsel). Nor can it justify
an exception to say that some police may knock at the door

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
614 HUDSON v. MICHIGAN
Breyer, J., dissenting
anyway (to avoid being mistaken for a burglar), for other
police (believing quick entry is the most secure, effective
entry) will not voluntarily do so. Cf. Mericli 130 (describ
ing Special Weapons and Tactics (SWAT) team practices);
R. Balko, No SWAT (Apr. 6, 2006), available at http://www.
cato.org/pub_display.php?pub_id=6344 (all Internet materials
as visited June 7, 2006, and available in Clerk of Court’s
case file).
Neither can the majority justify its failure to respect the
need for deterrence, as set forth consistently in the Court’s
prior case law, through its claim of “ ‘substantial social
costs’ ”—at least if it means that those “ ‘social costs’ ” are
somehow special here. Ante, at 596. The only costs it men
tions are those that typically accompany any use of the
Fourth Amendment’s exclusionary principle: (1) that where
the constable blunders, a guilty defendant may be set free
(consider Mapp itself); (2) that defendants may assert claims
where Fourth Amendment rights are uncertain (consider the
Court’s qualified immunity jurisprudence), and (3) that some
times it is difficult to decide the merits of those uncertain
claims. See ante, at 595–596. In fact, the “no-knock” war
rants that are provided by many States, by diminishing un
certainty, may make application of the knock-and-announce
principle less “ ‘cost[ly]’ ” on the whole than application of
comparable Fourth Amendment principles, such as deter
mining whether a particular warrantless search was justified
by exigency. The majority’s “substantial social costs” argu
ment is an argument against the Fourth Amendment’s exclu
sionary principle itself. And it is an argument that this
Court, until now, has consistently rejected.
III
The majority, Michigan, and the United States make sev
eral additional arguments. In my view, those arguments
rest upon misunderstandings of the principles underlying
this Court’s precedents.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
615 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
A
The majority first argues that “the constitutional violation
of an illegal manner of entry was not a but-for cause of ob
taining the evidence.” Ante, at 592. But taking causation
as it is commonly understood in the law, I do not see how
that can be so. See W. Keeton, D. Dobbs, R. Keeton, & D.
Owen, Prosser and Keeton on Law of Torts 266 (5th ed.
1984). Although the police might have entered Hudson’s
home lawfully, they did not in fact do so. Their unlawful
behavior inseparably characterizes their actual entry; that
entry was a necessary condition of their presence in Hudson’s
home; and their presence in Hudson’s home was a necessary
condition of their finding and seizing the evidence. At the
same time, their discovery of evidence in Hudson’s home was
a readily foreseeable consequence of their entry and their
unlawful presence within the home. Cf. 2 Restatement
(Second) of Torts § 435 (1963–1964).
Moreover, separating the “manner of entry” from the re
lated search slices the violation too finely. As noted,
Part I–A, supra, we have described a failure to comply with
the knock-and-announce rule, not as an independently unlaw
ful event, but as a factor that renders the search “constitu
tionally defective.” Wilson, 514 U. S., at 936; see also id.,
at 934 (compliance with the knock-and-announce requirement
is one of the “factors to be considered in assessing the rea
sonableness of a search or seizure” (emphasis added)); Ker
v. California, 374 U. S. 23, 53 (1963) (opinion of Brennan, J.)
(“[A] lawful entry is the indispensable predicate of a reason
able search”).
The Court nonetheless accepts Michigan’s argument that
the requisite but-for causation is not satisfied in this case
because, whether or not the constitutional violation occurred
(what the Court refers to as a “preliminary misstep”), “the
police would have executed the warrant they had obtained,
and would have discovered the gun and drugs inside the

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
616 HUDSON v. MICHIGAN
Breyer, J., dissenting
house.” Ante, at 592. As support for this proposition,
Michigan rests on this Court’s inevitable discovery cases.
This claim, however, misunderstands the inevitable discov
ery doctrine. Justice Holmes in Silverthorne, in discussing
an “independent source” exception, set forth the principles
underlying the inevitable discovery rule. See supra, at 607.
That rule does not refer to discovery that would have taken
place if the police behavior in question had (contrary to fact)
been lawful. The doctrine does not treat as critical what
hypothetically could have happened had the police acted
lawfully in the first place. Rather, “independent” or “inevi
table” discovery refers to discovery that did occur or that
would have occurred (1) despite (not simply in the absence
of) the unlawful behavior and (2) independently of that
unlawful behavior. The government cannot, for example,
avoid suppression of evidence seized without a warrant (or
pursuant to a defective warrant) simply by showing that it
could have obtained a valid warrant had it sought one. See,
e. g., Coolidge v. New Hampshire, 403 U. S. 443, 450–451
(1971). Instead, it must show that the same evidence “inevi
tably would have been discovered by lawful means.” Nix
v. Williams, 467 U. S., at 444 (emphasis added). “What a
man could do is not at all the same as what he would do.”
Austin, Ifs And Cans, 42 Proceedings of the British Academy
109, 111–112 (1956).
The inevitable discovery exception rests upon the princi
ple that the remedial purposes of the exclusionary rule are
not served by suppressing evidence discovered through a
“later, lawful seizure” that is “genuinely independent of an
earlier, tainted one.” Murray v. United States, 487 U. S.
533, 542 (1988) (emphasis added); see also id., at 545 (Mar
shall, J., joined by Stevens and O’Connor, JJ., dissenting)
(“When the seizure of the evidence at issue is ‘wholly
independent of ’ the constitutional violation, then exclusion
arguably will have no effect on a law enforcement officer’s
incentive to commit an unlawful search”).

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
617 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
Case law well illustrates the meaning of this principle. In
Nix, supra, police officers violated a defendant’s Sixth
Amendment right by eliciting incriminating statements from
him after he invoked his right to counsel. Those statements
led to the discovery of the victim’s body. The Court con
cluded that evidence obtained from the victim’s body was
admissible because it would ultimately or inevitably have
been discovered by a volunteer search party effort that was
ongoing—whether or not the Sixth Amendment violation had
taken place. Id., at 449. In other words, the evidence
would have been found despite, and independent of, the Sixth
Amendment violation.
In Segura v. United States, 468 U. S. 796 (1984), one of the
“trio of cases” Justice Scalia says “confirms [the Court’s]
conclusion,” ante, at 599–600 (plurality opinion), the Court
held that an earlier illegal entry into an apartment did not
require suppression of evidence that police later seized when
executing a search warrant obtained on the basis of informa
tion unconnected to the initial entry. The Court reasoned
that the “evidence was discovered the day following the
entry, during the search conducted under a valid war
rant”—i. e., a warrant obtained independently without use of
any information found during the illegal entry—and that “it
was the product of that search, wholly unrelated to the prior
[unlawful] entry.” Segura, supra, at 814 (emphasis added).
In Murray, supra, the Court upheld the admissibility of
seized evidence where agents entered a warehouse without
a warrant, and then later returned with a valid warrant that
was not obtained on the basis of evidence observed during
the first (illegal) entry. The Court reasoned that while the
agents’ “[k]nowledge that the marijuana was in the ware
house was assuredly acquired at the time of the unlawful
entry . . . it was also acquired at the time of entry pursuant
to the warrant, and if that later acquisition was not the re
sult of the earlier entry there is no reason why the independ

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
618 HUDSON v. MICHIGAN
Breyer, J., dissenting
ent source doctrine should not apply.” Id., at 541 (empha
sis added).
Thus, the Court’s opinion reflects a misunderstanding of
what “inevitable discovery” means when it says, “[i]n this
case, of course, the constitutional violation of an illegal man
ner of entry was not a but-for cause of obtaining the evi
dence.” Ante, at 592. The majority rests this conclusion
on its next statement: “Whether that preliminary misstep
had occurred or not, the police . . . would have discovered
the gun and drugs inside the house.” Ibid. Despite the
phrase “of course,” neither of these statements is correct.
It is not true that, had the illegal entry not occurred, “police
. . . would have discovered the gun and drugs inside the
house.” Without that unlawful entry they would not have
been inside the house; so there would have been no discovery.
See supra, at 615.
Of course, had the police entered the house lawfully, they
would have found the gun and drugs. But that fact is beside
the point. The question is not what police might have done
had they not behaved unlawfully. The question is what they
did do. Was there set in motion an independent chain of
events that would have inevitably led to the discovery and
seizure of the evidence despite, and independent of, that be
havior? The answer here is “no.”
B
The majority, Michigan, and the United States point out
that the officers here possessed a warrant authorizing a
search. Ante, at 592. That fact, they argue, means that
the evidence would have been discovered independently or
somehow diminishes the need to suppress the evidence. But
I do not see why that is so. The warrant in question was
not a “no-knock” warrant, which many States (but not Michi
gan) issue to assure police that a prior knock is not necessary.
Richards, 520 U. S., at 396, n. 7 (collecting state statutes).
It did not authorize a search that fails to comply with knock

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
619 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
and-announce requirements. Rather, it was an ordinary
search warrant. It authorized a search that complied with,
not a search that disregarded, the Constitution’s knock-and
announce rule.
Would a warrant that authorizes entry into a home on
Tuesday permit the police to enter on Monday? Would a
warrant that authorizes entry during the day authorize the
police to enter during the middle of the night? It is difficult
for me to see how the presence of a warrant that does not
authorize the entry in question has anything to do with the
“inevitable discovery” exception or otherwise diminishes
the need to enforce the knock-and-announce requirement
through suppression.
C
The majority and the United States set forth a policy
related variant of the causal connection theme: The United
States argues that the law should suppress evidence only
insofar as a Fourth Amendment violation causes the kind of
harm that the particular Fourth Amendment rule seeks to
protect against. It adds that the constitutional purpose of
the knock-and-announce rule is to prevent needless destruc
tion of property (such as breaking down a door) and to avoid
unpleasant surprise. And it concludes that the exclusionary
rule should suppress evidence of, say, damage to property,
the discovery of a defendant in an “intimate or compromising
moment,” or an excited utterance from the occupant caught
by surprise, but nothing more. Brief for United States as
Amicus Curiae 12, 28.
The majority makes a similar argument. It says that evi
dence should not be suppressed once the causal connection
between unlawful behavior and discovery of the evidence be
comes too “attenuated.” Ante, at 592. But the majority
then makes clear that it is not using the word “attenuated”
to mean what this Court’s precedents have typically used
that word to mean, namely, that the discovery of the evi
dence has come about long after the unlawful behavior took

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
620 HUDSON v. MICHIGAN
Breyer, J., dissenting
place or in an independent way, i. e., through “ ‘means suffi
ciently distinguishable to be purged of the primary taint.’ ”
Wong Sun v. United States, 371 U. S. 471, 487–488 (1963); see
Brown v. Illinois, 422 U. S. 590, 603–604 (1975).
Rather, the majority gives the word “attenuation” a new
meaning (thereby, in effect, making the same argument as
the United States). “Attenuation,” it says, “also occurs
when, even given a direct causal connection, the interest pro
tected by the constitutional guarantee that has been violated
would not be served by suppression of the evidence ob
tained. ” Ante, at 593. The interests the knock-and
announce rule seeks to protect, the Court adds, are “human
life” (at stake when a householder is “surprised”), “property”
(such as the front door), and “those elements of privacy and
dignity that can be destroyed by a sudden entrance,” namely,
“the opportunity to collect oneself before answering the
door.” Ante, at 594. Since none of those interests led to
the discovery of the evidence seized here, there is no reason
to suppress it.
There are three serious problems with this argument.
First, it does not fully describe the constitutional values,
purposes, and objectives underlying the knock-and-announce
requirement. That rule does help to protect homeowners
from damaged doors; it does help to protect occupants from
surprise. But it does more than that. It protects the occu
pants’ privacy by assuring them that government agents will
not enter their home without complying with those require
ments (among others) that diminish the offensive nature of
any such intrusion. Many years ago, Justice Frankfurter
wrote for the Court that the “knock at the door, . . . as a
prelude to a search, without authority of law . . . [is] in
consistent with the conception of human rights enshrined
in [our] history” and Constitution. Wolf, 338 U. S., at 28.
How much the more offensive when the search takes place
without any knock at all. Cf. Wilson, 514 U. S., at 931
(knock-and-announce rule recognizes that “the common law

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
621 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
generally protected a man’s house as ‘his castle of defence
and asylum’ ” (quoting 3 W. Blackstone, Commentaries
*288)); Miller, 357 U. S., at 313 (federal knock-and-announce
statute “codif[ied] a tradition embedded in Anglo-American
law” that reflected “the reverence of the law for the individu
al’s right of privacy in his house”).
Over a century ago this Court wrote that “[i]t is not the
breaking of his doors” that is the “essence of the offence,”
but the “invasions on the part of the government . . . of the
sanctity of a man’s home and the privacies of life.” Boyd,
116 U. S., at 630. And just this Term we have reiterated
that “ ‘it is beyond dispute that the home is entitled to special
protection as the center of the private lives of our people.’ ”
Georgia v. Randolph, ante, at 115 (quoting Minnesota v. Car
ter, 525 U. S. 83, 99 (1998) (Kennedy, J., concurring)). The
knock-and-announce requirement is no less a part of the
“centuries-old principle” of special protection for the privacy
of the home than the warrant requirement. See Randolph,
ante, at 115 (citing Miller, supra, at 307). The Court is
therefore wrong to reduce the essence of its protection to
“the right not to be intruded upon in one’s nightclothes.”
Ante, at 597; see Richards, 520 U. S., at 393, n. 5 (“[I]ndivid
ual privacy interest[s]” protected by the rule are “not incon
sequential” and “should not be unduly minimized”).
Second, whether the interests underlying the knock-and
announce rule are implicated in any given case is, in a sense,
beside the point. As we have explained, failure to comply
with the knock-and-announce rule renders the related search
unlawful. Wilson, supra, at 936. And where a search is
unlawful, the law insists upon suppression of the evidence
consequently discovered, even if that evidence or its posses
sion has little or nothing to do with the reasons underlying
the unconstitutionality of a search. The Fourth Amendment
does not seek to protect contraband, yet we have required
suppression of contraband seized in an unlawful search.
See, e. g., Kyllo v. United States, 533 U. S. 27, 40 (2001); Coo

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
622 HUDSON v. MICHIGAN
Breyer, J., dissenting
lidge, 403 U. S., at 473. That is because the exclusionary
rule protects more general “privacy values through deter
rence of future police misconduct.” James v. Illinois, 493
U. S. 307, 319 (1990). The same is true here.
Third, the majority’s interest-based approach departs from
prior law. Ordinarily a court will simply look to see if the
unconstitutional search produced the evidence. The major
ity does not refer to any relevant case in which, beyond that,
suppression turned on the far more detailed relation be
tween, say, (1) a particular materially false statement made
to the magistrate who issued a (consequently) invalid war
rant and (2) evidence found after a search with that warrant.
But cf. ante, at 601–602, n. 2 (plurality opinion) (citing New
York v. Harris, 495 U. S. 14 (1990), as such a case in section
of opinion that Justice Kennedy does not join). And the
majority’s failure does not surprise me, for such efforts
to trace causal connections at retail could well compli
cate Fourth Amendment suppression law, threatening its
workability.
D
The United States, in its brief and at oral argument, has
argued that suppression is “an especially harsh remedy
given the nature of the violation in this case.” Brief as
Amicus Curiae 28; see also id., at 24. This argument fo
cuses upon the fact that entering a house after knocking and
announcing can, in some cases, prove dangerous to a police
officer. Perhaps someone inside has a gun, as turned out to
be the case here. The majority adds that police officers
about to encounter someone who may try to harm them will
be “uncertain” as to how long to wait. Ante, at 595. It
says that, “[i]f the consequences of running afoul” of the
knock-and-announce “rule were so massive,” i. e., would lead
to the exclusion of evidence, then “officers would be inclined
to wait longer than the law requires—producing preventable
violence against officers in some cases.” Ibid.
To argue that police efforts to assure compliance with the
rule may prove dangerous, however, is not to argue against

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
623 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
evidence suppression. It is to argue against the validity of
the rule itself. Similarly, to argue that enforcement means
uncertainty, which in turn means the potential for dangerous
and longer-than-necessary delay, is (if true) to argue against
meaningful compliance with the rule.
The answer to the first argument is that the rule itself
does not require police to knock or to announce their pres
ence where police have a “reasonable suspicion” that doing so
“would be dangerous or futile” or “would inhibit the effective
investigation of the crime by, for example, allowing the de
struction of evidence.” Richards, supra, at 394; see Banks,
540 U. S., at 36–37; Wilson, 514 U. S., at 935–936.
The answer to the second argument is that States can, and
many do, reduce police uncertainty while assuring a neutral
evaluation of concerns about risks to officers or the destruc
tion of evidence by permitting police to obtain a “no-knock”
search warrant from a magistrate judge, thereby assuring
police that a prior announcement is not necessary. Rich
ards, 520 U. S., at 396, n. 7 (collecting state statutes). While
such a procedure cannot remove all uncertainty, it does pro
vide an easy way for officers to comply with the knock-and
announce rule.
Of course, even without such a warrant, police maintain
the backup “authority to exercise independent judgment con
cerning the wisdom of a no-knock entry at the time the war
rant is being executed.” Ibid. “[I]f circumstances support
a reasonable suspicion of exigency when the officers arrive
at the door, they may go straight in.” Banks, supra, at 37.
And “[r]easonable suspicion is a less demanding standard
than probable cause . . . .” Alabama v. White, 496 U. S. 325,
330 (1990); see Terry v. Ohio, 392 U. S. 1, 21–22 (1968) (no
Fourth Amendment violation under the reasonable suspicion
standard if “the facts available to the officer at the moment
of the seizure or the search ‘warrant a man of reasonable
caution in the belief ’ that the action taken was appropriate”).
Consider this very case. The police obtained a search
warrant that authorized a search, not only for drugs, but also

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
624 HUDSON v. MICHIGAN
Breyer, J., dissenting
for guns. App. 5. If probable cause justified a search for
guns, why would it not also have justified a no-knock war
rant, thereby diminishing any danger to the officers? Why
(in a State such as Michigan that lacks no-knock warrants)
would it not have justified the very no-knock entry at issue
here? Indeed, why did the prosecutor not argue in this very
case that, given the likelihood of guns, the no-knock entry
was lawful? From what I have seen in the record, he would
have won. And had he won, there would have been no sup
pression here.
That is the right way to win. The very process of arguing
the merits of the violation would help to clarify the contours
of the knock-and-announce rule, contours that the majority
believes are too fuzzy. That procedural fact, along with no
knock warrants, back up authority to enter without knocking
regardless, and use of the “reasonable suspicion” standard
for doing so should resolve the government’s problems with
the knock-and-announce rule while reducing the “uncer
tain[ty]” that the majority discusses to levels beneath that
found elsewhere in Fourth Amendment law (e. g., exigent
circumstances). Ante, at 595. Regardless, if the Court
fears that effective enforcement of a constitutional require
ment will have harmful consequences, it should face those
fears directly by addressing the requirement itself. It
should not argue, “the requirement is fine, indeed, a serious
matter, just don’t enforce it.”
E
It should be apparent by now that the three cases upon
which Justice Scalia relies—Segura v. United States, 468
U. S. 796; New York v. Harris, 495 U. S. 14; and Ramirez,
523 U. S. 65—do not support his conclusion. See ante, at
599–602. Indeed, Justice Kennedy declines to join this
section of the lead opinion because he fails to see the rele
vance of Segura and Harris, though he does rely on Ramirez.
Ante, at 604 (opinion concurring in part and concurring in
judgment).

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
625 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
Justice Scalia first argues that, if the “search in Segura
could be ‘wholly unrelated to the prior entry,’ . . . when the
only entry was warrantless, it would be bizarre to treat more
harshly the actions in this case, where the only entry was
with a warrant.” Ante, at 600. Then he says that, “[i]f the
probable cause backing a warrant that was issued later in
time could be an ‘independent source’ for a search that pro
ceeded after the officers illegally entered and waited, a
search warrant obtained before going in must have at least
this much effect.” Ante, at 600–601. I do not understand
these arguments. As I have explained, the presence of a
warrant that did not authorize a search that fails to comply
with knock-and-announce requirements is beside the point.
See Part III–B, supra. And the timing of the warrant in
Segura made no difference to the case. The relevant fact
about the warrant there was that it was lawfully obtained
and arguably set off an independent chain of events that led
the police to seize the evidence. 468 U. S., at 814; see also
ibid. (“The valid warrant search was a ‘means sufficiently
distinguishable’ to purge the evidence of any ‘taint’ arising
from the entry”). As noted, there is no such independent
event, or intervening chain of events that would purge the
taint of the illegal entry, present here. See supra, at 618.
The search that produced the relevant evidence here is the
very search that the knock-and-announce violation rendered
unlawful. There simply is no “independent source.”
As importantly, the Court in Segura said nothing to sug
gest it intended to create a major exclusionary rule excep
tion, notwithstanding the impact of such an exception on de
terrence. Indeed, such an exception would be inconsistent
with a critical rationale underlying the independent source
and inevitable discovery rules, which was arguably available
in Segura, and which is clearly absent here. That rationale
concerns deterrence. The threat of inadmissibility deters
unlawful police behavior; and the existence of an exception
applicable where evidence is found through an untainted in
dependent route will rarely undercut that deterrence. That

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
626 HUDSON v. MICHIGAN
Breyer, J., dissenting
is because the police can rarely rely upon such an excep
tion—at least not often enough to change the deterrence cal
culus. See Murray, 487 U. S., at 540 (“We see the incentives
differently. An officer with probable cause sufficient to ob
tain a search warrant would be foolish to enter the premises
first in an unlawful manner. By doing so, he would risk sup
pression of all evidence on the premises . . . ”); Nix, 467 U. S.,
at 445 (“A police officer who is faced with the opportunity to
obtain evidence illegally will rarely, if ever, be in a position
to calculate whether the evidence sought would inevitably
be discovered”); id., at 444 (“If the prosecution can establish
by a preponderance of the evidence that the information ulti
mately or inevitably would have been discovered by lawful
means—here the volunteers’ search—then the deterrence
rationale has so little basis that the evidence should be
received”).
Segura’s police officers would have been foolish to have
entered the apartment unlawfully with the ex ante hope that
an independent causal chain of events would later occur and
render admissible the evidence they found. By way of con
trast, today’s holding will seriously undermine deterrence
in knock-and-announce cases. Officers will almost always
know ex ante that they can ignore the knock-and-announce
requirement without risking the suppression of evidence dis
covered after their unlawful entry. That fact is obvious, and
this Court has never before today—not in Segura or any
other post-Weeks (or post-Mapp) case—refused to apply the
exclusionary rule where its absence would so clearly and so
significantly impair government officials’ incentive to comply
with comparable Fourth Amendment requirements.
Neither does New York v. Harris, supra, support the
Court’s result. See ante, at 593, 601; but see ante, at 604
(opinion of Kennedy, J.) (declining to join section relying on
Harris). In Harris, police officers arrested the defendant
at his home without a warrant, in violation of Payton v. New
York, 445 U. S. 573 (1980). Harris made several incriminat

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
627 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
ing statements: a confession in his home, a written inculpa
tory statement at the station house, and a videotaped inter
view conducted by the district attorney at the station house.
495 U. S., at 16. The trial court suppressed the statements
given by Harris in the house and on the videotape, and the
State did not challenge either of those rulings. Ibid. The
sole question in the case was whether the written statement
given later at the station house should also have been sup
pressed. The Court held that this later, outside-the-home
statement “was admissible because Harris was in legal cus
tody . . . and because the statement, while the product of an
arrest and being in custody, was not the fruit of the fact that
the arrest was made in the house rather than someplace
else.” Id., at 20. Immediately after the Court stated its
holding, it explained:
“To put the matter another way, suppressing the
statement taken outside the house would not serve the
purpose of the rule that made Harris’ in-house arrest
illegal. The warrant requirement for an arrest in the
home is imposed to protect the home, and anything in
criminating the police gathered from arresting Harris
in his home, rather than elsewhere, has been excluded,
as it should have been; the purpose of the rule has
thereby been vindicated.” Ibid. (emphasis added).
How can Justice Scalia maintain that the evidence
here—a gun and drugs seized in the home—is “ ‘not the
fruit’ ” of the illegal entry? Ante, at 601. The officers’ fail
ure to knock and announce rendered the entire search unlaw
ful, Wilson, 514 U. S., at 936, and that unlawful search led to
the discovery of evidence in petitioner’s home. Thus, Har
ris compels the opposite result than that reached by the
Court today. Like the Payton rule at issue in Harris, the
knock-and-announce rule reflects the “reverence of the law
for the individual’s right of privacy in his house.” Miller,
357 U. S., at 313; cf. Harris, 495 U. S., at 17 (“Payton itself

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
628 HUDSON v. MICHIGAN
Breyer, J., dissenting
emphasized that our holding in that case stemmed from the
‘overriding respect for the sanctity of the home that has been
embedded in our traditions since the origins of the Repub
lic’ ”). Like the confession that was “excluded, as it should
have been,” in Harris, id., at 20, the evidence in this case
was seized in the home, immediately following the illegal
entry. And like Harris, nothing in petitioner’s argument
would require the suppression of evidence obtained outside
the home following a knock-and-announce violation, precisely
because officers have a remaining incentive to follow the rule
to avoid the suppression of any evidence obtained from the
very place they are searching. Cf. ibid. (“Even though we
decline to suppress statements made outside the home fol
lowing a Payton violation, the principal incentive to obey
Payton still obtains: the police know that a warrantless
entry will lead to the suppression of any evidence found, or
statements taken, inside the home”).
I concede that United States v. Ramirez, 523 U. S. 65, of
fers the plurality its last best hope. Ante, at 602. But not
even that case can offer the plurality significant support.
The plurality focuses on the Court’s isolated statement that
“destruction of property in the course of a search may violate
the Fourth Amendment, even though the entry itself is law
ful and the fruits of the search are not subject to suppres
sion.” Ramirez, supra, at 71 (emphasis added). But even
if I accept this dictum, the entry here is unlawful, not lawful.
Wilson, supra, at 931, 934. It is one thing to say (in an
appropriate case) that destruction of property after proper
entry has nothing to do with discovery of the evidence, and
to refuse to suppress. It would be quite another thing to
say that improper entry had nothing to do with discovery of
the evidence in this case. Moreover, the deterrence analysis
for the property destruction cases (where, by definition,
there will almost always be quantifiable damages) might
well differ.

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
629 Cite as: 547 U. S. 586 (2006)
Breyer, J., dissenting
IV
There is perhaps one additional argument implicit in the
majority’s approach. The majority says, for example, that
the “cost” to a defendant of “entering this lottery,” i. e., of
claiming a “knock-and-announce” violation, “would be small,
but the jackpot enormous”—namely, a potential “get-out-of
jail-free card.” Ante, at 595. It adds that the “social costs”
of applying the exclusionary rule here are not worth the
deterrence benefits. Ante, at 599. Leaving aside what I
believe are invalid arguments based on precedent or the ma
jority’s own estimate that suppression is not necessary to
deter constitutional violations, one is left with a simple
unvarnished conclusion, namely, that in this kind of case, a
knock-and-announce case, “[r]esort to the massive remedy of
suppressing evidence of guilt is unjustified.” Ibid. Why
is that judicial judgment, taken on its own, inappropriate?
Could it not be argued that the knock-and-announce rule, a
subsidiary Fourth Amendment rule, is simply not important
enough to warrant a suppression remedy? Could the major
ity not simply claim that the suppression game is not worth
the candle?
The answer, I believe, is “no.” That “no” reflects history,
a history that shows the knock-and-announce rule is impor
tant. See Wilson, supra, at 931–936. That “no” reflects
precedent, precedent that shows there is no pre-existing
legal category of exceptions to the exclusionary rule into
which the knock-and-announce cases might fit. See supra,
at 612–613. That “no” reflects empirical fact, experience
that provides confirmation of what common sense suggests:
without suppression there is little to deter knock-and
announce violations. See supra, at 608–610.
There may be instances in the law where text or history
or tradition leaves room for a judicial decision that rests
upon little more than an unvarnished judicial instinct. But
this is not one of them. Rather, our Fourth Amendment
traditions place high value upon protecting privacy in the

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
630 HUDSON v. MICHIGAN
Appendix to opinion of Breyer, J.
home. They emphasize the need to assure that its constitu
tional protections are effective, lest the Amendment ‘‘sound
the word of promise to the ear but break it to the hope.’’
They include an exclusionary principle, which since Weeks
has formed the centerpiece of the criminal law’s effort to
ensure the practical reality of those promises. That is why
the Court should assure itself that any departure from that
principle is firmly grounded in logic, in history, in precedent,
and in empirical fact. It has not done so. That is why, with
respect, I dissent.
APPENDIX TO OPINION OF BREYER, J.
Fourth Amendment decisions from 1914 to present requir
ing suppression of evidence seized (or remanding for lower
court to make suppression determination) in a private home
following an illegal arrest or search:
1. Weeks v. United States, 232 U. S. 383 (1914) (war
rantless search)
2. Amos v. United States, 255 U. S. 313 (1921) (warrant
less arrest and search)
3. Agnello v. United States, 269 U. S. 20 (1925) (war
rantless search)
4. Byars v. United States, 273 U. S. 28 (1927) (invalid
warrant)
5. United States v. Berkeness, 275 U. S. 149 (1927) (in
valid warrant; insufficient affidavit)
6. Taylor v. United States, 286 U. S. 1 (1932) (warrant
less search)
7. Grau v. United States, 287 U. S. 124 (1932) (invalid
warrant; insufficient affidavit)
8. Nathanson v. United States, 290 U. S. 41 (1933) (in
valid warrant; insufficient affidavit)
9. McDonald v. United States, 335 U. S. 451 (1948)
(warrantless arrest and search)
10. Kremen v. United States, 353 U. S. 346 (1957) (per
curiam) (warrantless search)

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
631 Cite as: 547 U. S. 586 (2006)
Appendix to opinion of Breyer, J.
11. Elkins v. United States, 364 U. S. 206 (1960) (search
beyond scope of warrant)
12. Silverman v. United States, 365 U. S. 505 (1961)
(warrantless use of electronic device)
13. Chapman v. United States, 365 U. S. 610 (1961) (war
rantless search)
14. Mapp v. Ohio, 367 U. S. 643 (1961) (warrantless
search)
15. Wong Sun v. United States, 371 U. S. 471 (1963)
(warrantless search and arrest)
16. Fahy v. Connecticut, 375 U. S. 85 (1963) (warrant
less search)
17. Aguilar v. Texas, 378 U. S. 108 (1964) (invalid war
rant; insufficient affidavit)
18. Stanford v. Texas, 379 U. S. 476 (1965) (invalid war
rant; particularity defect)
19. James v. Louisiana, 382 U. S. 36 (1965) (per curiam)
(warrantless search)
20. Riggan v. Virginia, 384 U. S. 152 (1966) (per cu
riam) (invalid warrant; insufficient affidavit)
21. Bumper v. North Carolina, 391 U. S. 543 (1968) (lack
of valid consent to search)
22. Recznik v. City of Lorain, 393 U. S. 166 (1968) (per
curiam) (warrantless search)
23. Chimel v. California, 395 U. S. 752 (1969) (invalid
search incident to arrest)
24. Von Cleef v. New Jersey, 395 U. S. 814 (1969) (per
curiam) (invalid search incident to arrest)
25. Shipley v. California, 395 U. S. 818 (1969) (per cu
riam) (invalid search incident to arrest)
26. Vale v. Louisiana, 399 U. S. 30 (1970) (invalid search
incident to arrest)
27. Connally v. Georgia, 429 U. S. 245 (1977) (per cu
r iam) (invalid warrant; magistrate judge not
neutral)

547US2 Unit: $U65 [04-03-09 11:55:41] PAGES PGT: OPIN
632 HUDSON v. MICHIGAN
Appendix to opinion of Breyer, J.
28. Michigan v. Tyler, 436 U. S. 499 (1978) (warrant
less search)
29. Mincey v. Arizona, 437 U. S. 385 (1978) (warrant
less search)
30. Franks v. Delaware, 438 U. S. 154 (1978) (invalid
warrant; obtained through perjury)
31. Payton v. New York, 445 U. S. 573 (1980) (warrant
less arrest)
32. Steagald v. United States, 451 U. S. 204 (1981) (war
rantless search)
33. Michigan v. Clifford, 464 U. S. 287 (1984) (warrant
less search)
34. Welsh v. Wisconsin, 466 U. S. 740 (1984) (warrantless
entry into home without exigent circumstances)
35. Thompson v. Louisiana, 469 U. S. 17 (1984) (per cu
riam) (warrantless search)
36. Arizona v. Hicks, 480 U. S. 321 (1987) (unreason
able search)
37. Minnesota v. Olson, 495 U. S. 91 (1990) (warrantless
entry into home)
38. Flippo v. West Virginia, 528 U. S. 11 (1999) (per cu
riam) (warrantless search)
39. Kyllo v. United States, 533 U. S. 27 (2001) (warrant
less use of heat-imaging technology)
40. Kirk v. Louisiana, 536 U. S. 635 (2002) (per curiam)
(warrantless arrest and search)
41. Kaupp v. Texas, 538 U. S. 626 (2003) (per curiam)
(warrantless search)

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
633 OCTOBER TERM, 2005
Syllabus
KIRCHER et al. v. PUTNAM FUNDS TRUST et al.
certiorari to the united states court of appeals for
the seventh circuit
No. 05–409. Argued April 24, 2006—Decided June 15, 2006
The Securities Litigation Uniform Standards Act of 1998 (Act) specifies
that private state-law “covered” class actions alleging untruth or manip
ulation “in connection with the purchase or sale” of a “covered” security
may not “be maintained in any State or Federal court,” 15 U. S. C.
§ 77p(b), and authorizes removal to federal district court of “[a]ny cov
ered class action brought in any State court involving a covered secu
rity, as set forth in subsection (b),” § 77p(c). “A ‘covered class action’ is
a lawsuit in which damages are sought on behalf of more than 50 people.
A ‘covered security’ is one traded nationally and listed on a regulated
national exchange.” Merrill Lynch, Pierce, Fenner & Smith Inc. v.
Dabit, ante, at 83.
Petitioners, mutual fund investors, filed separate state-court actions,
each seeking to assert state-law claims on behalf of a class of investors
allegedly injured by devaluation of their holdings by respondent mutual
funds. The funds filed notices of removal in each case stating, among
other things, that the actions were removable under and precluded by
the Act. Once removed, however, the Federal District Court remanded
each case to state court on the ground that it lacked subject-matter
jurisdiction on removal because the Act did not preclude the investors’
claims. Since they were said to have been injured as “holders” of mu
tual fund shares, not purchasers or sellers, the court reasoned, their
claims did not satisfy § 77p(b)’s “in connection with the purchase or sale”
requirement, and the claims could therefore proceed in state court.
The Seventh Circuit acknowledged that 28 U. S. C. § 1447(d) bars review
of district court orders remanding removed cases for lack of subject
matter jurisdiction, but decided that the District Court had the last
word neither on the characterization of its decision as jurisdictional nor
on the correctness of its conclusion that remand was required. The
appeals court considered all covered class actions involving covered
securities, whether precluded or not, to be removable under the Act,
and therefore thought the preclusion issue distinct from the jurisdic
tional issue whether the case belonged in federal court at all. It held
that orders remanding “properly removed” suits as not precluded are
substantive and unaffected by § 1447(d), and therefore reviewable. Pro
posing that the Act reserves to the Federal Judiciary the exclusive au
thority to make the preclusion decision, the court said that treating re

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
634 KIRCHER v. PUTNAM FUNDS TRUST
Syllabus
mand orders in this context as immunized from appeal by § 1447(d)
would mean that a major substantive issue would escape review, since
it would not be open to resolution in the state court subject to review by
this Court. The Seventh Circuit subsequently consolidated the funds’
appeals and decided, on the merits, that the Act precludes the inves
tors’ claims.
Held: Orders remanding for want of preclusion under the Act are subject
to § 1447(d) and its general rule of nonappealability. Pp. 640–648.
(a) Section 1447(d), which states that an “order remanding a case to
the State court from which it was removed is not reviewable on appeal,”
applies to all remands based on the grounds specified in § 1447(c), includ
ing lack of subject-matter jurisdiction. Thermtron Products, Inc. v.
Hermansdorfer, 423 U. S. 336, 343–345. It applies equally to cases re
moved under the general removal statute, § 1441, and to those removed
under other provisions, see Things Remembered, Inc. v. Petrarca, 516
U. S. 124, 128, and its force is not subject to any statutory exception
that might cover this case. The District Court said that it was remand
ing for lack of jurisdiction, an unreviewable ground. Where a remand
order is based on one of § 1447(c)’s grounds, review is unavailable no
matter how plain the legal error in ordering the remand. Briscoe v.
Bell, 432 U. S. 404, 413, n. 13. The Seventh Circuit did not overlook
cases like Briscoe, but relied instead on cases such as Kontrick v. Ryan,
540 U. S. 443, which observed that some rulings loosely called jurisdic
tional are patently not jurisdictional in the strict sense. Viewing this
as such a case, the appeals court understood the District Court’s preclu
sion decision to be substantive, not jurisdictional, and consequently sub
ject to review. But the District Court was correct in understanding its
remand order to be dictated by a finding that it lacked removal jurisdic
tion. Section 77p(c)’s authorization for removal, on which district-court
jurisdiction depends, is confined to cases “set forth in subsection (b),”
i. e., those with claims of untruth or manipulation. That phrase imme
diately follows the § 77p(c) language describing removable cases as
covered class actions involving covered securities, and the language
has no apparent function unless it limits removal to covered class ac
tions involving claims like untruth or deception. Legislative history
tends to show that this was just what Congress understood. The pre
clusion determination is jurisdictional, as is the order implementing
it. Pp. 640–644.
(b) The Seventh Circuit’s reading was in part motivated by the court’s
erroneous assumption that the Act gives federal courts exclusive juris
diction to decide the preclusion issue. A covered action is removable if
it is precluded, and a defendant can enlist the Federal Judiciary to decide
preclusion, but he can elect to leave the case where the plaintiff filed it
and trust the state court to make the preclusion determination. What

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
635 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
a state court could do in the first place it may also do on remand; here,
the funds can ask for dismissal on preclusion grounds when they return
to state court. Collateral estoppel should be no bar to such a revisita
tion, given that § 1447(d) prevents the funds from appealing the District
Court’s decision. While the state court cannot review the decision to
remand in an appellate way, it is free to reject the remanding court’s
reasoning. Missouri Pacific R. Co. v. Fitzgerald, 160 U. S. 556, 583.
There is no reason to doubt that the state court in this litigation will
duly apply Dabit’s holding that holder claims are embraced by § 77p(b),
but this Court can review any claim of error on that point. Pp. 645–648.
403 F. 3d 478, vacated and remanded.
Souter, J., delivered the opinion of the Court, in which Roberts, C. J.,
and Stevens, Kennedy, Thomas, Ginsburg, Breyer, and Alito, JJ.,
joined, and in which Scalia, J., joined as to Parts I, III, and IV. Scalia,
J., filed an opinion concurring in part and concurring in the judgment,
post, p. 648.
Dav id C. Freder ick argued the cause for petitioners.
With him on the briefs were Scott K. Attaway, Robert L.
King, and Klint L. Bruno.
Mark A. Perry argued the cause for respondents. With
him on the brief were Miguel A. Estrada, Amanda M. Rose,
John D. Donovan, Jr., Thomas B. Smith, Steven B. Feirson,
Stephen J. McConnell, Nory Miller, Christopher P. Hall,
Todd D. Brody, Dale R. Harris, Phil C. Neal, Mark A. Rabi
nowitz, John W. Rotunno, Kenneth E. Rechtoris, James R.
Carroll, David S. Clancy, Charles F. Smith, Lee P. Garner,
and Robert Y. Sperling.*
Justice Souter delivered the opinion of the Court.
Title 28 U. S. C. § 1447(d) limits appellate review of a dis
trict court order remanding a case from federal to state
*Briefs of amici curiae urging affirmance were filed for the Chamber
of Commerce of the United States of America by Charles A. Rothfeld,
Robin S. Conrad, and Amar D. Sarwal; for the Securities Industry Associ
ation et al. by Carter G. Phillips and Richard D. Bernstein; and for the
Washington Legal Foundation by W. Reece Bader, James A. Meyers, Mi
chael C. Tu, and Daniel J. Popeo.
Brian Wolfman and Arthur R. Miller filed a brief for Law Professors
et al. as amici curiae.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
636 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of the Court
court. The question here is whether an order remanding a
case removed under the Securities Litigation Uniform
Standards Act of 1998 is appealable, notwithstanding
§ 1447(d). We hold it is not.
I
The Private Securities Litigation Reform Act of 1995 (Re
form Act), 109 Stat. 737, targeted “perceived abuses of the
class-action vehicle in litigation involving nationally traded
securities,” Merrill Lynch, Pierce, Fenner & Smith Inc. v.
Dabit, ante, at 81, and put limits on federal securities class
actions. But Congress soon discovered that “[r]ather than
face the obstacles set in their path by the Reform Act, plain
tiffs and their representatives [were] bringing class actions
under state law, often in state court,” ante, at 82. To block
this bypass of the Reform Act, Congress enacted the Securi
ties Litigation Uniform Standards Act of 1998 (Act), 112
Stat. 3227; see Dabit, ante, at 81–82.
The Act has a preclusion provision 1 and a removal provi
sion: 2 it provides that private state-law “covered” class ac
1 “No covered class action based upon the statutory or common law of
any State or subdivision thereof may be maintained in any State or Fed
eral court by any private party alleging—
“(1) an untrue statement or omission of a material fact in connection
with the purchase or sale of a covered security; or
“(2) that the defendant used or employed any manipulative or deceptive
device or contrivance in connection with the purchase or sale of a covered
security.” 112 Stat. 3228 (codified at 15 U. S. C. § 77p(b)).
The preclusion provision is often called a preemption provision; the Act,
however, does not itself displace state law with federal law but makes
some state-law claims nonactionable through the class-action device in fed
eral as well as state court. See Merrill Lynch, Pierce, Fenner & Smith
Inc. v. Dabit, ante, at 87 (“The Act does not deny any individual plaintiff,
or indeed any group of fewer than 50 plaintiffs, the right to enforce any
state-law cause of action that may exist”).
2 “Any covered class action brought in any State court involving a cov
ered security, as set forth in subsection (b) of this section, shall be remov
able to the Federal district court for the district in which the action is
pending, and shall be subject to subsection (b) of this section.” 112 Stat.
3228 (codified at 15 U. S. C. § 77p(c)).

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
637 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
tions alleging untruth or manipulation in connection with the
purchase or sale of a “covered” security may not “be main
tained in any State or Federal court,” 112 Stat. 3228 (codified
at 15 U. S. C. § 77p(b)),3 and it authorizes removal to federal
district court of “[a]ny covered class action brought in any
State court involving a covered security, as set forth in sub
section (b),” 112 Stat. 3228 (codified at § 77p(c)). “A ‘covered
class action’ is a lawsuit in which damages are sought on
behalf of more than 50 people. A ‘covered security’ is one
traded nationally and listed on a regulated national ex
change.” Dabit, ante, at 83 (footnotes omitted).
Petitioners are eight groups of investors holding mutual
fund shares, who filed separate actions in Illinois state
courts, each group seeking to represent a class of investors
allegedly injured by devaluation of their holdings by re
spondents (mutual funds, investment advisors, and an insur
ance company) (hereinafter collectively the funds).4 The
eight complaints asserted only state-law claims, such as neg
ligence and breach of fiduciary duty.
The funds filed notices of removal to federal district court
in each case stating, among other things, that the actions
were removable under and precluded by the Act. Once in
the District Court, however, the investors argued that the
cases should be remanded for lack of subject-matter jurisdic
tion, and in separate orders the District Court for the South
3 The Act amends “in substantially identical ways,” Dabit, ante, at 82,
n. 6, both the Securities Act of 1933, 48 Stat. 74, and the Securities Ex
change Act of 1934, 48 Stat. 881. For the sake of simplicity, the Seventh
Circuit relied exclusively on the amendments to the Securities Act of 1933,
and for ease of reference we will do the same.
4 The investors claim that the funds facilitated the practice of “market
timing,” whereby traders of mutual fund shares exploit brief discrepancies
between the stock prices used to calculate the shares’ value once a day,
and the prices at which those stocks are actually trading in the interim.
Brief for Petitioners 6. The investors say that market timing is harmful
to long-term holders of mutual fund shares and that the funds negligently
or recklessly failed to adopt procedures to protect the value of the inves
tors’ long-term investments.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
638 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of the Court
ern District of Illinois remanded each case to state court
on the ground that the District Court lacked subject-matter
jurisdiction on removal because the Act did not preclude the
investors’ claims. Since the investors were said to have
been injured as “holders” of mutual fund shares, not purchas
ers or sellers, the District Court reasoned, their claims did
not satisfy the “in connection with the purchase or sale” re
quirement of the Act’s preclusion provision, § 77p(b),5 and the
claims could therefore proceed in state court. The District
Court did not decide whether the claims otherwise met the
Act’s conditions for preclusion.
The funds filed notices of appeal from the remand orders,
and in one of the cases, 373 F. 3d 847 (2004), the Seventh
Circuit issued an opinion addressing the threshold question
of its appellate jurisdiction. The Court of Appeals acknowl
edged that 28 U. S. C. § 1447(d) bars review of district court
orders remanding for lack of subject-matter jurisdiction, 373
F. 3d, at 849 (citing Gravitt v. Southwestern Bell Telephone
Co., 430 U. S. 723 (1977) (per curiam)), but decided that the
District Court had the last word neither on the characteriza
tion of its decision as jurisdictional nor on the correctness of
its conclusion that remand was required, see 373 F. 3d, at 849.
The Court of Appeals considered all covered class actions
involving covered securities, whether precluded or not, to be
removable under the Act, and for that reason thought the
preclusion issue to be distinct from the jurisdictional issue of
whether the case belonged in federal court at all. Id., at
849–850. In the view of the Court of Appeals, if the District
Court remanded because, for example, the class comprised
too few investors to make the case a covered class action,
that would be a jurisdictional decision that the case had been
removed improperly, and the order would therefore be unre
viewable in accordance with § 1447(d). Id., at 849. But the
court held that orders remanding “properly removed” suits
5 As discussed in Part III, infra, we have since rejected this reasoning,
see Dabit, ante, at 88–89.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
639 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
as not precluded by the Act are substantive, “unaffected by
§ 1447(d),” id., at 851, and therefore subject to appellate ju
risdiction in the normal course.
As the Court of Appeals put it, once the District Court
had made that substantive decision of no preclusion in this
case, it was time for the court to bow out, not because it had
lacked “adjudicatory competence” to begin with but because
it had completed its work: “Once a court does all that the
statute authorizes, there is no adjudicatory competence to do
more. That is not the ‘lack of subject-matter jurisdiction’
that authorizes a remand. Otherwise every federal suit,
having been decided on the merits, would be dismissed ‘for
lack of jurisdiction’ because the court’s job was finished.”
Id., at 850. This remand, the court concluded, was therefore
not for want of jurisdiction, and review was not barred by
§ 1447(d).
To satisfy itself that its decision made “practical sense,”
the court proposed that the Act reserves to the Federal Judi
ciary the exclusive authority to make the preclusion decision.
Ibid. Treating remand orders in this context as immunized
from appeal by § 1447(d) would thus mean that “a major sub
stantive issue in the case [would] escape review,” since it
would not be open to resolution in the state court subject to
review by this Court. Ibid.
The Seventh Circuit subsequently consolidated the funds’
appeals and decided, on the merits, that the Act does pre
clude the investors’ claims. 403 F. 3d 478 (2005). We
granted certiorari to resolve a split of authority on the ques
tion whether § 1447(d) bars review of remand orders in cases
removed under the Act,6 546 U. S. 1085 (2006), and we now
vacate for want of jurisdiction on the part of the Court of
Appeals.
6 Compare 373 F. 3d 847 (CA7 2004) (case below) with Spielman v. Mer
rill Lynch, Pierce, Fenner & Smith, Inc., 332 F. 3d 116 (CA2 2003); Abada
v. Charles Schwab & Co., 300 F. 3d 1112 (CA9 2002); Williams v. AFC
Enterprises, Inc., 389 F. 3d 1185 (CA11 2004).

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
640 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of the Court
II
The policy of Congress opposes “interruption of the litiga
tion of the merits of a removed cause by prolonged litigation
of questions of jurisdiction of the district court to which the
cause is removed,” United States v. Rice, 327 U. S. 742, 751
(1946), and nearly three years of jurisdictional advocacy in
the cases before us confirm the congressional wisdom. For
over a century now, statutes have accordingly limited the
power of federal appellate courts to review orders remand
ing cases removed by defendants from state to federal court,
see id., at 748–752; Thermtron Products, Inc. v. Hermans
dorfer, 423 U. S. 336, 346–348 (1976). The current incarna
tion is 28 U. S. C. § 1447(d), which provides that an “order
remanding a case to the State court from which it was re
moved is not reviewable on appeal or otherwise.” 7 In
Thermtron, we held that the bar of § 1447(d) applies only to
remands based on the grounds specified in § 1447(c), that is,
a defect in removal procedure or lack of subject-matter juris
diction. 423 U. S., at 343–345; see also Things Remembered,
Inc. v. Petrarca, 516 U. S. 124, 127–128 (1995). So, we have
approved appellate review of a remand expressly based on
the District Court’s crowded docket, see Thermtron, supra,
at 340–341, and one based on abstention under Burford v.
Sun Oil Co., 319 U. S. 315 (1943), see Quackenbush v. All
state Ins. Co., 517 U. S. 706, 710–712 (1996). But we have
relentlessly repeated that “any remand order issued on the
grounds specified in § 1447(c) [is immunized from all forms of
appellate review], whether or not that order might be
deemed erroneous by an appellate court.” Thermtron, 423
U. S., at 351; see also id., at 343 (“If a trial judge purports to
remand a case on the ground that it was removed ‘improvi
dently and without jurisdiction,’ his order is not subject to
7 Title 28 U. S. C. § 1447(d) specifically excepts certain civil rights actions
from its bar; cf. § 1443.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
641 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
challenge in the court of appeals” (quoting § 1447(c) (1970
ed.))).
The bar of § 1447(d) applies equally to cases removed under
the general removal statute, § 1441, and to those removed
under other provisions, see Things Remembered, supra, at
128, and the force of the bar is not subject to any statutory
exception that might cover this case.8 Ostensibly, then,
§ 1447(d) stands in the way of reviewing the District Court’s
orders of remand in the present cases. The District Court
said that it was remanding for lack of jurisdiction, an unre
viewable ground, and even if it is permissible to look beyond
the court’s own label, the orders are unmistakably premised
on the view that removal jurisdiction under 15 U. S. C.
§ 77p(c) is limited to cases precluded by § 77p(b); on the Dis
trict Court’s understanding that “holder” claims are not sub
ject to preclusion under § 77p(b), the court had no subject
matter jurisdiction.9 Since there was no indication that
removal jurisdiction might exist on some ground other than
§ 77p(c) (complete diversity, for example),10 the remand or
8 “Absent a clear statutory command to the contrary, we assume that
Congress is aware of the universality of th[e] practice of denying appellate
review of remand orders when Congress creates a new ground for re
moval,” Things Remembered, 516 U. S., at 128 (internal quotation marks
omitted), like 15 U. S. C. § 77p(c). Congress has, when it wished, ex
pressly made 28 U. S. C. § 1447(d) inapplicable to particular remand orders.
See, e. g., § 1447(d); 12 U. S. C. § 1441a(l)(3)(C); § 1819(b)(2)(C); 25 U. S. C.
§ 487(d); cf. n. 7, supra. There is no such “clear statutory command” here,
and that silence tells us we must look to 28 U. S. C. § 1447(d) to determine
the reviewability of remand orders under the Act.
9 We take a pass on Justice Scalia’s position that we may not look
beyond the label, see post, at 650 (opinion concurring in part and concur
ring in judgment); the result here is the same whether we look near or far.
10 These cases raise exclusively state-law claims seeking damages insuf
ficient to satisfy the amount-in-controversy requirement of 28 U. S. C.
§ 1332; in those instances in which the funds asserted diversity as a basis
for subject-matter jurisdiction, the District Court determined that no
named plaintiff had a claim that met § 1332’s $75,000 threshold. See, e. g.,
Parthasarthy v. T. Rowe Price Int’l Funds, Inc., No. 03–CV–0673–DRH

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
642 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of the Court
ders were necessarily based on the trial court’s conclusion
that jurisdiction under § 77p(c) was wanting. And “[w]here
the order is based on one of the [grounds enumerated in 28
U. S. C. § 1447(c)], review is unavailable no matter how plain
the legal error in ordering the remand,” Briscoe v. Bell, 432
U. S. 404, 413–414, n. 13 (1977).
The Court of Appeals did not, of course, overlook the cases
holding that even a remand premised on an erroneous conclu
sion of no jurisdiction is unappealable; it relied instead on
cases like Kontrick v. Ryan, 540 U. S. 443 (2004), and Scar
borough v. Principi, 541 U. S. 401 (2004), which observed
that some rulings loosely called jurisdictional are patently
not jurisdictional in the strict sense, see 373 F. 3d, at 849
(citing Kontrick, supra; Scarborough, supra). The appeals
court saw this as such a case; it understood that a district
court had removal jurisdiction over any covered action under
subsection (c), with the consequence that a subsequent order
dismissing because of preclusion under subsection (b), or re
manding because the action was not precluded, rested simply
on an application of substantive law under subsection (b), law
that was not jurisdictional at all.
We think, however, that the District Court was correct in
understanding its remand order to be dictated by its find
ing that it lacked removal jurisdiction. Unlike the Court of
Appeals, we read authorization for the removal in subsec
tion (c), on which the District Court’s jurisdiction depends,
as confined to cases “set forth in subsection (b),” § 77p(c),
namely, those with claims of untruth, manipulation, and so
on. The quoted phrase immediately follows the subsection
(c) language describing removable cases as covered class ac
tions involving covered securities, and the language has no
apparent function unless it limits removal to covered class
actions involving claims like untruth or deception. And leg
(SD Ill., Jan. 30, 2004), App. to Pet. for Cert. 34a–37a; Spurgeon v. Pacific
Life Ins. Co., No. 04–CV–0355–MJR (SD Ill., June 24, 2004), id., at
59a–60a.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
643 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
islative history tends to show that this was just what Con
gress understood. See S. Rep. No. 105–182, p. 8 (1998)
(§ 77p(c) “provides that any class action described in Subsec
tion (b) that is brought in a State court shall be removable
to Federal district court, and may be dismissed pursuant to
the provisions of subsection (b)”); H. R. Rep. No. 105–640,
p. 16 (1998) (same).11
The funds argue that removal jurisdiction is broader by
emphasizing the adjective that introduces subsection (c):
“Any” covered action. § 77p(c). But that suggestion would
be persuasive only if we stopped reading right there, and we
do not stop there; we do not read statutes in little bites.
And, as just noted, if we did read the removal power that
broadly there would be no point to the phrase “as set forth in
subsection (b),” for subsection (b) cases would be removable
anyway as a subset of covered class actions. Ibid. The
funds purport to counter this objection with their argument
that on our reading the last phrase of subsection (c) is redun
dant in providing that removed cases “shall be subject to
subsection (b),” since subsection (b) cases would in any event
be so subject. Ibid. The funds are in fact right about that
redundancy, but the point does not count for their side, be
cause the phrase is redundant on their reading, too: any sub
section (b) case removed as falling within the broad category
of covered class actions would be treated in accordance with
subsection (b) if the subsection applied to that case. In sum,
we see no reason to reject the straightforward reading: re
moval and jurisdiction to deal with removed cases is limited
to those precluded by the terms of subsection (b).
Once removal jurisdiction under subsection (c) is under
stood to be restricted to precluded actions defined by subsec
11 Like the Court of Appeals here, we said in Dabit that a “key provision
of the [Act] makes all ‘covered class actions’ filed in state court removable.”
Ante, at 83, n. 7 (quoting 112 Stat. 3230). We sketched the removal provi
sion in broad strokes then because the question of its scope was not before
us. Now that it is, we speak more cautiously.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
644 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of the Court
tion (b), a motion to remand claiming the action is not pre
cluded must be seen as posing a jurisdictional issue. If the
action is precluded, neither the district court nor the state
court may entertain it, and the proper course is to dismiss.
If the action is not precluded, the federal court likewise has
no jurisdiction to touch the case on the merits, and the
proper course is to remand to the state court that can deal
with it. In either event, as the Court of Appeals said, the
district court’s order comes because its adjudicatory power
has been exercised and its work is done. But its adjudica
tory power is simply its authority to determine its own juris
diction to deal further with the case, see United States v.
Shipp, 203 U. S. 563, 573 (1906) (opinion for the Court by
Holmes, J.) (A federal court “necessarily ha[s] jurisdiction to
decide whether the case [is] properly before it”). The work
done is jurisdictional,12 as is the conclusion reached and the
order implementing it.13
12 The funds argue 15 U. S. C. § 77p confers jurisdiction greater than that
necessary to render the preclusion decision, analogizing § 77p(c) to the fed
eral officer removal statute, 28 U. S. C. § 1442(a). If there is any colorable
claim that an action is precluded, the argument goes, the district court can
keep the case for adjudication, even after concluding on the merits that
the state-law claims are not precluded; but because it has discretion to
keep the case or remand to state court, a remand is not jurisdictional
and hence is reviewable. The argument is flawed for two reasons. The
District Court here did not indicate it thought there was any basis to keep
the case for further development; right or wrong, it understood that it was
making a jurisdictional ruling. Nor is the analogy with federal officer
cases sound.
Section 1442(a) is an exception to the “well-pleaded complaint” rule,
under which (absent diversity) “a defendant may not remove a case to
federal court unless the plaintiff ’s complaint establishes that the case
‘arises under’ federal law.” Franchise Tax Bd. of Cal. v. Construction
Laborers Vacation Trust for Southern Cal., 463 U. S. 1, 10 (1983) (empha
sis deleted). The federal officer removal statute allows “suits against fed
eral officers [to] be removed despite the nonfederal cast of the complaint,”
Jefferson County v. Acker, 527 U. S. 423, 431 (1999), and reflects a congres
[Footnote 13 is on p. 645]

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
645 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
III
We have yet to deal with one objection to our application
of § 1447(d), which if well taken would be a serious one. The
sional policy that “federal officers, and indeed the Federal Government
itself, require the protection of a federal forum,” Willingham v. Morgan,
395 U. S. 402, 407 (1969). An officer’s federal defense need be only color
able to assure the federal court that it has jurisdiction to adjudicate the
case, see Acker, supra, at 431.
The funds assert that a preclusion defense need be only colorable as
well, but the Act is different. It avails a defendant of a federal forum in
contemplation not of further litigation over the merits of a claim brought
in state court, but of termination of the proceedings altogether, and a
merely colorable claim of preclusion does not satisfy a district court that
it may dismiss a case as precluded by the Act. There is no room for such
a case to exist in a limbo of colorable preclusion; if a claim is precluded, it
“may [not] be maintained,” 15 U. S. C. § 77p(b), and if the claim is not, the
federal courts no longer have any business being involved, as there is no
longer any federal question on which to moor the district court’s jurisdic
tion. Nor has Congress expressed in the Act, as it did with 28 U. S. C.
§ 1442(a), any policy of having particular suits tried in a federal court;
there is no indication whatsoever in the Act that, apart from its purpose
to preclude certain vexing state-law class actions, Congress intended to
add other state-law cases to the federal dockets, and there is no apparent
federal interest in spending time on such cases akin to the interest in
adjudicating suits against federal officers.
13 The funds suggest, in the alternative, that appellate jurisdiction in
this case was proper under Waco v. United States Fidelity & Guar
anty Co., 293 U. S. 140 (1934). Without passing on the continued vitality
of that case in light of § 1447(d), we note that on its own terms it is
distinguishable.
In Waco, a case was removed to federal court on an invocation of diver
sity jurisdiction, id., at 141, and the District Court thereafter “entered a
single decree embodying . . . separate orders,” id., at 142. In one order,
the District Court dismissed a cross-complaint against one party. In an
other, the District Court concluded that because of the dismissal there
was no diversity of citizenship and it thus lacked jurisdiction, and so it
remanded the case to state court. An appeal was taken from the order
of dismissal. This Court determined that the appeal would lie, because
“the decree of dismissal preceded that of remand,” and because the Dis
trict Court’s order of dismissal was conclusive upon the parties. Id., at
143. We noted that a “reversal [of the dismissal] cannot affect the order

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
646 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of the Court
Seventh Circuit’s reading of subsection (c) so as to treat the
application of the preclusion rule as nonjurisdictional was in
part motivated by its assumption that the Act gives federal
courts exclusive jurisdiction to decide the preclusion issue.
If that is so, and § 1447(d) applies, a remand order based on
a finding that an action is not precluded would arguably be
immune from review. This is what the funds in effect con
tend here when they say that a district court’s finding of no
subsection (b) preclusion would collaterally estop the state
court on remand; the district court would have the last word.
And of course the funds’ discomfort is made acute by our
recent decision in Dabit, which expressly disavows the
district court’s limited view of the scope of subsection (b)
preclusion.
But a district court does not have the last word on preclu
sion under the Act, for nothing in the Act gives the federal
courts exclusive jurisdiction over preclusion decisions. A
covered action is removable if it is precluded, and a defend
ant can enlist the Federal Judiciary to decide preclusion, but
a defendant can elect to leave a case where the plaintiff filed
it and trust the state court (an equally competent body, see
Missouri Pacific R. Co. v. Fitzgerald, 160 U. S. 556, 583
(1896)) to make the preclusion determination.
And what a state court could do in the first place it may
also do on remand; in this case, the funds can presently argue
the significance of Dabit and ask for dismissal on grounds of
of remand, but it will at least, if the dismissal . . . was erroneous, remit
the entire controversy, with the [previously dismissed party] still a party,
to the state court for . . . further proceedings.” Id., at 143–144.
The order appealed in Waco was not a remand order; the order here is,
and thus falls within § 1447(d)’s bar on appeals of “[a]n order remanding a
case” to state court. Moreover, the funds do not explain how to reconcile
their argument with Waco’s acknowledgment that the order of remand
“cannot [be] affect[ed]” notwithstanding any reversal of a separate order,
id., at 143. The District Court’s remand order here cannot be disaggre
gated as the Waco orders could, and if the Seventh Circuit’s preclusion
decision stands, there is nothing to remand to state court.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
647 Cite as: 547 U. S. 633 (2006)
Opinion of the Court
preclusion when they return to the state court. Collateral
estoppel should be no bar to such a revisitation of the pre
clusion issue,14 given that § 1447(d) prevents the funds from
appealing the District Court’s decision. See Standefer v.
United States, 447 U. S. 10, 23 (1980) (“[C]ontemporary prin
ciples of collateral estoppel . . . strongly militat[e] against
giving an [unreviewable judgment] preclusive effect” (citing
Restatement (Second) of Judgments § 68.1 (Tent. Draft No. 3,
1976))); see also Restatement (Second) of Judgments § 28(1)
(1980) (“Although an issue is actually litigated and deter
mined by a valid and final judgment, and the determination
is essential to the judgment, relitigation of the issue in a
subsequent action between the parties is not precluded
[when t]he party against whom preclusion is sought could
not, as a matter of law, have obtained review of the judgment
in the initial action”). While the state court cannot review
the decision to remand in an appellate way, it is perfectly
free to reject the remanding court’s reasoning, as we ex
plained over a century ago in Missouri Pacific R. Co.: “[A]s
to applications for removal on the ground that the cause
arose under the Constitution, laws, or treaties of the United
States,” the finality accorded remand orders is appropriate
because questions of this character “if decided against the
claimant” in state court are “open to revision . . . , irrespec
tive of the ruling of the [federal court] in that regard in the
matter of removal.” 160 U. S., at 583. Nor is there any
reason to see things differently just because the remand’s
basis coincides entirely with the merits of the federal ques
tion; it is only the forum designation that is conclusive.
Here, we have no reason to doubt that the state court will
duly apply Dabit’s holding that holder claims are embraced
14 Modern usage calls for the descriptive term, “issue preclusion,” in
place of “collateral estoppel.” But we are backsliders out of pity for the
tired reader; “preclusion” by statutory fiat is enough preclusion for one
opinion.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
648 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of Scalia, J.
by subsection (b),15 but any claim of error on that point can
be considered on review by this Court. See Franchise Tax
Bd. of Cal. v. Construction Laborers Vacation Trust for
Southern Cal., 463 U. S. 1, 12, n. 12 (1983) (“If the state
courts reject a claim of federal pre-emption, that decision
may ultimately be reviewed on appeal by this Court” (citing
Fidelity Fed. Sav. & Loan Assn. v. De la Cuesta, 458 U. S.
141 (1982))).
IV
We hold that the Act does not exempt remand orders from
28 U. S. C. § 1447(d) and its general rule of nonappealability.
We therefore vacate the judgment of the Court of Appeals
and remand the case with instructions to dismiss the appeal
for lack of jurisdiction.
It is so ordered.
Justice Scalia, concurring in part and concurring in the
judgment.
I join the judgment of the Court, and Parts I, III, and IV
of the Court’s opinion; I do not join Part II for the reasons
set forth below.
The District Court ordered these cases remanded to state
court for want of jurisdiction. We know this because the
orders say so: “Because the Court lacks subject matter juris
diction, the Court REMANDS this action to the Madison
County, Illinois Circuit Court.” App. to Pet. for Cert. 27a;
see also id., at 30a, 40a, 46a, 51a, 57a, 64a. Even if those
decisions were incorrect, the Court of Appeals lacked juris
15 The parties further dispute whether the investors’ claims satisfy the
other 15 U. S. C. § 77p(b) preclusion prerequisites, particularly the allega
tion of fraud; the investors take issue with the Seventh Circuit’s character
ization of their claims as charging fraud or manipulation, not mismanage
ment. Because the Court of Appeals lacked appellate jurisdiction, its
reading of the investors’ litigation position is not binding in future pro
ceedings and is open to consideration on remand.

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
Cite as: 547 U. S. 633 (2006) 649
Opinion of Scalia, J.
diction to review them because 28 U. S. C. § 1447(d) bars ap
pellate review of remand orders based on lack of subject
matter jurisdiction. See, e. g., Things Remembered, Inc. v.
Petrarca, 516 U. S. 124, 127–128 (1995). The Court correctly
concludes that the Seventh Circuit’s review of the remand
orders overstepped its appellate authority. I disagree with
the Court’s reasoning in Part II, however, because it holds
only that the Court of Appeals’ recharacterization was incor
rect, and not (as I believe) that recharacterization—being a
form of review—is categorically forbidden.
The Court of Appeals rejected the District Court’s de
scription of its orders because it believed the District Court
had been too loose in its use of the term “jurisdiction.” 373
F. 3d 847, 849–850 (2004). What the District Court actually
did, the Court of Appeals concluded, was to remand on non
jurisdictional grounds (not subject to the appellate-review
bar of § 1447(d)) after deciding that petitioners’ suits were
not precluded. Such recharacterization seems to me flatly
inconsistent with § 1447(d). Under that section, an “order
remanding a case to the State court from which it was re
moved is not reviewable on appeal or otherwise.” Ibid.
(emphasis added). But appellate review is exactly what is
involved in looking behind the face of an order to determine
its true basis: In order to reject a district court’s own charac
terization, a court of appeals must decide, as the Seventh
Circuit did here, that the district court was wrong. We have
therefore held, in language that makes plain the correct out
come here, that “[i]f a trial judge purports to remand a case
on the ground that it was removed ‘improvidently and with
out jurisdiction,’ his order is not subject to challenge in the
court of appeal, by mandamus, or otherwise.” Thermtron
Products, Inc. v. Hermansdorfer, 423 U. S. 336, 343 (1976)
(quoting § 1447(c) (1970 ed.); emphasis added). Whether the
District Court was right or wrong—even if it was so badly
mistaken that it misunderstood the true basis for its or

547US2 Unit: $U66 [03-16-09 16:25:54] PAGES PGT: OPLG
650 KIRCHER v. PUTNAM FUNDS TRUST
Opinion of Scalia, J.
ders—it purported to remand for lack of jurisdiction, and
§ 1447(d) bars any further review.*
Review of the sort engaged in by the Court of Appeals
threatens to defeat the purpose of § 1447(d). As we recog
nized in Thermtron Products, the appellate-review bar was
enacted “to prevent delay in the trial of remanded cases by
protracted litigation of jurisdictional issues.” Id., at 351.
Such delay can be created just as easily by asking whether
the district court correctly characterized the basis for its
order as it can by asking whether that basis was correct—
which even the Court of Appeals recognized was beyond its
jurisdiction, 373 F. 3d, at 849. See also Thermtron Prod
ucts, supra, at 343 (noting that § 1447(d) “prohibits review of
all remand orders [based on lack of subject-matter jurisdic
tion] whether erroneous or not”). The remand orders in
these cases date back to early 2004; over two years later,
federal courts are still engaged in appellate review.
The Court should end this delay by holding that appellate
courts cannot look behind the stated basis for the district
court’s remand order. Instead, it concludes that “the Dis
trict Court was correct in understanding its remand order to
be dictated by a finding that it lacked removal jurisdiction.”
Ante, at 642 (emphasis added). It seems to me no more
within our authority to declare the District Court’s views
correct than it was within the Court of Appeals’ authority
to reject them. Either decision is an exercise of appellate
review barred by the plain terms of § 1447(d).
*To say that we cannot recharacterize the District Court’s remand for
lack of jurisdiction is not to say that the basis for the remand is forever
insulated from review. Part III of the Court’s opinion makes clear that
the underlying legal issue of preclusion remains open in state court, and
need not be resolved in accordance with the (unreviewable) views of the
District Court.