We are presented here with a casebook-ready fact pattern implicating an area of Fourth Amendment law that has long been a source of confusion. Today we explain, as clearly as we can, how the exclusionary rule applies in cases in which evidence obtained during an illegal traffic stop is introduced against a passenger with no possessory interest in the vehicle.

We hold that when a vehicle is illegally stopped by the police, no evidence found during the stop may be used by the government against any occupant of the vehicle unless the government can show that the taint of the illegal stop was purged. The metaphorical bubble of causation encapsulates the entire vehicle and links the illegality of the stop to the Fourth Amendment rights of all of the occupants.

In so holding, we join all of our sister circuits that have directly faced this issue. We will canvass that caselaw, and explain why we agree with it.

On the night of October 28, 2003, Robert Mosley went to the Diamond Dolls nightclub in Philadelphia with his nephew Jerome Small, who drove. While they were at the club, Small received a telephone call from a romantic acquaintance, and told Mosley that he was leaving the club to go meet her. Not wanting to leave Mosley without a ride home, Small introduced Mosley to his friend Julian Hayes, who agreed to drop Mosley off on his way home. At around 1:30 a.m., Mosley left the club with Hayes and Erica Scott, a dancer at the club who was accompanying Hayes. Hayes and Scott got in the front seat of Hayes' vehicle, a green Suzuki SUV, and Mosley got in the back.

At about the same time, a police radio call went out advising officers to be on the lookout for a black man with dreadlocks driving a green SUV. The source of the information relayed in the radio call is not reflected in the record. Police officers on patrol in the neighborhood of the nightclub heard the call and shortly thereafter saw Hayes' SUV, a green SUV with a black driver, as it was pulling away from the nightclub. They immediately pulled the car over. Upon approaching the car, the responding officers observed a gun on the floor under the driver's seat. They then ordered Hayes, Scott, and Mosley to get out of the car, and searched it, recovering a second gun from the front seat, two from the floorboards of the back seat area, and one from the back seat itself. Hayes and Mosley were arrested and charged with gun possession.

However, the Supreme Court has held that anonymous tips do not provide sufficient justification for an investigatory stop, see Florida v. J.L., 529 U.S. 266, 120 S.Ct. 1375, 146 L.Ed.2d 254 (2000), and the officers did not observe Hayes committing any traffic violation that would have justified the stop under Whren v. United States, 517 U.S. 806, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996). The government conceded that the stop was illegal, and dropped all charges against Hayes.

The government proceeded, however, with the gun possession case against Mosley, arguing that because he was a passenger in the vehicle, he could not seek to suppress the guns, notwithstanding the illegality of the stop. Mosley contended that insofar as he had been illegally seized by the traffic stop, he should have the same suppression claim as Hayes. The District Court agreed with the government, and admitted into evidence the guns found in the back seat of the vehicle.1 Mosley was convicted under 18 U.S.C. § 922(g) for possessing a firearm following a felony conviction.2 He appeals on several grounds; we will decide the case on the suppression issue.3 "We review the denial of a motion to suppress for clear error as to the underlying factual determinations and exercise plenary review over the application of the law to those facts." United States v. Williams, 417 F.3d 373, 376 (3d Cir.2005).

When one peruses the traffic-stop suppression caselaw, one is struck by how rarely a traffic stop is found to have been illegal. In Whren v. United States, 517 U.S. 806, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996), the Supreme Court established a bright-line rule that any technical violation of a traffic code legitimizes a stop, even if the stop is merely pretext for an investigation of some other crime. And once a car has been legally stopped, the police may "escalate" the encounter by visually inspecting the interior of the car, and checking credentials and asking questions of the occupants. See United States v. Givan, 320 F.3d 452, 458 (3d Cir.2003) ("After a traffic stop that was justified at its inception, an officer who develops a reasonable, articulable suspicion of criminal activity may expand the scope of an inquiry beyond the reason for the stop and detain the vehicle and its occupants for further investigation."). Courts give considerable deference to police officers' determinations of reasonable suspicion, see, e.g., United States v. Nelson 284 F.3d 472, 482 (3d Cir.2002), and the cases are steadily increasing the constitutional latitude of the police to pull over vehicles.4

Passengers in cars, unlike owners or licensees, have no reasonable expectation of privacy in the interior of the vehicle in which they are riding. Because the Fourth Amendment's protection against unreasonable searches is predicated on the invasion by the government of a person's reasonable expectation of privacy, passengers are generally held to lack "standing"5 to object to evidence discovered in a search of a vehicle. See Rakas v. Illinois, 439 U.S. 128, 99 S.Ct. 421, 58 L.Ed.2d 387 (1978). Fourth Amendment rights are personal rights, and a search of a car does not implicate the rights of non-owner passengers: the car is treated conceptually like a large piece of clothing worn by the driver.

But we should not be distracted by the fact that this case involves evidence found in a car. This is not an "auto search" case. The search of the car is not before us; the seizure of Mosley is. This case is about an illegal seizure by the police of the defendant, pursuant to which evidence was discovered. The violation of Mosley's Fourth Amendment rights was the traffic stop itself, and it is settled law that a traffic stop is a seizure of everyone in the stopped vehicle, see Delaware v. Prouse, 440 U.S. 648, 653, 99 S.Ct. 1391, 59 L.Ed.2d 660 (1979).6 Thus passengers in an illegally stopped vehicle have "standing" to object to the stop,7 and may seek to suppress the evidentiary fruits of that illegal seizure under the fruit of the poisonous tree doctrine, as expounded in the line of cases following Wong Sun v. United States, 371 U.S. 471, 83 S.Ct. 407, 9 L.Ed.2d 441 (1963). The dispositive legal issue is the causal relationship between the traffic stop and the discovery of the evidence: whether the evidence found in the car was "fruit" of the illegal stop. The question that bedeviled the proceedings below is whether the evidence found in the car was causally linked to the illegal seizure of Mosley, the passenger.

In analyzing that causal connection, the District Court relied on a Tenth Circuit case, United States v. DeLuca, 269 F.3d 1128 (10th Cir.2001), in which a passenger sought to suppress evidence found during a traffic stop. Taking its cue from that case, the District Court proposed various hypotheticals to test the "factual nexus" between the violation of Mosley's Fourth Amendment rights and the discovery of the evidence. Specifically, the District Court pondered whether, if Mosley had asked for and received permission to leave the scene immediately after the car was stopped, the evidence in the car would still have been discovered. Since the answer to that question is obviously "yes," the District Court ruled that Mosley had failed to demonstrate an adequate "factual nexus" between the illegal government action and the discovery of the challenged evidence. Although the government had clearly violated Mosley's Fourth Amendment rights, the District Court determined that the specific violation of Mosley's rights—as opposed to Hayes' rights—did not have a sufficiently close causal connection, or "factual nexus," to the discovery of the evidence to support Mosley's suppression claim.

The general requirement of a "factual nexus" between a specific Fourth Amendment violation and a specific piece of evidence derives from two Supreme Court cases on wiretap evidence. In those cases, the government had gathered thousands of discrete pieces of evidence over many months of investigation. Over the course of the investigation, the government committed various illegal acts, and the question for the Court was how to determine which pieces of evidence were tainted by the particular illegal actions. See Nardone v. United States, 308 U.S. 338, 341, 60 S.Ct. 266, 84 L.Ed. 307 (1939); Alderman v. United States, 394 U.S. 165, 183, 89 S.Ct. 961, 22 L.Ed.2d 176 (1969). Because of the multiplicity and complexity of the investigation and the evidence, there was no commonsense causal relationship between any given Fourth Amendment violation and any given piece of evidence. Rather than hold all the evidence to have been tainted by the violations, the Court held that the proper course is for district courts to probe more deeply, and employ thought experiments to determine which violation was causally connected to which piece of evidence.8 These cases express the commonsense proposition that a single Fourth Amendment violation does not taint an entire case, but only the evidence uncovered as a result of that violation.

Prior to DeLuca, courts had not generally thought that traffic stops presented sufficiently complex investigatory contexts to warrant such an added layer of causal analysis. In DeLuca, however, a panel of the Tenth Circuit, over a heated dissent by former Chief Judge Seymour, applied that counterfactual—"What if?"—approach to a suppression motion brought by a passenger who was illegally detained during a traffic stop. The application of "factual nexus" hypotheticals to a traffic stop suppression case was a novel idea. The majority view in the circuits was and remains that in a traffic stop, there will always be a sufficient "nexus" between the stop and the search, unless there are significant intervening events that sever or attenuate the causal chain. There is generally no "nexus" problem in illegal traffic stop cases, as the leading treatise puts it, because the connection between the illegal action and the discovery of the evidence is straightforward: If the police had not pulled over the vehicle, they would not have discovered the evidence. See 6 Wayne R. LaFave, Search and Seizure § 11.4(d) (4th ed.2004) (summarizing caselaw).

We can understand why the DeLuca majority thought that the case required some revision to the traditional fruits analytical apparatus. DeLuca did not involve an illegal traffic stop. It involved instead a legal traffic stop that became illegal when the police continued to detain the vehicle and its occupants after the legitimate purposes of the stop had been completed.9 Because the stop itself was legal, the court reasoned that whatever government actions violated the defendant's rights must be analytically separable from the traffic stop. The Fourth Amendment violation in DeLuca was the detention of the defendant passenger, along with the other occupants of the car, beyond the period necessary to complete the legitimate purposes of the stop. But at the point that the detention of the defendant passenger became illegal, the police already had control over the vehicle. The court therefore tried to tease out the causal efficacy of the precise illegality raised by the defendant passenger. The court asked whether, if the passenger had been allowed to leave the scene before the detention became illegal, the incriminating evidence would still have been discovered. The answer was "yes": it wasn't his car, so his leaving would not have removed the evidence.

The question in DeLuca was thus whether the police would still have discovered the challenged evidence if they had let the passenger go before doing anything illegal to him. On the DeLuca facts, if the police had let the passenger go after the initial legal stop but before the subsequent illegal prolongation of the stop, then nothing illegal would have been done to him— and the evidence would still have been discovered. Therefore the passenger had no suppression claim. So reasoned the Tenth Circuit.10

But we agree with Mosley that DeLuca is inapposite here. The hypothetical, and the holding, are relevant only to situations in which the initial traffic stop is legal, and that is not our case. We express no opinion on the viability in this Circuit of the DeLuca test on DeLuca facts; we decide the case before us, not a different one.11 But whatever its viability on appropriate facts, this case does not present such facts. If the initial traffic stop is illegal, then even if the passenger is allowed to leave the scene before the search, it will not be the case that the police have not violated his Fourth Amendment rights. Where the traffic stop itself is illegal, it is simply impossible for the police to obtain the challenged evidence without violating the passenger's Fourth Amendment rights. DeLuca says nothing about such situations.

The Tenth Circuit itself has addressed the question whether DeLuca applies to traffic stops illegal from their inception, and stated clearly that it does not. It applies, the court explained in a recent suppression case applying DeLuca, only to cases in which the illegal police conduct occurred after the police had legally gained control of the vehicle.12 The temporal sequence of events makes all the difference, said the court:

[Wong Sun ] and its progeny, see, e.g., United States v. Melendez-Garcia, 28 F.3d 1046, 1053-54 (10th Cir.1994), are readily distinguishable insofar as in those cases the illegal police conduct preceded the means by which the evidence was obtained, thus establishing the requisite factual nexus between the evidence and the illegal conduct. By contrast, any unlawful police activity here occurred after voluntary consent had been obtained. Consequently, [the defendant] could not establish that, but for the alleged illegal seizure, the evidence would not have come to light as required by DeLuca.

United States v. Roberts, 91 Fed.Appx. 645, 648 (10th Cir.2004) (emphasis added).

The District Court's application of DeLuca to the case at bar, in which the illegal police conduct preceded the means by which the evidence was obtained, was error, because even under the Tenth Circuit's heightened "factual nexus" test, such a temporal scenario would appear to clearly supply the requisite "factual nexus."13

It is clear, so far, that an illegal traffic stop constitutes a seizure of all occupants of a vehicle, including passengers, and that whatever the utility of the DeLuca counterfactual analysis to fact patterns not before us, it is not the proper analytical model for the case that is before us.

Fourth Amendment rights are personal rights. That proposition is not in dispute. The problem is that from that proposition, the result in any particular case is not always immediately obvious. Mosley does not dispute that he had no personal expectation of privacy in the interior of the car; the government does not dispute that he did have a reasonable expectation that he would not be seized without probable cause. The problem is whether the evidence in this case was discovered as a result of his seizure.

As we will detail below, most courts treat evidence found during an illegal traffic stop as the fruits of that stop, and see no conceptual difficulties in suppressing such evidence when introduced against passengers. Our most recent decision to face such a scenario, United States v. Williams, 413 F.3d 347 (3d Cir.2005), concerned a non-owner passenger in a parked van that was approached by two police officers, who subsequently observed contraband in the vehicle. The defendant claimed that the approach constituted a seizure. We framed the suppression issue as follows:

The central issue on appeal ... is whether Williams was seized within the meaning of the Fourth Amendment when the police ... approached the van. If Williams was "seized" by the police when they approached the parked van without probable cause or reasonable suspicion to do so, then the District Court correctly suppressed all evidence obtained in connection with the ensuing arrest under the "fruits of the poisonous tree" doctrine.

413 F.3d at 351. Our approach can be summarized as follows: We ask, first, whether there was a seizure. If so, then we ask whether there was reasonable suspicion or probable cause. If not, then suppression is proper unless the government can show that the taint was purged. We did not even comment in Williams on the "causal nexus" between the alleged seizure and the discovery of evidence; on these facts we took that nexus to be self-evident.

However, the quoted passage is dicta because we held both that the approach was not a seizure and that the officers had reasonable suspicion sufficient for a Terry stop in any event. The vehicle was parked in a public place and the officers did not make a show of force, and "law enforcement officers do not violate the Fourth Amendment's prohibition of unreasonable seizures merely by approaching individuals on the street or in other public places." Id. at 352 (quoting United States v. Drayton, 536 U.S. 194, 200, 122 S.Ct. 2105, 153 L.Ed.2d 242 (2002)).

Though dicta, the analytical approach suggested by Williams is consistent with the approach proposed by the leading Fourth Amendment treatise.

In the typical case in which an illegal arrest is followed by a search no "fruits" problem of any magnitude is presented.... This is direct rather than derivative evidence, and there is no occasion to be concerned about the limits of the fruit of the poisonous tree doctrine.... Problems of this kind arise in the context of passengers in vehicles attempting to have evidence found therein suppressed.... [T]he passenger will claim that the subsequent discovery of the evidence in the vehicle was a fruit of his prior illegal detention. It is very often the case that such a connection is readily made out when the initial stopping of the vehicle (and thus all of the occupants) was illegal.

6 LaFave, supra, § 11.4(d) (emphasis added).

We have not, however, yet had occasion to decide a case requiring a precise articulation of the application of the fruit of the poisonous tree doctrine to evidence found during illegal traffic stops when introduced against vehicle passengers. We will therefore state the theoretical issue as perspicuously as we can: Is an illegal traffic stop of a car occupied by a driver and a passenger a single constitutional violation, with two victims, each of whom can seek to suppress all fruits of that violation? Or is it analytically separable into two individual constitutional violations, each with one victim, each of whom may seek to suppress only the fruits of the violation of his individual right? Today we endorse the former proposition, but as we acknowledge the logical appeal of the latter proposition, we will set out the best arguments in its favor, and then explain why we are rejecting them.

The "analytic separation" position, the position we are rejecting, can be defended both on logical grounds and via a reductio ad absurdum. Here are the logical grounds. Because Fourth Amendment rights are personal rights, when the police pull over a car with a driver and a non-owner passenger, two individual seizures occur simultaneously: seizure of the driver, and seizure of the passenger.14 Thus, we should not see the traffic stop as one discrete constitutional violation (with two victims), but rather as two separate constitutional violations (each with one victim). Because fruits analysis always proceeds from the specific constitutional violation inflicted on the defendant, this analytic separation of violations necessitates a concomitant separation of the respective fruits suppression analyses. Because the violation of the passenger's constitutional rights was the seizure of the passenger (analytically separated from the seizure of the driver), we must ask, in any fruits analysis, whether the challenged evidence was seized as a causal result of the seizure of the passenger. To get at the interior of the car, where the evidence is located, the police must perforce seize the driver and the passenger, by stopping the car. But that one act constitutes two separate seizures. The passenger was not in control of the car, so it is the seizure of the driver, and not the passenger, that is required in order for the police to obtain control over the car. The passenger's presence has no effect on the ability of the police to seize the driver and control the car. Thus the evidence in the car would be found whether or not the passenger was present in the car at all. Seizure of the passenger is, accordingly, logically ancillary to the search of the vehicle. It is logically necessary only for a search of the passenger's person. On this view, therefore, evidence found in the car, but not on the passenger's person, cannot be suppressed by the passenger, even though the seizure of the passenger was unconstitutional.

Some simple hypotheticals bring out the force of this argument. Suppose that the car stops and X, a passenger, gets out at the corner. Just as the car begins to pull away, the police, acting on an anonymous tip, roar up and stop the car, simultaneously ordering X, on the sidewalk, to freeze. Evidence is found in the car but is suppressed as to the driver because the stop was illegal. It is then introduced against X. May X suppress? Certainly not. The two seizures are clearly separate causal events for Fourth Amendment purposes. Nor can X challenge a search of the car, no matter how illegal, because he has no privacy interest in the car.

Now put X back in the car. What is the legal significance of that change of place? The most obvious difference is that it becomes physically impossible for the police to stop the car (and detain the driver) without also seizing and detaining X. But X's legal situation with respect to the car and driver remains the same: as a passenger, his privacy interest in the interior of the car is precisely what it was when he was standing on the sidewalk—zero. Thus, just as X's seizure on the sidewalk cannot logically or legally have been a cause of the discovery of evidence inside the car, neither can his seizure while inside the car as a passenger be a logical or legal cause of the discovery of evidence inside the car.

The discovery of evidence inside a car simply cannot be a violation of a passenger's Fourth Amendment rights, according to this argument. The passenger has a personal right not to be seized, and that right is violated when the car is illegally pulled over. But the illegal seizure of the passenger's person cannot be a legal cause of the discovery of evidence that is not on the passenger's person. The police may have performed both acts at the same time, but because there was no logical connection between the two the illegal seizure was not a but-for cause of the discovery of the evidence.

Here is the reductio ad absurdum. It is settled law that if the police illegally enter a house and search it, the owner or tenant of the house, or any long-term guests, may suppress evidence found during the search.15 Short-term guests, however, have no expectation of privacy in the house and therefore cannot suppress the fruits of the illegal search. This is true even if the short-term guests were seized pursuant to the illegal entry into the house. If, for example, the police barricaded the exits to a residence, entered, and ordered everyone present to freeze, they would have thereby seized everyone present in the residence. If their entry was illegal, for example if effected without a warrant and absent exigent circumstances, then all of those seizures were illegal. However, only the owner, tenant, or long-term guests could suppress evidence found during the search. Short-term guests would face the full evidentiary weight of any evidence discovered. See Minnesota v. Olson, 495 U.S. 91, 110 S.Ct. 1684, 109 L.Ed.2d 85 (1990).

Now make the house a mobile home, and drive it down the street until it is pulled over illegally. The police again illegally seize all the occupants and search the interior. Under the holding we are announcing today—which conforms, as we observed above, with those of our sister circuits—those short-term guests are now passengers in an illegally seized vehicle, and thus can suppress all evidence discovered during the search. This result appears anomalous insofar as the Fourth Amendment has been repeatedly characterized by the Supreme Court as affording enhanced protection to the home, and diminished protection to vehicles. Compare, e.g., Georgia v. Randolph, 547 U.S. ___, 126 S.Ct. 1515, 164 L.Ed.2d 208 (2006) (citing Wilson v. Layne, 526 U.S. 603, 610, 119 S.Ct. 1692, 143 L.Ed.2d 818 (1999)) (invoking the "centuries-old principle of respect for the privacy of the home"); Minnesota v. Carter, 525 U.S. 83, 99, 119 S.Ct. 469, 142 L.Ed.2d 373 (1998) (Kennedy, J., concurring) ("[I]t is beyond dispute that the home is entitled to special protection as the center of the private lives of our people."); Miller v. United States, 357 U.S. 301, 307, 78 S.Ct. 1190, 2 L.Ed.2d 1332 (1958) (noting common law tradition that a poor man in a humble cottage might nonetheless bid defiance to all the forces of the crown (if it is his cottage)), with, e.g., United States v. Martinez-Fuerte, 428 U.S. 543, 561, 96 S.Ct. 3074, 49 L.Ed.2d 1116 (1976) (explaining that Fourth Amendment protections are lessened for occupants of cars because cars searches do not implicate "the sanctity of private dwellings, ordinarily afforded the most stringent Fourth Amendment protection."). Yet in our hypothetical, passengers in a vehicle are afforded greater Fourth Amendment protection than guests in a home.

We have set out a coherent logical argument for an analytic separation rule, and highlighted an apparently anomalous contrast between vehicle and home seizure cases that rejection of that rule might seem to entail. But we are not persuaded. We will respond to the argument in two ways. First, in Part III.B, we will survey the caselaw of our sister circuits, all of which uphold passenger suppression claims following illegal stops. Second, in Part III.C, we will reject the apparent logical plausibility of the analytic separation argument as inconsistent with the more pragmatic methodology exemplified by the Supreme Court's exclusionary rule jurisprudence.

The Ninth and Eleventh Circuits have recently decided cases with precisely analogous facts. See United States v. Twilley, 222 F.3d 1092 (9th Cir.2000); United States v. Chanthasouxat, 342 F.3d 1271 (11th Cir.2003). In each case, the traffic stop was illegal from its inception,16 and evidence found during the traffic stop was introduced against a passenger in the car. In each case, the government contested the passenger's motion to suppress by arguing that the passenger lacked "standing" to challenge the evidence because he did not have a reasonable expectation of privacy in the interior of the car. In each case, the court applied a standard fruits analysis, and suppressed the evidence as the fruit of the illegal seizure.17

The two courts treated this fact pattern as an utterly straightforward and unremarkable application of the fruit of the poisonous tree doctrine. Their analyses proceed precisely as Mosley argues ours should, in the following three steps. First, the traffic stop was a violation of the defendant's Fourth Amendment rights; second, the evidence was found as a direct consequence of the stop (that is, there were no significant intervening events); third, the government did not establish any of the various available exceptions (attenuation, independent source, inevitable discovery). The Twilley court held that "[b]ecause we conclude that the stop was not supported by reasonable suspicion, and because the subsequent search was a product of the stop, the evidence leading to Twilley's conviction should have been suppressed." Id. at 1097.18 The Chanthasouxat court held that "[b]ecause we conclude that the initial traffic stop violated the Fourth Amendment, the violation was not cured by voluntary consent, and that Chanthasouxat's [incriminating] statements [made during the stop] were fruits of the poisonous tree, we hold that the drug evidence and Chanthasouxat's statements must be suppressed." 342 F.3d at 1281.

What is noteworthy about these cases is that they include no discussion whatsoever of the "factual nexus" between the illegal stop and the discovery of the evidence. The courts see the chain of events between the illegal stop and the subsequent discovery of evidence to constitute such a self-evidently close "factual nexus" that, as Professor LaFave puts it, see 6 LaFave, supra, § 11.4(d), there is no need to probe the limits of the fruit of the poisonous tree doctrine.19

In United States v. Reed, 349 F.3d 457 (7th Cir.2003), the defendant passenger confessed to drug distribution after spending several hours with police (whether in custody or not was disputed) following a traffic stop. He contended that the police lacked probable cause for the stop,20 and sought suppression of the confession on fruits grounds. The district court ruled that even if the arrest was illegal, his subsequent confession was still voluntary. Thus the district court did not rule on the legality of the initial stop. On appeal, the Seventh Circuit, applying Brown v. Illinois, 422 U.S. 590, 95 S.Ct. 2254, 45 L.Ed.2d 416 (1975), vacated and remanded, holding that, notwithstanding the lapse of time and the voluntariness of Reed's confession, the confession would have to be suppressed as fruit of the poisonous tree if the initial stop had been illegal. The Reed court explicitly rejected a "counterfactual purgation" analysis of the sort employed by the Tenth Circuit in DeLuca and by the District Court in the case at bar:

The dissent opines that Reed would have confessed even if he had been allowed to go home rather than being held for hours, because his confession stemmed from a desire to win favor and reward and was not a consequence of the illegality. We cannot make that determination as a matter of law, and the government has the burden to prove that his confession was attributable to a factor other than the prolonged detention following the allegedly illegal arrest. That Reed determined that it was in his best interest to cooperate does not somehow divorce his decision from the unlawful detention.

Reed, 349 F.3d at 466.

The Reed court rejects both components of the DeLuca counterfactual analysis: first, the shifting of the explanatory burden to the defendant, and second, the artificial "divorcing" of the (legal) act by which the evidence was physically obtained from its (illegal) necessary causal antecedent. If that divorce is going to occur, the court insists, it will be through the standard fruits exceptions, and they must be shown by the government.

In United States v. Guevara-Martinez, 262 F.3d 751 (8th Cir.2001), the defendant was a passenger in a car pulled over in an illegal traffic stop. During the stop, the officers searched the car and discovered methamphetamine. On that basis they arrested Guevera-Martinez. That night, in jail, he told an INS agent that he was in the country illegally. Guevera-Martinez was charged in separate indictments with both drug and immigration violations. Guevera moved to suppress the drugs, arguing that the traffic stop was illegal. The district court granted the motion, and the drug charges were dismissed. On the immigration charge, Guervera-Martinez sought to suppress both the statements he made and the fingerprints taken while he was in jail. The district court granted the suppression motion on the same grounds, holding that both the statements and prints were fruit of the illegal traffic stop, and the Eighth Circuit affirmed.

The Eighth Circuit applied a straightforward fruits analysis: "[O]fficers obtained Guevera-Martinez's fingerprints [and statements] by exploiting his unlawful detention, instead of by means sufficient to have purged the taint of the initial illegality." Id. at 755. The court deemed it irrelevant that Guevera-Martinez "will inevitably face" civil deportation proceedings in which new fingerprint evidence would easily be obtained, and which would support the reinstatement of the criminal immigration charges: "[T]he government asks us to ignore its use of tainted evidence in this case. We decline to [do so]... The important thing is that those administering the criminal law understand that they must obtain the evidence the right way." Id. at 756 (quoting Davis v. Mississippi, 394 U.S. 721, 726 n. 4, 89 S.Ct. 1394, 22 L.Ed.2d 676 (1969)) (internal quotes, brackets omitted).

United States v. Dortch, 199 F.3d 193 (5th Cir.1999), is also very close in its facts to the case at bar. In Dortch, the defendant was the driver rather than a passenger, but this difference is irrelevant because Dortch's suppression claim was based solely on the detention itself. Dortch was stopped for a traffic infraction, and he did not challenge the legality of the stop. However, the police continued to detain him (and, obviously, his vehicle) after the legitimate purpose of the traffic stop had been completed. Id. at 197. Instead of allowing him to leave, the police detained him for fifteen minutes while they summoned a drug-sniffing dog team.

The bulk of the Fifth Circuit's analysis is devoted to the question of whether Dortch's continued detention, once the purposes of the traffic stop had been completed, was legal. The court concluded that it was not—that the police had no probable cause to continue to detain him. The government argued that Dortch consented to the subsequent search once the dog team arrived, but the court held that even if he did consent, the evidence must still be suppressed: "even if Dortch's consent was voluntarily given ... the consent was not valid. Instead, because the causal chain between the illegal detention and Dortch's consent ... was not broken, the search was nonconsensual." Id. at 202.

The court thus ordered suppression of all the evidence—which was considerable —that had been amassed against Dortch, and remanded not just for retrial, but with the specific order that he be granted a judgment of acquittal. The justification was straightforward fruits analysis: the moment the legal purposes of the traffic stop were satisfied, any further detention of Dortch was illegal. When the dog team arrived, the dogs were sent around the car and alerted to it. On that basis, the police searched the car and found drugs. And on that basis they secured a warrant for Dortch's home, searched it and found more drugs and paraphernalia.

The Dortch case is particularly significant as an analogue to the case at bar because it was governed by the twin propositions that, first, an exterior investigation of a car by a drug-sniffing police dog is not a search as a matter of law, see Dortch, 199 F.3d at 198; (citing United States v. Seals, 987 F.2d 1102, 1106 (5th Cir.1993)); see also Illinois v. Caballes, 543 U.S. 405, 409, 125 S.Ct. 834, 160 L.Ed.2d 842 (2005), and second, that the dog's alerting to the car constitutes probable cause for a search, as a matter of law. See Dortch, 199 F.3d at 198 (citing United States v. Zucco, 71 F.3d 188, 191-92 (5th Cir.1995)). Thus, in Dortch's case, as in Mosley's, there was no search to which he can object. If the dog team had arrived while Dortch was legally detained, there would have been no Fourth Amendment violation.21 The only basis for Dortch's suppression motion, as for Mosley's, is that the challenged evidence is the fruit of an illegal detention. Thus it is irrelevant that Dortch was the driver rather than a passenger, because that distinction applies only to challenges to the search of the car, and Dortch, like Mosley, made no such challenge.

United States v. Jones, 234 F.3d 234 (5th Cir.2000), is similar to Dortch. In Jones, the car was validly stopped for speeding, but the police continued to detain both driver and passenger without probable cause after the legitimate purposes of the stop—issuance of a citation, and computer criminal history check— were completed. During the illegal prolongation of the detention, the driver consented to a search of the car, and contraband was discovered. The Fifth Circuit, upon determining that the consent was not sufficiently attenuated from the illegal detention to have been purged, ordered suppression of the evidence as to both the driver and the passenger, reasoning that although the passenger had no standing to challenge the search of the car, he could challenge the illegal extension of the traffic stop, and that extension was a single event which was causally connected to the discovery of the evidence. Id. at 244.

Dortch and Jones suggest very strongly that had Mosley's case arisen in the Fifth Circuit, the gun would be suppressed. In Dortch, just as here, there was no invasion of the defendant's reasonable expectation of privacy in the search that produced the evidence. The invasion came solely in the detention (seizure of the person), and the court applied a straightforward fruits analysis, pursuant to which it took it as self-evident that a car search following an illegal traffic stop is the fruit of that stop.

In United States v. Kimball, 25 F.3d 1 (1st Cir.1994), the defendant was a passenger in a car that was stopped by the police. All three occupants were arrested on suspicion of burglary, and the police performed an "inventory" search of the car upon towing it to the station. Kimball sought to suppress evidence recovered from the car, and the court framed the issue as follows:

A police officer's act of stopping a vehicle and detaining its occupants constitutes a seizure within the meaning of the Fourth Amendment.... When a police officer effects an investigatory stop of a vehicle, all occupants of that vehicle are subjected to a seizure, as defined by the Fourth Amendment. The fact that a defendant is a passenger in a vehicle as opposed to the driver is a distinction of no consequence in this context. The interest in freedom of movement and the interest in being free from fear and surprise are personal to all occupants of the vehicle, and an individual's interest is not diminished simply because he is a passenger as opposed to the driver when the stop occurred.... Thus, if the initial stop of the vehicle was illegal, evidence seized by virtue of that stop, such as the tools in this instance, may be subject to exclusion under the "fruit of the poisonous tree" doctrine.

Kimball, 25 F.3d at 5-6 (citations omitted).

The court then added, in a very significant footnote:

The Government's reliance on Rakas v. Illinois, in the context of a stop, is misplaced. In Rakas, the United States Supreme Court held that a mere passenger in an automobile ordinarily does not have the legitimate expectation of privacy necessary to challenge the search of that automobile. The Supreme Court's decision, however, was limited to the issue of whether the passenger's legitimate expectation of privacy was invaded by a search of the vehicle, and not the stop thereof.

Id. at 6 n. 3 (emphasis added, internal cites omitted).

The footnote is significant because it helps to clarify what we think was a source of confusion in the proceedings below. Rakas held simply that a passenger in a car has no reasonable expectation of privacy in the interior of the car, and thus that a search of that car does not violate any of the passenger's Fourth Amendment rights. But Rakas says absolutely nothing about the scope of the exclusionary rule with respect to the fruits of an illegal stop. The Kimball court makes it clear that the passenger's reasonable expectation of privacy vis-a-vis the search is completely irrelevant to the fruits inquiry with respect to the stop.

We read Kimball as holding that the determinative question in passenger suppression is the legality of the stop: if the stop is illegal, then standard fruits suppression will apply.22

The government brings to our attention United States v. Carter, 14 F.3d 1150 (6th Cir.1994), a case in which evidence found during a search of a vehicle was suppressed as to the driver but admitted as to the passenger. Carter was a passenger in a van, which the police pulled over for failure to have a valid license plate. The crucial factual and procedural point, which makes Carter inapposite, is that Carter did not challenge the legality of the initial stop. Thus Carter, like DeLuca, cannot guide the analysis in our case, because unlike in our case, the court did not find that the initial traffic stop itself was illegal:

[W]hether or not the original traffic stop was unconstitutional—an issue that was not preserved in Carter's objection to the magistrate's report and that we do not reach here—we shall assume for purposes of analysis, not only that the subsequent arrest of the driver was unconstitutional, but also that the detention of Mr. Carter, if not illegal from the outset, became illegal when the driver was arrested.

Carter, 14 F.3d at 1154.

Note that the court does not hold that Carter's suppression motion would fail even if the initial traffic stop had been illegal. The court does not assume that the initial stop was illegal, but only that Carter's detention "became illegal when the driver was arrested." That distinction (carefully preserved by the "if not" construction) makes all the difference. As long as the police initially obtained control over the vehicle legally, then (on the DeLuca reasoning) no search of the vehicle after that point, no matter how unconstitutional, will be subject to challenge by a passenger. But Carter does not address the situation in which the initial stop was illegal; to so treat it is to stretch its holding not only beyond its facts but beyond its language.

As the preceding survey establishes, the prevailing rule in the courts of appeals is that an illegal traffic stop entails a suppression remedy for all occupants of the car. The authority of our sister circuits is persuasive. However, we think the interests of justice and jurisprudential clarity are best served by giving this issue the fullest possible airing, so we think it incumbent on us to rebut the logical argument for analytic separation we presented above. We think that argument is rebuttable independently of the persuasive weight of our sister circuits' decisions, for the following reasons.

First, we do not think it quite so self-evident that the seizure of a passenger in a vehicle is not a but-for cause of the discovery of evidence in the vehicle. The simplest statement of the concept of but-for causation is that event A is a but-for cause of event B if event B could not happen without event A happening first. But-for causation is an inference drawn from regularly observed correlation.

But it must be stressed that causation is an inference, not an observation, as philosophers since at least Hume have reminded us. The only empirical facts that we can discover about the world are facts about correlation. We cannot observe causal relationships, whether of cue ball to eight ball, of moon to tides, or of diet-pill ingestion to heart failure. What we observe is correlation, and when we see it regularly enough, we hypothesize causation. Science progresses by repeated testing and attempted invalidation of causal hypotheses. Those that survive the process persist.

The relationship between the seizure of a passenger in a moving vehicle, which necessarily occurs when that vehicle is stopped by the police, and the subsequent discovery of evidence during that stop, is one of ineluctable and undeniable correlation. The day has not dawned when a police officer can effectuate a traffic stop without seizing all the occupants of the vehicle.23 The correlation between that seizure and the discovery of evidence during the stop is a perfect 100 percent. As an empirical matter, therefore, the facts on the ground meet the definition of but-for causation: you never see the one without seeing the other first.

So should we hypothesize—or, given our institutional position, perhaps "impose" is the better term—a causal relationship between the seizure and the discovery of the evidence? As detailed above, the dominant view in the circuits is that the causal nexus between the traffic stop and the discovery of evidence is self-evidently sufficient to support suppression.

We agree that the causal relationship seems close, and clear. But we will not pretend that we have no role in imposing, as a legal matter, that causal relationship onto the fact pattern before us, or that such judgments can be made with iron-clad logical or empirical precision. We do not ignore the fact that the analytic separation of individual constitutional violations is a plausible logical deduction from the proposition that Fourth Amendment rights are personal. We do, however, reject such separation.

We think the better view is that a traffic stop is a single act, which affects equally all occupants of a vehicle. To us, that description of traffic stops comports with the commonsense experience of everyone who has ever ridden in a car; we agree with the First Circuit that the distinction between passenger and driver "is a distinction of no consequence in this context." Kimball, 25 F.3d at 5. A police officer who pulls over a vehicle does not, in the act of pulling over the vehicle, interact separately with each occupant; rather, the officer undertakes one action—turning on the siren and lights—which instantly affects everyone in the targeted vehicle, signaling to them that their freedom of movement has been restricted. It defies common sense and common experience to transmute one action into three, and we will not endorse a Fourth Amendment approach that relies on such a transmutation. The government insists that "Fourth Amendment rights are personal rights," and we do not contest the abstract validity of that proposition. But general propositions do not decide concrete cases, and we reject the practical implications the government seeks to derive from that proposition. We reject "blind adherence to a phrase which at most has superficial clarity and which conceals underneath that thin veneer all of the problems of line drawing which must be faced in any conscientious effort to apply the Fourth Amendment." Rakas v. Illinois, 439 U.S. 128, 147, 99 S.Ct. 421, 58 L.Ed.2d 387 (1978).

Furthermore, the relationship between a traffic stop and the discovery of evidence during that stop is not like the relationship between two balls on a billiard table. Law is not physics. In even its most aspirationally scientific manifestations, the law recognizes and embraces the infusion of a multiplicity of values into judicial determinations of causal relationships, see, e.g., Carl F. Cranor & David A. Eastman, Scientific Ignorance and Reliable Patterns of Evidence in Toxic Tort Causation, 64 L. & Contemp. Prob. 5 (2001) (assessing the extent to which courts are willing or able to assimilate and evaluate scientific models of causation in complex cases), and the Supreme Court has done so even more emphatically, and recently, in the context of the exclusionary rule. Fourth Amendment rules cannot be derived from deductive reasoning from first principles, nor do they crystallize, unaided, from the naked facts of disputed searches.

The Supreme Court has just this Term reiterated that the exclusionary rule was founded on, and is grounded in, the continuing exercise of pragmatic judicial supervision of the law enforcement activities of the executive branch, effectuated by expansion and contraction of the bubble of proximate cause as courts face particular concrete factual situations. See Hudson v. Michigan, 547 U.S. ___, 126 S.Ct. 2159, ___ L.Ed.2d ___ (2006) (eliminating suppression remedy for knock-and-announce violations because, inter alia, as compared with the 1950s, police departments are more likely to be "staffed with professionals," and lawyers are more likely to take on § 1983 cases). The test by which our models of constitutional causation are measured is not empirical validation by experiment, but rather the march of social progress, refracted through continual judicial evaluation of constitutional purposes and social consequences. As the social context of law enforcement evolves, so too does the exclusionary rule. The debate about the constitutional consequences of traffic stops is akin, in this respect, to the debate about the constitutional consequences of electronic surveillance or computer copying. See, e.g., Orin S. Kerr, Searches and Seizures in a Digital World, 119 Harv. L.Rev. 531, 561 (2005) (pondering whether electronic copying of computer files constitutes a "seizure," and if so what the appropriate Fourth Amendment remedy should be).

The Supreme Court stressed in Hudson that in determining whether a particular Fourth Amendment violation is causally related to a particular challenged piece of evidence in such a way as to trigger the exclusionary rule, we must look not only to the logical relationship between the violation and the discovery of the evidence, but also to the nature of the personal and social interests the Constitution protects, the prevalence of the illegal police practice at issue, the deterrent value of the suppression remedy, and the likely practical effects of a particular rule. See Hudson, at 2166-71 (determining constitutional suppression requirements based on evolving social realities).

Applying these tests, we find that the purposes of the Fourth Amendment are best served by extending the bubble of proximate causation to vehicle passengers. Passengers, no less than drivers, have a constitutional interest in protection from unreasonable seizures. Two of the more constitutionally troubling varieties of unreasonable seizures, both implicated in this case, are those that occur when police stop vehicles based on anonymous tips, or based on the race of the vehicle's occupants. While the Supreme Court may be right about the increased professionalism of police and the robustness of the § 1983 plaintiffs' bar, we cannot say that either racial profiling or reliance on anonymous tips has declined in frequency in recent years, or that civil lawsuits will adequately deter such practices. Nor can we say that the various other categories of cases that give rise to passenger suppression motions are rare, decreasing, sufficiently internally disciplined, or otherwise deterred. Furthermore, Americans spend more time in cars with each passing year, see, e.g., Dep't of Transportation, 2001 National Household Travel Survey, available at http://www.fhwa.dot.gov/policy/ohpi/nhts (reporting "dramatic increases" in time spent driving between 1969 and 2000), and as the circuit cases described above demonstrate, the use of traffic stops as investigatory tools is a widespread and standard police practice.24

Furthermore, were we to allow the government to use against Mosley the evidence recovered in this case, we would weaken, indeed nearly eviscerate, the Supreme Court's clear and unanimous command in Florida v. J.L. that an anonymous tip is a constitutionally deficient basis for an investigatory detention. See 529 U.S. at 273-74, 120 S.Ct. 1375. The Supreme Court did not, we presume, contemplate that its plain constitutional holding could be simply ignored as to all occupants of a vehicle other than the driver, and we will not lightly countenance such a result.

Nor, finally, is the reductio ad absurdum presented above (stationary mobile home versus moving mobile home) really all that absurd. The level of justification required to support a valid car search is extremely low, as we explained above, as compared with that required to support a valid home search. Most obviously, in contradistinction to house searches, there is no general warrant requirement for car searches. See, e.g., Maryland v. Dyson, 527 U.S. 465, 467, 119 S.Ct. 2013, 144 L.Ed.2d 442 (1999). Thus all occupants of cars, drivers and passengers alike, are overwhelmingly more likely to be subject to police scrutiny when on the road than when at home. Of course, illegal police entries into and searches of homes do occur. But the frequency with which police effect forcible entries into homes is incomparably less than the frequency with which they pull over cars, and the teaching of Hudson is that such social facts matter. If warrantless house searches were as common as roadside traffic-stop searches, for example, then perhaps the protection afforded houseguests in Minnesota v. Olson, 495 U.S. 91, 110 S.Ct. 1684, 109 L.Ed.2d 85 (1990), might have been broader. The exclusionary rule expresses, inherently and always, a standard of reasonableness that evolves along with police practices and social norms.

The exclusionary rule is a judge-made remedy designed to deter illegal police conduct. The Hudson decision has made it clear, if there was ever any doubt, that decisions about both the application of the rule are pragmatic decisions requiring practical wisdom rather than syllogisms. Justice Scalia's opinion epitomizes such pragmatic balancing, "interpreting the Constitution in light of its own practical concern for an active liberty that is itself a practical process," Stephen Breyer, Active Liberty 74 (2005). It is in that spirit that we decide this case.

The car in which Mosley was riding was pulled over illegally. Mosley was illegally seized the moment the car was pulled over. The stopping of the vehicle was a but-for cause of the discovery of the guns. The bubble of causation which links a traffic stop to a subsequent search extends to all occupants of the stopped vehicle. To overcome Mosley's suppression motion, therefore, the government would have had to establish one of the traditional exceptions to the Wong Sun rule: attenuation, inevitable discovery, independent source, or some intervening act or event sufficient to purge the taint of the illegal stop. The government raised no such exception at trial, and raises none on appeal. We therefore hold that the guns were fruit of the poisonous tree and must be suppressed.

Accordingly, we will vacate Mosley's conviction, and remand the case to the District Court for further proceedings consistent with this opinion.