¶ 1. The question presented in this case is whether law-enforcement officers may routinely search a motor vehicle without a warrant, after its occupant has been arrested, handcuffed, and secured in the back seat of a police cruiser, absent a reasonable need to protect the officers’ safety or preserve evidence of a crime. We hold that such warrantless searches offend the core values underlying the right to be free from unreasonable searches and seizures embodied in Chapter I, Article 11 of the Vermont Constitution. Accordingly, the trial court judgment to the contrary is reversed.
¶ 2. During the early morning hours of September 23, 2003, South Burlington police officer David Solomon observed a vehicle on Shelburne Road that appeared to be traveling at a speed of forty-five to fifty miles per hour in a thirty-five mile-per-hour zone. The officer *394followed the vehicle, which weaved several times and continued to travel in excess of the speed limit. Based on these observations, the officer activated his blue lights. The vehicle, in response, pulled into the lot of a service station on Shelburne Road.
¶ 3. While speaking with the driver, later identified as defendant, the officer detected a faint odor of intoxicants and observed defendant’s eyes to be watery and bloodshot. At the officer’s request, defendant exited the vehicle and performed a number of field sobriety tests. Based on his further observations, the officer arrested defendant for driving under the influence (DUI), handcuffed him, and placed him in the rear of his police cruiser. A woman passenger in the vehicle was identified, released, and left the scene. Defendant produced an unsigned bill of sale that purported to vest title to the vehicle in himself, but a check of the vehicle registration failed to identify defendant as the vehicle’s owner. A further records check disclosed that defendant’s Texas driver’s license was suspended.
¶ 4. After defendant was arrested and placed in the police cruiser, Officer Solomon and another officer who had arrived as backup searched defendant’s car. Officer Solomon later testified that he routinely searches the vehicles of drivers arrested for DUI under the “incident-to-arrest” doctrine, confining his search to what he described as the “lungeable” area of the vehicle, i.e., the area that the driver or passengers could potentially reach. The officer acknowledged, however, that he did not feel in any danger from defendant, who was handcuffed and seated in the back of the police cruiser at the time of the search. Nor did the officer harbor any concern that evidence in the vehicle might be removed or destroyed.
¶ 5. In their initial search of the vehicle, the officers discovered the head of a parking meter behind the driver’s seat, a pipe with burnt residue in an open compartment attached to the driver’s door, and an empty beer can and a glass jar containing fragments of a green leafy substance under the driver’s seat. The officers opened the jar and smelled the contents, confirming their suspicion that it had contained marijuana. Officer Solomon also detected a very faint odor of marijuana in the vehicle, although he acknowledged in his affidavit that the odor was not consistent with having been freshly smoked.
¶ 6. Having previously concluded that they would not permit the vehicle to be driven from the scene absent proof of ownership and insurance, the officers further determined — based on their initial search — to impound the car, tow it to the police station, and apply for a search warrant. A warrant was granted, and the subsequent search of *395a backpack on the back seat of the vehicle uncovered a clear plastic bag containing a white powdery substance, later determined to be 7.2 grams of the drug ecstasy.1
¶ 7. Defendant was charged with possession of marijuana, possession of ecstasy, and possession of stolen property. He moved to suppress all of the evidence on the ground that it had been discovered pursuant to an illegal search incident to arrest. In his memorandum in support of the motion, defendant urged rejection of the federal Fourth Amendment standard set forth in New York v. Belton, 453 U.S. 454 (1981), which automatically permits the warrantless search of a motor vehicle following the arrest of its operator under the search-incident-to-arrest doctrine. Defendant argued for a more protective standard under Chapter I, Article 11 of the Vermont Constitution, to require a showing by the government that exigent circumstances justified the warrantless search to secure the officers’ safety or preserve evidence of a crime.2
¶ 8. Following a hearing in which Officer Solomon testified to the circumstances of the stop and search, the court issued a written decision denying the motion to suppress. The court found that the warrantless search comported with both state and federal law as a search incident to arrest. Defendant later entered a conditional plea of guilty to one count of possession of ecstasy, and received a suspended sentence of two to five years and an order of restitution, all stayed pending the outcome of this appeal.
*396¶ 9. A motion to suppress evidence presents a mixed question of fact and law. While we uphold the trial court’s factual findings absent clear error, we review the trial court’s conclusions of law de novo. State v. Simoneau, 2003 VT 83, ¶ 14, 176 Vt. 15, 833 A.2d 1280.
¶ 10. As noted, this appeal presents a fundamental question concerning the extent to which Article 11 authorizes a search incident to arrest following a motorist’s arrest for DUI. In addressing this issue, we do not write on a clean slate. While we have recognized that the Fourth Amendment and Article 11 both seek to protect our “‘freedom from unreasonable government intrusions into ... legitimate expectations of privacy,”’ State v. Kirchoff, 156 Vt. 1, 6, 587 A.2d 988, 992 (1991) (quoting Oliver v. United States, 466 U.S. 170, 187 (1984) (Marshall, J., dissenting)), we have also long held- that our traditional Vermont values of privacy and individual freedom — embodied in Article 11 — may require greater protection than that afforded by the federal Constitution. See State v. Rheaume, 2005 VT 106, ¶ 8 n.*, 179 Vt. 39, 889 A.2d 711 (recalling the extensive case law holding that Article 11 “affords individuals greater privacy rights than its federal counterpart in certain circumstances”). Recently, for example, we held that law-enforcement officers must have a reasonable basis to believe that their safety is at risk or a crime requires investigation, to order a driver stopped for a motor vehicle violation out of his or her vehicle. State v. Sprague, 2003 VT 20, ¶ 16, 175 Vt. 123, 824 A.2d 539. Although the United States Supreme Court has ruled — to the contrary — that the Fourth Amendment permits routine exit orders in such circumstances, Pennsylvania v. Mimms, 434 U.S. 106, 111 (1977), we concluded in Sprague that “a rule requiring a minimal level of objective justification ... strikes the proper balance ... between the need to ensure the officer’s safety and the constitutional imperative of requiring individualized, accountable decisionmaking for every governmental intrusion upon personal liberties.” Sprague, 2003 VT 20, ¶ 16.
¶ 11. Spragus is especially instructive for our purposes here because it illustrates the principles that this Court applies in weighing the competing interests of individual freedom and effective law enforcement that invariably underlie Article 11 cases. In Mimms the Supreme Court embraced a “bright-line” rule for officers to follow by allowing them to order drivers out of their vehicles without any particularized suspicion or safety concern. In Sprague, however, we rejected administrative simplicity as an adequate basis for a seizure when weighed against the individual’s right to be free from arbitrary police intrusions. “[Dispensing entirely with the requirement that an officer provide *397some reasoned explanation for an exit order,” we observed, “invites arbitrary, if not discriminatory, enforcement.” Id. ¶ 19. Hence, we required an individualized showing of some “objective circumstance” that would cause a reasonable officer to believe the order was necessary to protect the officer’s safety or to investigate a suspected crime. Id. ¶ 20.
¶ 12. Although the specific holding in Sprague was new, its basic reasoning was consistent with many of our earlier decisions. A similar balance was struck, for example, in Kirchoff, where we rejected a Supreme Court ruling that privacy in land may not extend beyond the immediate area surrounding the home, observing that “[t]his per se approach cannot be squared with Article 11.” 156 Vt. at 8, 587 A.2d at 993. State v. Savva similarly stands for the principled rejection of “bright-line” rules or administrative efficiency as adequate grounds for dispensing with the constitutionally based warrant requirement. 159 Vt. 75, 616 A.2d 774 (1991). Confronted, as in Kirchoff, with several longstanding Supreme Court precedents — in this case granting police authority to automatically search closed containers within a vehicle — we nevertheless rejected the high court’s “bright-line tests ... because these tests fail to do justice to the values underlying Article 11.” Savva, 159 Vt. at 87, 616 A.2d at 781 (quotation omitted).
¶ 13. The values illustrated by these and many other decisions of this Court rest — at their core — on the fundamental principle of limited government. Article ll’s warrant requirement represents one of the essential checks on unrestrained government determined by the framers — and confirmed through hard experience — to be necessary to the preservation of individual freedom. The warrant requirement serves as a check on the executive power by guaranteeing review by a neutral and detached magistrate before a search is carried out, thereby deterring “searches on doubtful grounds” and assuring the people of “an impartial objective assessment” prior to a governmental invasion. Id. at 86-87, 616 A.2d at 780; see also State v. Geraw, 173 Vt. 350, 352, 795 A.2d 1219, 1221 (2002) (observing that the warrant requirement “reflects a deeply-rooted historical judgment that the decision to invade ... privacy... should normally be made by a neutral magistrate, not by the agent of the search itself”).3
*398¶ 14. Searches outside the normal judicial process are, therefore, presumptively unconstitutional, and permissible only pursuant to a few narrowly drawn and well-delineated exceptions. Savva, 159 Vt. at 86, 616 A.2d at 780; State v. Meunier, 137 Vt. 586, 588, 409 A.2d 583, 584 (1979). Such rare exceptions are allowed “only in those extraordinary circumstances which make the warrant and probable-cause requirement impracticable.” State v. Petrucelli, 170 Vt. 51, 62, 743 A.2d 1062, 1070 (1999) (quotation omitted). As we explained in Petrucelli, “[e]xceptions to the warrant requirement ‘must be factually and narrowly tied to exigent circumstances and reasonable expectations of privacy.’” Id. (quoting Savva, 159 Vt. at 87, 616 A.2d at 781).4
¶ 15. One such exception is the search-incident-to-arrest doctrine. Although its scope has varied over time, the essential elements of the doctrine were settled by the United States Supreme Court in the landmark case of Chimel v. California, 395 U.S. 752 (1969). Reconciling years of debate, the high court held that, when a suspect has been lawfully arrested, the police may conduct a warrantless search of the person arrested for “any weapons that the latter might seek to use” to resist arrest or facilitate an escape, and “any evidence on the arrestee’s person in order to prevent its concealment or destruction.” Id. at 763. In a famous subsequent passage, the Court observed further that “the area into which an arrestee might reach in order to grab a weapon or evidentiary items must, of course, be governed by a like rule.” Id.
¶ 16. This so-called “grab rule” defined and limited the doctrine for more than a decade, and was routinely applied in every state including Vermont. See, e.g., Meunier, 137 Vt. at 588, 409 A.2d at 584 (citing Chimel for the principle that a search incident to arrest must be “reasonable in time and scope”); State v. Mayer, 129 Vt. 564, 567, 283 A.2d 863, 865 (1971) (citing Chimel to uphold a warrantless “protective *399search” of defendant for weapons at the time of his arrest); see generally 3 W. LaFave, Search and Seizure § 7.1, at 502-14 (4th ed. 2004) (reviewing history and development of search-incident-to-arrest doctrine). In Belton, 453 U.S. at 460, however, the Supreme Court revisited the doctrine in the context of a motor-vehicle search, explaining that police officers remained uncertain after Chimel about the precise scope of their authority and required a more “workable rule.” To provide such a bright-line rule, the Court held that when police officers have arrested the occupant of a vehicle, they may routinely search its passenger compartment and the contents of any containers found therein as a “contemporaneous incident of that arrest.” Id. at 460-61. More recently, in Thornton v. United States, 541 U.S. 615, 623-24 (2004), the Supreme Court reaffirmed the rule announced in Belton, holding that it applied even where the driver had been arrested, handcuffed, and secured in the back seat of a police cruiser.
¶ 17. Belton was the subject of sharp criticism when it was decided, and it has remained controversial ever since. Justice Brennan, writing in dissent, observed that the rule was “analytically unsound and inconsistent with every significant search-incident-to-arrest case” with similar facts in the Court’s recent history. 453 U.S. at 468. The Court had always required that exceptions to the warrant clause be firmly grounded in, and narrowly tailored to, the extraordinary circumstances justifying the exception. Plainly, however, an arrestee who has been secured away from the vehicle is in no position to seize a weapon or evidence from its interior. See id. at 465-66 (Brennan, J., dissenting) (“When the arrest has been consummated and the arrestee safely taken into custody, the justifications underlying Chimel’s limited exception to the warrant requirement cease to apply: at that point there is no possibility that the arrestee could reach weapons or contraband.”). Nor, as Justice Brennan observed, had the Court ever held that mere administrative simplicity was a sufficient basis for a warrant exception. See id. at 469 (“[T]he mere fact that law enforcement may be made more efficient can never by itself justify disregard of the Fourth Amendment.” (quotation omitted)). Furthermore, as Justice Brennan noted, the need for so-called “bright lines” was simply unsupported; the search-incident-to-arrest doctrine under Chimel placed no greater demands on law enforcement officers than other Fourth Amendment rules requiring the exercise of considered police judgment in light of the facts and circumstances, as when deciding whether reasonable suspicion justifies an investigatory stop and frisk, or whether probable cause supports a warrantless arrest. Id. at 471 *400(“The standard announced in Chimel is not nearly as difficult to apply as the Court suggests.”). Indeed, Justice Brennan observed, the bright-line rule forged by the Belton majority was not even likely to eliminate the continued need for the exercise of police judgment in determining, for example, the exact nature of a “contemporaneous” search incident to arrest. Id. at 470 (“Would a warrantless search incident to arrest be valid if conducted five minutes after the suspect left his car? Thirty minutes? Three hours?”).
¶ 18. The concerns identified in the Belton dissent have continued to gather support from courts and commentators alike. Professor LaFave and others have questioned the warrantless search rationale based on either safety or simplicity, particularly as studies have shown that the police almost invariably handcuff and remove arrested drivers from the area of the vehicle. See 3 LaFave, supra, § 7.1(c), at 525; see also M. Moskowitz, A Rule in Search of a Reason: An Empirical Reexamination of Chimel and Belton, 2002 Wis. L. Rev. 657, 697 (suggesting that auto searches following arrest should require a showing of “particular and unusual facts” that hinder the police from their usual procedure of “restraining and removing the suspect from any area that might contain a weapon or evidence”); A. Alschuler, Bright Line Fever and the Fourth Amendment, 45 U. Pitt. L. Rev. 227, 274 (1984) (“If any bright line rule had been necessary to resolve issue in Belton, it would have been the opposite of the rule that the Court announced.”); E. Shapiro, New York v. Belton and State Constitutional Doctrine, 105 W. Va. L. Rev. 131, 137 (2002) (noting that “[cjriticism of Belton has been vigorous and sustained,” based principally on the lack of support for the Court’s rationale that “existing law had proven to be so unworkable that it was necessary to forego Chime’ s approach in favor of a bright-line rule”).
¶ 19. In addition, while a majority of states continue to apply the rule in Belton, a number have either rejected or modified it under their state constitutions. See Shapiro, supra, 105 W. Va. L. Rev. at 141-42 (listing and discussing the state decisions that have declined to follow Belton or have applied a modified federal approach). New Jersey, Pennsylvania, New Mexico, and Nevada have all unequivocally rejected Belton under their state constitutions, applying instead the familiar standard predicated upon a showing of necessity to secure the officer’s safety or preserve evidence. See Camacho v. State, 75 P.3d 370, 373-74 (Nev. 2003) (rejecting Belton and concluding that “under the Nevada Constitution, there must exist both probable cause and exigent circumstances for police to conduct a warrantless search of an automo*401bile incident to a lawful custodial arrest”); State v. Eckel, 888 A.2d 1266, 1276-77 (N.J. 2006) (declining to adopt Belton and holding that under the New Jersey Constitution the search-incident-to-arrest doctrine applies only “to ensure police safety or to avoid the destruction of evidence”); State v. Pittman, 127 P.3d 1116, ¶ 16 (N.M. Ct. App. 2005) (“Because of New Mexico’s strong preference for a warrant, we hold that even after a valid arrest, one of Chimel’s two rationales must be present before an officer may search a vehicle without a warrant.”), cert. granted, 131 P.3d 660 (N.M. Jan. 10, 2006) (No. 29,584); Commonwealth v. White, 669 A.2d 896, 902 (Pa. 1995) (invalidating warrantless vehicle search where the arrestee was a secure distance from his vehicle, and holding that under the Pennsylvania Constitution the police may search only “the arrestee’s person and the area in which the person is detained in order to prevent the arrestee from obtaining weapons or destroying evidence”).
¶ 20. In our judgment, these decisions more closely reflect the principles and values underlying Article 11 as expressed in numerous opinions of this Court than the “abrupt shift in the standard of fourth amendment protections” represented by the Belton decision. C. Hancock, State Court Activism and Searches Incident to Arrest, 68 Va. L. Rev. 1085, 1085 (1982). As earlier explained, we have consistently rejected bright-line rules — however laudable their purpose in easing the burden on law-enforcement officers — as an adequate basis for relaxing the fundamental limitation on governmental power represented by the warrant requirement. Indeed, we have scrupulously maintained the principle — even, as here, in the face of contrary United States Supreme Court holdings — that any exception to the warrant requirement must be factually and narrowly tied to the exigencies that rendered a warrant application impracticable under the circumstances. Absent such circumstances, Article 11 simply forbids a warrantless search. As the New Jersey Supreme Court explained in admirably clear and unambiguous terms in Eckel, a warrantless automobile search based “solely on the arrest of a person unable to endanger the police or destroy evidence cannot be justified under any exception to the warrant requirement and is unreasonable.” 888 A.2d at 1277.
¶ 21. The State here offers no serious argument that the warrantless search in this case was justified as a search incident to arrest on any basis other than the blanket authority of Belton. Although our dissenting colleague claims that the search was somehow necessary to protect the officer’s safety or preserve evidence, no persuasive evidence *402or argument is offered to demonstrate how defendant — handcuffed in the back seat of the police cruiser — or his passenger who had left the scene, presented any form of threat. The dissent’s further assertion that the search here was actually consistent with pre-Belton decisional law is equally unsound. One need only read the impassioned Belton dissent to understand how fundamentally at odds that decision was with prior law. Contrary to the dissent’s additional claim, moreover, it is clear that under Chimel and its progeny a showing of exigent circumstances in the form of a threat either to officer safety or to the preservation of evidence is essential to justify a warrantless vehicle search.
¶22. Having rejected Belton in favor of the traditional rule requiring that officers demonstrate a need to secure their own safety or preserve evidence of a crime, and finding no evidence of either need in this case, we are compelled to conclude that the trial court order denying defendant’s motion to suppress must be reversed.
¶ 28. Although, in our view, the reasons that compel rejection of Belton apply with equal and obvious force to the so-called “Belton variation” adopted by several states, and although the State has not argued otherwise, we defer closing this portion of the discussion to consider this alternative in light of the dissent’s strong endorsement of it. As the dissent notes, several states have allowed the police to conduct warrantless searches of automobiles after the occupant has been arrested in order to obtain evidence related to the crime that formed the basis of the arrest. As the dissent observes, the rationale of these decisions appears to be that “the arrest itself provides the probable cause basis for the search.” Post, ¶ 90. The dissent would adopt this approach so long as the search was for “evidence related to the crime” and limited to the passenger compartment of the vehicle. Post, ¶ 90.5
f 24. The so-called Belton variation endorsed by the dissent is just that, a variation of Belton. Although the rationale is different — the arrest purportedly provides the probable cause to search — the *403reasoning remains essentially the same, based on a perceived need to authorize routine warrantless searches absent any particularized showing that the delay attendant upon obtaining a warrant is impracticable under the circumstances. As earlier observed, however, such an approach is fundamentally at odds with Article 11, under which warrantless searches are presumptively unconstitutional absent a showing of specific, exigent circumstances justifying circumvention of the normal judicial process. As we explained in State v. Trudeau, “no amount of probable cause can justify a warrantless search or seizure absent exigent circumstances.” 165 Vt. 355, 360, 683 A.2d 725, 729 (1996) (quoting Horton v. California, 496 U.S. 128, 137 n.7 (1990) (quotation omitted)). Surely this principle applies with equal or greater force where the probable cause is merely presumed from the fact of an arrest.
¶ 25. Inherent, too, in the Belton variation are a number of assumptions that simply do not withstand scrutiny. First, as earlier discussed, support for the assumption that case-by-case evaluations are unworkable in the context of warrantless vehicle searches is simply lacking. Second, the assumption that an arrest automatically provides probable cause for a search is highly questionable. The finding of probable cause is a decidedly fact-specific determination, turning on whether the particular circumstances establish a “nexus between the crime, the suspect, and the place to be searched.” State v. Towne, 158 Vt. 607, 616, 615 A.2d 484, 489 (1992). A driver arrested for DUI may have been drinking at home, at a friend’s, in a restaurant or bar, or at a sporting event, but not necessarily in his or her car. While the facts — e.g., the strong odor of intoxicants coming from inside the vehicle or an actual admission by the suspect — might indicate the presence of alcohol in the vehicle, the arrest itself does not invariably establish the requisite nexus to search. Nothing about the fact that the search occurs in a vehicle, moreover, would justify a reduced probable-cause standard. Indeed, while we have acknowledged that vehicles support a somewhat diminished expectation of privacy, this is not to say — and we have never held — that they carry no expectation of privacy, or that an arrest of the driver obviates the need to establish specific probable cause to search.
¶ 26. The dissent’s additional assumption of administrative simplicity is equally questionable. The dissent would permit searches only for evidence “related to the crime” for which the suspect was arrested. Post, ¶ 90. Would this permit a vehicle search following an arrest of the driver on an outstanding warrant for failure to appear? What if the *404underlying charges on the outstanding warrant related to possession of cocaine? Would an arrest for assaulting an officer during a routine vehicle stop authorize a search, and if so, for what? Does the nature of the arrest define the scope of the search, i.e., would an arrest based on possession of stolen televisions authorize a search under the car seat? The so-called bright-line rule advocated by the dissent raises as many questions as it answers. It most assuredly does not, however, commend itself as superior to the traditional search-ineident-to-arrest rule in any respect.
¶ 27. Finally, in view of the dissent’s strenuous claims to the contrary, we take the opportunity to explain the necessity of today’s holding. Our dissenting colleague proffers essentially three separate doctrinal exceptions to the warrant requirement as more suitable “independent grounds” of decision. Post, ¶ 40. It is, of course, a fundamental tenet of judicial restraint that courts will not address constitutional claims — least of all novel or unresolved constitutional claims — when adequate lesser grounds are available. See In re Sealed Documents, 172 Vt. 152, 156, 772 A.2d 518, 523 (2001) (noting “[o]ur tradition of addressing issues of constitutional significance only when the matter is squarely and necessarily presented”).
¶ 28. First, it is asserted that the parking-meter head discovered behind the driver’s seat was “in plain view” and therefore — as patent contraband — provided an independent basis to search the car under the well-settled plain-view exception to the warrant requirement. Post, ¶¶ 42-51. The claim is predicated upon the investigating officer’s statement, in response to a question from the trial court, that the parking meter was visible from outside the vehicle. As noted, however, the search here did not proceed from a plain-view observation of the parking meter. Indeed, the officer repeatedly acknowledged that he did not see the parking meter during his initial contact with defendant outside the vehicle; he became aware of its existence only during the more probing search inside the car. The trial court addressed this seeming anomaly by finding unequivocally that the officer discovered the parking meter during the search incident to arrest, while noting that it “was arguably exposed to plain view.”6
*405¶ 29. Thus, the facts underlying the dissent’s proposed plain-view analysis may be characterized, at best, as uncertain. The legal basis, however, can only be described as dubious. The dissent relies on a single statement in Trudeau, 165 Vt. at 358, 683 A.2d at 727, quoting Horton v. California, 496 U.S. at 136, to the effect that an “essential predicate” underlying the plain-view doctrine is that “the officer did not violate the Fourth Amendment in arriving at the place from which the evidence could be plainly viewed.” (Emphasis added.) Nothing in either decision, however, remotely suggests that the underscored language was intended by this Court or the United States Supreme Court to establish a constructive plain-view standard, to be satisfied whenever an officer asserts in hindsight that the evidence could have been plainly viewed, although in fact it was not. On the contrary, in both cases, as indeed in virtually every case dealing with the doctrine that we have uncovered, the plain-view exception was based on the officer’s actual observation of the evidence in question.
¶ 30. This is hardly surprising, as it is the police officer’s perception of the object which establishes, in each case, its “plain-view” status. As the high court explained in Terns v. Brown, the plain-view doctrine is predicated on two principles: first, “that when a police officer has observed an object in plain view” from a legal vantage point the owner’s privacy interests are forfeited; and second, that requiring a warrant once the police “have obtained a first-hand perception of [the object] would be a needless inconvenience.” 460 U.S. 730, 739 (1983) (emphasis added, quotation omitted). Thus, as the Court observed, “our decisions have come to reflect the rule that if, while lawfully engaged in an activity in a particular place, police officers perceive a suspicious object, they may seize it immediately.” Id. (emphasis added). This basic rule has been applied in every case to come before the Court, including those where the objects in question were observed through aerial surveillance, or with the aid of illumination. See, e.g., Florida v. Riley, 488 U.S. 445, 448-49 (1989) (search upheld where police in helicopter were able to observe with the “naked eye” marijuana growing in greenhouse); California v. Ciraolo, 476 U.S. 207, 213-15 (1986) (police observed marijuana visible to “naked eye” from aircraft); Brown, 460 U.S. at 739 (use of flashlight to enhance visibility did not invalidate seizure of drugs observed by officers). To modify the doctrine by allowing the seizure of objects which the officers did not observe — as advocated by the dissent — would eviscerate its fundamental evidentiary and legal grounding.
*406¶ 31. In essence, therefore, the dissent proposes that we forgo addressing an issue — the scope of the search-incident-to-arrest doctrine in the context of a vehicle search — that the police officers here expressly relied on, that the parties briefed and argued at trial and on appeal, that formed the core of the trial court’s decision, and that — as explained earlier — has been the subject of extensive discussion and debate among courts and commentators. Instead, the dissent urges that we address a novel constitutional issue based on questionable facts and even less legal support. With respect, we fail to see how this proposed alternative makes any sense, or serves any sound jurisprudential purpose.
¶ 32. The dissent also claims that defendant’s failure to provide a valid driver’s license, registration, or insurance card, coupled with irregularities in the vehicle’s plates and bill of sale, authorized the police to conduct a warrantless search for proof of ownership. The argument is unpersuasive. It relies, essentially, on the so-called “automobile exception” to the warrant requirement, which — as we have elsewhere explained — requires a showing of both probable cause that the vehicle contains evidence of a crime, and exigent circumstances suggesting that the evidence may be lost during the delay attendant upon obtaining a warrant. See Savva, 159 Vt. at 89-90, 616 A.2d at 782 (holding that warrantless search of bags found within car “was not supported by exigent circumstances because a less intrusive option was available” and therefore must be invalidated); State v. Girouard, 135 Vt. 123, 129, 373 A.2d 836, 840 (1977) (describing the “well-delineated preconditions” to the automobile exception as “1) probable cause to believe that the vehicle contains evidence of crime and 2) exigent circumstances”).
¶ 33. Neither requirement was satisfied here. Despite the officer’s suspicion that the car might have been stolen, he did not arrest defendant on that basis and identified no ground, much less probable cause, to believe that proof of ownership might be discovered behind or underneath the driver’s seat, where the parking meter and glass jar containing marijuana were found. Even if it were assumed, however — as the dissent urges — that the inadequate proof of ownership established probable cause to believe that the car was stolen, the circumstances did not establish that element of urgency essential to the execution of a warrantless search. The officer readily acknowledged that he had no concerns about the possibility of evidence inside the vehicle being removed or destroyed. Indeed, prior to the search, the officers had not *407observed any evidence of a crime in the vehicle, let alone evidence that might conceivably be lost or destroyed.7
¶ 34. Furthermore, defendant was under arrest, the car was not on a public highway but safely parked in a commercial lot, and the police had determined that it would be grounded, i.e., locked and kept there until they determined its ownership. Hence, there was no exigency compelling an immediate search rather than a subsequent warrant application. In Trudeau, the principal case on which the dissent relies, the police had observed evidence in plain view within the vehicle that related directly to the offense for which defendant was arrested. Indeed, we analyzed Trudeau as a plain-view case, not an automobile-exception case, emphasizing that the officers violated no privacy rights of the defendant when they observed an open beer can in plain view on the floor of the defendant’s car before arresting him for DUI. 165 Vt. at 358, 683 A.2d at 727-28. Here, in contrast, the officers had no indication that defendant’s vehicle contained any contraband or evidence of a crime. Furthermore, the record in Trudeau revealed the presence of two additional passengers in the vehicle who also appeared to be intoxicated and who had remained near the vehicle during the police encounter, although they had not been arrested. This was sufficient to suggest that they might have had not only the opportunity, but the incentive, to seek access to the vehicle to remove the evidence the police had observed therein, and thus established the exigency necessary to forgo a warrant. Trudeau, 165 Vt. at 357, 361, 683 A.2d at 726, 729. Neither circumstance was present here. The police had not observed any evidence of a crime in the vehicle, and there was nothing to indicate that the passenger, who had been questioned by the police and had departed, would have any reason to return to the vehicle or ability to remove its contents. Accordingly, we are not persuaded that the automobile exception provides a viable basis to uphold the trial court decision.8
*408¶ 35. Finally, the dissent proposes in a footnote that the search here could be validated as an inventory search under the inevitable-discovery doctrine. Courts have approved inventory searches of lawfully impounded vehicles to protect the owner’s property while in police custody, see, e.g., Colorado v. Bertine, 479 U.S. 367, 372-73 (1987), and have upheld the admission of evidence that the police would have “inevitably discovered” during such a search. United States v. Seals, 987 F.2d 1102, 1107-08 (5th Cir. 1993). The doctrine has no application here because, prior to the illegal search, the officer testified that they had determined only to “ground” the vehicle, i.e., to leave it in place in the private lot where it was parked. The decision to impound the vehicle was not made until after the warrantless search, and was based on the evidence obtained during that illegal search. Accordingly, there was no legal basis to impound the vehicle, and hence no grounds for applying the inevitable-discovery doctrine.
¶ 36. In closing, we believe that it is essential to be as clear about what this case concerns as what it does not. Although the dissent repeatedly and emphatically asserts that our holding somehow removes important safety protections for law-enforcement officers, it cites not one shred of evidence in the record nor a single statistic, relevant public-safety study, or other empirical evidence outside the record to support the claim. Indeed, as we have explained, the evidence and authorities demonstrate that, far from removing safety protections, our holding is entirely consistent with existing, standard police procedures and removes no essential safeguards. We yield to no one on this Court in our commitment to the safety of Vermont law-enforcement officers in the field. Strident assertions, however, are no substitute for proof. In the absence of a demonstrated need, we are not at liberty to disregard the fundamental constitutional requirement of a search warrant. By limiting the exercise of arbitrary governmental *409power, this constitutional safeguard protects the police no less than the public.9
¶ 37. Justice Robert Jackson once observed that “[w]hen the right of privacy must reasonably yield to the right of search is, as a rule, to be decided by a judicial officer, not by a policeman or government enforcement agent.” Johnson v. United States, 333 U.S. 10, 14 (1948). Any other rule, he explained, “would reduce the [right] to a nullity” and leave us secure in our homes and persons “only in the discretion of [law-enforcement] officers.” Id. Where, as here, the sole justification for dispensing with the fundamental safeguard of personal liberty represented by the warrant requirement is law-enforcement efficiency, we have consistently ruled in favor of liberty. As our own Justice Larrow once observed, “[t]his seems a slight price to pay for the fundamental rights preserved by” the Constitution. State v. Connolly, 133 Vt. 565, 571, 350 A.2d 354, 368 (1975).
Reversed.
The officer testified several times to the effect that “the car wasn’t going to be driven because we had no documentation of who it belonged to, that it was registered or that it was insured.” In other words, the decision to “ground” the car was made before the initial search, based on the lack of proof of ownership. Later, based on the evidence obtained during the search, the officers determined that the vehicle would be impounded and a warrant obtained for a more thorough search. With respect, the dissent is simply mistaken in asserting that the decision to ground the vehicle was made after the search.
The full text of Article 11 reads:
That the people have a right to hold themselves, their houses, papers, and possessions, free from search or seizure; and therefore warrants, without oath or affirmation first made, affording sufficient foundation for them, and whereby by any officer or messenger may be commanded or required to search suspected places, or to seize any person or persons, his, her or their property, not particularly described, are contrary to that right, and ought not to be granted.
Vt. Const. ch. I, art. 11.
The dissent’s assertion that State v. Martin, 145 Vt. 562, 496 A.2d 442 (1985), represents “essentially a bright-line rule” adopted by this Court, post, ¶ 68, is well wide of the mark. There, we rejected the claim that DUI roadblocks “constitute a per se *398violation of the Fourth Amendment,” id. at 565, 496 A.2d at 445, adopting instead a balancing test “directly related to the characteristics of the DUI roadblock in each case.” Id. at 570, 496 A.2d at 448 (emphasis added). This is the opposite of a bright-line standard.
Although the word “unreasonable” does not appear in the text of Chapter I, Article 11 of the Vermont Constitution, see supra, note 2, we have consistently construed the provision to forbid only unreasonable searches and seizures. State v. Record, 150 Vt. 84, 85, 548 A.2d 422, 423 (1988). As discussed above, we have also consistently held that warrantless searches are presumptively unreasonable unless justified by a well-recognized exception. State v. Mountford, 171 Vt. 487, 493, 769 A.2d 639, 646 (2000), abrogated on other grounds by Brigham City, Utah v. Stuart, 547 U.S. 398, 126 S. Ct. 1943 (2006).
As the dissent notes, this variation also appears to have been endorsed by Justice Scalia in a concurring opinion in Thorton. While sharply criticizing Belton, Justice Scalia nevertheless opined that, “[i]f Belton searches are justifiable, it is not because the arrestee might grab a weapon or evidentiary item from his car, but simply because the car might contain evidence relevant to the crime for which he was arrested.” 541 U.S. at 629 (Scalia, J., concurring).
The dissent asserts that we mischaracterize the record “with respect to whether the parking meter was in plain view.” Post, ¶ 43. Not so. The officer’s testimony was clear, unequivocal, and undisputed that he did not observe the parking meter from outside the vehicle, and was unaware of its existence until it was discovered during the vehicle search.
Contrary to the assertion of the dissent, we neither “emphasize” nor “repeatedly” rely on the office’s subjective perception that he did not feel threatened or pressed to preserve evidence. We merely note the officer’s testimony in this regard as further proof of the absence of evidence of exigent circumstances in this case.
To be sure, other courts have held that, under the traditional automobile exception to the warrant requirement, a driver’s failure to produce documentation of ownership may establish a reasonable suspicion that the vehicle is stolen and thereby establish the basis for a limited search of the vehicle in those places, such as the glove compartment or sun visor, where such documents are normally stored. See, e.g., State v. Holmgren, 659 A.2d 939, 940 (N.J. Super. Ct. App. Div. 1995) (holding that failure to produce registration *408allows search of vehicle for evidence of ownership “confined to the glove compartment or other area where a registration might normally be kept in a vehicle”) (quotations omitted); State v. Barrett, 406 A.2d 198, 200 (N. J. Super. Ct. Law Div. 1979) (invalidating search of vehicle for registration where there was “no expectation that any indicia of title would be found in the rear of the vehicle”). Other courts have even held that such proof of ownership might be found in places other than the glove compartment, such as under seats. In re Arturo D., 38 P.3d 433, 446-47 (Cal. 2002). These cases rely, however, on either the Fourth Amendment or a state equivalent under which exigent circumstances have not been deemed to be an essential element of a warrantless automobile search. As noted, our law is directly to the contrary.
The study to which the dissent refers, post, ¶ 85, and which has been cited by the United State Supreme Court on several occasions, shows the high frequency of shootings of police officers as they “approach a suspect seated in an automobile.” Adams v. Williams, 407 U.S. 143, 148 n.3 (1972). That is not the situation here. Indeed, the study in question is particularly inapposite in the search-incident-to-arrest context, where studies have shown that, in fact, police officers invariably remove suspects from anywhere near their vehicles and often — as here — handcuff and place them in the back seat of the police cruiser, where there is no risk of their gaining access to a weapon or evidence in the detained vehicle. See M. Moskovitz, A Rule in Search of a Reason: An Empirical Reexamination of Chimel and Belton, 2002 Wis. L. Rev. 657, 676 (observing that a survey of police practices reveals that “Belton’s generalization that articles inside the relatively narrow compass of the passenger compartment of an automobile are in fact generally, even if not inevitably, within the area into which an arrestee might reach in order to grab a weapon or evidentiary item is — at least in general — false” (quotation omitted)); 3 LaFave, supra, § 7.1(c), at 525 (observing that, because “the police can, and typically do, immediately remove the arrestee from the vehicle,” close and lock his or her vehicle, and place him or her in handcuffs, “the ‘difficulty’ and ‘disarray' the Belton majority alluded to has been more a product of the police seeing how much they could get away with (by not taking the above-mentioned procedures) than their being confronted with inherently ambiguous situations”).