dissenting.
The main issue presented by State v. Owens, 302 Or 196, 729 P2d 524 (1986), State v. Forseth, 302 Or 233, 729 P2d 545 (1986), and State v. Westlund, 302 Or 225, 729 P2d 541 (1986), is whether a warrant is required to open a transparent container and chemically analyze its contents. That issue could have been resolved in the affirmative with a simple reference to State v. Lowry, 295 Or 337, 667 P2d 996 (1983), but today the court holds that a warrant is not required.
Such a reversal deserves explanation, but it is not to be found in the majority’s opinion.1 The majority confronts neither Lowry nor, more importantly, the warrant issue. Instead, it sidesteps both with its holding, unsupported by either written reason or citation to any other court’s reasoning, that opening a transparent container and chemically analyzing its contents does not infringe any rights under Article I, section 9, of the Oregon Constitution.2
I am confident that the majority does not and cannot mean what it has written; rather, unwilling to confront the warrant requirement, it has created a de facto exception to the requirement through the only other means available to it. Apart from its contraction of constitutional rights, however, the majority’s failure to explain its departure from precedent *216is in itself a threat to constitutional rights. Without a consistent and reasoned analysis, rights protected by Article I, section 9, will disappear whenever four members of the court are persuaded that it is expedient that they do so.
The relevant facts in Owens, Forseth, and Westlund do not significantly differ. In each case, during a search for other items, police discovered a small, transparent glass vial or plastic bag containing an unidentified powder. The police seized the container because they suspected that the powder was a controlled substance. Later, without obtaining a warrant, police opened the containers and chemically analyzed their contents.
I assume, as did this court in Lowry, 295 Or at 346, that the containers were lawfully seized.3 The question, then, is whether a warrant was required before opening the containers and testing their contents. Under Lowry, the answer is clearly yes. 295 Or at 348.
I.
The majority holds that no warrant is required for opening a transparent container because such an action is not a “search” under Article I, section 9. The conclusion that opening a transparent container is not a search is based on the following sequence of assertions: (1) “[a] ‘search’ occurs when a person’s privacy interests are invaded”; (2) “containers * * * that by their very nature announce their contents * * * do not support a cognizable privacy interest under Article I, section 9”; (3) “[transparent containers * * * announce their contents [because their contents are] visible virtually to the same extent as if the contents had been discovered in ‘plain view,’ outside the confines of any container.” While the first assertion can be accepted as a truism, the second assertion is merely an ipse dixit, and the third is patently incorrect.
*217If the sense of sight were alone capable of unerringly identifying substances, the majority’s assertion that transparent containers “announce their contents” might be valid. But, obviously, this is not true. For example, a white powder might be baby powder, cocaine, confectioners’ sugar, or arsenic. While a transparent container does not exclude the sense of sight, it does exclude other senses. A clear liquid in a closed bottle might be mineral water or gin. If the bottle were opened, there would be no doubt. The notion that a transparent container is inherently devoid of any privacy interest because it does not exclude the sense of sight is simply indefensible.
More importantly, even assuming that containers, transparent or not, can “announce their contents,” the majority’s assertion that such containers do not support a cognizable privacy interest under Article I, section 9, does not follow. The majority’s rationale apparently is that, since the contents of such a container are by definition known, opening the container does not infringe any privacy interest. But this rationale confuses privacy with secrecy.4 Privacy certainly includes secrecy, but this is a relatively recent development that can be traced to Katz v. United States, 389 US 347, 88 S Ct 507, 19 L Ed 2d 576 (1967). Prior to Katz, a search was defined in terms of a physical trespass to a “constitutionally protected area,” i.e., a trespass to “persons, houses, papers, and effects,” regardless whether anything “secret” was at risk of discovery. Amsterdam, Perspectives on the Fourth Amendment, 58 Minn L Rev 349, 356-57 (1974); 1 LaFave, Search and Seizure § 2.1, at 223-24; see Olmstead v. United States, 277 US 438, 465, 48 S Ct 564, 72 L Ed 944 (1928); State v. Hilton, 119 Or 441, 444-45, 249 P 1103 (1926). Physical invasions of these “areas” were the only ways in which the eighteenth century framers of the Fourth Amendment and the mid-nineteenth century framers of Article I, section 9, could conceive of an invasion of privacy. See Landynski, Search and Seizure and the Supreme Court 198-205 (1966). In Katz, the Court recognized that a new doctrine was needed to protect *218privacy from technological developments such as electronic surveillance. Katz v. United States, supra, 389 US at 348-353. Such surveillance was not in itself perceived to be objectionable in the same manner as a trespass; rather, it was objectionable because of the information that it could uncover. Hence, the Court in Katz and subsequent cases developed an analysis that enlarged the scope of protection afforded by the Fourth Amendment based on the expectation of privacy in information. This expectation of privacy analysis was not meant to supplant, but to supplement, the older trespass analysis. State v. Louis, 296 Or 57, 60, 672 P2d 708 (1983). Although this court has on occasion employed the Katz analysis, see State v. Holt, 291 Or 343, 630 P2d 854 (1981), the older trespass analysis retains its validity, Louis, 296 Or at 60.
Opening a closed container is nothing less than a trespass upon an “effect.” As such, it is a “search” under Article I, section 9. See, e.g., State v. Keller, 265 Or 622, 625-26, 510 P2d 568 (1973) (opening fishing tackle box a search); State v. O’Neal, 251 Or 163, 444 P2d 951 (1968) (opening wallet and observing contents described as a search). Knowledge of the contents of a protected “area” is irrelevant to the question whether an invasion of that “area” is a “search.” Such knowledge may provide justification for the invasion, but it cannot negate the constitutional character of the invasion. For example, police may lawfully be able to observe the entire contents of a small garage through a window that opens onto a public street, but this fact is irrelevant to whether an entry into the garage is a “search.” See Coolidge v. New Hampshire, 403 US 443, 468, 91 S Ct 2022, 29 L Ed 2d 564 (1971) (“plain view” alone cannot justify a warrantless search or seizure); State v. Tremaine, 56 Or App 271, 641 P2d 637 (1982) (same); cf. State v. Olson, 287 Or 157, 164-65, 598 P2d 670 (1979) (probable cause to arrest does not justify forced entry into home). If the police observe marijuana plants in the garage and, based on their experience, they are virtually certain that the plants are in fact marijuana plants, an entry into the garage will nonetheless be a search that will require a warrant unless some exception applies. Similarly, virtually the entire passenger compartment of a car is generally within the view of anyone who wishes to look. This does not mean that the passenger compartment cannot support a cognizable privacy interest under Article I, section 9. Even if the police observe *219obvious contraband on the front seat and, therefore, are justified in seizing the car, an entry into the car to seize the contraband is a “search” that must be made in a manner consistent with Article I, section 9.
I doubt that the majority would seriously quarrel with the preceding analysis. Their difficulty is not with the scope of the words “search” or “effects” in Article I, section 9, but with that section’s warrant requirement. I suspect that in back of the notion that transparent containers do not support a cognizable privacy interest is the feeling that the police know what they will find, that the issuance of the warrant is practically inevitable, and that requiring a judge to issue a warrant under these hypothesized circumstances is pointless. See, e.g., State v. Lowry, supra, 295 Or at 359 (Jones, J., specially concurring); State v. Flores, 68 Or App 617, 627, 634, 685 P2d 999 (1984) (opinion of Gillette, J.).
I am not so convinced of the unerring ability of the police, or judges for that matter, to divine which containers “announce their contents.” Moreover, Article I, section 9, is clear that determinations of probable cause are to be made by magistrates prior to the invasion of privacy unless one of the well-established exceptions to the warrant clause is applicable. If the majority believes that a warrant should not be required to open a transparent container, it should justify its conclusions in terms of the warrant clause, rather than through tinkerings with the scope of Article I, section 9, based on muddled or nonexistent analysis.5 If it cannot justify its conclusions in terms of that clause, then it should follow that clause’s mandate.
*220II.
The majority also concludes that, when a substance is lawfully seized and there is probable cause to believe that it is a controlled substance, a chemical test of the substance, for the sole purpose of determining whether or not it is a controlled substance, is not a “search” under Article I, section 9. This is a narrow holding, particularly in its limitation of testing to “whether or not” the substance is a controlled substance. If this language is to be taken at face value, a test that discloses that a substance is, for example, talcum powder, rather than simply that it is not a controlled substance, would be a “search.”
The holding may be narrow, but it is nothing more than an ipse dixit. The majority’s entire explication consists of three sentences without any citation to authority. The first sentence states the holding. The second sentence restates the holding as applied to the “search” issue. The third sentence alone provides a clue. That sentence suggests that a chemical test is not a “search” because its purpose is limited. The majority does not explain the basis for such a de minimis argument, does not explain why it would be applicable to this case, and does not explain the possible scope of its application to other cases.
There are a number of problems with this holding. First, the holding simply does not make any internal sense. In a commendable effort to limit the scope of its pronouncement, the majority would require probable cause to believe that the substance tested was a controlled substance. But, as noted above, “probable cause” is irrelevant to whether the action justified by probable cause is a “search.” Moreover, if the test is not a “search” or “seizure” under Article I, section 9, there is no constitutional basis for imposing on the police a probable cause requirement, or indeed any requirement, even that of “reasonableness.” So long as access to the substance was not obtained unlawfully, the police could test the substance for anything and for any reason or no reason.
Second, if the majority’s holding permits testing that will tell the police precisely what the substance is, even if it is not contraband, then the majority’s de minimis argument is somewhat difficult to comprehend. Any number of substances *221look like contraband. For many of these substances, particularly legitimate medicines, knowledge that a person possessed them could be extremely damaging. Persons with medicines to control chronic health problems, such as epilepsy, heart disease or manic depression, might find their careers or relationships jeopardized. A test that could disclose such information could hardly be characterized as de minimis.
If the majority’s holding is limited to tests that disclose only the presence or absence of contraband, its de minimis argument appears to be similar to that advanced by the United States Supreme Court in United States v. Jacobsen, 466 US 109, 104 S Ct 1652, 80 L Ed 2d 85 (1984). In Jacobsen, the Court held that an individual does not have a legitimate expectation of privacy in the possession of contraband. Id. at 123. The Court further held that a “field test” of a substance for the presence of cocaine was not a “search” because the test could reveal only whether the substance was cocaine and thus could not compromise any legitimate interest in privacy. Id.
There is, however, no “contraband exception” to Article I, section 9. Such an exception is unjustifiable in any case, either under Article I, section 9, or the Fourth Amendment. There is no constitutional differentiation in items that the police may search for and seize. “On its face [the Fourth Amendment] assures the ‘right of the people to be secure in their persons, houses, papers, and effects * * *,’ without regard to the use to which any of those things are applied.” Warden v. Hayden, 387 US 294, 301, 87 S Ct 1642, 18 L Ed 2d 782 (1967). So does the Oregon constitutional text.
The notion that officers may seize “contraband” upon probable cause alone but do not need a warrant is wholly illogical. If the argument is that “contraband” is something that may not legally be possessed and, therefore, nothing is seized that belongs to the person from whom it is taken, that argument places the taking altogether outside the constitutional guarantee. On that theory officers also would not need probable cause to seize “contraband.”
Understandably, the majority shies away from that result. It wants officers to have probable cause for the seizure but not to need a warrant to seize “contraband.” But, as so often in search and seizure cases, the majority is blind to the *222trap of reasoning backward from what the search or seizure ultimately reveals. If a warrant ordinarily would be required for a search or a seizure, its omission cannot be justified by the fact that the seized effects are “contraband.” What if, despite probable cause to believe otherwise, the materials seized without a warrant prove to be something other than controlled drugs, or what if the owner turns out to have a prescription for them, so that a warrant was required after all? Does the failure to have obtained a warrant then turn out to have been unlawful and make the officer a trespasser liable for invasion of the possessor’s rights? The majority has no answer for these intrinsic contradictions in its opinion.
The law of search and seizure largely came into existence to protect against searches for contraband. In England, this contraband was primarily unlicensed (and generally “seditious”) publications; in America, it was primarily smuggled goods. Landynski, supra, at 19-48. John Hancock, for example, was a smuggler of goods. Landynski, supra, at 30 n 49. This is not to suggest that either Article I, section 9, or the Fourth Amendment gives criminals a substantive right to keep their activities secret from the government. Rather, both provisions are limitations on the power of executive officers to investigate the private affairs of individuals. Whether those “private affairs” include murder, drug smuggling, eccentric behavior, political activities, sexual relations, or simply reading a book or watching television, the limitation on the government is the same. We may not approve of some or all of these activities, but that is the point. The Star Chamber did not approve of anti-monarchical tracts. George III did not approve of colonial free trade. Richard Nixon did not approve of antiwar activists. The government’s power to declare certain activities or possession of certain articles illegal does not give it the power to define the scope of constitutional safeguards against warrantless search and seizure.
The majority, of course, is not wholly insensitive to these concerns. That much is obvious from its requirement that there be probable cause to test. The majority’s real bete noire is the warrant requirement. In these cases decided today, the majority has, in effect, simply upheld warrantless searches based upon probable cause. Unfortunately, the majority does not explain why warrants are unnecessary in these cases or reconcile these cases with our many cases that hold that *223warrantless searches must be justified under some exception to the warrant requirement. In an effort to avoid such a difficult task, they have further muddied the requirements of Article I, section 9, and unjustifiably contracted its scope.
For years this court followed the rule that a warrantless search was per se unreasonable in the constitutional sense, and to be valid a warrantless search had to fall within one of the narrowly defined and circumscribed exceptions. Now, rather than finding an exception that fits, or even reasoning its way to a new exception, the majority holds that the trespass to these containers is simply not a search. That, at least, is to hold that the Katz analysis supplanted rather than supplemented prior doctrine, a holding contrary to what we so recently said in State v. Louis, supra.
This and other recent decisions of this court, such as State v. Brown, 301 Or 268, 721 P2d 1357 (1986), and State v. Bennett, 301 Or 299, 721 P2d 1375 (1986), mean that the present majority has reversed the longstanding rule that warrantless searches are limited to a few carefully delineated exceptions based on necessity. The majority’s general rule now is that police officers almost always may seize and search a person or personal effects without a warrant on probable cause alone, except when they enter a house to do so and cannot persuade a judge that they obtained prior consent to do so, or in very limited circumstances such as in State v. Kock, 302 Or 29, 725 P2d 1285 (1986).
This reversal of direction is as unnecessary as it is tragic. It may seem futile to dissent at length when the members of the majority so readily abandon analyses that they themselves joined, presumably after attentive reading and consideration, as has recently happened also in State v. Smith, 301 Or 681, 725 P2d 894 (1986). The source of future protection of the rights of Oregonians, however, “remains the constitutional guarantee of a judicial warrant stated in Article I, section 9, enacted by the people in 1859, not what judges write about it.” State u. Brown, supra, 301 Or at 298 (Linde, J., dissenting). It, therefore, is important once more to recall that what the majority has written is not the only, the best, or the final meaning to be given to that guarantee.
Finally, I feel constrained to say that I do not quite understand the purpose of the concurring opinion. That I have *224written a dissent makes clear that I have been unable to convince the majority of the unsoundness of its opinion. Yet, the concurring opinion seems to be nothing more than a “dissent” from my dissent. That the concurring justice does not merely concur in the majority’s result implies that he accepts the majority’s analysis. If he does not, but would rather rest the result on a theory of search incident to arrest, he should concur only in the result and make it clear that only four members of this court believe that opening these containers and analyzing the contents were not searches.
Linde, J., joins in this dissenting opinion.cannot fathom the concurring Justice’s contention that this dissent castigates the majority simply for not following State v. Lowry, 295 Or 337, 667 P2d 996 (1983). If the court errs in construing a constitutional provision, I certainly agree that the court should correct the error. The court has an obligation, however, to state why the court was then in error and why the court is now correct; self-satisfaction with the new result is not sufficient. To point to such an obligation is hardly an effort “to make some sort of juridical virtue out of intellectual intransigence.” 302 Or at 208.
In addition, the concurring opinion’s reference to “the mathematical model to constitutional law,” 302 Or at 212, is bizarre. This “model” bears no resemblance to anything that I can find in Justice Linde’s dissent in State v. Brown, 301 Or 268, 279, 721 P2d 1357 (1986), much less this dissent.
This dissent applies to the majority’s approach to all of the above-mentioned cases, as well as to State v. Herbert, 302 Or 237, 729 P2d 547 (1986).
The majority states that if a person is arrested, no search warrant ever is needed to search a container taken from the person for evidence relating to the crime for which the arrest is made or later is said to have been made. State v. Owens, 302 Or at 201-205. Justice Linde and I have maintained that this is not the correct reading of the warrant requirement. State v. Caraher, 293 Or at 741, 770, 653 P2d 942 (1982) (Lent, J., dissenting). It attributes exaggerated importance to the choice whether or not to arrest a suspect, a decision which itself ordinarily rests only on an officer’s determination of probable cause and which, therefore, does not give the warrantless search any greater legal basis than it would have without the arrest. See State v. Brown, supra, 301 Or at 289-90 (Linde, J., dissenting).
If a 'police officer on a social visit to someone’s home took away a small sample of pills or of a powder from his hosts’ toilet cabinet to determine its composition, I doubt we would hold that he had not trespassed upon an interest recognized by the tort of invasion of privacy. Cf. McClain v. Boise Cascade Corp., 271 Or 549, 533 P2d 343 (1975). Trespass, of course, was and is the common law’s characterization of an unwarranted official entry on private premises or seizure of persons or effects.
The concurring opinion, perhaps implicitly conceding that opening a container and testing its contents are searches under Article I, section 9, seeks to justify the actions of the police in this case under the search-incident-to-arrest doctrine. But that opinion itself expresses doubts, as well it might, about the cogency of extending that doctrine to containers not under the control of an arrested person. After having castigated the present dissenting opinion for relying on a recent precedent, State v. Lowry, supra, the concurring opinion, like the majority, simply cites Caraher as if that decision resolved all issues concerning search-incident-to-arrest. It did not. Moreover, the concurring opinion does not explain in what manner these searches were “close in time and space to the arrest,” State v. Caraher, supra, 293 Or at 760. Even if the searches were somehow “close in time and space to the arrest,” there is no justification for extending the doctrine of search incident to arrest to situations in which there is no practical necessity for a warrantless search. See State v. Caraher, supra, 293 Or at 770-71 (Lent, J., dissenting).