JAMES, J.

Eric Johansen, Deputy Public Defender, argued the cause for appellant. Also on the briefs was Ernest G. Lannet, Chief Defender, Criminal Appellate Section, Office of Public Defense Services.

Peenesh Shah, Assistant Attorney General, argued the cause for respondent. Also on the briefs were Ellen F. Rosenblum, Attorney General, and Benjamin Gutman, Solicitor General.

Before Ortega, Presiding Judge, and Shorr, Judge, and James, Judge.

JAMES, J. Defendant appeals a judgment of conviction on one count of compelling prostitution, ORS 167.017, raising three assignments of error. In his first assignment, he argues that the trial court abused its discretion in denying his request for new court-appointed counsel; we reject that argument without discussion. In another assignment, he asserts that the court committed structural error by instructing the jury that it could return a nonunanimous verdict; but, in light of the fact that the verdict was unanimous, any instructional error in that regard was harmless. State v. Flores Ramos , 367 Or. 292, 319, 478 P.3d 515 (2020) (holding that nonunanimous jury instruction was not a structural error); State v. Ciraulo , 367 Or. 350, 354, 478 P.3d 502 (2020), cert. den. , ––– U.S. ––––, ––– S.Ct. ––––, ––– L.Ed.2d ––––, 2021 WL 2519399 (June 21, 2021) (holding that an erroneous nonunanimous jury instruction was "harmless beyond a reasonable doubt" when the verdict was unanimous).

Defendant's remaining assignment of error, and the focus of our analysis in this opinion, concerns the trial court's denial of defendant's motion to suppress evidence discovered during a search of his cell phone. Defendant argues that the warrant to search the phone was not supported by probable cause and, even if supported by probable cause, lacked the particularity required under Article I, section 9, of the Oregon Constitution.

For the reasons explained below, we conclude that the warrant to search the phone was supported by probable cause but that many of the search commands in the warrant failed to satisfy the particularity requirement under Article I, section 9. Accordingly, the final portion of our opinion focuses on what has, thus far, been little discussed in our ever-evolving case law pertaining to digital warrants: What should a court do when it concludes that some, but not all, of a digital data warrant is insufficiently particular? As we explain, in such situations the court must hold a hearing wherein the state, as the party with the burden, must establish that the evidence sought to be utilized was discovered through a search or forensic analysis responsive to the surviving, constitutionally particular, portion of the warrant, and not derived from the unconstitutional portion of the warrant. In this case, because we cannot tell, on this record, whether the evidence at issue was discovered as a result of the lawful search commands or based on the overbroad or nonspecific commands, we vacate and remand for the trial court to conduct such a hearing and to resolve that question.

I. BACKGROUND

Beaverton Police Detective Opitz was investigating prostitution-related activity and found advertisements on backpage.com offering J for sex individually or as a part of a group of women, which included Gregg, someone police knew from an earlier prostitution sting. In September 2017, Opitz arranged a date with J, who was 17 years old, over text message. Opitz watched defendant drive up in a silver Mercedes and drop off J at the arranged time near the arranged location.

After defendant drove away, Opitz had a conversation with J in which she first claimed to have been dropped off by Uber but later admitted that defendant had dropped her off; J had two cell phones and a box of condoms with her. Opitz told officers to stop defendant's car and to arrest him, which they did. A search of defendant's car yielded some cell phones, a pack of condoms, and an EconoLodge motel room key.

Opitz subsequently prepared a 21-page affidavit in support of a warrant to search the contents of the seized phones, including defendant's cell phone. The beginning of the warrant application set forth Opitz's training and experience, including his experience with investigations related to sex-trafficking and exploitation of children. Among other things, the affidavit stated:

"I know that the internet contains many websites that are used to aid and assist in the advertising of prostitution. Some of the websites used are www.backpage.com, www.eroticmugshots.com, and www.cityvibe.com. I know that minors (subjects under the age of 18) will post or be posted on these websites using photos that are not actually them.

"I also know it is common for those involved in internet based sex trafficking to exclusively utilize cellular telephones to run the prostitution enterprise. An internet capable mobile device is used to take the pictures of the victims, connect to the desired website, generate the advertisement, upload the images, and enter payment information. I know, as a certified cell phone examiner and from discussions with other computer forensics examiners,

each of these activities can be detectable through a forensic examination of the device."

Opitz's affidavit further stated that, based on his experience as a sex trafficking investigator, he knew it "to be common for someone involved in commercial sexual exploitation to maintain virtual constant contact with the victim" such that, "[w]hen not physically together, the cellular phone becomes the primary method of communication between ‘pimp’ and victim," which "serves to both maintain a level of control over the victim, while also serving the practical purpose of directing the girl in the prostitution business," including being able to "tell the girl when to post ads, when and where to be for a ‘date,’ and how much to charge." The affidavit then described, in a general way, the kinds of information that can be discovered by a forensic search of a cell phone.

The next sections of the affidavit concerned Opitz's investigation, the arrests of defendant and Gregg, and an interview conducted with J once she was transported to the Beaverton Police Department. Among other averments, Opitz stated that he received information in August 2017 that defendant and Gregg were trafficking J for sex, and that he learned from police reports that J had recently been residing with defendant and Gregg in Vancouver, Washington. He stated that he was able to locate a number of prostitution-related postings on backpage.com for both Gregg and J, that he knew that backpage.com is used by prostitution providers, traffickers, and customers, and that he found postings for the Portland and Vancouver areas associated with Gregg dating back to June 15, 2017, and postings incorporating J as part of a "2 for 1" deal with Gregg beginning on July 1, 2017. Opitz also described postings under a different phone number, beginning August 29, 2017, advertising J alone and as part of dates with one or two other women; one of the postings included a photo of J, Gregg, and another female posing in underwear "in front of what appears to be a silver colored Mercedes Benz with no front license plate."

According to the affidavit, J told Opitz that she and defendant were "boyfriend/girlfriend," that defendant knew that she was 17, and that J had engaged in prostitution during the time that she, Gregg, and defendant had resided together. She told Opitz that she used "the Pinger app to text/communicate with johns," which was how she had communicated with Opitz. J also admitted to Opitz that she posted ads on backpage.com but did not pay for the ads, and that Gregg had taught her what to put in the ads, told her what to say to customers and how to act with them, and that Gregg had been the one who introduced her to defendant. She also told Opitz that defendant had rented a motel room at an EconoLodge, and that J had at least one prostitution date there, on September 3, 2017.

After the recitation of facts related to the investigation, the affidavit set forth Opitz's conclusion that he had probable cause to believe that defendant had committed the crimes of prostitution, promoting prostitution, and compelling prostitution in Washington County, Oregon, and probable cause to believe that evidence of those crimes would be found on, among other items, defendant's cell phone. Opitz identified nine categories of evidence to be "searched, seized, and analyzed":

"(1) Any and all communications (voice, email, text, or otherwise) between [J, Gregg, and/or defendant].

"(2) Evidence related to the relationship between [J, Gregg, and/or defendant].

"(3) Evidence regarding any communications (voice, email, text, or otherwise) involving prostitution related activities.

"(4) Any photos of [J, defendant, or Gregg] that show an association with prostitution including any profiting from prostitution.

"(5) Images, videos and/or data which depict [J or Gregg] in sexually explicit positions or conduct that relate to internet postings or advertisements.

"(6) Any evidence related to use of internet sites associated with prostitution, including backpage.com for a period of time 06/15/2017 to 09/06/2017.

"(7) Any evidence related to the use of Uber or other ride-sharing or taxicab companies.

"(8) Any evidence regarding the locations, including geolocation information, of

the phones for a period of time from 06/15/2017 to 09/06/2017.

"(9) Any other evidence related to the crimes of Prostitution ( ORS 167.007 ), Promoting Prostitution ( ORS 167.012 ) and/or Compelling Prostitution ( ORS 167.017 )."

A magistrate issued a warrant authorizing the search of defendant's phone that included all nine of those search commands. Police executed that warrant, and the search of defendant's phone yielded photographs from defendant's phone (State's Exhibit 14), text messaging from defendant's phone to a contact named "baby," which Opitz determined to be J (State's Exhibit 19), and text messaging from defendant's phone to J (State's Exhibit 20).

Defendant moved to suppress "any and all objects, information, statements, and observations obtained by and resulting from the search" of defendant's phone, arguing that the affidavit in support of the warrant failed to establish probable cause that incriminating evidence would be found in his cell phone, and, in any event, that the warrant lacked specificity and was overbroad under Article I, section 9. The parties’ arguments in their briefing and at the hearing on the suppression motion were framed in light of our decision in State v. Mansor , 279 Or. App. 778, 801, 381 P.3d 930 (2016) ( Mansor I ), the controlling authority at the time, which addressed the particularity requirement in the case of a search for digital evidence on a computer. The trial court denied defendant's motion to suppress, the evidence from defendant's phone (State's Exhibits 14, 19, and 20) was offered at trial, and defendant was convicted of one count of compelling prostitution. II. DISCUSSION

On appeal, defendant argues that the trial court erred in denying his motion to suppress because Opitz's affidavit failed to establish probable cause to believe that incriminating evidence would be found on defendant's phone, and because the issued search warrant lacked particularity for purposes of Article I, section 9.

A. Probable Cause

Initially, defendant argues that Opitz's affidavit fails to establish more than a speculative possibility that evidence would be found on defendant's phone. See State v. Cazee , 308 Or. App. 748, 482 P.3d 140 (2021) (explaining that we review a trial court's determination of probable cause to support a search warrant for legal error, evaluating whether the affidavit alleged sufficient facts to permit a neutral and detached magistrate to determine that seizable evidence probably would be found in the place to be searched—i.e. , the cell phone). We reject defendant's probable cause argument.

The probable cause analysis is the same for warrantless searches as it is for those conducted with a warrant. State v. Foster , 350 Or. 161, 169, 252 P.3d 292 (2011). Probable cause exists if the facts on which the officers relied would "lead a reasonable person to believe that seizable things will probably be found in the location to be searched." State v. Anspach , 298 Or. 375, 381, 692 P.2d 602 (1984). That is, there must exist "a nexus between three things: (1) that a crime has been, or is currently being, committed, and that (2) evidence of that crime (3) will be found in the place to be searched." State v. Van Osdol , 290 Or. App. 902, 908, 417 P.3d 488 (2018) ; State v. Goennier , 291 Or. App. 694, 698, 422 P.3d 391, rev. den. , 363 Or. 481, 424 P.3d 729 (2018) (holding same). To place that in visual terms—we can conceive of probable cause as three overlapping Venn diagram circles, shown below:

One circle is the specific crime of investigation. There must be facts that support a specific crime, not just generalized concerns of criminal activity. The second circle is the location to be searched. And, the third circle is the evidence of the specific crime of investigation that is expected to be found in the location searched. Where those three Venn circles overlap is the nexus—the point at which constitutional probable cause exists. The standard is one of probability, not certainty. Anspach , 298 Or. at 380-81, 692 P.2d 602.

Applying that here, the affidavit in support of the warrant was sufficient to establish that J was using a cell phone to communicate via text message to arrange prostitution dates; that J had been offered as part of duo dates with Gregg, and engaged in prostitution for defendant while living with defendant and Gregg; that Gregg had taken nude photographs of J; that defendant had dropped off J for the prostitution date arranged by Opitz; that J had two cell phones in her possession at the time she was dropped off; and that defendant was in possession of a cell phone in his car at the time that he dropped off J. The affidavit also recites Opitz's training and experience that it is "common for someone involved in commercial sexual exploitation to maintain virtual constant contact with the victim" such that "the cellular phone becomes the primary method of communication between ‘pimp’ and victim" for purposes of control and practical purposes of directing the business (including being able to "tell the girl when to post ads, when and where to be for a ‘date,’ and how much to charge"). Those recitations, taken together, are sufficient to establish that it is more likely than not that defendant, Gregg, and J were using cell phones to conduct and facilitate prostitution activities, and that evidence, such as text messages (the evidence circle of the Venn diagram), of the crime of promoting prostitution (the crime of investigation circle of the Venn diagram) would be found on defendant's phone (the location circle of the Venn diagram). Accordingly, the facts set forth in the affidavit in this case place us at the center of those overlapping Venn diagrams—the nexus—where constitutional probable cause exists. Goennier , 291 Or. App. at 698, 422 P.3d 391. Thus, defendant's phone was lawfully seized and subject to search for evidence of the crimes of prostitution, promoting prostitution, and compelling prostitution.

B. Particularity

The fact that defendant's phone was lawfully seized and subject to search does not, by itself, permit the state to analyze and use all of the information found on the phone. Under Article I, section 9, a search warrant must "particularly describ[e] the place to be searched, and the person or thing to be seized." The warrant must serve to guide the hand of the executing officer. In the case of warrants for data, because the typical data storage device will contain more unresponsive data than responsive data, particular care in guiding the executing officer's hand must be given. We therefore turn to defendant's assertion that the warrant lacked particularity, which presents a question of law that we review for legal error. State v. Paye , 310 Or. App. 408, 413, 486 P.3d 808 (2021).

As noted earlier, the trial court's ruling in this case was based on our analysis of particularity in Mansor I . However, the Supreme Court subsequently issued its decision in State v. Mansor , 363 Or. 185, 421 P.3d 323 (2018) ( Mansor II ), which affirmed our decision in that case but involved an analysis that differed from ours "in some fairly significant ways." Paye , 310 Or. App. at 413, 486 P.3d 808. In State v. Savath , 298 Or. App 495, 447 P.3d 1, rev. den. , 365 Or. 722, 453 P.3d 548 (2019), we summarized Mansor II before applying it in the context of a search of a cell phone. We repeat some of that summary here:

"Both the Supreme Court of the United States and the Oregon Supreme Court have recognized that, with regard to searches for data on cell phones or similar electronic devices, the particularity requirement takes on special significance. ‘[U]nlike most other "things" that may be seized in a search, a computer or other digital device is a repository with a historically unprecedented capacity to collect and store a diverse and vast array of personal information.’ [Mansor II , 363 Or.] at 208, 421 P.3d 323 (recognizing that a cell phone might be better viewed as a ‘place’ to be searched than a ‘thing’ to be examined); see Riley v. California , 573 U.S. 373, 396-97, 134 S. Ct. 2473, 189 L. Ed. 2d 430 (2014) (observing that ‘a cell phone search would typically expose to the government far more than the most exhaustive search of a house’ (emphasis in original)). At the same time, however, the application of the particularity requirement to warrants authorizing electronic searches creates challenges not usually present with physical searches. ‘[A] category of information that is a likely source of evidence * * * may be composed of many types of data and files, and the * * * software's organization of those data and files may be unrelated to the user's perception of how their data is organized.’ [ Mansor II , 363 Or. at 197, 421 P.3d 323 ] (rejecting, for that reason, requirement of ex ante limitations based on file or data type or specific application). Further, there is typically ‘no way to know what data a file contains without opening it,’ as specific files may be hidden or disguised, either intentionally or inadvertently. Id. at 198 . Therefore, an electronic search ‘likely will need to examine, at least briefly, some information or data beyond that identified in the warrant.’ Id. at 218, 421 P.3d 323. * * *.

"In light of those considerations, a warrant for an electronic search ‘must identify, as specifically as reasonably possible in the circumstances, the information to be searched for.’ [ Mansor II , 363 Or. at 218, 421 P.3d 323 ] (emphasis added). As [ Mansor II ] explains, the essential ‘thing’ about which a warrant must be particular is the probative information, not types of files or their location within the computer's

file-management system: ‘[T]he "what" is a description of the information related to the alleged criminal conduct which there is probable cause to believe will be found on the computer.’ Id. at 216 (emphasis in original). To further narrow the search and specify the information sought, the warrant must include, ‘if relevant and available, the time period during which that information was created, accessed, or otherwise used.’ Id. at 218 ."

298 Or. App. at 499-501, 447 P.3d 1.

In Savath , we considered whether the following search command was sufficiently particular:

"[Evidence of the crimes of unlawful possession and delivery of methamphetamine and oxycodone] located on [the defendant's phone], to wit: all names and telephone numbers that have been recorded on the cell phone to include all outgoing calls, incoming calls, missed calls, phone

contact lists and address items; all messages both voice and text, text drafts and emails; all photos, videos; and downloaded items related to controlled substance offenses that may be on the phone."

Id. at 497-98, 447 P.3d 1. We agreed with the defendant that the command said "almost nothing about the information they may seek—the ‘what’ as to which the warrant must be particular," and that "the warrant's summary characterization of the information sought—‘related to controlled substances offenses’—was insufficient to apprise the executing officer of which items were or were not subject to the warrant." Id. at 501-02, 447 P.3d 1.

Relying on the Supreme Court's decision in Mansor II , defendant argues that the nine search commands were insufficiently specific and therefore unlawfully permitted officers to "rummage through the entirety of defendant's phone and decide what they believed was relevant," thereby making it a general warrant. According to defendant, the warrant failed to limit the search to relevant dates, even though the police knew the time period when relevant documents were created or internet sites visited, and it did not include any time-based description of the information sought or time period during which such information was created, accessed, or otherwise used. The state, for its part, defends only a subset of the search commands under Mansor II . The state argues:

"Its command to search communications was limited to those between defendant, Gregg, and the victim, or those discussing prostitution generally. The same is true for its command to search photographs, which was limited to photographs either depicting a connection to prostitution or suitable for use in an advertisement for prostitution. And although the warrant commanded a search for other data, any overbreadth in those commands was severable."

The state expressly does not concede that the remaining commands were overbroad, but neither does it develop any argument as to how they satisfy the particularity requirements described in Mansor II .

In assessing whether the commands are sufficiently particular, we consider not only the warrant itself but Opitz's affidavit in support of the warrant, which the parties below treated as a package. See Mansor II , 363 Or. at 203-04, 421 P.3d 323 (holding that the defendant bore the burden of establishing facts pertaining to the challenge to the validity of the warrant, including controverting the inference that affidavit physically accompanied the warrant and was incorporated). In Mansor II , the warrant authorized police to "seize and search and forensically examine" certain items listed in an attachment, including the defendant's computers. Id. at 204-05, 421 P.3d 323. Yet, the only description of any relevant information that the affiant believed might be found on the computers was that of internet search history for a particular date. Id.

The court viewed "that description of the information to be searched for as a limitation on the search, analysis, and forensic examination authorized by the warrant." Id. at 205, 421 P.3d 323. It concluded:

"The warrant, read in conjunction with and limited by the affidavit, met the particularity requirement of Article I, section 9, as we have articulated it above. It sufficiently described the ‘what’ to be searched for and the

relevant time frame: The June 12 internet search history. It informed those executing the warrant as to what they were to look for with a reasonable degree of certainty. And, because that description limited the extent of the search that was authorized by the warrant, as we read it, the permitted search was not broader than the supporting affidavit supplied probable cause to justify."

Id. at 219, 421 P.3d 323 (internal citations, alterations, and quotation marks omitted).

As we understand Mansor II , when an affidavit is considered as part of the warrant, the search commands must be read in conjunction with and limited by the descriptions in the affidavit. A search warrant cannot be broader than a supporting affidavit supplies probable cause to justify, so the supporting affidavit can operate to inform those executing a warrant as to what they are authorized to search for under a specific command—i.e. , limit the parameters of the search command. Id.

Viewing the nine search commands in light of Opitz's attached affidavit, we agree with defendant—and with the state's partial and implicit acknowledgment—that many of the commands in the warrant do not comport with the level of particularity demanded by Article I, section 9, as articulated in Mansor II .

The first two search commands—"[a]ny and all communications (voice, email, text, or otherwise) between [J, Gregg, and/or defendant]" and "[e]vidence related to the relationship between [J, Gregg, and/or defendant]"—include no restrictions on the time or subject matter of the information that is sought. Calling forth the image of the Venn diagram referenced earlier, one circle concerned the specific crime of investigation. Here, the search for texts or communications between two people, without time limitation, is disconnected from the specific crime of investigation, morphing into a search for more generalized nefarious activity. Further, "evidence related to the relationship" is so broad that nearly anything could be contemplated, thus failing to articulate the specific evidence to be sought—another circle in our Venn diagram. For these reasons, they were insufficient to apprise an executing officer of which information was or was not subject to the warrant. Nothing in the affidavit would serve to provide reasonable parameters on what type of information was being sought, so the command amounts to a general warrant for a search of anything incriminating.

The seventh command (Uber or ride-sharing) and ninth command (a catchall for any other evidence related to prostitution crimes) similarly lack the requisite specificity to allow an executing officer to identify the information that is sought. Although the affidavit mentions J claiming to have been dropped off by Uber, that was in the context of a lie to Opitz about whether defendant had dropped her off. Neither of those commands supplies dates, subject matter limitations, or other parameters that would provide a reasonable degree of specificity to an officer executing those commands, despite the fact that police were investigating prostitution activity that occurred during discrete periods of time, at known physical locations (like the EconoLodge and defendant's residence), by known persons, J and Gregg. Nothing in the attached affidavit can be understood to cabin the commands or identify the information sought with any reasonable degree of certainty.

The eighth command—for all geolocation information over a three-month span—likewise lacks specificity, given the nature of the prostitution activity that was being investigated, and Opitz's affidavit does not include descriptions of locations or activities that would limit and provide any reasonable degree of certainty as to what the eighth command sought; it amounted to a general hunt through the phone for its whereabouts for three months, without any guidance to an executing officer as to what geolocation data would be evidence of any criminal activity. See State v. Bock , 310 Or. App. 329, 334, 485 P.3d 931, 935 (2021) ("Any interpretation of the search command broad enough to permit the use of any material discovered on the cell phone relevant to establish the device owner or user's identity is impermissibly nonspecific. A warrant without clear limitations on the material subject to search and seizure requires the executing officer to employ discretion in deciding what to search or seize." (Emphasis in original.)).

That leaves the middle four commands:

"(3) Evidence regarding any communications (voice, email, text, or otherwise) involving prostitution related activities.

"(4) Any photos of [J, defendant, or Gregg] that show an association with prostitution

including any profiting from prostitution.

"(5) Images, videos and/or data which depict [J or Gregg] in sexually explicit positions or conduct that relate to internet postings or advertisements.

"(6) Any evidence related to use of internet sites associated with prostitution, including backpage.com for a period of time 06/15/2017 to 09/06/2017."

We have little trouble concluding that the latter two of those commands are sufficiently particular to satisfy Article I, section 9. The fifth command specifically describes the evidence that is sought by identifying the persons in the photos, videos, or other data, and the activities that are depicted: sexually explicit positions or conduct that relate to internet postings or advertisements. Given that level of specificity, we agree with the state that the absence of a temporal limitation was not fatal to the search command. See Mansor II , 363 Or. at 218, 421 P.3d 323 ("[A]nalytically, ‘temporal limitations’ are more accurately seen as a way of identifying with greater specificity the ‘what’ that is being searched for, rather than as a separate, independently required element, in meeting the particularity requirement for a computer search.").

The sixth command is likewise sufficiently particular. Although the phrase "related to use of internet sites associated with prostitution" provides little direction to an executing officer, the rest of the command identifies a specific site, backpage.com, and Opitz's affidavit lists two other sites (www.eroticmugshots.com and www.cityvibe.com). Given the time limitations in the command (use during the "period of time 06/15/2017 to 09/06/2017"), as well as the specific websites listed in the warrant and affidavit, the search command is sufficiently specific and not overbroad.

The third and fourth commands present closer questions. As for the third command, the phrase "involving prostitution related activities" is open-ended and, in the abstract, gives little guidance to an officer executing those commands. On the other hand, in Paye , we recently upheld a similar command in the context of a warrant to search a computer for evidence of a defendant's involvement in prostitution-related crimes.

In Paye , the affidavit in support of the warrant stated that one of the women whose services the defendant promoted had reported that the defendant " ‘had other women working for him’ " and that the woman had " ‘observed "listings" on his computer.’ " 310 Or. App. at 415, 486 P.3d 808. The search command, in turn, was broadly worded:

" ‘Any and all evidence documenting the [crimes of promoting prostitution and compelling prostitution], to include: any and all digital images, digital video clips, and or photographs depicting [C], and/or any other as-of-yet unidentified females; Contact information: to include, telephone numbers, names, and electronic mail (email) addresses; private messages; data storage identifying information; SMS/text messages and history; emails[.]’ "

Id.

In analyzing the question of particularity, we explained that "those specifications—which indicate that the evidence expected to be found on defendant's computer consists of images and videos of women he prostituted, contact information for people connected with the crimes, and communications about the crimes—satisfy the standard set by Mansor II for describing what evidence reasonably can be expected to be found on a computer." Id. We then expressly rejected the defendant's argument that the absence of temporal limitations on the search rendered it overbroad:

"[T]he warrant was supported by probable cause that defendant was engaged in an ongoing enterprise promoting prostitution, and that evidence of that enterprise would be found on his computer. The warrant identified with specificity what the evidence was. There is no indication that

specific relevant time frames pertaining to that ongoing activity were ‘available to the police,’ and, given the ongoing nature of defendant's prostitution business, such limitations, to the extent relevant, are not highly so. Thus, under Mansor II , this is not a situation in which the omission of

temporal limitations renders the warrant unconstitutionally overbroad."

Id. at 416-17.

In light of Paye , we conclude that the command to search for evidence "regarding any communications (voice, email, text, or otherwise) involving prostitution related activities" was sufficiently particular in the context of the supporting affidavit and crimes involved. As explained earlier, Opitz's affidavit gives rise to probable cause to believe that defendant, J, and Gregg were using cell phones to communicate about and arrange prostitution dates among one another and with potential johns, including through text messages about those dates. Even without specific date ranges supplied in the command itself, it sufficiently indicates what —that is, messages and communications related to prostitution activities—that reasonably would be expected to be found on defendant's cell phone beginning around June 2017 through the fall of 2017; given the ongoing nature of defendant's activities, strict time parameters were not necessarily available to police, even if they knew when some of the activity took place.

As was the case in Paye , the affidavit created probable cause to believe that the prostitution crimes were part of an ongoing business involving defendant and Gregg and that predated their involvement with J and involved other unknown women as well. Given the nature of the crimes, as supported by probable cause in the attached affidavit, the command identified the information to be searched for as specifically as reasonably possible in the circumstances. See Paye , 310 Or. App. at 416, 486 P.3d 808 (explaining that the "omission of such limits is not, ultimately, problematic if such limitations are not ‘relevant and available to the police,’ and if the warrant otherwise adequately identifies with specificity the evidence expected to be found based on the probable cause determination"). We reach a different conclusion, however, with regard to the fourth command, which authorizes a search for "[a]ny photos of [J, defendant, or Gregg] that show an association with prostitution including any profiting from prostitution." The command identifies specific persons in the photos, but it uses the vague phrase "association with prostitution including any profiting from prostitution." That phrase provides little, if any, guidance as to what reasonably could be expected to be found on the phone. And, unlike the third search command, the fourth cannot be saved by the broader context of the supporting affidavit. To the extent photos are mentioned in the affidavit, they are photos that are sexually suggestive or part of advertisements—the subject of a separate search command (the fifth command). An officer executing the fourth search command would not have any clear limitations on the material subject to search and seizure, amounting to the type of rummaging that violates Article I, section 9.

In sum, we conclude that only the third, fifth, and sixth of the nine search commands were valid. The remainder lacked the particularity required by Article I, section 9.

C. Severability Issues

The remaining question is what to do about the fact that some search commands were valid but others were not. Ordinarily, "[i]f a portion of a search warrant fails to describe the items sought with sufficient particularity, that portion may be excised and the balance of the warrant upheld." State v. Vermaas , 116 Or. App. 413, 416, 841 P.2d 664 (1992), rev. den. , 316 Or. 142, 852 P.2d 838 (1993) (citing State v. Sagner , 12 Or. App. 459, 471, 506 P.2d 510 (1973) (adopting the court's severance reasoning in Aday v. Superior Court of Alameda Cty. , 55 Cal 2d 789, 797, 362 P.2d 47, 52, 13 Cal.Rptr. 415 (1961) )). See also State v. Burnham , 289 Or. App. 783, 785, 412 P.3d 1233 (2018) ("[T]he trial court only erred insofar as it did not suppress evidence seized under the overbroad portion of the warrant, and it did not err by admitting evidence covered by the valid portions of the warrant.").

The exception to that rule is where the warrant is essentially general but meets the particularity requirement "as to minor items." Sagner , 12 Or. App. at 471, 506 P.2d 510 ("Both [ Aday ] and United States v. Ketterman , 276 A.2d 243 (D.C. App. 1971), caution that invalid portions of a warrant will not be considered severable in all circumstances. If a warrant is essentially general, but meets the requirement of particularity as to minor items, ‘[s]uch an abuse of the warrant procedure, of course, could not be tolerated.’ "); Aday , 55 Cal 2d at 797, 13 Cal.Rptr. 415, 362 P.2d at 52 ("We recognize the danger that warrants might be obtained which are essentially general in character but as to minor items meet the requirement of particularity, and that wholesale seizures might be made under them, in the expectation that the seizure would in any event be upheld as to the property specified. Such an abuse of the warrant procedure, of course, could not be tolerated."). Accord Wayne R. LaFave, 2 Search and Seizure: A Treatise on the Fourth Amendment § 4.6(f) (6th ed Sept 2020 Update) (addressing partial invalidity and explaining that, "At least when there is no problem of pretext and the valid portions of the warrant are not an ‘insignificant or tangential part of the warrant,’ no problem is presented in cases like Aday , where the items seized are those for which probable cause existed and which were particularly described." (Footnotes omitted.)). This is not a case in which the invalid commands could be said to be an insignificant or tangential part of a general warrant.

Here, as the preceding analysis suggests, the invalid commands can be assessed independently and can be severed from the valid portions. The difficulty is the next step: determining what evidence was the product of one of the invalid, rather than valid, search commands. As noted earlier, the state discovered and offered at trial text messages and photos (which included backpage.com screenshots) that had been discovered on defendant's phone. However, the suppression record before us does not include evidence about how the search of the phone was conducted; that is, we cannot tell what search command was being executed when the text messages and photos were discovered.

In this case, the state argues that we can nevertheless affirm by simply working backward from the evidence that was discovered, tracing it to a lawful command. In other words, the state's view is that we look to whether the evidence that is subject to the suppression motion is the kind that police would be expected to discover from one of the lawful commands, regardless of how the electronic search was actually conducted.

We disagree with that approach. The question before us under Article I, section 9, is not whether the state could have searched and seized evidence under a valid warrant; the question is whether it did . The answer to that question depends not on how a forensic search of the phone might have been conducted; rather, it depends on how the search was actually conducted . As the court explained in Mansor II , "when the state conducts a reasonably targeted search of a person's computer for information pursuant to a warrant that properly identifies the information being sought, the state has not unreasonably invaded the person's privacy interest, and the state may use the information identified in the warrant in a prosecution or any other lawful manner." 363 Or. at 221, 421 P.3d 323. But, "when the state looks for other information or uncovers information that was not authorized by the warrant, Article I, section 9, prohibits the state from using that information at trial, unless it comes within an exception to the warrant requirement." Id.

In Bock , we recently rejected the idea that the "plain view" doctrine plays any role in the context of an electronic search. That is, we rejected the state's argument that it could use evidence discovered outside the scope of a lawful search command so long as it was not "far beyond" the scope of the warrant. 310 Or. App. at 340, 485 P.3d 931. Consequently, any evidence that is not discovered pursuant to a lawful search command must be suppressed; otherwise, we would "sanction the sort of general warrant that the plain view doctrine was never meant to authorize." Id.

This case ultimately presents the same problem that we encountered in State v. Frischman , 298 Or. App. 186, 188-89, 445 P.3d 946, rev. den. , 365 Or. 721, 453 P.3d 546 (2019), and it requires the same disposition. In Frischman , we agreed with the state that the invalid portions of the warrant could be excised, allowing for admission of the evidence obtained pursuant to the valid portions. But the parties did not "appear to have developed the facts as to what evidence, in particular, was seized as a result of the invalid portion of the warrant or to have understood fully the need for that factual development following the determination that the warrant was partly good and partly bad." Id. at 188, 445 P.3d 946. And, at oral argument, the parties "presented competing views of whether the seizure of certain items authorized by valid portions of the warrant may, nonetheless, have been tainted as a result of the officers’ execution of the invalid portion of the warrant." Id. at 188-89, 445 P.3d 946. In that circumstance, we concluded that "a remand is required so that the trial court can determine what evidence is admissible pursuant to the valid portions of the warrant and what evidence must be suppressed because it was obtained based on the invalid portion of the warrant," and we left it to the trial court to determine in the first instance whether the officers’ execution of the invalid portion of the warrant tainted the execution of the valid portion. Id.

Here, as in Frischman , the record created below is insufficient for us to sort out what parts of the evidence admitted at trial (State's Exhibits 14, 19, and 20) was discovered while police were executing one of the lawful search commands as opposed to one of the invalid commands. We therefore must remand for development of a record as to how the forensic search of the phone was conducted.

In remanding this case, we observe that an adequate record may likely require testimony from the officers or forensic technicians who conducted the search; an explanation of the steps that were undertaken while executing the warrant; and, importantly, evidence of what search commands were being carried out at each step of the search of the phone, and what those commands generated. As the Supreme Court noted in Mansor II , "although a computer search may need to be broad, it must be reasonably executed." 363 Or. at 218 n 15, 421 P.3d 323 (citing Or. Const., Art I, § 9). On remand, the parties can further address the manner in which this warrant was executed and trace the discovery of the evidence that is the subject of defendant's motion to suppress; and, depending on whether all or part of the evidence at issue was lawfully discovered, the trial court can determine, in the first instance, whether a new trial is necessary.

Vacated and remanded.