Defendant, who was convicted on two counts of violating a stalking protective order (SPO), ORS 163.750, argues that the trial court erred in denying his motion for a judgment of acquittal. For the reasons set forth below, we reverse.
Because the jury found defendant guilty, we state the facts in the light most favorable to the state. State v. Gibson, 338 Or 560, 562, 113 P3d 423, cert den, 546 US 1044 (2005). Defendant had been pursuing the victim, V, since 2005. We described the pertinent background facts surrounding the issuance of the SPO in V’s favor in Van Buskirk v. Ryan, 233 Or App 170, 172-73, 225 P3d 118, rev dismissed, 348 Or 218 (2010):
“[Petitioner [V] writes for the Portland Tribune. She met respondent once at an open house held by petitioner’s employer. Over the next two years, respondent never saw petitioner again, but communicated with her through letters and e-mails, attempting to establish a romantic relationship with her. In many of respondent’s correspondences, he told petitioner that, if the contact was inappropriate, ‘please let me know and I will stop.’ However, petitioner or those acting on her behalf repeatedly asked respondent to stop attempting to contact her, to no avail.
“In his letters, respondent commented on petitioner’s articles and on her appearance. He extended invitations to meet for lunch or dinner, shared his project ideas, expressed gratitude for her time and support, and often concluded with such statements as ‘[wjishing you a joyful day!’ or ‘[e]njoy the day!’ Petitioner has a young son whom she had mentioned in some of her newspaper articles, and the letters also made references to him. The letters reflected that respondent was experiencing irrational and delusional thinking about his relationship with petitioner. In a letter to petitioner’s parents, he compared them to Romeo and Juliet. The letters repeatedly referred to people who had been ‘contaminated’ and who were against him and his desire not to expose petitioner to ‘danger or contamination.’ Petitioner received a few letters from respondent in 2005, 11 letters from respondent in 2006, and 12 letters in 2007.
*320“Respondent also attempted to contact petitioner by telephone and in person. He called her on her home and work telephones and left messages for her but never spoke to her, because she screened his calls.
“Respondent attempted many times to visit petitioner at work to drop off letters for her. He never succeeded in seeing petitioner at her workplace, because petitioner’s coworkers, supervisor, and receptionist intervened; however, respondent’s visits were troubling to petitioner and to her coworkers. Respondent began attending public events sponsored by the Tribune or its sister company.
“Respondent attempted to contact petitioner through her parents, although he never succeeded in meeting them in person.”
(Third and fourth brackets in original.) Based on that evidence, and evidence that V feared for her safety and the safety of her son, we concluded that the trial court properly issued an SPO. In doing so, we agreed with the respondent (defendant in the present case) that his communication-based contacts with V did not provide a basis for issuing an SPO because, under the test announced in State v. Rangel, 328 Or 294, 977 P2d 379 (1999), those communications did not contain “an unequivocal threat that was objectively likely to be followed by unlawful acts and that instilled in petitioner a fear of imminent and serious personal violence.” Van Buskirk, 233 Or App at 177. We concluded, nonetheless, that the communications provided context for evaluating the petitioner’s noncommunicative contacts with V, which included seeking her out at work on multiple occasions and locating and trying to visit her parents. Id. Those noncommunicative contacts, we concluded, demonstrated “a pattern of behavior that made petitioner’s apprehension reasonable.” Id.
The SPO issued in March 2007. The charges at issue in the present case were based on events that occurred on or about May 7 and May 15, 2007. There is no significant dispute about the nature of the contacts that occurred on those dates. On or about May 7, 2007, defendant mailed a letter to V’s father at his home address. That letter, like those described in our previous opinion, reflected defendant’s ongoing irrational and delusional belief that he had a personal *321relationship with V, and contained numerous incomprehensible statements. In the letter, defendant expressed a desire to meet with Vs father and also asked him to thank V for him. The letter contained no threats of any sort. The letter contained an enclosure — a letter to a pastor of a church that defendant apparently believed V attended — that also reflected defendant’s obsessive belief that he had a personal relationship with V and that contained nothing that could be described as a threat.
On or about May 15, 2007, defendant mailed a package to Vs father at his place of work. The package contained a blank Mother’s Day card, a music CD, and another letter. In the letter, defendant again asked Vs father to thank V for him and indicated that defendant thought that V was expecting him to fulfill a promise. He expressed a desire to get to know V better, but noted that he would “not put her in any danger.” Again, the letter, while rambling and incoherent, contained no threats.
Defendant moved for a judgment of acquittal on the ground that the contacts that the state alleged to have violated ORS 163.750 were constitutionally protected communications under Article I, section 8, of the Oregon Constitution as interpreted in Rangel. The trial court denied defendant’s motion for a judgment of acquittal without explanation, and a jury subsequently convicted defendant on the two charges at issue here.
On appeal, defendant reiterates his argument that, because the two contacts alleged as the basis for his convictions were constitutionally protected speech under Article I, section 8, the trial court erred in denying his motion for a judgment of acquittal. For the reasons set forth below, we agree with defendant.
We begin with a description of the pertinent statutes relating to stalking. Two statutes allow for the issuance of SPOs and identically describe stalking as follows:
“(a) The person intentionally, knowingly or recklessly engages in repeated and unwanted contact with the other person or a member of that person’s immediate family or household thereby alarming or coercing the other person;
*322“(b) It is objectively reasonable for a person in the victim’s situation to have been alarmed or coerced by the contact; and
“(c) The repeated and unwanted contact causes the victim reasonable apprehension regarding the personal safety of the victim or a member of the victim’s immediate family or household.”
ORS 30.866(1) (defining circumstances in which a plaintiff may initiate a civil action to obtain a stalking protective order); ORS 163.735(1) (defining circumstances in which a law enforcement officer may issue a citation to show cause why a court should not enter a stalking protective order). As described in our previous opinion, V brought a civil action and obtained a stalking protective order pursuant to ORS 30.866(1). Under ORS 30.866(2) and ORS 163.738(2)(b), an SPO may prohibit “all contact listed in ORS 163.730.” The definition of “contact” found in ORS 163.730, which specifically applies to ORS 30.866 as well as to ORS 163.732 and ORS 163.750, provides, in part:
“(3) ‘Contact’ includes but is not limited to:
“* * * * *
“(d) Sending or making written or electronic communications in any form to the other person;
“(e) Speaking with the other person by any means;
“(f) Communicating with the other person through a third person;
“* * * * *
“(h) Communicating with a third person who has some relationship to the other person with the intent of affecting the third person’s relationship with the other person;
“(i) Communicating with business entities with the intent of affecting some right or interest of the other person.”
Regardless of whether an SPO is issued based on an action initiated pursuant to ORS 30.866 or ORS 163.735, the violation of any such SPO is criminalized by ORS 163.750, which provides:
*323“(1) A person commits the crime of violating a court’s stalking protective order when:
“(a) The person has been served with a court’s stalking protective order as provided in ORS 30.866 or 163.738 or if further service was waived under ORS 163.741 because the person appeared before the court;
“(b) The person, subsequent to the service of the order, has engaged intentionally, knowingly or recklessly in conduct prohibited by the order; and
“(c) If the conduct is prohibited contact as defined in ORS 163.730(3)(d), (e), (f), (h) or (i), the subsequent conduct has created reasonable apprehension regarding the personal safety of a person protected by the order.”
(Emphasis added.)
In the present case, the contacts alleged to have violated the SPO both involved “communicating with [V] through [V's parents],” thus implicating the “contact” described in ORS 163.730(3)(f), “[communicating with the other person through a third person.” Because the alleged unlawful conduct was of the type described in ORS 163.730(3)(f), the state was required to prove that it “created a reasonable apprehension regarding the personal safety of a person protected by the order.” ORS 163.750(1)(c).
With that statutory background in mind, we turn to Rangel, the case that is central to both parties’ arguments. In Rangel, the court considered the question whether a conviction for the crime of stalking under ORS 163.732,1 was constitutionally infirm under Article I, section 8, on the ground that the statute criminalized conduct that included protected expression. The court stated in Rangel:
*324“Article I, section 8, does not prohibit the enactment of statutes that focus on forbidden effects of expression, if they are not directed at the substance of expression. If the proscribed means include speech or writing, however, even a law written to focus on a forbidden effect must be scrutinized to determine whether it appears to reach privileged communication or whether it can be interpreted to avoid such ‘overbreadth.’ ”
328 Or at 299 (emphasis in original) (quoting State v. Moyle, 299 Or 691, 695-97, 702, 705 P2d 740 (1985)). The court concluded that ORS 163.732 “identifies expression as one means that may produce those forbidden effects,” e.g., repeated and unwanted contacts. Rangel, 328 Or at 299.
The court then considered whether “a narrowing construction is possible to save it from overbreadth.” Id. (footnote omitted). Reviewing its prior case law, the court concluded that, to survive an overbreadth challenge, it had required “a showing that the communicative act itself is unprotected because, for example, it is a prelude to imminent and serious proscribable harm.” Id. at 301 (citations omitted). The court also concluded that the stalking laws were susceptible to a narrowing construction that satisfied that standard:
“The gist of the crime of stalking is knowingly alarming or coercing another through repeated and unwanted ‘contacts.’ Where the state relies on one or more ‘contacts’ that constitute speech or writing, rather than physical force or other behaviors that are beyond the scope of Article I, section 8, the definition of‘coerce’ in ORS 163.730(2) expressly requires proof of a threat. We conclude that in defining ‘alarm’ in ORS 163.730(1), the legislature also contemplated, as a logical necessity, that a speech-based contact would be punishable as an element of stalking only if it constitutes a threat. If the contact in question amounts to communication by speech or writing, only a threat will be sufficient to ‘cause apprehension or fear resulting from the perception of danger,’ as ORS 163.730 requires.”
Id. at 302-03 (emphasis added); see also id. at 303 (“[A] proscribable threat is a communication that instills in the addressee a fear of imminent and serious personal violence *325from the speaker, is unequivocal, and is objectively likely to be followed by unlawful acts.”).
There is no suggestion in the present case that defendant’s communications with V contained any sort of “threat,” much less an unequivocal one that could instill “a fear of imminent and serious personal violence * * * and is objectively likely to be followed by unlawful acts.” Id. Rather, the state asserts that Rangel has no application in this case, because Rangel involved the crime of stalking under ORS 163.732, not a prosecution for violation of a stalking protective order under ORS 163.750 and that the definition of “alarm” that the court employed in Rangel does not apply here because ORS 163.750 does not use the term “alarm.” It follows, the state reasons, that a prosecution under ORS 163.750 may be based on expressive contacts that do not constitute “threats.” In addition, the state argues that, in a prosecution under ORS 163.750, it is not required to prove that the protected person’s alarm is objectively reasonable for a person in the protected person’s situation.
As explained below, we reject both of the state’s premises and hold that, in order to survive an Article I, section 8, challenge, an expressive “prohibited contact,” for purposes of ORS 163.750, must contain an unequivocal threat that instills a “fear of imminent and serious personal violence * * * and is objectively likely to be followed by unlawful acts.” Rangel, 328 Or at 303. We address the state’s arguments in reverse order.
As noted, the court in Rangel interpreted the term “alarm,” defined as “caus[ing] apprehension or fear resulting from the perception of danger,” ORS 163.730(1), to require that, if the alarm is caused by an expressive contact, the expression must contain an “unequivocal” threat, the victim’s “fear of danger” must be a fear of imminent and serious physical harm and, most importantly, that the threat “is objectively likely to be followed by unlawful acts.” Rangel, 328 Or at 303. Put another way, there must be an unequivocal threat of the sort that makes it objectively reasonable for the victim to believe that he or she is being threatened with imminent and serious physical harm.
*326That narrowing interpretation, as the state notes, applies to the term “alarm,” which is not used in ORS 163.750, but is used in ORS 163.732. The comparable provision in ORS 163.750 is that, if the prohibited contact is expressive contact, it must have “created reasonable apprehension regarding the personal safety or a person protected by the order.” As a purely textual matter, we reject the state’s contention that that phrase can be interpreted to mean that the apprehension is measured only subjectively and that the state is not required to prove that the apprehension is objectively reasonable. Had the legislature intended to punish contacts that merely cause “subjective” apprehension, it would not have qualified the term “apprehension” with the adjective “reasonable.” In short, we conclude that, for purposes of ORS 163.750(1)(c), the “apprehension” that an expressive contact causes must be objectively reasonable.
We turn to the more nuanced question — whether the “apprehension” described in ORS 163.750(1)(c), like the “alarm” described in ORS 163.730(1) and ORS 163.732(1), must be caused by a threat “that instills in the addressee a fear of imminent and serious personal violence from the speaker, is unequivocal, and is objectively likely to be followed by unlawful acts.” Rangel, 328 Or at 303. As a starting point, we note that the text of ORS 163.750(1)(c) indicates that the apprehension must be “regarding the personal safety of a person protected by” the previously issued SPO. So, from the text alone, we know that the legislature intended that the apprehension must be objectively reasonable and that it must concern the “personal safety” of a person protected by the SPO. The remaining question is whether, in order to avoid the overbreadth problems identified in Rangel, that objectively reasonable apprehension concerning personal safety must be based on an expressive contact that contains an “unequivocal threat” of “imminent” unlawful acts. Id.
In State v. Maxwell, 165 Or App 467, 998 P2d 680 (2000), rev den, 334 Or 632 (2002), which, like this case, involved a prosecution under ORS 163.750, we applied the Rangel standard in assessing expressive and nonexpressive contacts that were alleged to have violated an SPO. We described the applicable analytical framework as follows:
*327“In Rangel, the Supreme Court held that ‘a contact based on communication must consist of a threat that convincingly expresses to the addressee the intention that it will be carried out, and that the actor has ability to do so.’ [328 Or] at 306. However, there is no constitutional violation if none of the contacts on which the state relies for a conviction involves communication. Id. at 300. We must therefore determine whether defendant’s convictions were based on contacts that involved communication and, if so, whether the evidence was sufficient for the jury to find beyond a reasonable doubt that defendant communicated a threat of the type required by Rangel.”
Maxwell, 165 Or App at 475 (emphasis added). In Maxwell, we carefully parsed out which portions of the contacts at issue were “expressive” and which were not and concluded that the nonexpressive aspects of the contacts were sufficient to support the convictions. Were the state correct that Rangel had no application to SPO prosecutions under ORS 163.750, we would not have analyzed the contacts in Maxwell under the rubric of the test set forth in Rangel. Thus, we reject the state’s premise that our case law provides no guidance on whether the Rangel narrowing construction must apply to SPO stalking cases, as well as other stalking cases; we know from Maxwell that we have, in the past, applied that narrowing construction in a prosecution under ORS 163.750.
The state’s argument may be better understood to suggest that that aspect of our analysis in Maxwell was incorrect. If that is the state’s argument, we disagree. It is true that we did not explain in detail why the Rangel narrowing construction applied in the context of a prosecution under ORS 163.750. Nonetheless, as explained below, we adhere to our conclusion in Maxwell that, in prosecutions under ORS 163.750, like in prosecutions for the crime of stalking (and indeed, like in cases concerning the issuance of SPOs), expressive “contacts” must be evaluated in light of the constitutional protections provided by Article I, section 8.
It is telling that the state does not make any detailed argument as to why the protections of Article I, section 8, are somehow altered or abridged in a prosecution for violation of a court’s stalking protective order, where they are not so altered or abridged in an action for issuance of an SPO or in a *328prosecution for the crime of stalking. The state merely asserts that “[b]oth Rangel and [State v. Ciancanelli, 339 Or 282, 121 P3d 613 (2005),] recognize that, where the harmful expression is directed at a particular addressee, that individual is entitled to be free from that harm.”2 However, neither Rangel nor Ciancanelli stands for the sweeping proposition that people necessarily are entitled to be free from any harm caused by expression. To the contrary, as the court went to great lengths to observe in Rangel, its prior decisions establish that many varieties of expression that may be perceived as causing harm are, in fact, constitutionally protected. See Rangel, 328 Or at 303 (quoting Moyle, 299 Or at 705, for the proposition that a threat excludes “ ‘the kind of hyperbole, rhetorical excesses, and impotent expressions of anger or frustration that in some contexts can be privileged even if they alarm the addressee’ ”). In light of the interrelationship among the statutes setting out the various prohibitions on stalking, as described in ORS 30.866, ORS 163.732, ORS 163.738, and ORS 163.750, the Rangel analysis logically applies with equal force to prosecutions for violation of a stalking protective order.
Accordingly, we conclude that Article I, section 8, as interpreted by the court in Rangel, requires that, in order to be actionable as stalking, expressive contacts such as those described in ORS 163.730(3)(d), (e), (f), (h), and (i), must not only cause reasonable apprehension or alarm in the victim, but must involve unequivocal threats that instill a “fear of imminent and serious personal violence * * * and [are] objectively likely to be followed by unlawful acts.” Rangel, 328 Or at 303.
The state failed to prove that defendant’s two expressive communications directed toward V^s father satisfied that standard. Accordingly, the trial court erred in denying defendant’s motion for judgment of acquittal.
Reversed.
The crime of stalking occurs when:
“(a) The person knowingly alarms or coerces another person or a member of that person’s immediate family or household by engaging in repeated and unwanted contact with the other person;
“(b) It is objectively reasonable for a person in the victim’s situation to have been alarmed or coerced by the contact; and
“(c) The repeated and unwanted contact causes the victim reasonable apprehension regarding the personal safety of the victim or a member of the victim’s immediate family or household.”
ORS 163.732(1).
The court’s opinion in Ciancanelli, does not appear to have any direct bearing on the arguments at issue here. That case concerned a statute, held unconstitutional, that was directed at expression. Most of the analysis pertained to whether there existed some “historical exception” to protecting the type of expression at issue in that case. 339 Or at 320-23.