dissenting.

I believe that parts of the harassment statute implicate First Amendment freedoms and that those parts violate the Constitution. I therefore respectfully dissent.

A. Prior Texas Cases

Our decisions in Long v. State1 and May v. State,2 and the Fifth Circuit’s decision in Kramer v. Price,3 set the backdrop for the *672claim we now face.4 These decisions invalidated statutes that contained the terms “annoy” and “alarm” as implicating First Amendment freedoms and being unduly vague.5 In its opinion today, the Court’s only attempt to distinguish these cases is to say that they involved different versions of the harassment statute than the one that is currently before us.6

But the harassment provision at issue in the present case is similar in pertinent ways to the stalking provision that we invalidated in Long. The part of the harassment statute at issue provides:

A person commits an offense if, with intent to harass, annoy, alarm, abuse, torment, or embarrass another, he ... makes repeated telephone communications ... in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend another.7

The stalking provision in Long provided in relevant part:

A person commits an offense if, with intent to harass, annoy, alarm, abuse, torment, or embarrass another, he ... on more than one occasion engages in conduct ... that is reasonably likely to harass, annoy, alarm, abuse, torment, or embarrass that person.8

A comparison of these passages reveals only three features in the current statute that were not in the former stalking provision: (1) the use of “repeated” instead of “on more than one occasion,” (2) the additional emotional state of “offend” in the current statute, and (3) the limitation of the current statute to telephone communications.

B. “Repeated”

The term “repeated” introduces an ambiguity not present with the “on more than one occasion” language in the former stalking statute. Is conduct “repeated” if it occurs twice? If the two occurrences are a year apart? Does it matter whether the circumstances show that the multiple instances of conduct are connected to the same scheme or course of conduct? 9 With respect to the facial constitutionality of a statute, this ambiguity may be acceptable if First Amendment considerations are absent. In many cases, it would be clear that a person’s conduct was “repeated,” so we could adjudicate on a case-by-case basis whether a particular person was afforded adequate notice under the statute.10 But if First Amendment freedoms are at stake, we cannot allow the ambiguity in a statute to chill protected expression.11

*673The Court contends that we can read “repeated” to mean “more than one call in close enough proximity to properly be termed a single episode.”12 The Court fails to cite a single source, from a dictionary or otherwise, that limits “repeated” to actions in a single episode. Although a court has a general duty to employ “reasonable narrowing constructions” to avoid constitutional violations, it may not “assume the legislative prerogative and rewrite a statute” that is not “readily subject” to being narrowly construed.13 Would once a day for a month constitute “a single episode?” Under any common understanding it would not, but it is hard to believe the Legislature did not intend to reach such conduct. As discussed below, the Legislature has a legitimate interest in proscribing conduct over the telephone that is intentionally harassing, abusing, and tormenting. In my opinion, that interest would legitimately extend to multiple instances of conduct that could not reasonably be considered part of a single episode.

C. Offend

The term “offend” does nothing to obviate the concerns expressed in Long. We explained there that a person’s intent to inflict a more intense emotional state than “annoy” might serve to take the First Amendment out of the picture.14 But “offend” is no more intense an emotional state than “annoy” and “embarrass,” which we concluded “probably describe the least intense emotional states available under the [stalking] statute.”15

D. Telephone Communications

That leaves the remaining distinctive feature of the current harassment provision: the fact that it is limited to telephone communications. I agree with the Court that the First Amendment is not implicated when the “intent” and the “reasonably likely” effect of a person’s conduct is to “harass,” “abuse,” or “torment.” But I would hold that the First Amendment is implicated when the “intent” or “reasonably likely” effect of a person’s conduct is to “annoy,” “alarm,” “embarrass,” or “offend.”

1. Binding Precedent

In Long, we said that the terms “harass,” “abuse,” and “torment” carried a greater emotional intensity than “annoy” or “embarrass,” though they still implicated First Amendment freedoms under the stalking statute.16 The former stalking statute expansively covered “any conduct in which a person could possibly engage.” 17 But the context in which the harassing conduct occurs can be important. Government can act, consistent with the First Amendment, to prevent “substantial privacy interests” from “being invaded in an essentially intolerable manner.” 18 The State has the greatest leeway in regulating expression in favor of privacy *674interests when it seeks to “prohibit intrusion into the privacy of the home.”19 At the same time, the Supreme Court has stressed that “we are often ‘captives’ outside the sanctuary of the home and subject to objectionable speech.”20 Telephone harassment under the current statute can sometimes occur in the home, but the statute does not limit it in that manner.

2. Other Jurisdictions

In United States v. Bowker, the Sixth Circuit addressed the constitutionality of the federal telephone harassment statute, which proscribed phone calls made by a person “without disclosing his identity and with intent to annoy, abuse, threaten, or harass.”21 The court identified two factors in upholding the constitutionality of the federal statute against a First Amendment overbreadth challenge.22 First, the court explained that the federal statute focuses “not simply [on] annoying telephonic communications” but also prohibits “abusive, threatening or harassing communications.”23 Second, the court said, the statute operates in a more private realm in which the listener has to “deal with much more inconvenience to avoid” the unwanted speech, and the unidentified nature of the caller was more likely to instill fear in the listener and make it more difficult for the listener to confront the caller.24

The New Hampshire Supreme Court distinguished its state’s telephone harassment statute from that in Bowker on two grounds: the state statute prohibited “all telephone calls placed with the intent to annoy or alarm” and it did not require anonymity.25 The court found that the New Hampshire statute was not directed toward only those communications intended to instill fear in the victim, but instead swept far more broadly.26 “Simply put,” the court explained, “the First Amendment is made of sterner stuff.”27 The court concluded that the New Hampshire statute was overbroad because it covered a substantial amount of protected First Amendment speech, and there was a real likelihood that it might discourage citizens from exercising that speech.28

In Galloway v. State,29 Maryland’s court of last resort addressed the constitutionality of its harassment statute, which provided as follows:

a) Course of conduct defined. — In this section “course of conduct” means a persistent pattern of conduct, composed of a series of acts over a period of time, that evidences a continuity of purpose.

(b) Applicability. — This section does not apply to any peaceable activity intended to express political views or provide information to others.

(c) Prohibited Conduct. — A person may not follow another person in or about a public place or maliciously engage in a *675course of conduct that alarms or seriously annoys another person:

(1) With intent to harass, alarm, or annoy the other person;

(2) After reasonable warning or request to desist by or on behalf of the other person; and

(3) Without a legal purpose.30

The Galloway court initially stressed the significance of the adverb “seriously” that preceded “annoy,”31 but the court also noted a number of other restrictions that made the statute acceptable: the statute required “a reasonable warning to desist,” it did not apply to “any peaceable activity intended to express political views or provide information to others,” and it mandated that there be no “legal purpose” for the activity.32 Additionally, the court employed a narrowing construction by reading a “reasonable person” standard into the statute.33 Further, the court found that the presence of a specific intent requirement, though it might not be able by itself to save a statute from constitutional infirmity, helped to ensure that an accused had fair notice that he was violating a criminal statute.34 The court also quoted from a Connecticut case for the proposition that its harassment statute did not implicate overbreadth concerns under the First Amendment because the “prohibition is against purposeful harassment by means of a device readily susceptible to abuse as a constant trespasser upon our privacy.”35

3. Comparison with Other Statutes

The harassment provision before us lacks most of the restrictions found in the statutes that were analyzed in Bowlcer and Galloway. The portion of the statute at issue does not require that the calls be anonymous.36 Although the Texas statute proscribes “repeated” communications, it does not require a showing that these communications were pursuant to a course of conduct. No exception is made for the expression of political views or the dissemination of information, and no reasonable warning to desist is required before further conduct triggers prosecution. Furthermore, the Texas statute does not contain a “reasonable person” standard.37 And though specific intent is required by the Texas provision, that specific intent can attach to low intensity emotional states (annoy, alarm, embarrass).

4. Low Intensity versus High Intensity Emotional States

As we said in Long, “The First Amendment does not permit the outlawing of conduct merely because the speaker intends to annoy the listener and a reasonable person would in fact be annoyed.”38 *676The same could be said of the intent to alarm or embarrass. An intent to inflict one of these low intensity emotional states on the listener, even in the context of a telephone conversation, could easily be part of a legitimate communication protected by the First Amendment. A constituent might call his congressman to express a grievance, and annoying that congressman may be part of a legitimate attempt to spur him to act on that grievance. Political phone calls could be made to the homes of prospective voters with the intent to alarm those voters about a candidate’s opponent39 or about a particular government policy. A concerned neighbor or jilted lover might hope to embarrass a husband engaged in an extramarital affair by leaving a message on his answering machine for the wife to hear. Any of these behaviors might be criminal if the caller calls twice.

On the other hand, no one has the right to intentionally harass, abuse, or torment other people over a telephone. Specific intent is a limited mental state that is distinctly narrower than the mental states of knowledge and recklessness.40 The statute requires that the caller intend to harass, abuse, or torment the recipient of the call, not merely that the caller know or suspect that his call may have that effect. Moreover, the telephone is a comparatively personal and private method of communication in which messages can be difficult to screen. As the Galloway court explained, it is “a device readily susceptible to abuse” by a person who intends to be “a constant trespasser upon our privacy.” When the intent of the actor is to inflict one of the higher-intensity emotional states of harass, abuse, and torment in the relatively private, “captive-audience” telephone context, and the actor’s conduct is reasonably likely to achieve that end, the First Amendment provides no protection.

E. “Sole” Intent

Consequently, I would hold that the harassment provision at issue implicates the First Amendment with respect to the terms “annoy,” “alarm,” “embarrass,” and “offend,” but does not implicate the First Amendment with respect to the terms “harass,” “abuse,” and “torment.” The Court contends that the entire statute is outside the purview of the First Amendment because “in the usual case, people whose conduct violates § 42.07(a)(4) will not have an intent to engage in legitimate communications of ideas, opinion, or information; they will have only the intent to inflict emotional distress for its own sake.”41 But nothing in the statute limits its application to those occasions when the actor’s sole intent is to inflict emotional distress, and if the court is implying that situations are rare in which a person has more than one intent, I disagree. The mischief this statute can create is enormous, as some of the hypotheticals given above illustrate.

And it is not feasible to graft “sole intent” onto the harassment statute as a narrowing construction. As I have explained, a court has a general duty to *677employ reasonable narrowing constructions to avoid constitutional violations, but it may not assume the legislative prerogative and rewrite a statute that is not readily subject to being narrowly construed.42 The Legislature can legitimately punish a person who intentionally harasses, abuses, or torments someone over the telephone even if that is not the person’s sole intent. One can easily imagine an ex-boyfriend hounding someone over the telephone with the intent to harass, abuse, or torment, but also having a particular grievance, real or imagined, to communicate. Such actors still deserve to be punished and are legitimately covered by the statute. When intentionally harassing, abusing, and tormenting conduct is occurring over the telephone, the authorities should not have to refrain from acting because the caller has or might have another reason for the call. Moreover, at trial the State would be saddled with proving “sole intent.” A “sole intent” narrowing construction would do more to undermine the statute than simply severing (with respect to the part of the statute at issue) the problematic terms “annoy,” “alarm,” “embarrass,” and “offend.”43 And finally, there is no basis in the statute for applying a “sole intent” construction only to the emotional states of “annoy,” “alarm,” and “embarrass.” The Supreme Court has explained that a party “give[s] the game away” when he argues that an intent requirement applies in some circumstances but not others if there is no basis whatever in the text for doing so.44 The Supreme Court stated that, though it “will often strain to construe legislation so as to save it against constitutional attack, it must not and will not carry this to the point of perverting the purpose of a statute.”45

F. Spoken Words

The Court also says that the text of harassment statute “does not require that the actor use spoken words.”46 The portion of the statute at issue here proscribes “telephone communications.”47 Although someone may use a “telephone communication” for something other than words (e.g. silence or heavy breathing), the portion of the statute at issue here overwhelmingly implicates the spoken word. For example, ringing the phone repeatedly is proscribed by a part of the statute not at issue here.48 Moreover, although words are not required, that fact would not distinguish this case from Long, in which First Amendment freedoms were found to be implicated by the stalking statute.49

For the above reasons, I respectfully dissent.

delivered the opinion of the Court,

The court of appeals held that two subsections of our harassment statute, Texas Penal Code § 42.07, were unconstitutionally vague on their face. We reverse.

*665On April 24, 2006, in Bexar County, Samuel Scott was charged by information with one count of harassment, a misdemeanor offense. The information alleged, in relevant part, that:

“on or about the 5th Day of December, 2005, Samuel Scott, hereinafter referred to as defendant, with intent to harass, annoy, alarm, abuse, torment, and embarrass Yvette Scott, hereinafter referred to as complainant, did make repeated communications to the complainant, to wit: telephone calls, in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass and offend the complainant.”

On July 11, 2006, again in Bexar County, Scott was again charged by information with harassment. The information alleged, in relevant part, that:

“on or about the 12th Day of March, 2006, Samuel Scott, hereinafter referred to as defendant, with intent to harass, annoy, alarm, abuse, torment, and embarrass Yvette Scott, hereinafter referred to as complainant, did make repeated telephone communications to the complainant in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass and offend the complainant, to wit: the defendant called the complainant repeatedly by telephone while intoxicated, late at night, leaving abusive and harassing voice mail messages.”

Sometime thereafter (the record does not reveal the exact date), Scott filed a motion to quash the two charging instruments. In his motion, Scott, citing both the First and Fourteenth Amendments to the United States Constitution, argued that Texas Penal Code § 42.07, the statute under which he was charged, was unconstitutional on its face1 because it was both unduly vague and overbroad.2 He argued further that the statute’s vagueness and overbreadth threatened “to induce individuals to forgo [their] First Amendment rights [to free speech] for fear of violating an unclear law.” Scott, citing our decision in Long v. State, 931 S.W.2d 285, 288 (Tex.Crim.App. 1996), noted that “when [as here] a vagueness challenge involves First Amendment considerations, a criminal law may be held facially invalid even though it may not be unconstitutional as applied to the defendant’s conduct.” Notably, Scott did not argue that § 42.07 was vague as applied to his conduct, i.e., that the statute failed to give him a reasonable opportunity to know that the conduct for which he was charged was prohibited.3

*666On June 25, 2008, the trial court held a hearing on Scott’s motion to quash. Neither Scott nor the State presented any evidence at that hearing. Just after the start of the hearing, Scott explained to the trial court that he believed that the April 24, 2006 information had been brought under Texas Penal Code § 42.07(a)(4), whereas he believed that the July 11, 2006 information had been brought under § 42.07(a)(7).4 Toward the end of the hearing, Scott, when queried on the matter by the trial court, appeared to concede that both charging instruments had been brought under § 42.07(a)(4).

As for the substance of his motion to quash, Scott reiterated the arguments that he had made in his written motion. In addition, he argued that: (1) § 42.07(a)(4) and (a)(7) were both unconstitutionally vague and overbroad on their face because the terms “annoy” and “alarm” included in those statutory subsections were “inherently vague”; (2) neither statutory subsection clearly indicated “whose sensibilities must be offended”; (8) neither statutory subsection clearly indicated “the standard by which the [prohibited] conduct [was] to be assessed”; and (4) the specific intent elements included in the statutory subsections did not save them from vagueness.

The State, in response, did not address the merits of Scott’s arguments. Instead, the State argued simply that “[t]here is no court that has any jurisdiction over this specific [trial] court that has ruled that specific statute to be unconstitutionally vague.”

At the conclusion of the hearing, the trial court denied Scott’s motion to quash. Shortly thereafter, Scott pled nolo conten-dere, pursuant to a plea bargain, to each harassment charge. The trial court then assessed Scott’s punishment in each case *667at incarceration for two days and a $50 fine.

On direct appeal, Scott reiterated the arguments that he had made in the trial court. In addition, he argued, for the first time,5 that: (1) § 42.07(a)(4) and (a)(7) were both unconstitutionally vague as applied to his conduct (although he failed to explain how they were unconstitutionally vague as applied to his conduct); and (2) § 42.07(a)(4) and (a)(7) were both unconstitutionally vague and overbroad because the terms “abuse,” “torment,” “embarrass,” and “repeated” included in those statutory subsections were “inherently vague.”6

The State, in response to Scott’s arguments, argued that: (1) both charging instruments tracked the language of § 42.07(a)(4) and charged Scott under that statutory subsection; (2) § 42.07(a)(4) was neither unconstitutionally vague nor over-broad; (3) § 42.07(a)(4) did not implicate the free-speech guarantee of the First Amendment; and (4) “the fact that [Scott] chose [in the trial court] not to challenge the statute as applied to [his conduct] constitute[d] a failure to preserve [that] issue for appellate review.”

The court of appeals agreed with Scott’s arguments, reversed the trial court’s judgments, and rendered judgments of acquittal. Scott v. State, 298 S.W.3d 264 (Tex.App.-San Antonio 2009). More specifically, the court of appeals held that: (1) the July 11, 2006 information, which involved allegations of harassing voice-mail messages, could have been brought under § 42.07(a)(7) because voice mail messages fell within the statutory definition of “electronic communication”; (2) § 42.07(a)(4) and (a)(7) both implicated the free speech guarantee of the First Amendment;7 and (3) the terms “annoy,” “alarm,” “abuse,” “torment,” “embarrass,” and “repeated” contained in the two statutory subsections were all unconstitutionally vague.8

We granted the State’s petition for discretionary review, which raised six grounds,9 in order to determine whether *668the court of appeals erred in its analysis. See Tex.R.App. Proc. 66.3(d).

We turn first to the State’s fourth ground for review, in which the State argues that the court of appeals erred in addressing the constitutionality of § 42.07(a)(7). The State argues that, contrary to the court of appeals’s conclusion, the July 11, 2006 information against Scott charged him under § 42.07(a)(4), not (a)(7).

A careful reading of the July 11, 2006 information, which we quoted earlier, reveals that it tracks the language of § 42.07(a)(4), which we also quoted earlier, but not the language of (a)(7). Furthermore, the text of (a)(4) suggests that it covers ordinary voice (and therefore voice mail) communication involving an ordinary telephone, whereas the text of (a)(7) suggests that it covers various types of non-telephonic, “electronic” communication, e.g., e-mail, instant message, etc. Therefore, we conclude that the July 11, 2006 information against Scott was brought under § 42.07(a)(4) and that the court of appeals erred in addressing the constitutionality of § 42.07(a)(7). We sustain the State’s fourth ground for review.

We turn next to the State’s second ground for review, in which the State argues that the court of appeals erred in concluding that § 42.07(a)(4) implicated the free-speech guarantee of the First Amendment. The question of whether the statutory subsection implicates the free-speech guarantee is important in the cases before us, because if the statutory subsection does implicate the free-speech guarantee, then Scott, in making his vagueness challenge, is relieved of the usual requirement of showing that the statutory subsection was unduly vague as applied to his conduct. See footnote three, supra.

To answer the question of whether § 42.07(a)(4) implicates the free-speech guarantee of the First Amendment, we must first determine the protection afforded by the free-speech guarantee, and then we must determine the meaning of § 42.07(a)(4). The First Amendment provides, in relevant part, that “Congress shall make no law ... abridging the freedom of speech.” This guarantee of free speech, which was made applicable to the various states by the Due Process Clause of the Fourteenth Amendment, Gitlow v. New York, 268 U.S. 652, 666, 45 S.Ct. 625, 69 L.Ed. 1138 (1925), generally protects the free communication and receipt of ideas, opinions, and information, Red Lion Broadcasting Co. v. F.C.C., 395 U.S. 367, 390, 89 S.Ct. 1794, 23 L.Ed.2d 371 (1969); Chaplinsky v. New Hampshire, 315 U.S. 568, 571-72, 62 S.Ct. 766, 86 L.Ed. 1031 (1942). In a nation of ordered liberty, however, the guarantee of free speech cannot be absolute. The State may lawfully proscribe communicative conduct (i.e., the communication of ideas, opinions, and information) that invades the substantial pri*669vacy interests of another in an essentially intolerable manner. Cohen v. California, 403 U.S. 15, 21, 91 S.Ct. 1780, 29 L.Ed.2d 284 (1971).10

Turning to the meaning of § 42.07(a)(4), we find that, given the allegations in the charging instruments, the relevant portion of (a)(4) reads as follows: “A person commits an offense if, with intent to harass, annoy, alarm, abuse, torment, or embarrass another, he ... makes repeated telephone communications ... in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend another.” We notice several things about that text.11 First, the text requires that the actor have the specific intent to harass, annoy, alarm, abuse, torment, or embarrass the recipient of the telephone call. That is, the text requires that the actor have the intent to inflict harm on the victim in the form of one of the listed types of emotional distress. Second, the text requires that the actor make repeated telephone calls to the victim; one telephone call will not suffice.12 Third, the text requires that the actor make those telephone calls in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend an average person.13 Fourth, the text does not require that the actor use spoken words.

Having examined the text of § 42.07(a)(4), we conclude that it is not susceptible of application to communicative conduct that is protected by the First Amendment. In other words, the statutory subsection does not implicate the free-speech guarantee of the First Amendment. The statutory subsection, by its plain text, is directed only at persons who, with the specific intent to inflict emotional distress, repeatedly use the telephone to invade another person’s personal privacy and do so *670in a manner reasonably likely to inflict emotional distress. Given that plain text, we believe that the conduct to which the statutory subsection is susceptible of application will be, in the usual case, essentially noncommunicative, even if the conduct includes spoken words.14 That is to say, in the usual case, persons whose conduct violates § 42.07(a)(4) will not have an intent to engage in the legitimate communication of ideas, opinions, or information; they will have only the intent to inflict emotional distress for its own sake. To the extent that the statutory subsection is susceptible of application to communicative conduct, it is susceptible of such application only when that communicative conduct is not protected by the First Amendment because, under the circumstances presented, that communicative conduct invades the substantial privacy interests of another (the victim) in an essentially intolerable manner.15 We sustain the State’s second ground for review.

Given our disposition of the State’s second and fourth grounds for review, we need not reach the State’s remaining grounds for review. We dismiss them.

Because § 42.07(a)(4) does not implicate the free-speech guarantee of the First Amendment, Scott, in making his *671vagueness challenge16 to that statutory subsection, was required to show that it was unduly vague as applied to his own conduct. He has not done that. Therefore, his vagueness challenge fails.

We reverse the judgments of the court of appeals and affirm the judgments of the trial court.

JOHNSON, J., filed a concurring opinion, in which COCHRAN, J., joined.

concurring opinion in which COCHRAN, J., joined.

Harassment, as defined in this statute, requires an intent to harass, annoy, alarm, abuse, torment, embarrass, or offend another in a manner reasonably likely to attain its goal; it does not require that the intended result actually occur. Because the intent is that of the speaker, that person must intend to do something that he or she believes would harass, annoy, alarm, abuse, torment, embarrass, or offend another and to do so repeatedly. While communication of some political view or expression of a personal opinion is protected speech, it is completed with one contact. If expressed to the same hearer repeatedly, such a “communication” may become harassment, as the pertinent information has already been communicated and no longer has a legitimate purpose.

I find it odd to divide the various actions into two groups based on the dubious distinction of being more or less “intense” emotionally. Why are “annoy,” “alarm,” “embarrass,” and “offend” less invasive of privacy than “harass,” “abuse,” and “torment”? Given the context, “alarm” is directed at the hearer’s alarm over the possibility of personal harm, a reasonable reaction, not alarm that the neighbor is having an affair or that a candidate for mayor has purportedly converted to a religion that the hearer does not approve of. Some hearers might experience extremely emotional embarrassment from being publicly exposed as a liar. Other hearers might not be even unsettled by a threat of bodily injury. One has only to keep up with current events to view the emotional extremes to which some persons will go when they are “offended.”

Harassment is in the mind of the speaker, not the hearer. The speaker who intends to harass, annoy, alarm, abuse, torment, embarrass, or offend another has himself defined, for that purpose, both the applicable term and the word “repeatedly.” They are not vague or over-broad for the speaker; they are clearly and precisely known. There is no ambiguity of intent in the mind of the speaker, and intent.under-girds the offense.

I join the opinion of the Court.