KEVIN OWENS v. THE STATE OF TEXAS (2025)

CONCURRING AND DISSENTING OPINION

Appellant says that an acquittal should be granted while the State says that the convictions should be affirmed.¹ Although they come to opposite conclusions, both Appellant and the State view the First Amendment issue as an all or nothing inquiry. I see a middle ground between these extremes because I think the focus of our inquiry should be different. The question we should ask is: What would a reasonable jury think it could do? I agree that count one in the information must be dismissed because a reasonable jury could convict of that count only on a basis that would violate the First Amendment. But count two in the information presents a more complex question because that count permits a reasonable jury to convict on a constitutionally permissible basis and also permits such a jury to convict on a basis that would violate the First Amendment. As I will explain, the possibility of a conviction on an unconstitutional basis could have been eliminated by striking the statutory terms “annoy, alarm, embarrass, and offend” from the jury charge. Because count two permits conviction on a constitutional basis, dismissal of that count is inappropriate. But because count two also permits conviction on an impermissible basis, and that problem could have been eliminated by striking some of the statutory methods of committing the offense from the jury instructions, the as-applied challenge to count two boils down to jury charge error that, if harmful, would require a new trial. Consequently, I would remand count two to the court of appeals to conduct a harm analysis.

I. BACKGROUND

The Court has detailed much of the factual background of this case, but I wish to add a few things. There were two electronic harassment counts in the information. The first count covered email messages from on or about May, 13, 2018 to June 14, 2018. The second count covered email and text messages from on or about July 1, 2018 to July 17, 2018. Both counts alleged all the statutory methods of committing harassment, and the jury charge alleged all of these methods in the disjunctive: “with the intent to harass, annoy, alarm, abuse, torment, or embarrass another,” the defendant sent the messages “in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend another.”²

The evidence at trial showed only three messages (all emails) sent before July 1, 2018. The first message was sent on May 13 and the second two messages were sent on June 14. The remaining emails or text messages were sent on or after July 1.

At trial, the State introduced the victim's July 9 email response from the “Office Manager” account, which warned that Appellant would receive no responses to further emails and that his emails would be forwarded to the police:

I manage and filter Dr. Bira's emails and monitor and respond to risk issues. Your recent email flagged as risk. Please keep in mind that as you are a former patient of Dr. Bira's, she does not have responsibility for your safety nor does she receive your emails directly. If you are considering suicide: call 911, take yourself to the nearest ER, or call the suicide hotline at 1-800-273-8255. After you take steps to ensure your immediate safety, I advise you to find another provider to work with (www.psychologytoday.com - therapist finder). Per protocol for the type of email you sent, your contact has been reported to SAPD to ensure your safety. As you have been provided with safety measures and notification has been given, you will not receive any additional emails from Dr. Bira's office. Keep in mind that any contact from you will continue to be filtered and forwarded to police and legal personnel in order to ensure safety and maintain appropriate use of this email line. Please do take care of yourself, per the suggestions above, and we wish you the best.3

The victim also testified that her attorney sent Appellant two “cease and desist” letters, at least one of which was sent before July 4, but she did not specify exactly when those letters were sent.

On appeal, Appellant challenged the electronic harassment statute on its face and as applied and also raised a jury charge complaint for failing to include a First Amendment defense. In his brief on the “as applied” claim, Appellant argued that the claim was preserved in several ways, including defense counsel asking “that the factual question of whether Mr. Owens's speech fell into an unprotected category be presented to the jury.” In his petition for discretionary review, Appellant asked how an “as applied” complaint should be handled and laid out three subparts: (1) whether the statute was used to punish the content of his speech, (2) whether speech can be integral to criminal conduct absent a nonspeech crime, and (3) whether a defensive instruction on protected speech was required. The Court refused Appellant's petition but, on its own motion, granted review of the following issue: “Was Texas Penal Code section 42.07(a)(7) unconstitutional as applied to Appellant?”⁴

In addition to arguing that he engaged in protected speech, Appellant's brief on discretionary review continues to argue that the applicability of the First Amendment should have been submitted to the jury. And in his reply brief, Appellant argues that, while it is possible that the jury convicted him without regard to the content of the speech, such a result is highly improbable under the jury instructions:

The present case is not about essentially noncommunicative conduct, but about essentially communicative conduct—the sending of emails and messages containing images or words. Still, it is imaginable that Mr. Owens's conviction was unconnected to his expression—that the jury took the content of his communications out of the equation when deciding whether the manner of his communications was reasonably likely to cause emotional harm. But the trial court did not interpret “manner” to exclude “content,” and it is highly unlikely that the jury did. The State argued to the jury that it was the content of the communications that made them harmful. Please see below at 24. So while it is imaginable that the jury decided the “reasonably likely to ․” issue without reference to the content of the communications, it is highly improbable.

II. ANALYSIS

A. We are not bound by the parties’ characterization of the “as applied” issue and can resolve the case on a middle ground between their positions.

This Court is the “caretaker of Texas law.”⁵ In deciding the proper legal standard or legal outcome in a case, we are not bound by any concessions made by the parties ⁶ and are not constrained to choose between extreme positions the parties may have taken.⁷ We cannot allow the parties’ positions in a particular case to warp our legal holdings because “our judgments are precedents.”⁸ Sometimes a “middle ground” between extreme legal positions is the proper one.⁹

B. An “as applied” holding should not be overly expansive and can legitimately focus on the jury charge.

In an “as applied challenge,” a party “asserts that the statute is unconstitutional as applied to his particular facts and circumstances.”¹⁰ Such an assertion requires a recourse to evidence and cannot be based solely on the charging instrument.¹¹ And in a jury trial, it is the jury that applies the law to facts.¹² In Texas, trial courts help juries accomplish this task by crafting application paragraphs in the jury charge.¹³ It should be no surprise that an “as applied” violation of the First Amendment can be a result of the jury instructions.¹⁴

We must keep in mind that we have already held the electronic harassment statute to be facially constitutional.¹⁵ Any “as applied” holding this Court makes concerning the statute ought not to be so expansive as to effectively negate our facial-constitutional holding. We should also keep in mind that there is a presumption that statutes are constitutional, even as applied.¹⁶ That presumption is reversed “when the government seeks to restrict and punish speech based on its content,”¹⁷ but that does mean the defendant must at least show that he is being punished on the basis of the content of his speech before the burden shifts to the State to justify such punishment. If the defendant makes a showing that he is being punished based on content, then the State has the burden to show that the statute, as it is being applied, is necessary to serve a compelling state interest and is narrowly drawn to do so.¹⁸

In a given case, a defendant could show that a statute's application to his conduct is necessarily unconstitutional—i.e. that no reasonable jury could convict on a constitutional basis.¹⁹ In such a case, dismissal of the prosecution would be appropriate. But in a given case, a defendant may show only that a reasonable jury could have convicted on an unconstitutional basis, meaning that it could also have convicted him on a constitutional basis. Under those circumstances, dismissal of the prosecution for an “as applied” violation is not an appropriate remedy because the State has the right to punish a defendant for criminal conduct when doing so would not violate constitutional prohibitions. Nevertheless, it is still impermissible for the jury charge to apply the law in such a way as to permit a First Amendment violation.²⁰ If that occurs, and if the error is harmful, then the appropriate remedy is a new trial.²¹

C. How a jury interprets the electronic-harassment offense depends in part on the content of the speech at issue in the case.

The content of the electronic messages at issue in a particular case will influence whether the jury perceives the electronic harassment statute to be directed at the content of speech. Consider the following four situations:

(1) the sending of emails that are blank or contain only gibberish,

(2) the sending of emails that, individually, are not offensive,

(3) the sending of emails that, individually, are offensive,

(4) the sending of political emails.

A jury considering the sending of blank or gibberish emails would understand that speech had nothing to do with whether the messages were sent in a manner likely to harass, annoy, alarm, abuse, torment, embarrass, or offend the recipient. The fact that the content of the emails was blank or gibberish might actually have been annoying, but it wouldn't have been speech, so there would be no First Amendment problem. This scenario was essentially why one of this Court's prior opinions held that the electronic harassment statute was facially constitutional—because it did not necessarily implicate speech at all.²²

In the second category, a jury would understand that the content of the individually inoffensive messages is not why the defendant is being prosecuted. The jury in this situation would perceive the statute to be content neutral and would focus solely on the frequency of the messages to determine whether the messages were sent in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend the recipient.

In the third category, however, a jury could reasonably conclude that the offensive content of the messages could be a basis for finding that they were sent in a manner reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend the recipient. The more messages there were, the more likely the jury would also view frequency as also being part of the inquiry. But, at least for the terms “annoy, alarm, embarrass, and offend,” nothing requires the jury to place more importance on repetition than content. “Repeated” in the statute merely means that more than one message was sent.²³ There is no requirement of “periodic frequency” or “temporal relationship.”²⁴

And for the fourth category, political messages, a jury could likewise convict someone for what would constitute protected political speech. And even if the frequency of the messages were part of the inquiry, in a political context, that frequency itself might be part of the point of the communication.²⁵

How a reasonable jury would perceive the electronic harassment statute is kind of like quantum mechanics, where observing a particle can influence its behavior.²⁶ The content of the messages has a profound impact on what a reasonable jury would think the statutory allegations authorize it to do.²⁷

D. Under the terms “annoy,” “alarm,” “embarrass,” and “offend,” the jury charge created an “as applied” violation of the First Amendment by permitting conviction solely on the basis of protected speech.

The present case involved the third scenario, offensive messages.²⁸ There were over 30 of them, though, as I discuss later, only three applied to the first count.²⁹ At least as to the 30 messages applicable to the second count, each juror could have picked any two to support that count—because two was all that was required ³⁰ and the jurors did not have to be unanimous.³¹ Even if the jurors had to believe that the emails and text messages all came from Appellant, they did not necessarily have to believe that all were sent with the requisite intent to harass, annoy, alarm, abuse, torment, or embarrass, nor did they have to believe that any particular combination of messages was reasonably likely to harass, annoy, alarm, abuse, torment, embarrass, or offend.³²

Importantly, the jury charge alleged the various statutory methods of committing the offense in the disjunctive, which is generally allowed,³³ but, for this statute and this case, creates First Amendment problems. Speech can be intentionally annoying without losing First Amendment protection.³⁴ That can be true even if the speech is “persistently annoying.”³⁵ And the same is true of speech that is alarming,³⁶ embarrassing,³⁷ or offending.³⁸ If all that is required is two or more messages that, by their content, annoy, alarm, embarrass, or offend another, with the intent to do so, then the First Amendment is in serious jeopardy. Two annoying, alarming, embarrassing, or offending messages would be enough to support a prosecution under the statutory language, so long as accompanied by an intent to annoy, alarm, or embarrass.³⁹

The First Amendment was designed “to protect the freedom to think as you will and to speak as you think.”⁴⁰ Punishing the occasional use of annoying, alarming, or embarrassing language runs contrary to that design. Such punishment could make people self-conscious about whether to speak their mind and could turn what ought to be a vibrant environment for the exchange of ideas into one where people tread lightly when they talk, as if walking on eggshells. The victim's business email and business phone number were open to the public. Although these avenues of communication had obvious business purposes, it would not be feasible to require that all communications through these avenues be cordial. Caustic communications could arise within even the business context. For example, a client could have a fee dispute with her. Or, a former client could give her negative feedback about therapy. Appellant's speech implicated both of these issues. Dissatisfaction with a service a business provides is an all too common event in our society. Subjecting occasional expressions of dissatisfaction to criminal prosecution could chill large amounts of protected speech.

E. But, under the terms “harass,” “torment,” and “abuse,” the jury charge supplied constitutional ways to convict.

The words “harass, abuse, and torment” do not completely eliminate First Amendment concerns, but they come much closer to doing so. The government can shut off speech solely to protect someone else from hearing it “upon a showing that substantial privacy interests are being invaded in an essentially intolerable manner.”⁴¹ Such interests are not invaded by someone merely engaging in multiple instances of communication that intentionally annoy, alarm, embarrass, or offend another.

But harassing, abusing, and tormenting is different.⁴² Those terms embody higher intensity emotional states.⁴³ And those terms all include the idea of a connected, substantial repetition.⁴⁴ The verb “harass” involves conduct that is “persistent,” “continual,” “pestering,” “frequent,” or “chronic.”⁴⁵ The verb “torment” involves conduct that causes “extreme anguish,” “agony,” or severe distress of body or mind, and such conduct is usually “persistent” or “recurrent.”⁴⁶ The verb “abuse” involves conduct that “takes unfair or undue advantage of” someone or treats someone in a way so as to “injure, hurt, or damage” that person.⁴⁷ “Harass” can be seen as a severe form of “annoy” while “torment” and “abuse” can be seen as severe forms of “alarm.” A few negative emails might cause someone to be annoyed or alarmed (or embarrassed), but pervasive activity is what gives rise to being harassed, tormented, or abused.

To some degree, the repetition inherent in the terms “harass,” “torment,” and “abuse,” make content less important to the offense. To the extent content still matters, these terms ensure that it is considered only when needed to show an intolerable violation of privacy. It is one thing to say that people should not be required to “walk on eggshells” when they speak and quite another to say that a person can systematically target another in a way that significantly disrupts the target's life or causes the target to feel like she has to constantly look over her shoulder. There is a point at which unwanted communication is pervasive enough to be an intolerable violation of privacy, and that point is captured by the words “harass,” “torment,” and “abuse.” Commonly encompassed by those terms are statements like, “He won't leave me alone,” and “I am afraid he is going to hurt me.” And although the victim's business email and business phone number were available to the public, they both still embodied personal forms of communication, as opposed to, say, an online message board.

And to whatever extent conduct that “annoys, alarms, embarrasses, or offends” might be thought to rise to the level of conduct not protected by the First Amendment—based on the frequency and severity of the messages—such conduct would necessarily rise to the level of “harassing, tormenting, or offending.” Thus, we need not try to ascertain whether the State could narrow its prosecution to a particular frequency or combination of messages to avoid violating the First Amendment under the lesser terms of “annoy, alarm, embarrass, or offend.” The higher-intensity terms of “harass, torment, and abuse” already subsume any such scenarios, and so, we need focus only on the ability of these higher-intensity terms to ensure that no First Amendment violation occurs.

F. The ability to block or ignore messages was not a panacea.

The Court says that the victim was not powerless to avoid the messages because she could have blocked them or deleted them without reading them. The victim testified that she did not block Appellant because the police told her not to, suggesting she knew how to block him.⁴⁸

Assuming that the victim could have easily blocked Appellant's emails, there were some reasons not to do so. The victim was a therapist. Her profession meant she would encounter people with various mental problems, some of whom could pose a danger to themselves or others. She might at least feel ethically obligated not to block a former client in case he needed to be referred elsewhere for help with suicide counseling. And in fact, Appellant sent her a message that could be interpreted as a suicide threat. It is also possible for her to have felt ethically obligated to refer a potentially dangerous client for treatment for anger or mental illness issues.

Moreover, even if the victim felt no obligation toward a former client or had not been a therapist, she might have seen the email as a sign that Appellant possibly posed a physical danger to her. She might have felt that blocking him would be unwise because later emails could warn her if he later became an imminent threat. Forewarned is forearmed. The victim testified that she was “glad” she had not blocked him and that she no longer lived in Texas because she was scared for her safety. That testimony suggests that, at least in retrospect, the victim thought it wise not to block Appellant so that she would be warned that he posed a physical threat to her.

G. Appellant's jury-charge remedy is not feasible.

Appellant proposes that a defensive instruction be given to the jury telling it to “determine whether the State has proven beyond a reasonable doubt that the speech falls into some category of unprotected speech, and if the State has not, the jury must acquit.” Appellant's proposed solution is not feasible. Such a generic defensive instruction would create a constitutional vagueness problem by requiring jurors (as well as law enforcement and the public at large) “to be First Amendment scholars” in construing the electronic harassment statute.⁴⁹

H. Count one must be dismissed; count two should be remanded for a harm analysis. If harm were found in connection with count 2, the proper remedy would be a new trial.

The information said that count one took place between May 13 and June 14 and that count two took place between July 1 and July 17. The phrase “on or about” was used to introduce these time periods. Legally, “on or about” means any date “anterior to the presentment of the [charging instrument] and within the statutory limitation period.”⁵⁰ If this legal meaning were applied to both counts, it could create an overlap between them, which would pose a double jeopardy problem. But the jury charge did not define what “on or about” means. Without such a definition, a reasonable jury looking at the description of the two counts in the jury charge (which tracked the information) would necessarily construe count one to embrace only messages sent before July 1. Such a jury would conclude that the remaining messages were embraced by count two.

That means that a reasonable jury would have only three messages supporting count one, with one of those messages being sent on May 13, and the remaining two messages being sent over a month later. There is simply no way that a reasonable jury would think that these three messages were annoying, alarming, harassing, tormenting, abusing, embarrassing, or offending based on the number of messages alone, and certainly not based on that number spread over a month-long time period. The only way the messages could meet the statutory annoying, etc., elements would be based on the content of those messages. Moreover, although the three messages could, by their content, reasonably be said to be annoying, alarming, or offending, no rational jury—considering both the content and the number of messages—could find those messages together to be harassing, tormenting, or abusing. Consequently, count one violates the First Amendment.

But count two was a mixed bag, permitting conviction on multiple bases, some of which violated the First Amendment and some of which did not. Count two encompassed 30 messages, many of which occurred after Appellant was asked to cease and desist and after he was warned that the messages were being sent to law enforcement and that the victim would not be responding to them. This combination of messages, many sent after Appellant was warned, could easily be seen as rising to the level of harassing, tormenting, and abusing the victim. But it was also possible that the jury could view the messages differently and base its decision to convict on a combination of messages that would be insufficient to rise to that level.

Because many of the messages relating to count two were offensive by their content, the jury charge should have limited the jury to the “harass,” “torment,” and “abuse” methods of committing the offense, to ensure that the jury did not convict Appellant on a basis that violated the First Amendment. Consequently, count two should be remanded to the court of appeals to conduct a harm analysis.⁵¹ In doing so, the court of appeals could decide whether error was preserved and apply the appropriate harm standard.⁵² If the error were found harmful, then the remedy would be a reversal and a remand for a new trial, with the jury charge on retrial being limited as I have proposed.⁵³

I concur in the Court's judgment as to count one and respectfully dissent from the Court's judgment as to count two.

FOOTNOTES

1.   Appellant alternatively suggests a new trial on the basis of jury charge error. I will comment on that claim later.

2.   It is proper to allege different statutory methods of committing the same offense in the conjunctive in the charging instrument and in the disjunctive in the jury charge. Floyd v. State, 2024 WL 4757855, * 5 (Tex. Crim. App. Nov. 13, 2024) (quoting Jourdan v. State, 428 S.W.3d 86, 94 (Tex. Crim. App. 2014)) (“[D]ifferent modes of commission may be presented in a jury instruction in the disjunctive when the charging instrument, in a single count, alleged the different means in the conjunctive.”).

3.   The record shows that Appellant responded to this email, saying that he was not contemplating suicide.

4.   Owens v. State, No. PD-0075-24 (Tex. Crim. App. June 5, 2024) (order) (not designated for publication).

5.   Hunter v. State, 954 S.W.2d 767, 769 (Tex. Crim. App. 1995).

6.   Oliva v. State, 548 S.W.3d 518, 520 (Tex. Crim. App. 2018) (“We, of course, are not bound by any agreement or concessions by the parties on an issue of law.”); Long v. State, 931 S.W.2d 285, 289 (Tex. Crim. App. 1996) (“Appellant concedes that (a)(7)(A) contains a ‘reasonable person’ standard, which was absent from the earlier statute. But we are not bound by such a concession. Whether the stalking provision contains a ‘reasonable person’ standard is a question of law, and we are under no obligation to accept a concession on an issue of law even if all parties agree.”); see also Colorado Republican Fed. Campaign Comm. v. Fed. Election Comm'n., 518 U.S. 604, 622 (1996) (“[W]e are not bound to decide a matter of constitutional law based on a concession by the particular party before the Court as to the proper legal characterization of the facts.”).

7.   Williams v. State, 585 S.W.3d 478, 481 (Tex. Crim. App. 2019) (“It is entirely appropriate for us to decide this issue, even if we ultimately construe ‘substantial compliance’ to mean something other than what the parties or the court of appeals understood it to mean.”); see also Taylor v. Illinois, 484 U.S. 400, 410 (1988) (“Petitioner's claim that the Sixth Amendment creates an absolute bar to the preclusion of the testimony of a surprise witness is just as extreme and just as unacceptable as the State's position that the Amendment is simply irrelevant.”); Houser v. Folino, 927 F.3d 693, 698 (3d Cir. 2019) (“The parties stake out the extreme positions on this question․ Our precedents, however, do not support either extreme.”) (ellipsis inserted); Brotherhood of Locomotive Engineers v. Union Pacific R.R., 879 F.3d 754, 760 (7th Cir. 2017) (“The parties have advanced extreme positions in this case․ Neither extreme persuades us.”); McDonnell Douglas Corp. v. United States, 323 F.3d 1006, 1015 (Fed. Cir. 2003) (“Rather, the proper interpretation of the default provision lies somewhere between the parties’ extreme positions.”).

8.   See Young v. United States, 315 U.S. 257, 259 (1942) (“Furthermore, our judgments are precedents, and the proper administration of the criminal law cannot be left merely to the stipulation of parties.”); Estrada v. State, 313 S.W.3d 274, 286 (Tex. Crim. App. 2010) (“This Court must still independently examine the error confessed because ‘our judgments are precedents, and the proper administration of the criminal law cannot be left merely to the stipulation of parties.’ ”).

9.   See, e.g., Vandyke v. State, 538 S.W.3d 561, 571 (Tex. Crim. App. 2017) (“[U]ndue influence test” for determining a Separation of Powers violation “takes the middle ground between those who would seek rigid compartmentalization and those who would find no separation of powers violation until one branch completely disrupted another branch's ability to function.”) (internal quotation marks omitted). In addition to arguing that his case should be dismissed, Appellant does in fact argue a middle ground. As I explain later, Appellant's particular middle-ground solution is not feasible.

10.   Irsan v. State, 708 S.W.3d 584, 631 (Tex. Crim. App. 2025) (quoting from State ex rel. Lykos v. Fine, 330 S.W.3d 904, 910 (Tex. Crim. App. 2011)).

11.   Id. (citing Lykos, supra).

12.   Sparf v. United States, 156 U.S. 51, 102 (1895) (“We must hold firmly to the doctrine that in the courts of the United States it is the duty of juries in criminal cases to take the law from the court and apply that law to the facts as they find them to be from the evidence.”); United States v. Manzano, 945 F.3d 616, 627 (2d Cir. 2019) (citing Sparf); State v. Sayles, 472 Md. 207, 234 (2021) (same).

13.   Yzaguirre v. State, 394 S.W.3d 526, 530 (Tex. Crim. App. 2013) (“The application paragraph is what explains to the jury, in concrete terms, how to apply the law to the facts of the case.”); see also Campbell v. State, 664 S.W.3d 240, 247-48 (Tex. Crim. App. 2021) (citing cases that refer to “the facts, as applied to the law in the application paragraph”).

14.   See United States v. Lee, 6 F.3d 1297, 1302 (8th Cir. 1993) (Gibson, J., concurring, with majority of the Court remanding for retrial in accordance with concurring opinion's instructions) (“As applied under the jury instructions of this case, section 241 targeted conduct which, though expressive of a discriminatory idea or philosophy, is nevertheless protected expressive conduct. We conclude that section 241, as applied in the prosecution against Lee, violated the First Amendment, and Lee's conviction on Count I must be reversed. We also conclude that the indictment need not be dismissed, but that there must be a new trial with the following principles in mind.”); Hasty v. United States, 669 A.2d 127, 131-32 (D.C. App. 1995) (narrowing construction needed to make statute constitutional must be included in jury charge); Bismarck v. Schoppert, 469 N.W.2d 808, 810 (N.D. 1991) (“On appeal, Schoppert advances several arguments, two of which, taken together, are dispositive. Schoppert first challenges the jury instructions defining disorderly conduct and then argues there was insufficient evidence to support the jury's verdict. The sum of his argument is that the ordinance, as applied to him, violates the first amendment․ The inclusion of the underlined phrase ‘inflicts injury,’ Schoppert says, under the circumstances of this case, allowed the jury to convict him for conduct that is constitutionally protected. We agree.”); cf. Sanchez v. State, 209 S.W.3d 117, 119 (Tex. Crim. App. 2006) (“In the course of rejecting the appellant's contention that the statute is unconstitutional, we resolved an ambiguity in the scope of the requirement that the sexual conduct be ‘unwelcome.’ ․ The jury charge abstractly defined sexual harassment in the same ambiguous terms that the statute utilizes. Thus, it did not clearly inform the jury that it must find that, not only the appellant's ‘sexual advances,’ but also his ‘requests for sexual favors, or other verbal or physical conduct of a sexual nature,’ must be ‘unwelcome’ in order to support a guilty verdict.”).

15.   Ex parte Sanders, 663 S.W.3d 197 (Tex. Crim. App. 2022); Ex parte Barton, 662 S.W.3d 876 (Tex. Crim. App. 2022).

16.   Faust v. State, 491 S.W.3d 733, 743-44 (Tex. Crim. App. 2015).

17.   Ex parte Lo, 424 S.W.3d 10, 15 (Tex. Crim. App. 2013).

18.   Texas v. Johnson, 491 U.S. 397, 412 (1989) (subjecting a statute that was content-based as applied to “the most exacting scrutiny”) (quoting Boos v. Barry, 485 U.S. 312, 321 (1988)); Boos, supra (defining “the most exacting scrutiny” as showing “that the “regulation is necessary to serve a compelling state interest and that it is narrowly drawn to achieve that end”).

19.   See, e.g., Johnson, supra at 419.

20.   See supra at n.14.

21.   See id.

22.   See Sanders, 663 S.W.3d at 215-16 (“The bare fact that data of any nature is sent electronically does not mean that anything has been expressed. The statute is equally violated by the repeated sending of communications containing expressive speech as it is by the repeated sending of communications containing no speech at all. A person intending to harass another could violate the statute by sending several e-mails containing only the letter ‘B’ (arguably a ‘writing’) or e-mails containing nothing (some minimal level of ‘data’). Or the person could violate the statute by sending computer code (‘signals’ or ‘data’) that would be a readable sequence of machine language understood by a computer but entirely indecipherable and meaningless to humans. And there is no requirement that the data be actually usable. Entirely meaningless data understandable by neither man nor machine could just as well be sent, repeatedly, in a manner reasonably likely to harass, etc., with the specific intent to harass, etc.”).

23.   Wilson v. State, 448 S.W.3d 418, 424 (Tex. Crim. App. 2014) (“[W]e need not go any further than we did in Scott, that ‘one telephone call will not suffice’ and a conviction secured by evidence of a single communication will not stand.”) (discussing meaning of “repeated” in similarly worded telephone harassment statute).

24.   Id. (“The communications’ periodic frequency or the temporal relationship of each communication are characteristics that may further describe the communications’ nature, but we do not find those characteristics necessary to the definition of repeated.”)

25.   Political speech involves special concerns that might make the entire electronic-harassment statute unconstitutional as applied to a particular case. See United States v. Sryniawski, 48 F.4th 583, 587-88 (8th Cir. 2022). Since political speech is not at issue here, I need not try to explore the contours of the electronic-harassment statute's interaction with it.

26.   See Ethan Siegel, Observing The Universe Really Does Change The Outcome, And This Experiment Shows How, Forbes (via Forbes.com) (May 26, 2020) (“The most puzzling fact about quantum mechanics is that the answer you get depends on how you look at the individual quanta that are part of the experiment. If you make certain classes of measurements and observations, they behave like particles; if you make other choices, they behave like waves. Whether and how you observe your own experiment really does change the outcome, and the double-slit experiment is the perfect way to show how.”).

27.   The Court misunderstands my argument when it says that I argue that part of the statute is facially unconstitutional. To the contrary, I acknowledge that the statute can be constitutionally applied—under all of the methods of committing electronic harassment—to the first two of the four general situations I have outlined. Only situations three and four pose problems for some of the methods in the statute.

28.   There is no question that all of Appellant's emails and text messages were “speech” as contemplated by the First Amendment. See 303 Creative v. Elenis, 600 U.S. 570, 587 (2023) (“All manner of speech—from ‘pictures, films, paintings, drawings, and engravings,’ to ‘oral utterance and the printed word’—qualify for the First Amendment's protections.”).

29.   There was also a Facebook message, but it was not included in either count of the information, which focused only on emails and text messages.

30.   See supra at nn.23-24.

31.   See Johnson v. State, 364 S.W.3d 292, 296 (Tex. Crim. App. 2012) (“The Supreme Court has explained that a ‘jury need not always decide unanimously which of several possible sets of underlying brute facts make up a particular element.’ ”).

32.   See Garcia v. State, 667 S.W.3d 756, 762 (Tex. Crim. App. 2023) (“The jury acts as the sole judge of the credibility of the witnesses and may choose to believe all, some, or none of the testimony presented․ If the record supports conflicting inferences, the reviewing court must presume that the factfinder resolved the conflicts in favor of the prosecution and defer to the jury's factual determinations.”) (ellipsis inserted, internal quotation marks omitted).

33.   See supra at n.2.

34.   Norwell v. Cincinnati, 414 U.S. 14, 14-16 (1973); Long, 931 S.W.2d at 293 (“[C]onduct does not lose First Amendment protection merely because the actor intends to annoy the recipient.”); United States v. Dennis, 132 F.4th 214, 228 (2d Cir. 2025) (quoting Sryniawski, 48 F.4th at 587: “The First Amendment ․ ‘protects a variety of speech that is intended to trouble or annoy, or to make another timid or fearful.’ ”).

35.   United States v. Yung, 37 F.4th 70, 78 (3d Cir. 2022) (“The First Amendment protects at least some speech that persistently annoys someone and makes him fearful or timid.”).

36.   See 303 Creative, 600 U.S. at 572 (Speech does not lose protection merely because it causes “anguish or incalculable grief.”)

37.   NAACP v. Claiborne Hardware Co., 458 U.S. 886, 910 (1982) (“Speech does not lose its protected character, however, simply because it may embarrass others or coerce them into action.”).

38.   Johnson, 491 U.S. at 414 (“If there is a bedrock principle underlying the First Amendment, it is that the government may not prohibit the expression of an idea simply because society finds the idea itself offensive or disagreeable.”).

39.   The word “offend” appears only in the conduct part of the statute, with no counterpart in the “intent” part of the statute.

40.   303 Creative, 600 U.S. at 584.

41.   Snyder v. Phelps, 562 U.S. 443, 459 (2011) (quoting Cohen v. California, 403 U.S. 15, 21 (1971)).

42.   See Ojogwu v. Rodenburg Law Firm, 26 F.4th 457, 463 n.4 (8th Cir. 2022) (“Most circuits to consider the issue have concluded that a consumer debtor has Article III standing to assert an FDCPA claim that a debt collector's harassing calls or letters invaded a privacy interest protected by the well-established tort of ‘intrusion upon seclusion.’ ”); Charvat v. NMP, LLC, 656 F.3d 440, 453 (6th Cir. 2011) (referring to Ohio authority suggesting that a privacy action can be had when a debt collector initiates a campaign to “harass and torment” the debtor). See also infra at nn. 43, 45-47.

43.   See Scott v. State, 322 S.W.3d 662, 676 (Tex. Crim. App. 2010) (Keller, P.J., dissenting); infra at nn.45-47.

44.   See infra at nn.45-47.

45.   See Wagner v. State, 539 S.W.3d 298, 310 (Tex. Crim. App. 2018) (defining “harassing manner,” based on definitions of “harass,” to mean “persistently disturb, bother continually, or pester another person”) (also quoting from Harass, Webster’s New International Dictionary (3d ed. 2002): “to vex, trouble, or annoy continually or chronically”) (commenting that this set of definitions “necessarily requires multiple events of harassing communication” and “troubling or annoying someone with frequent or persistent requests or interruptions”).

46.   See State v. Eagle Hawk, 411 N.W.2d 120, 123 n.5 (S.D. 1987) (quoting from Torment, Webster’s New Collegiate Dictionary (1980): “extreme pain or anguish of body or mind: agony,” and “to cause severe [usually] persistent or recurrent distress of body or mind”); State v. Mireles, 16 Wn. App. 2d 641, 654 n.5 (Wash. App. 2021) (quoting Torment, Webster’s Third New Int’l Dictionary (2002): “severe suffering of the body or mind”) (remarking that “[t]ormenting appears to be more severe conduct than intimidating or harassing”).

47.   United States v. Cortez-Cortez, 770 F.3d 355, 358 (5th Cir. 2014) (quoting Abuse, Webster’s Third New International Dictionary (1986): “take unfair or undue advantage of” or “to use or treat so as to injure, hurt, or damage”).

48.   Blocking texts on an Apple or Android phone is relatively straightforward, but blocking on email depends on the email program being used. See, e.g., https://support. mozilla.org/en-US/kb/blocking-sender (last page view May 22, 2025) (“Thunderbird does not have an option for blocking email messages from specific people or domains. However, you can use message filters to automatically dispose of unwanted messages.”) (providing instructions on how to set up message filters). Not everyone knows how to block an email, or whatever the effective equivalent is in the program being used. Only three of Appellant's messages were texts; the remaining thirty messages were emails. As for choosing to delete the emails without reading them, the subject line of an email could contain objectionable content, which would be difficult to avoid seeing before an email is manually deleted. Many of the subject lines of Appellant's emails included content that could have been construed as a violation of the statute (e.g., accusing the victim of raping him), and for some emails, the subject line was the only content of the messages.

49.   See Long, 931 S.W.2d at 295.

50.   State v. West, 632 S.W.3d 908, 913 (Tex. Crim. App. 2021) (bracketed material substituted for “indictment”).

51.   The Court says the only remedy for an “as applied” violation is dismissal of the prosecution, but the cases cited do not address a situation in which only some of several charged alternative statutory methods of committing the offense are unconstitutional as applied.

52.   See Reed v. State, 680 S.W.3d 620, 625-26 (Tex. Crim. App. 2023); Almanza v. State, 686 S.W.2d 157, 171 (Tex. Crim. App. 1985) (op. on reh'g).

53.   See supra at n.14.

Parker, J., filed a concurring and dissenting opinion in which Schenck, P.J., joined.