Mark Bennett

Mark Bennett has practiced trial and appellate law for criminal defendants in Houston and nationwide since 1995. He has made a hobby of slaying unconstitutional statutes, doing so in the high courts of Texas and Georgia, as well as in various intermediate and trial courts. He can best be reached at .

Free Speech in Texas

What is speech? What makes a restriction content based? What is substantial overbreadth? Can a substantially overbroad content-based restriction on speech ever satisfy strict scrutiny?

The answers to these questions may become clearer any Wednesday now. The Court of Criminal Appeals is currently considering four overbreadth challenges to the constitutionality of different penal statutes. These challenges may, before this article is published, change the face of free-speech law in Texas.

Meanwhile, this article will provide a broad overview of challenges to content-based restrictions on speech.

Vagueness vs. Overbreadth

First Amendment challenges to content-based restrictions come in two forms: vagueness and overbreadth.

Vagueness

A statute is void for vagueness under the First Amendment if:

  • A person of ordinary intelligence cannot know what is forbidden;
  • There are no determinate guidelines for law enforcement; or
  • The law is not sufficiently definite to avoid chilling protected expression.

Take, for example, Texas’s harassment statute, section 42.07 of the Texas Penal Code. The harassment statute forbids certain speech intended to “harass, annoy, alarm, abuse, torment, or embarrass” another. Expression is chilled not only by the statute’s threat of conviction (because a jury might find that the defendant intended to harass, annoy, alarm, abuse, torment, or embarrass) but also by the threat of arrest (because a police officer might think the defendant intended to harass, annoy, alarm, abuse, torment, or embarrass) and the threat of prosecution (because a prosecutor might think the defendant intended to harass, annoy, alarm, abuse, torment, or embarrass).

Courts assume that when dealing with statutes such as sexual assault of a child, which restrict non-speech conduct based on the actor’s intent, a jury can reliably read an actor’s mind.

But speech is special—the Free Speech Clause itself is proof of that. Society does not have the same interest in preventing the chilling of non-speech conduct that it has in ensuring that protected speech is not chilled.

Even assuming that a jury can reliably read a speaker’s mind, section 42.07 risks chilling speech because whether a person will be arrested or prosecuted for his speech depends not on a jury’s mindreading, but on a police officer’s and/or a prosecutor’s.

A person might beat the rap and not the ride, but the ride is expensive and no fun and for the ordinary speaker, both the threat of the ride and the threat of the rap can chill speech. A restriction criminalizing speech based on the speaker’s intent “…blankets with uncertainty whatever may be said. It compels the speaker to hedge and trim.” See Thomas v. Collins, 323 U.S. 516, 535 (1945). Under Texas Penal Code section 42.07, it is the threat of prosecution, and not only the threat of conviction, that will likely cause a speaker to hedge and trim.

The speaker who does not intend to annoy has no way to know that his words will not be misinterpreted. He is at the mercy of the inferences of others who might not wish him well. A distinction based on the intent of the speaker or the speech’s effect on its listeners “…offers no security for free discussion.” See Fed. Election Com’n v. Wisconsin Right To Life, Inc., 551 U.S. 449, 468 (2007) (external citation omitted).

Overbreadth

The First Amendment overbreadth doctrine is also intended to prevent restrictions chilling protected speech, but it is not dependent on a lack of clarity. While vagueness is about the person knowing what a statute forbids; overbreadth is about what the statute actually forbids. When it comes to speech, the vagueness argument is basically that speakers can’t determine whether  constitutionally protected communications are prosecutable under the statute at issue. The overbreadth argument is that protected communications are, in fact, prosecutable under the statute but should not be.

Lawyers often confuse vagueness with overbreadth. If a statute seems to restrict protected speech—if it could be used to restrict a real and substantial amount of protected speech—it is not vague, but overbroad.

The Court of Criminal Appeals itself has shown a lack of clarity regarding the difference between vagueness and overbreadth: “[The defendant’s] argument, as we understand it, is that § 42.07 is overbroad on its face because its inherent vagueness makes it unclear whether it prohibits a substantial amount of protected speech.” See Scott v. State, 322 S.W.3d 662, 665 fn.3 (Tex. Crim. App. 2010) (abrogated on other grounds by Wilson v. State, 448 S.W.3d 418 (Tex. Crim. App. 2014)). That is not the argument; that is not what overbroad means.

A content-based restriction that clearly restricts only unprotected speech is valid. A content-based restriction that leaves in doubt whether protected speech is punishable or not is vague. A content-based restriction that clearly restricts a real and substantial amount of protected speech (in relation to its legitimate sweep, the unprotected speech it restricts) is overbroad.

Protected speech is speech outside all recognized categories of historically unprotected speech. States are not free to invent new categories of unprotected speech; they must provide evidence of a longstanding tradition of restricting speech in that category.

So, the overbreadth argument is that the statute, regardless of how ambiguous it might be, forbids speech that is not unprotected.

Returning to the example of Texas Penal Code section 42.07 harassment, while the vagueness argument is that speakers can’t determine whether their speech will be construed as intended to harass, the overbreadth argument is that even speech that is intended to harass (not to mention “annoy, alarm, abuse, torment, or embarrass”) is not unprotected speech, so that the statute by its terms restricts a real and substantial amount of protected speech.

You can see how there might be interplay between these two arguments: one approach courts have taken to save overbroad statutes from overbreadth is to interpret them narrowly—to impose on the Legislature’s words a narrower meaning than the common or legal definition gives them.

Whether rewriting the statute thus is legitimate or not, doing so can convert a statute that was overbroad into one that is vague. While speakers are expected to be familiar with the words of a statute, they are not likewise expected to have read the cases interpreting a statute. So, the fact that the words of the statute alone do not limit the statute’s reach can make the statute itself vague.

Moreover, in Texas, a charging instrument is drafted and a criminal trial jury is instructed according to the language of the statute; there is no mechanism for incorporating narrowing definitions into an indictment, much less a jury charge. Even a defendant who is familiar with the narrowing of an overbroad statute has no guarantee that he will not still be prosecuted and even convicted for his protected speech.

Facial Overbreadth vs. As-Applied Overbreadth

First Amendment overbreadth challenges to statutes can themselves be divided into facial (or as-written) and as-applied challenges.

An as-written challenge makes the argument that, regardless of whether the speaker’s own speech is protected, the statute restricts (based on its content) a real and substantial amount of speech that is protected. An as-applied challenge to a statute is that the speech for which the speaker is being prosecuted (based on its content) is itself protected (that is, not in a recognized category of historically unprotected speech).

As an illustration, defamatory speech is unprotected. If a person is being prosecuted under a facially overbroad or vague statute, but their speech is for some reason unprotected, they might have an as-written but not an as-applied challenge. If a person is being prosecuted under a valid statute, but their speech is protected, he has an as-applied but not an as-written challenge. And if a person is being prosecuted under an overbroad or vague statute for protected speech, they have both an as-written and an as-applied challenge.

As-written challenges can be made pretrial, through either an application for writ of habeas corpus (which can be appealed pretrial) or through a motion to quash (which can only be appealed after a judgment).

Historically, as-applied challenges could only be made in trial because the facts that must be developed to determine whether speech is protected are the same as the facts that must be developed to determine whether the speech violates the statute. Generally, as-applied unconstitutionality challenges were not cognizable in pretrial habeas.

In the Rick Perry case, however, the Texas Court of Criminal Appeals considered the appeal of the then-Governor’s pretrial habeas as-applied challenge to the Abuse of Official Capacity statute. See Ex parte Perry, 483 S.W.3d 884 (2016). The Perry opinion offers two justifications for the cognizability of Perry’s pretrial as-applied challenge. Presiding Judge Keller’s opinion held that Perry’s challenge was cognizable because Perry’s rights would be “effectively undermined if not vindicated prior to trial.” Id. at 895 (footnote omitted).

It is unclear what rights, other than a governor’s right to exercise his veto power, would be cognizable under this rule. For example, intermediate courts have held that the right to free speech is not a right effectively undermined if not vindicated before trial, but it is not obvious that the chilling effect of a pending criminal prosecution on the accused’s right to speak freely does not effectively undermine that right.

Judge Newell, joined by Judges Keasler and Hervey, suggested a different test for pretrial cognizability—a test that might allow a defendant to terminate the chilling effect of a content-based prosecution without the cost and expense of trial. In Judge Newell’s view, “addressing the constitutional claim because the violation is apparent from the pleadings resolves the matter much more cleanly.” Id. at 924.

The Perry split on this issue appears to be the following: Presiding Judge Keller and Judges Yeary and Alcala for the “effectively undermined” rule; Judges Newell, Keasler, and Hervey for the “apparent from the pleadings” rule; Judge Meyers against cognizability for either reason; and Judges Johnson and Richardson not weighing in on the question.

An as-applied challenge under the Free Speech Clause is appropriate when the statute is not facially overbroad, but the defendant’s own speech is constitutionally protected (that is, not in an unprotected category)—in other words, when the statute restricts some protected speech, as exemplified by the defendant’s, but not a real and substantial amount of protected speech. The lesson of Perry’s uncertainty is that, where the constitutional violation is clear from the pleadings (for example, where the sworn Complaint gives enough detail that the court can tell that the accused’s speech is not in fact within any unprotected category), the accused may be able to pursue an as-applied challenge before trial.

Or maybe not. But what does the accused have to lose?

Plenty to Gain

A continuance can be as good as an acquittal for as long as it lasts. If you file a pretrial application for writ of habeas corpus challenging a content-based restriction either as written or as applied to the speech of the accused and the trial court denies relief, you can appeal to the intermediate court of appeals, and from there to the Texas Court of Criminal Appeals. But what happens to the trial-court case in the interim?

As a practical matter, most trial courts want to see legitimate constitutional questions answered before trial and so will not proceed to trial while a habeas appeal is pending. Whether the trial court may proceed to trial while the habeas appeal is pending is an unanswered question.

There have been recent cases from Austin and Beaumont supporting the proposition that a trial court may proceed, but neither case addressed section 11.32 of the Texas Code of Criminal Procedure which provides that once the defendant has appeared in court on the habeas, “he is no longer detained on the original warrant or process, but under the authority of the habeas corpus” until the habeas has been finally determined. See Tex. Code Crim. Proc. § 11.32. If the “original process” is the indictment or information  and a person can only be tried on an indictment or information, if the person is no longer detained on the original process, it stands to reason that he cannot be tried until the habeas is resolved.

It’s an open question, though, and it’s going to have to be litigated in mandamus when a trial court attempts to put a defendant to trial while a viable free-speech challenge is pending, and then possibly taken up to the Texas Court of Criminal Appeals on a Petition for Discretionary Review (PDR).

Voir Dire: Method Follows Model

Imagine you are repairing a car. How you perform the task depends on your understanding of how the car works. If you think internal combustion is powered by magic hamsters, you’re going to do the job differently than if you understand the suck-squeeze-bang-blow of a four-stroke engine. And, because there are no hamsters, feeding the hamsters is not going to work as well as making sure that the fuel, air, compression, spark, combustion, and exhaust necessary to make an internal-combustion engine work are all present.

Or imagine you are writing a song. If you understand music theory, you’re going to write a better song than if you think dissonance is more pleasant to the ear than consonance and make up your own scales.

You may be a natural-born mechanic or songwriter, with an intuitive or even subconscious understanding of your subject, but most people don’t have that sort of talent. Most mechanics and songwriters study their subject matter, and do their work according to their understanding—their model—of how it works. Method follows model.

Because method follows model, a better—closer to the truth—model leads to better—more successful—methods. You don’t have to understand music theory perfectly, but the better you understand it the better your songs might be.

This is no less true of trying cases than of tuning V8s. Trying a case to a jury is about getting jurors to adopt the beliefs that you want them to adopt. Unless we are naturally talented (and almost none of us are), we need to form a model of how people adopt beliefs and develop our trial skills to work with this model.

How jurors decide cases determines how you try cases. If your model of how jurors decide cases is more accurate, your method will be better. If your understanding is closer to the truth, your results are going to be better.

Voir dire1 is our first contact with the jury. I contend, for reasons that I’ll lay out below, that it is our most important contact with the jury. To have a method of voir dire, we first have to have a model of how jurors decide cases—that is, of how people adopt beliefs. The less accurate the model of how people adopt beliefs—the less it reflects reality—the less effective your voir dire will be.

So. We want a good method for causing people to adopt beliefs that are helpful to our clients. We need a good model. How do people adopt beliefs?

Our jury trial system is based on one model we can read in the instructions given to the jury by the court: Jurors are instructed to wait until they have heard all the evidence to begin deliberating, and then to decide the case based only on the evidence and not on emotion or sympathy. In this model, people withhold judgment until the facts are in, then adopt beliefs by weighing those facts and rationally deciding what is most likely correct.

This is a nice way to look at the decision-making process. It gives people a lot of credit for rationality and gives comfort that the world can be a safe and predictable place. Its only shortcoming is that it is laughably wrong.2

As evidence of its wrongness, I offer cognitive biases. Cognitive biases are unconscious mental processes, ingrained systems to save our brain’s decision-making work. The cognitive bias of confirmation bias, which is our tendency to filter information in a way that confirms our preexisting beliefs, is one such example. Due to confirmation bias, we seek out information that confirms our beliefs, and we discard information that refutes them.

Or consider the cognitive bias of fundamental attribution error, which causes us to attribute others’ failings (as we see them) to their character and choices and our own failings to external factors.

Another cognitive bias is affinity bias, which is our tendency to be biased toward people who we perceive to be like us.3

We are not aware of these cognitive biases in our day-to-day decision-making. What their existence tells us is that we aren’t the rational decision-making machines that we imagine ourselves to be.4

And neither are jurors.

So how do we—and jurors—adopt beliefs? If the classic rational-decision-making model is less than perfectly accurate, what model can we adopt that is more accurate and so will guide us to a better way to try cases?

The existence of cognitive biases suggests that we adopt beliefs irrationally, based on things like affinity, and then resist changing them and rationalize—find plausible rational justifications for—them. This model better reflects real-world observation and predicts real-world behavior than the classic model.5 It is, in other words, a better model.

The better model will lead us to better methods. And as long as we have a better model, we don’t need a perfect model. If we have a better model and therefore better methods than our adversary, we are at an advantage.6

Our next step in developing those better methods is to apply this better model, which is very general (“how people adopt beliefs”), specifically to trial (“how jurors, in trial, adopt beliefs”).

Consider the life of a juror. Pulled out of her usual routine, she comes to an unfamiliar environment—the courthouse—where she sits on an uncomfortable pew, and people talk at her about things she doesn’t understand. Worse, they try to make her talk. There’s a judge and there’s a prosecutor—she’s watched “Law and Order,” so she knows that’s the good guy—and a defense lawyer (bad guy!) with the defendant, who’s probably guilty.

Given the better model, at what point does this juror reach a belief about whether our client is guilty?7 The answer is disheartening: She has reached a belief before anyone starts talking. But there’s nothing we can do about that since that’s the first time we even see her, so let’s concentrate on what we can do something about. At what point does a juror reach a belief about whether our client is guilty that we can do something about?

So that we don’t lose heart—and we needn’t lose heart, because our method is better than our adversary’s—let’s treat the beliefs that the juror reaches before we get up to talk as tentative beliefs. Assume that we can change them and instill our own beliefs. Our first opportunity to do so is when we begin our voir dire. Our second opportunity is when we give our opening statement.

By tradition—this is not a written rule—we cannot talk about the facts of the case in voir dire. This unwritten rule probably arises from the proscription against improper commitment questions: If we tell the jurors what the facts of our case are, we are close to committing them to acting on those facts.

But if we can’t talk about the facts, then the prosecutor hasn’t talked about the facts. The jurors’ tentative beliefs, when we begin our voir dire, are based on the prosecutor’s framing of the issue and the prosecutor’s credibility. In our voir dire, our objectives (aside from and superior to the mechanics of eliminating unfavorable jurors) are to reframe the issue and to build credibility with the jury.

We reframe the issue with hypotheticals. Ideally, we will come up with a hypothetical that will allow the jurors to come up with our theory of the case on their own. If the jurors feel that our theory of the case is their idea, they are invested in it and will more readily accept it than if we have to tell them what it is. If the jurors can’t come up with our theory of the case on their own, it’s probably not a good defense.8

Our theory of the case might be “she never intended to meet for sex but was only offering to introduce him to someone his own age,” or “she cried rape when he said he wouldn’t marry her,” or “when his wife found the money missing, he had to claim that it was stolen so that she wouldn’t find out about the affair.”

A hypothetical might be the following: “I’m a criminal-defense lawyer. If I told you, ‘text me if you need a family lawyer,’ what might I mean?” Or: “Why would a woman cry rape after consensual sex?” Or: “If a man gave a woman money, why might he claim that she had stolen it?”9 This will certainly be a different frame than the State has put on the case. Jurors will give several answers; record them all, and don’t commit to any of them yet.10

Before coming in to the courtroom, the jury has started forming a group, and the lawyers are not part of it. An important thing for you to know about groups is that a group feels things that happen to one member as happening to all members.11 Here, if one member of the group comes up with your theory of the case, the rest of the group will treat it as the group’s idea.

The State had the advantage of primacy in its framing, but we have the advantage of recency. We also have the advantage of having trusted the jury to come up with our theory of the case. Not only will they be more committed to the theory for having come up with it themselves, but they will reciprocate the trust we put in them with trust for us.

That—trust—brings us to our other primary voir dire task (other than framing) which is to build credibility with the jury. Other descriptions of what we’re building include:

  • Trust;
  • Liking;
  • Attraction; and
  • Rapport.

Generally: If we like someone, we trust them. If we trust them and like them, we are attracted to them—we are willing to spend time with them and hear what they have to say.12 If we trust and like someone and they trust and like us, we have rapport.

What is more, if we trust someone they are more likely to trust us. If we like someone they are more likely to like us. If we are attracted to someone they are more likely to be attracted to us. These are feedback loops, which you will start noticing everywhere in human relationships.

Most two-way interpersonal communications involve feedback loops. If I smile warmly and sincerely at you, you may feel good and smile warmly and sincerely at me, and I will feel good: a feedback loop. If you scowl at me, I may feel bad and scowl at you, and you will feel bad: also a feedback loop.13

The personal traits that make other people trust, like, be attracted to, and feel rapport with you are charisma. Charisma is in part about opening feedback loops with people that make them feel good. If you make people feel good, they will want to spend time with you and hear what you have to say.

Charisma in jury selection is also about another kind of loop: the Zeigarnik Effect loop. The Zeigarnik Effect is the tendency of our brains to pay attention to unresolved things. If we hear “Once upon a time . . .” our brains pay attention until “. . . happily ever after.” If our audience is curious, they will want to spend time with us and hear what we have to say.

There is a common misconception that charisma is something that we just have or don’t have, and can’t do anything about. Nothing could be further from the truth. We can take concrete actions to increase our charisma. Paying attention to positive feedback loops and Zeigarnik Effect loops is a good start. Beyond that, there are various descriptions of the components of charisma. For example:

  • Power + presence + warmth;14
  • Affability + influence;15 or
  • Makes people feel comfortable, smiles at people often, can get along with anyone, has a presence in a room, has the ability to influence people, knows how to lead a group.16

One researcher has dissected charisma as “both verbal and nonverbal”:

Verbal aspects involve use of metaphor, story, and emotionally appealing language to communicate an inspiring vision and increase self-efficacy. Nonverbal components include paralinguistics (aspects of speech such as variability in volume, rate, pitch, articulation, fluency, and emphasis), kinesics (body involvement such as posture shifts or head movements), gestural fluency, facial expressivity, and eye contact.17

Here, I count fourteen separate competencies—use of story, variability in rate of speech, gestural fluency, and so on—any one of which you could work on to increase your charisma.

By increasing your charisma, you make yourself more credible to the jury. By making yourself more credible to the jury, you improve the chances that they will adopt and act upon your story. And by improving the chances that they will adopt your story, you give your client a fighting chance in a system that is rigged against him. This is all because you were willing to consider that the classic model of how jurors make decisions might be improved upon—that what you learned in law school, and in your practice so far, might be based on assumptions that are simply wrong.

CCA Strikes Down Portion of Online Solicitation of Minor Statute (Other Texas Statutes Vulnerable)

When Grant Scheiner first approached me about challenging the online-solicitation-of-a-minor statute before trial, we thought we were probably just postponing an inevitable trial on the merits. The client was accused of “with the intent to arouse and gratify the sexual desire of the defendant, intentionally communicating in a sexually explicit manner with [the complainant] . . . whom the defendant believed to be younger than 17 years of age, by text message.” It seemed intuitively obvious that an adult couldn’t talk dirty to a child with the intent to arouse or gratify his own sexual desire. But I had challenged Texas Penal Code Sec. 33.021(b) before, with some success in trial courts, in cases involving complainants representing themselves to be children—so we knew there was something wrong with the statue. We just didn’t realize how much was wrong with it.

Procedure

There are two ways to attack the constitutionality of a statute: with a motion to quash or with a writ of habeas corpus. The denial of a motion to quash is not appealable before trial, so if we had challenged the constitutionality of the statue with a motion to quash, we would have had to try the case before appealing.

If the client had been convicted of online solicitation of a minor, he would immediately have had to register as a sex offender. His registration would have spread across websites that revel in and profit from republishing the misfortunes of others; that stain would have been indelible.

The facial (as opposed to “as applied”) unconstitutionality of a statute can be asserted with a pretrial writ of habeas corpus, and the denial of such a writ can be appealed before trial. See Ex Parte Ellis, 309 S.W.3d 71, 79 (Tex. Crim. App. 2010). It seemed obvious to us that the proper procedure in this case would be to file a writ of habeas corpus alleging that Section 33.021(b) violated the First Amendment. Then we could appeal the trial court’s denial of the writ, assuming 248th District Court Judge Joan Campbell in fact denied our writ.

So we did.

And she did.

We appealed to the First Court of Appeals in Houston. We took our time filing the brief, making certain our argument was thoughtful and well-researched. (As the late Percy Foreman said, “A continuance is as good as an acquittal, for as long as it lasts.”) The State briefed the issue. The court didn’t hear argument; it issued an opinion upholding the statute. Lo v. State, 393 S.W.3d 290 (Tex. App.—Houston (1st Dist.) 2011, rev’d, Ex Parte Lo, __S.W.3d__ (Tex. Crim. App. 2013). We filed a motion for rehearing. It was then—when addressing the First Court’s opinion—that I realized that Grant and I were right.

I remember excitedly calling Grant when I finished the first draft for our rehearing: “I think we’re right. We might win this.”

Ten months later the First Court denied our motion for rehearing. It had been 33 months since our client had been charged, and two years since we had filed our writ on his behalf.

Once we realized that we were right—that Section 33.021(b) did not pass muster under current Supreme Court standards—the pace picked up. It took 364 days from filing PDR to getting a unanimous 25-page opinion.

The Law

Traditionally, a content-based speech restriction is presumptively invalid, and the State has the burden of showing that the statute meets strict scrutiny. In this case the Court of Criminal Appeals held that the State had failed to show that Section 33.021(b) survived strict scrutiny.

“Sexually explicit” communications are defined as anything that relates to sex. It’s hard to imagine something that doesn’t, in some way, relate to sex. The State argued that coupling “sexually explicit” with “the intent to arouse or gratify” narrowed the statute sufficiently to satisfy the First Amendment. But the intent to arouse or gratify sexual desire is not forbidden thought, and sexually explicit communications are not forbidden speech; the court found that coupling the two did not render the restriction constitutional:

“Section 33.021(b) prohibits constitutionally protected speech when that speech is coupled with constitutionally protected thought. . . . [C]onsistent with the First Amendment, it is conduct designed to induce a minor to commit an illegal sex act with titillating talk that may be proscribed, not the titillating talk itself.”

I say that strict scrutiny is “traditionally” the test because this is in doubt after U.S. v. Stevens (559 U.S. 460 (2010)), in which the Supreme Court did not mention strict scrutiny, rejected any sort of balancing test as “startling and dangerous,” and applied a categorical test: If the forbidden speech falls within a category of traditionally unprotected speech, the statute is valid; otherwise it is not. The result in our case would have been the same had the Texas Court of Criminal Appeals applied a strict categorical test, but it’s something to keep in mind for future challenges to Texas statutes regulating speech.

What Happens Next

At the time of writing, the Harris County District Attorney’s Office’s public position is that the office will have to review what to do about anyone convicted under the voided law. We hope that Harris County will agree to relief, and that other counties will follow their lead. It’s inherently unfair for people to remain in prison, remain on probation, keep registering as sex offenders, or have criminal records for violating an unconstitutional law. But even if the State agrees to relief, there will be no one-size-fits-all solution; a single lawyer taking the wrong approach for a particular defendant might screw things up for many other defendants.

Other Statutes

For the same reasons that Section 33.021(b) is unconstitutional, Texas’ improper-photography and online-impersonation statutes are unconstitutional under either strict scrutiny or the categorical test. There may be others, as well. Unless it has already been vetted by the Court of Criminal Appeals, any statute that forbids speech based on its content should be scrutinized and tested—even if intermediate courts have already upheld it—and such litigation, as Grant and I learned, is highly specialized, with a long learning curve. If you are fighting charges for a statute that you think might violate the First Amendment, please drop me an email at .