OPINION

Bob Pemberton, Justice.

This appeal arises from an ongoing criminal prosecution that, as the district court observed, involves "unique circumstances" that "have been widely reported, argued, and discussed by many with no standing in the case."1 Whatever the focus of such commentary, our disposition of this appeal turns on legal issues — primarily procedural in nature — that may be of somewhat less public renown.

To summarize the proceedings below, the appellant — James Richard "Rick" Perry, who until recently served as Governor of Texas — sought dismissal, through a pretrial writ of habeas corpus, of two pending criminal charges ("abuse of official capacity" and "coercion of a public servant") that are predicated on alleged acts preceding or relating to his line-item veto of a proposed legislative funding appropriation. In seeking dismissal, Perry has contended chiefly that the statutes on which each charge is based, "as applied" to him, violate constitutional protections related to free expression and the separation of powers. Even while terming these "as applied" constitutional challenges "compelling,"2 the district court determined that it could not decide their merits at that juncture, let alone grant relief, due to procedural limitations the Court of Criminal Appeals has imposed on the ability of lower courts to address such "as applied" challenges when raised through pretrial habeas corpus, as Perry has attempted here. While Perry contends this ruling was error, we reach the same conclusion that the district court did — under the Court of Criminal Appeals's binding precedents, Perry cannot bring his "as applied" constitutional challenges through pretrial habeas corpus.

Perry has also asserted that the statute on which the "coercion of a public servant" charge is based "facially" violates the First Amendment to the United States Constitution. While recognizing that defendants may bring such facial constitutional challenges through pretrial habeas corpus, the district court rejected Perry's claims on the merits. As to this ruling we respectfully disagree with the district court — the statute on which the "coercion of a public servant" is based, as written, and as we are bound to construe it, violates the First Amendment and, accordingly, cannot be enforced.

As a consequence of these holdings, we affirm the district court's denial of relief as to the "abuse of official capacity" charge, because Perry's "as-applied" constitutional challenges cannot be addressed through pretrial habeas corpus under current Texas law. However, because the First Amendment bars enforcement of the statute on which the "coercion of a public servant" charge is based, that charge must be dismissed.

BACKGROUND

The indictment

Through an indictment returned by a Travis County grand jury, appellant Perry has been charged with two counts. Count I alleges that Perry violated section 39.02, subsection (a)(2) of the Texas Penal Code — "Abuse of Official Capacity" — which in pertinent part makes it an offense for a "public servant," "with intent to harm... another," to intentionally or knowingly "misuse government property ... that has come into the public servant's custody or possession by virtue of the public servant's office or employment."3 Count II alleges that Perry violated Penal Code section 36.03(a)(1) — "Coercion of Public Servant" — which makes it an offense for a person, "by means of coercion" (a term defined under the Penal Code to include "a threat, however communicated, ... to take or withhold action as a public servant") to "influence" or attempt to "influence" a "public servant" to certain ends.4 Although the parties are continuing to litigate the precise content of the indictment before the district court, it suffices for present purposes to note their general agreement that both charges relate to alleged conduct by Perry preceding or relating to his line-item veto of a legislative appropriation to fund the "Public Integrity Unit," then housed within the Travis County District Attorney's Office, while he was serving as Governor of Texas.

In response to the indictment, Perry has vigorously disputed that the alleged conduct actually equals a crime under either of the statutes under which he was charged, properly construed, and he urges that any prosecution for such conduct would infringe not only his personal free-speech rights under the Texas and federal constitutions, but also core powers vested in all Texas governors by the Texas Constitution. Were the pending criminal charges civil claims instead, Perry would have in his defensive arsenal an array of procedural mechanisms through which he could raise such challenges to the legal viability of the State's theories at the outset and possibly obtain dismissal prior to trial — e.g., special exceptions,5 motions for summary judgment,6 and the recently created motion to dismiss a cause of action that "has no basis in law or fact,"7 with possible immediate recourse to appellate courts in the event such relief was denied at the trial level.8 But the pretrial remedies presently available to defendants in Texas criminal cases are more limited9 — as the Texas Court of Criminal Appeals has put it, "trial on the merits" remains "`the main event.'"10 Whether it is sound policy to draw such a distinction between defendants faced with loss of property versus loss of liberty is not a question that this intermediate appellate court is empowered to address — we are instead duty-bound to follow that existing law unless and until the Legislature or the Court of Criminal Appeals instructs us otherwise.11

Proceedings below

This appeal arises from Perry's attempt to invoke one of the pretrial remedies that potentially may be available to defendants in Texas criminal cases — an application for a pretrial writ of habeas corpus. Generally described, a criminal defendant may obtain pretrial habeas corpus relief from a trial court when he or she is (1) "restrained" or "confined" (2) illegally and (3) does not possess what is deemed an "adequate remedy by appeal" against such restraint or confinement.12 Further, an applicant who is denied that relief at the trial level, such as Perry, has a right to appeal that order,13 a potential avenue for obtaining what amounts to interlocutory appellate review regarding a pending prosecution.14

There has been no dispute that Perry is "restrained" in the sense required for pretrial habeas relief pursuant to each of the two charges alleged in the indictment.15 As for the remaining requirements for such relief, he has challenged the legality of his restraint pursuant to Count I of the indictment ("Abuse of Official Capacity") on nine distinct constitutional grounds:

Regarding Count II ("Coercion of Public Servant"), Perry has brought the following constitutional claims:

As the district court observed, the claims raised by Perry are "unique," "important," and "certainly deserve careful consideration in an appropriate forum."17 The court even went as far as to term "compelling" Perry's arguments regarding the Separation of Powers and the Speech and Debate Clause.18 However, the district court was bound, as are we, to adhere to Court of Criminal Appeals precedents instructing lower courts that we cannot reach the merits of any claim presented through pretrial habeas, however meritorious it may potentially be, without first determining, as a "threshold" matter, that the claim is properly "cognizable" through that procedural mechanism — that is, whether the court can properly decide the merits of the claim at that juncture or must leave it to be addressed through other trial-level proceedings and any subsequent post-conviction appellate remedy.19 This requirement rests on jurisprudential policies, frequently emphasized by the Court of Criminal Appeals in recent years, that "pretrial habeas, followed by interlocutory appeal, is an `extraordinary remedy'" and should not be "`misused to secure pretrial appellate review of matters that in actual fact should not be put before appellate courts at the pretrial stage.'"20 The court has similarly stressed that "[a] defendant may use a pretrial writ of habeas corpus only in very limited circumstances"21 and that the remedy "should be reserved for situations in which the protection of the applicant's substantive rights or the conservation of judicial resources would be best served by interlocutory review."22

The district court thus proceeded immediately to that threshold inquiry and determined, without hearing evidence, that, under "the current state of the law in Texas," only Perry's challenges to the constitutionality of the statutory basis for the "Coercion of Public Servant" charge "on its face" — i.e., claims 1-4 regarding Count II — were cognizable in pretrial habeas. As for the remaining claims — sixteen in all, including all of Perry's claims regarding Count I — the district court concluded each was not cognizable — as the district court put it, "the court's hands are tied" (emphasis in original)23 — because the claim asserted an "as applied" constitutional challenge. Accordingly, the district court denied relief as to the sixteen claims asserting "as applied" challenges without reaching their merits. Although it did reach the merits of the facial constitutional challenges it held to be cognizable, the court rejected them. On appeal, Perry insists that both sets of rulings were in error.

COGNIZABILITY ISSUES

We will begin with Perry's arguments relating to the sixteen "as applied" claims the district court declined to reach, because these arguments concern a threshold issue and because Perry devotes the majority of his issues on appeal to arguing the claims' cognizability and merits.24 Perry's specific contentions are best understood against the backdrop of the "the current state of the law in Texas" on which the district court relied.25

The Court of Criminal Appeals has held that "as applied" constitutional challenges are not cognizable in pretrial habeas corpus

The district court emphasized several recent Court of Criminal Appeals decisions addressing the extent to which an "as applied" constitutional challenge to a penal statute can be properly decided through pretrial proceedings, including pretrial habeas corpus, in lieu of awaiting trial on the merits and any post-conviction appeal. As the Court of Criminal Appeals has explained, an "as applied" constitutional challenge "concedes the general constitutionality of the statute, but asserts that the statute is unconstitutional as applied to [the defendant's] particular facts and circumstances."26 In contrast, a "facial" constitutional challenge seeks to establish that the statute is unconstitutional and unenforceable as to any person.27 As between these two classifications, the district court concluded that each of the sixteen claims it did not reach was an "as applied" challenge because it asserted that a penal statute violated constitutional limitations as it would operate against Perry's particular circumstances — with emphasis on his unique duties and powers as Governor — as opposed to contending that some constitutional defect rendered the statutes unenforceable as to anyone against whom it could be applied. Indeed, as the district court noted, Perry himself phrased virtually all of these claims in terms of the statutes' constitutionality "as applied" to him and his role as Governor at the time.

Having classified the sixteen claims as "as applied" challenges, the district court looked to Ex parte Weise,28 in which the Court of Criminal Appeals granted review to address "whether a pretrial writ of habeas corpus may issue on the ground that a penal statute is being unconstitutionally applied because of the allegations in the indictment or information."29 The applicant, Weise, had been charged with illegal dumping and sought pretrial habeas corpus on the basis that the illegal-dumping statute was unconstitutional "as applied" to him because the charging instrument did not allege a mens rea.30 In evaluating whether pretrial habeas was available as a remedy, the court emphasized a longstanding general rule that pretrial habeas corpus is not available to "test the sufficiency" of the charging instrument.31 The Court of Criminal Appeals acknowledged, however, that it had also recognized certain "exceptions" to this general rule that it had deemed to fall into a broader category of cases in which it had long permitted pretrial habeas relief where "the alleged defect would bring into question the trial court's power to proceed."32 One such "exception," the court noted, was where the face of the charging instrument showed that the prosecution was barred by limitations.33 Another recognized "exception" noted by the Weise court — and of more direct relevance here — is a complaint that a statute defining the charged offense is unconstitutional on its face, as this would mean that "there is no valid statute and the charging instrument is void."34 The Court of Criminal Appeals concluded that Weise's complaint did not fit under any of these recognized exceptions that would permit his claim to be raised through pretrial habeas corpus. It observed that Weise was not making a challenge to the facial constitutionality of the statute, nor "claiming that the trial court lacked the power to proceed."35 Instead, the court reasoned, "he is claiming that the statute as applied via the information is unconstitutional because it fails to allege a mens rea."36 It termed this sort of challenge, "in reality, simply an attack on the charging instrument."37 Absent any "alleged... deficiencies that we have recognized as cognizable on habeas corpus," the court held that the issue of whether the illegal-dumping statute required a culpable mental state "is not yet ripe for review."38

In addition to Weise, the district court also looked to the Court of Criminal Appeals's more recent decision in Ex parte Ellis.39 Ellis involved an attempt by defendants to raise by pretrial habeas and subsequent appeal what they had styled as a "facial" vagueness challenge to Texas's money-laundering statute. The thrust of this challenge, however, was not that the statute was vague in all of its applications, but that it was unconstitutionally vague if it was construed to apply to checks, an issue of eventual significance in the underlying prosecution.40 The district court denied relief, and that ruling was appealed to this Court. This Court affirmed the district court's ruling, concluding that the money-laundering statute was not unconstitutionally vague, but its reasoning turned on a holding that the statute did not apply to checks.41 The Court of Criminal Appeals granted the State's request for review and held that the defendants' challenge was not cognizable because, in the court's view, the defendants "did not really advance a facial challenge." Specifically, the court reasoned that the defendants had presumed there were some valid applications of the money-laundering statute and had thus "advanced an `as applied' claim that was disguised as a facial challenge."42

Citing Weise, the Ellis court emphasized the longstanding rule that "pretrial habeas is not available to test the sufficiency of the charging instrument."43 Nor, the Ellis court added, could the remedy be used "to construe the meaning and application of the statute defining the offense charged."44 Additionally, while acknowledging that "[p]retrial habeas can be used to bring a facial challenge to the constitutionality of the statute that defines the offense," the Court of Criminal Appeals stressed that the remedy "may not be used to advance an `as applied' challenge," again citing Weise for that proposition.45 "Addressing the `as applied' substance of the claim," the high court held, thus "resulted in a circumvention of the pretrial habeas cognizability limitations."46

Another Court of Criminal Appeals case of significance to the district court was Lykos v. Fine.47 Lykos arose out of a pending capital murder case in which the State had given notice of its intent to seek the death penalty.48 The defendant filed a novel pretrial motion seeking a ruling declaring the death-penalty sentencing statute unconstitutional "as applied" to him, and attempted to demonstrate by pleadings and proof that he faced what he claimed to be an unconstitutionally high systemic risk of wrongful conviction or other error at trial.49 After the trial court set an evidentiary hearing on the motion, the State sought to bar the proceedings via mandamus and prohibition, challenging the trial court's legal authority to conduct such a hearing before trial.50 Concluding that "there is no basis under Texas law to conduct a pretrial evidentiary hearing to determine the `as applied' constitutionality of a state penal or criminal procedural statute" and that the trial court, consequently, was "acting beyond the scope of his lawful authority," the Court of Criminal Appeals conditionally granted relief.51

In reaching that conclusion, the Court of Criminal Appeals relied heavily on the view that "as applied" constitutional challenges could only be brought "during or after a trial on the merits, for it is only then that the trial judge and reviewing courts could have the particular facts and circumstances of the case needed to determine whether the statute or law has been applied in an unconstitutional manner."52 The court also relied upon prior decisions in which it had emphasized that the Legislature had not seen fit to provide criminal courts procedural mechanisms by which they could grant "declaratory judgments"53 or test, prior to trial, the sufficiency of evidence supporting or defeating the elements of charged offenses.54 In that regard, the court cited with approval a 1950 case, Ex parte Hammonds,55 in which it had refused to entertain a pretrial habeas claim seeking determination of whether a set of agreed facts submitted by the parties constituted the crime of rape. In Hammonds, the court had observed that "[i]n Texas, procedure such as demurrer to the evidence, declaratory judgment or pre-trial judgment, in criminal cases, is not recognized," and refused to allow pretrial habeas to be used to achieve effectively the same ends.56

Emphasizing Weise, Ellis, and Lykos, the district court concluded that "[t]he Texas Court of Criminal Appeals has made it crystal clear that a trial court has no authority to consider the merits of a pretrial writ of habeas corpus based upon an `as applied' challenge to the constitutionality of a statute"57 and held that this principle barred it from reaching the merits of the sixteen claims it had classified as "as applied" constitutional challenges.

But Perry insists these precedents are distinguishable

Although not disputing that the sixteen claims each assert a form of "as applied" constitutional challenge rather than what is generally considered a facial one, Perry insists his particular "as applied" challenges are nonetheless cognizable in pretrial habeas, notwithstanding the Court of Criminal Appeals precedents discussed above, for two related sets of reasons. First, while acknowledging that the Court of Criminal Appeals in Ellis "has made the broad statement that pretrial habeas `may not be used to advance an `as applied' challenge,'" Perry insists that the high court's jurisprudence is actually more nuanced than the district court assumed. He reasons that the "as applied" constitutional challenges that have been the concern of the Court of Criminal Appeals in cases like Ellis and Lykos refer only to challenges that, as a practical matter, could not be determined without recourse to evidence. While such evidentiary development may be necessary to assert "typical as-applied challenges," Perry maintains, this is not true of all "as applied" challenges, because some "can be decided solely by reference to the indictment and the statute." In fact, Perry urges, the Court of Criminal Appeals has previously entertained on pretrial habeas this very sort of "as-applied-to-the-indictment" challenge, as he terms it, in a 1991 decision, Ex parte Boetscher.58

In that case, Boetscher, a resident of Michigan, was indicted in Lubbock County for criminal nonsupport of his children, which the then-applicable version of the Penal Code enhanced to a felony on the sole basis that he had resided out of state at the time the offense was allegedly committed.59 Boetscher sought relief by pretrial habeas corpus, and the Court of Criminal Appeals summarized his contentions as "argu[ing] that [the enhancement provision], as applied to the unusual circumstances of his case, denies him equal protection [under the Fourteenth Amendment] because it makes his alleged conduct a felony, rather than a misdemeanor, `solely because he is not a resident of Texas.'"60 There is no indication in the opinion that the State disputed the cognizability of Boetscher's equal-protection challenge in pretrial habeas, and the Court of Criminal Appeals did not address that issue beyond dropping a footnote stating, "We have previously entertained pretrial habeas corpus proceedings challenging the constitutionality of penal statutes."61 In support of that assertion, the court cited two cases, Ex parte Crisp,62 and Ex parte Psaroudis.63 Crisp involved a facial challenge to amendments to the Controlled Substances Act that was founded on an asserted defect in the bill caption.64 Psaroudis, on the other hand, addressed whether there was a "valid statute under which [the defendant] can be charged" without any explicit reference to an asserted constitutional defect; the court instead construed the Controlled Substances Act to determine whether the statute prohibited the acts with which the applicant had been charged, possession or delivery of hashish.65

The Boetscher court preceded its analysis of the claim by quoting the indictment "in relevant part," noting allegations that Boetscher had failed to provide child support for his children in Lubbock County and that "the defendant was then residing in another state, to-wit: Michigan."66 The court subsequently observed that statutory amendments had changed the enhancement provisions, which had previously applied to persons who committed the offense in Texas and then fled the state, to "provide a felony penalty for all defendants who commit the offense while simply residing in another state."67 This new enhancement scheme, the court concluded, "plainly implicates one of the basic rights of all Americans," the right to travel, and was thus deemed invalid under the Equal Protection Clause of the Fourteenth Amendment unless "the government can show it is truly necessary to the promotion of a compelling governmental interest."68 Because the State had not attempted to make any such showing, the court went on to hold that "the equal protection clause of the Fourteenth Amendment prohibits the application of [the enhancement provision] to appellant under this indictment."69 In closing, the court expressly disclaimed any opinion as to whether the statute could be applied against "nonsupport offenders who commit the offense in this state and then flee."70

Perry argues that his sixteen "as-applied-to-the-indictment" constitutional challenges are like those the Court of Criminal Appeals reached in Boetscher. As Perry sees it, his indictment, on its face, establishes that he "is being prosecuted for a veto and an alleged veto threat," infringing not only his personal free-speech rights and immunities he possesses under the Speech and Debate Clause, but also the separation of powers mandated by the Texas Constitution, amounting to unlawful interference with the policymaking branches by entities within the Judicial Department.71 In these ways, Perry reasons, his indictment is akin to the one in Boetscher, which facially revealed that the State was seeking to impose penalties on the applicant that hinged on the fact that he resided in Michigan, violating the right to travel. Just as Boetscher was permitted to challenge the constitutionality of the enhancement provision "as applied" to his circumstances via pretrial habeas, Perry urges, he is likewise permitted to bring his "as-applied-to-the-indictment" challenges that way.

Perry's second set of arguments draws upon additional jurisprudential policy justifications that the Court of Criminal Appeals has sometimes invoked when permitting pretrial habeas relief in other contexts. Foremost among these are cases where the high court has permitted the use of pretrial habeas to assert constitutional protections that it perceived would be effectively undermined otherwise. These have included claims regarding bail and those seeking to assert the protection against double jeopardy.72 Perry regards the double-jeopardy cases as particularly instructive here because that constitutional protection entails a "right not to be tried" that must be vindicated before trial if it is to be effective. The same is true of his claims under the Texas Constitution's Separation-of-Powers provision and Speech and Debate Clause, Perry insists, because these protections shield public officials not only from ultimately being held criminally liable based on acts within their scope, but also from being subjected to prosecution based on those acts at all.

Perry further asserts that pretrial habeas relief would not only be the best (if not exclusive) means of protecting his substantive rights as compared to trial and appeal, but would also serve the broader jurisprudential policy interest of conserving judicial resources.73 In this regard, Perry argues that his "as-applied-to-the-indictment" constitutional challenges are the "functional equivalent" of facial constitutional challenges, turning on questions of law that require no evidentiary development and could be resolved by appellate courts as well as trial courts could. He adds that favorable resolution of his claims would compel his immediate discharge from restraint, another consideration in the Court of Criminal Appeals's pretrial habeas jurisprudence.74 And besides these judicial policy considerations, Perry urges, immediate resolution of his claims via pretrial habeas and appeal is compelled by the separation-of-powers concerns he has invoked, "[t]he paramount public interest in the effective functioning of state institutions in general, and the untrammeled exercise of the line-item veto by the governor in particular." In short, Perry maintains that prudential limitations that the Court of Criminal Appeals has imposed on the pretrial habeas corpus remedy within the Judicial Department should not serve to facilitate a more serious ongoing infringement by that Department on the other governmental branches.

The State attacks Perry's proposed distinctions

The State responds that the district court was correct — the Court of Criminal Appeals's decisions in Ellis, Weise, and Lykos ultimately control and are fatal to Perry's attempt to assert his sixteen "as-applied-to-the-indictment" claims via pretrial habeas corpus. As for Perry's reliance on Boetscher, the State grants that the Court of Criminal Appeals permitted there an "as-applied equal protection challenge on the face of the indictment" via pretrial habeas. However, the State insists, "[f]ew cases are as simple or unusual as Ex parte Boetscher," observing that the challenged statute "[o]n its face ... treated in-state residents differently from out-of-state residents," that Boetscher's "out-of-state residence was the gravamen of the enhanced penalty," and that the indictment's allegation that Boetscher lived out of state was uncontroverted and "not subject to interpretation." Under those particular circumstances, in the State's view, Boetscher's equal-protection challenge amounted to a "`pure law' challenge that is justiciable using only the language in the indictment."

In this regard, the State draws comparisons to the Court of Criminal Appeals's jurisprudence regarding cognizability of limitations-based challenges to charging instruments. As the high court noted in Weise, it has long recognized an "exception" against the general rule barring pretrial habeas challenges to "test the sufficiency" of the charging instrument "when the pleading, on its face, shows that the offense charged is barred by limitations."75 But this principle has been further refined, as the State emphasizes, by a distinction between challenges based on charging instruments that establish an "incurable," "irreparable," and "absolute" limitations bar to prosecution, versus those that complain of asserted deficiencies in a "tolling paragraph," "explanatory averment," or "innuendo allegations" that would "suffice to show that the charged offense is not, at least on the face of the indictment, barred by limitations."76 Only the former category — which the high court has elsewhere termed a "pure law" limitations defense — is deemed cognizable in pretrial habeas; the latter — what the court has labeled a "limitations factual defense" — is not.77

The State urges that Perry's "as-applied-to-the-indictment" constitutional challenges are far removed from the "pure law" challenge addressed in Boetscher and analogous limitations cases, requiring factual development beyond what can be determined from the face of the indictment alone. In fact, as the State emphasizes, the parties have continued to litigate the precise content of the indictment — the predicate of Perry's "as-applied-to-the-indictment" challenges — before the district court in the interim since that court issued the order on appeal, with the State having since filed a "Bill of Particulars & Amendment of Indictment" that purports to add "further clarity" to the asserted factual bases for both counts, and Perry objecting to that filing on numerous grounds and moving to quash. Among other changes, the State's filing purports to modify Count I, which in its original form did not allege explicitly that Perry's supposed "misuse [of] government property" made the basis for the charge consisted of a gubernatorial veto — a central premise of Perry's "as-applied-to-the-indictment" constitutional challenges. The new language would include specific allegations that Perry's "misuse [of] government property" underlying Count I was his use of "the lawful power of gubernatorial veto for an unlawful purpose, to-wit: eliminating funding for the Public Integrity Unit after Ms. Lehmberg refused to resign from her elected position as Travis County District Attorney."78

Leaving aside any such uncertainties regarding the content of the indictment, the State contends more broadly that Perry's "as-applied-to-the-indictment" challenges rest upon myriad speculations, hypotheses, and interpretations as to what the underlying facts actually are and how the two statutes at issue would apply to them — e.g., precisely what then-Governor Perry said and did, whether these facts would constitute violations of Penal Code sections 36.03(a)(1) and 39.02(a)(2), and whether any such applications of the statutes would violate the constitutional provisions he has invoked. This factual record will exist only after evidence is presented at trial, the State insists, and Perry's attempts to adjudicate them via pretrial habeas amount ultimately to the sorts of impermissible pretrial challenges to the sufficiency of the indictment, attempts "to construe the meaning and application of the statute defining the offense charged," "declaratory judgment" claims, and "as applied" constitutional challenges that the Court of Criminal Appeals condemned in Ellis and Weise. The State further emphasizes additional Court of Criminal Appeals precedents that have refused pretrial habeas relief in other contexts when the court perceives evidentiary development at trial to be necessary or beneficial in resolving the claim.79

Perry's "as applied" constitutional challenges are not cognizable under the Court of Criminal Appeals's current precedents

Perry's arguments ultimately amount to assertions about what Texas pretrial habeas law currently is, on one hand, and what Perry contends it ought to be, on the other. With regard to the existing law, the Court of Criminal Appeals has sent us lower courts the following unequivocal messages through its current pretrial habeas jurisprudence:

As the district court concluded, these precedents would appear to make "crystal clear" that Perry's "as-applied-to-the-indictment" claims are not cognizable in pretrial habeas. The closest existing Court of Criminal Appeals authority that Perry offers to support a contrary conclusion is Boetscher. That case, which predates Weise, Ellis, and Lykos by a decade or more, is somewhat difficult to reconcile with those more recent decisions, and the Psaroudis case, on which Boetscher partly founded its cognizability holding,89 would appear to be squarely inconsistent with them.90 At least one respected secondary authority has suggested that some Court of Criminal Appeals cognizability decisions from that earlier era "fail[ed] at the time to have fully embraced the distinction [between] challenges to statutes on their faces and as applied" and would likely be decided differently by the high court today.91 As for the Court of Criminal Appeals's own pronouncements, it has never explicitly overruled Boetscher, but amid its synthesis of Texas pretrial habeas law in Weise, the court cited Boetscher as an illustrative case where it had held facial constitutional challenges to be cognizable in pretrial habeas.92 Whether this aspect of Weise would imply support for Perry's notion that "as-applied-to-the-indictment" challenges are deemed tantamount to facial challenges and are thus cognizable, represents an after-the-fact attempt by the Court of Criminal Appeals to recharacterize Boetscher in line with its current views regarding the cognizability of "as applied" challenges, or neither, is unclear.

Without more definite guidance from the Court of Criminal Appeals, we must continue to follow Boetscher and, as neither case purported to overrule it, view it as an unstated qualification to the holdings of Ellis and Weise that "as applied" constitutional challenges are not cognizable in pretrial habeas.93 However, we agree with the State that Perry's "as-applied-to-the-indictment" claims are distinguishable from those addressed in Boetscher. The factual bases for Boetscher's "as applied" challenge were straightforward and readily ascertainable from the face of his indictment: he was being prosecuted for child-support nonpayment, with a penalty enhancement because he resided in Michigan. The opinion reflects no dispute regarding either the underlying fact of Boetscher's Michigan residency or that the statute made him susceptible to higher penalties solely for that reason.94 Consequently, while employing "as applied" phrasing, the Court of Criminal Appeals's analysis resembled that in a conventional facial challenge — the inquiry centered on the language of the statute itself, concluding that the statute's "classification scheme... implicates one of the basic rights of all Americans," the right to travel.95

Perry's "as applied" challenges, by contrast, are intertwined not only with disputes about what the underlying facts are — indeed, disputes even about what underlying facts have been alleged in the indictment — but also with disputes about whether those facts would constitute violations of the statutes under which he is charged. These features of his "as-applied-to-the-indictment" claims, unlike those in Boetscher, implicate jurisprudential policies against using pretrial habeas to test the sufficiency of allegations in a charging instrument or to obtain a "declaratory judgment" or otherwise determine the legal effect of particular facts under a criminal statute prior to trial. In these ways, Perry's "as-applied-to-the-indictment" claims more closely resemble the "as-applied" claims addressed in Ellis and Weise, and we accordingly conclude that those decisions, rather than Boetscher, control here.96 Under Ellis and Weise, Perry's "as-applied-to-the-indictment" claims are not cognizable in pretrial habeas. While the subject matter of these claims may indeed be "important," and "worthy of careful consideration," as the district court acknowledged, that consideration alone has not been a controlling determinant for the Court of Criminal Appeals.97

Perry's remaining rationales for limiting or avoiding the effect of Ellis and Weise are ultimately rooted not in any existing controlling precedent of the Court of Criminal Appeals, but broader "factors" he identifies in what he terms the high court's "evolving jurisprudence regarding cognizability in pretrial habeas." As we have already emphasized, this Court is not empowered to "evolve" or otherwise alter the binding effect of the Court of Criminal Appeals's controlling precedents, even if we might perceive sound justifications for doing so.98 Such changes are the prerogative of the high court itself, and until it so acts, we remain bound by Ellis and Weise.99 Similarly, "`[a]s an intermediate appellate court, we are not free to mold Texas law as we see fit but must instead follow the precedents of the Texas Supreme Court [or the Texas Court of Criminal Appeals] unless and until the high court[s] overrule[] them or the Texas Legislature supersedes them by statute.'"100 In similar instances where that principle has required us to reject proposed modifications of the existing law that would conflict with higher court precedents, litigants have sometimes succeeded later in obtaining the desired change in a higher court or from the Legislature.101 But unless and until such changes occur here, Perry's "as-applied-to-the-indictment" challenges are not cognizable in pretrial habeas corpus.

As a final note regarding cognizability of these "as applied" challenges, we acknowledge a broader concern raised by Perry and also by amici who support him — in their view, this case represents an instance where a defendant has been made to face criminal charges of dubious legal viability (and/or politically motivated origins) and will inevitably be vindicated, the only question being at what point in the proceedings this will occur. In such circumstances, they urge, the opportunity and eventuality of obtaining a favorable judgment at trial or on appeal will often do little to rectify the harm the defendant suffers to reputation, professional standing, and the like during the meantime. In fact, they suggest, inflicting such harms may be the primary goal of those who pursue the charges. We express no opinion at this juncture as to whether these characterizations of the charges pending against Perry are accurate — the more relevant observation now is that similar things were said of the proceedings that gave rise to Ellis and related cases, yet it was only through trial and post-judgment appeals that the now-former public official at the center of the cases was able to refute the legal viability of the charges against him.102 Under Ellis and similar decisions of the Court of Criminal Appeals, such potential harms incurred by defendants in the meantime, however considerable they may be, are simply deemed insufficient in themselves to provide a basis for relief through pretrial habeas corpus. If the Texas criminal justice system should operate differently, that change must come from the Court of Criminal Appeals or the Legislature.

As the law stands now, the district court did not err in denying Perry pretrial habeas relief on all sixteen of his claims that assert "as applied" constitutional challenges. And because these include the sole claims Perry asserts regarding Count I ("Abuse of Official Capacity"), we affirm the district court's denial of relief as to that charge. But as to Count II ("Coercion of Public Servant"), there remain to be considered Perry's alternative facial constitutional challenges.

FACIAL CONSTITUTIONAL CHALLENGES

In claims 1-4 concerning Count II, Perry challenges the facial constitutionality of the statute on which that charge is based, section 36.03(a)(1) of the Penal Code, as it incorporates the relevant portion of the Penal Code's definition of "coercion." These claims are grounded in the established doctrines of "overbreadth" and "vagueness" that derive from the "freedom of speech" guaranteed by the First Amendment to the United States Constitution,103 which has been applied to the states through the Fourteenth Amendment.104 Although Perry has also based these claims on the Texas Constitution's free-speech provision, Article I, Section 8,105 his arguments, at least on appeal, rely entirely on case decisions applying the First Amendment, and he does not identify any feature of the Texas provision that would cause it to operate any differently than its federal counterpart. Accordingly, we will focus our analysis entirely on the First Amendment protection and jurisprudence.106

The First Amendment prohibits overbroad and vague speech restrictions

The First Amendment overbreadth doctrine holds that a statute is facially invalid if, as written, it sweeps within its coverage a "substantial" amount of First Amendment-protected expression as compared to any activity it proscribes constitutionally.107 The doctrine thus presumes that the challenged statute might otherwise have some legitimate applications, contrary to the typical character of facial challenges,108 and it follows that a party may bring an overbreadth challenge without regard to whether he would otherwise have suffered a constitutional violation from the statute's application, an exception to typical standing requirements.109 This result is deemed justified by concerns that the statute would otherwise suppress or "chill" the constitutionally protected free expression of numerous persons not before the court.110 The Court of Criminal Appeals, quoting the United States Supreme Court, recently summarized these underlying jurisprudential policies in Ex parte Lo:

But because the overbreadth doctrine departs from normal standing concepts and has broader effects than case-by-case adjudication — the constitutional defect is not merely a bar to the statute's enforcement against a particular defendant, but causes the statute to be invalidated altogether — the high courts have emphasized that the overbreadth doctrine is "strong medicine" that should be employed "sparingly" and "only as a last resort,"112 and "must be carefully tied to the circumstances in which facial invalidation of a statute is truly warranted."113 Consequently, before a court invalidates a statute for overbreadth, the statute's reach to impermissible applications "`must be not only real, but substantial as well, judged in relation to the statute's plainly legitimate sweep.'"114 Otherwise, "whatever overbreadth may exist should be cured through case-by-case analysis of the fact situations to which its sanctions, assertedly, may not be applied."115

The vagueness doctrine, on the other hand, derives in part from due-process notice concerns that require a criminal law to be sufficiently clear to afford a person of ordinary intelligence a reasonable opportunity to know what is prohibited while also establishing determinate guidelines for law enforcement.116 According to the Texas Court of Criminal Appeals in Ellis, a facial vagueness challenge ordinarily must demonstrate that a criminal law fails this standard "in all of its applications."117 However, "`[w]hen a statute is capable of reaching First Amendment freedoms, the doctrine of vagueness `demands a greater degree of specificity than in other contexts'" for reasons similar to those underlying the overbreadth doctrine — "to preserve adequately the right of free expression because `uncertain meanings inevitably lead citizens to steer far wider of the unlawful zone than if the boundaries of the forbidden areas were clearly marked.'"118 Thus, "where First Amendment freedoms are implicated, the law must be sufficiently definite to avoid chilling protected expression" or be struck down.119 Likewise, when a vagueness challenge involves First Amendment considerations, a criminal law may be held invalid on its face, as with an overbroad law, "even if the law has some valid application" and "even though it may not be unconstitutional as applied to that defendant's conduct."120 However, "`perfect clarity and precise guidance have never been required even of regulations that restrict expressive activity.'"121 Furthermore, the concern of the vagueness doctrine "`is not the possibility that it will be sometimes difficult to determine whether the incriminating fact [the statute] establishes has been proved; but rather the indeterminancy of precisely what that fact is.'"122

The State does not dispute that Perry's claims asserting that Penal Code section 36.03(a)(1) is unconstitutionally overbroad and vague "on its face" are, in both form and substance, facial constitutional challenges, as opposed to being "disguised" as-applied claims of the sort Ellis condemned. Nor does the State dispute that facial constitutional challenges like these are cognizable in pretrial habeas corpus123 — in fact, the Court of Criminal Appeals has recently decided both facial overbreadth and vagueness challenges to criminal statutes that had been raised through that procedural mechanism124 — or that the district court could properly reach the merits of both challenges here. On appeal, however, the State urges us to decline to reach Perry's appeal of the district court's decision rejecting those challenges on the merits. Instead, the State argues, we should remand those claims to the district court, allow proceedings there to run their course, and weigh in only when and if warranted by an appeal from a judgment after trial. To support that notion, the State cites language from Supreme Court decisions emphasizing the jurisprudential policies favoring judicious use of the overbreadth doctrine as an alternative to awaiting "as applied" constitutional challenges raised through case-by-case adjudication.125 It also argues more generally that analysis of First Amendment facial challenges tends to benefit from development of a full evidentiary record at trial, as opposed to relying on more abstract analysis of the statutory language and its hypothetical implications.

While such considerations might inform application of the overbreadth doctrine here, we cannot agree that they permit us to defer or avoid addressing Perry's facial challenges altogether. The State does not cite, nor are we aware of, any Court of Criminal Appeals decision where the court has deemed these considerations to impact the cognizability of a facial overbreadth or vagueness challenge on pretrial habeas or the subsequent availability of appellate relief. To the contrary, the Court of Criminal Appeals's holdings instruct us that trial and appellate courts should reach facial constitutional challenges at the pretrial habeas stage because they implicate "the trial court's power to proceed," a jurisdiction-like concern.126 These decisions leave us no more discretion to defer addressing Perry's facial challenges at this juncture, as the State urges, than they permit us to decide his "as-applied-to-the-indictment" challenges now, as he insists.

Accordingly, we proceed to the merits of Perry's facial constitutional challenges. Whether a statute is facially unconstitutional is a question of law that we review de novo.127 We begin by inquiring whether Penal Code section 36.03(a)(1), as it incorporates the relevant portion of the Code's "coercion" definition, reaches a substantial amount of First Amendment-protected activity.128 The "first step" in that analysis "is to construe the challenged statute," as "it is impossible to determine whether a statute reaches too far without first knowing what the statute covers."129

Section 36.03(a)(1) is an expansive criminal prohibition of speech

When construing statutes, we ordinarily focus on the literal text and determine the objective meaning of the words the Legislature used, as "the text of the statute is the law, in the sense that it is the only thing actually adopted by the legislators" and it "is the only definitive evidence of what the legislators ... had in mind when the statute was enacted into law."130 Penal Code section 36.03(a), titled "Coercion of a Public Servant or Voter," states:

The focus of the offense created by section 36.03(a)(1) is thus the use of "coercion" to "influence" (i.e., bring about or cause132), or attempt to so "influence," certain behavior of a "public servant."133 Other Penal Code provisions require that these actions must be committed intentionally or knowingly.134 Consequently, a person violates section 36.03(a)(1) if he employs "coercion" with the conscious objective or desire, or actual awareness of, bringing about, or in an attempt to bring about, the behavior of a "public servant" described in the statute.135

In both the legal arena and ordinary usage, the precise parameters of "coercion" and when it occurs have often been elusive,136 although the concept frequently denotes some sort of impingement on a listener's free will or autonomy through force or threat.137 As "coercion" is used in the Texas Penal Code, however, the Legislature has defined the term more specifically. This definition currently appears in paragraph (9) of the Code's general definitions provision, section 1.07(a):

Thus, the defining feature of "coercion" under the Penal Code is "a threat" to do or perform one or more of the six types of acts specified in paragraphs (A)-(F) of the definition. And, as confirmed by the phrase "however communicated" that immediately follows "threat," the Legislature has limited "coercion" to speech and excluded means of compulsion that might be considered non-communicative in nature, such as bare physical force.139

Further guidance regarding the nature of these "threats" is found in Olivas v. State, where the Court of Criminal Appeals examined "the meaning of the term `threaten' as used in the Penal Code" in order to identify the elements of the Code's assault-by-threat prohibition.140 After observing that "[t]he word `threaten' is not statutorily defined in the Penal Code," the court turned to, and cited approvingly, the following definition of "threaten" from Webster's:

The court similarly quoted a definition of "threat" from Black's Law Dictionary: "A communicated intent to inflict harm or loss on another or another's property."142 Because the court couched its analysis in terms of determining the meaning of "threaten" "as used in the Penal Code," we must presume that it would ascribe a similar meaning to "threat" as used in the Penal Code's "coercion" definition. Accordingly, "threat" as used there would denote a declared or expressed intention "of hurting or punishing" or "to inflict injury," a "menacing indication of ... something dangerous or evil," or "a source of danger [or] harm" that is communicated to another through speech. It also seems implicit in the notion of such a "threat" that the communication would be sufficiently specific to generate expectation that the actor could and would actually carry out particular harm on a particular person, thereby excluding more generalized complaints or advocacy.143

In turn, paragraphs (A)-(F) of the definition specify six categories of threatened harm or loss that would distinguish "coercion" under the Penal Code.144 If the threatened harm or loss would itself constitute a felony offense, the offense created by section 36.03(a)(1) would be a third-degree felony; otherwise, it is a Class A misdemeanor.145 The specific focus of Perry's facial challenges is the portion of paragraph (F) that references "threats ... to take or withhold action as a public servant."146 "Public servant" as used in the Penal Code is defined as:

In short, "public servant" encompasses public officials and employees in every permutation of Texas state or local government, without distinction or limitation regarding branch or department (legislative, executive, judicial), level (county, city, district, etc.), or how the position is selected (elected, appointed, etc.). The definition also extends to certain other persons who provide a specific or limited form of governmental service or function — jurors, grand jurors, attorneys and notaries "participating in performing a governmental function," and arbitrators and other private adjudicators of causes or controversies. Even candidates for public office, others not yet having qualified or assumed their official duties, and persons performing a governmental function under a claim of right come within the definition.

As is implicit in the phrase "threat ... to take or withhold action as a public servant," the focus of paragraph (F), as it relates to Perry's facial challenges, is on threats made by a public servant that he will take or withhold action in his official capacity or under color of his delegated authority.149 We can also discern that paragraph (F) contemplates a taking or withholding of action by a public servant that would "hurt," "punish," inflict injury" upon, amount to "danger" to, or "harm" another person, as this would follow from the definition of "threat" endorsed in Olivas.150

With this understanding of the "coercion" element in mind, we now return to the remaining text of Penal Code section 36.03(a)(1). Combining the two, section 36.03(a)(1), as it operates against threats made by a public servant concerning his own actions, prohibits that public servant from:

• communicating an intent to take or withhold action in his official capacity or under color of his delegated authority that would "hurt," "punish," inflict injury" upon, amount to "danger" to, or "harm" another person ...

• with the conscious objective, desire, or actual awareness of bringing about, or in an attempt to bring about ...

• any of the following actions by another "public servant" (defined the same way as with the actor; i.e., expansively):

The required mens rea element, as previously indicated, relates to the actor's intent or awareness of using the threat to induce, or in an attempt to induce, the desired conduct of the other public servant. Neither section 36.03(a)(1) nor paragraph (F) requires that the actor make any explicit demand or quid-pro-quo offer regarding the desired action on the part of the public servant sought to be influenced; it is enough that a specific threat to take or withhold action was made with the requisite mens rea.151 While proof of an explicit demand for action would certainly be probative of an aim to "influence" the second public servant to those ends, nothing in the statute forecloses proof of the prohibited aim through other circumstances. Further, neither section 36.03(a)(1) nor paragraph (F) requires that the public servant sought to be influenced necessarily be the same person who would be harmed if the threat were carried out.152

As for the action desired of the public servant sought to be "influenced," the exact phrase "influence [a public servant] to violate his known legal duty" appears in section 240.2 of the Model Penal Code, the analog to both section 36.03 and section 36.04 (the improper-influence prohibition) in the Texas Penal Code.153 As the drafters of the model code (the American Law Institute) explained, "influence ... to violate his known legal duty" denotes an attempt to secure action by a public servant that would "violate a specific obligation of his office."154 Such "legal duties," the Institute elaborated, are those "specific[ally] and clearly delineated" by constitutional provisions, statutes, rules, or other legal authorities governing his performance of official functions, in contrast to discretionary decisions or any "general obligation of fidelity to the public interest or other implicit moral obligation."155 Furthermore, according to the Institute, "the requirement of knowledge" of the legal duty applies both to the public servant sought to be influenced and the actor seeking to influence him.156 The import of this language, in other words, is to proscribe threats that are calculated to bring about action by a public servant whom the threatener knows does not have lawful authority or discretion to so act. To this extent, the language has some parallels to section 36.04's improper-influence prohibition, which criminalizes certain private communications made to public servants who exercise official discretion in adjudicatory proceedings (e.g., courts, administrative law judges, or prosecutors) made "with an intent to influence the outcome of the proceedings on the basis of considerations other than those authorized by law."157

Standing in contrast to this portion of section 36.03(a)(1) are the other two types of desired public-servant conduct that the statute addresses, "specific exercise of ... official power" and "specific performance of ... official duty." These components would extend more broadly to capture threats calculated to cause action by a public servant who would possess lawful authority and discretion to so act based on the threat,158 as well as those aimed at public servants whom the threatener does not know to lack such authority. The modifiers "specific exercise of ... official power" and "specific performance of ... official duty" make clear, however, that some particular action or conduct by the public servant must be sought and not merely some effect on the public servant's general comportment in office.159

The range of threats made by public servants that are ultimately criminalized by section 36.03(a)(1) is limited somewhat by a subsection (c) of section 36.03. That provision states:

Negation of this exception has been held to be an element of the offense.161 In other words, in cases where the asserted violation of section 36.03(a)(1) is predicated on a threat by a public servant, the State must prove, in addition to the other elements of the offense, that (1) the public servant who made the threat is not "a member of the governing body of a governmental entity," and (2) the threat was not an "official action" of that member, a phrase that denotes an act performed by the member in his official capacity and under color and by virtue of his office.162

A key effect of subsection (c), as Perry acknowledges and the district court agreed, is to exclude from section 36.03(a)(1)'s proscription threats to take or withhold action made by members of the Legislature (who would be part of the "governing body" of a "governmental entity," the State of Texas) made in the course of their wranglings over legislation and policy. Subsection (c) would have similar application to other elected officials and policymakers who serve on city councils, school boards, and county commissioners' courts, to name but a few examples. In fact, subsection (c) originated in a package of three amendments impacting section 36.03 (which originated in the 1974 Penal Code163) that were made by the Seventy-First (1989) Legislature following controversy regarding the existing statute's potential ramifications for policymakers at the county level.164 The documented legislative history reflects that proponents advocated the amendments "[t]o protect public officials from criminal prosecution arising from the execution of their legal duties," and cited a purported incident where "certain county commissioners" had come under grand jury investigation regarding possible violations of the existing statute "in response to budget cuts" they had imposed.165 The timing of the 1989 amendments also corresponds to the prosecution that gave rise to State v. Hanson.166 Hanson (which we will explore in more detail when analyzing section 36.03(a)(1)'s First Amendment implications) arose from charges brought under the original versions of section 36.03(a)(1) and paragraph (F) against the then-constitutional county judge for Bosque County, Hon. Regina Hanson.167 As the Tenth Court of Appeals summarized the charges, it had been alleged that Judge Hanson had "intentionally or knowingly threatened to terminate the county's funding of the salaries of a deputy district clerk and an assistant district attorney in an attempt to coerce the district judge into firing the county auditor and the county attorney into revoking a misdemeanant's probation."168

Another of the 1989 amendments added the phrase "influences or attempts to influence a public servant to violate the public servant's known legal duty,"169 thereby specifying that section 36.03(a)(1)'s prohibition extends to threats aimed at causing action by another public servant whom the threatener knows does not have lawful authority or discretion to so act. The final change consisted of an amendment to the "coercion" definition (which was then located in Penal Code section 36.01(1)) to require in paragraph (F) that the threatened taking or withholding of action by a public servant must be "unlawful,"170 a term that the Penal Code defines as conduct that would be either criminal or tortious absent a defense amounting to justification or privilege.171 However, the Legislature removed the "unlawfully" qualifier from paragraph (F) in 1993, amid sweeping changes to the Penal Code in which the Legislature also moved the "coercion" definition to its current location in section 1.07.172 The remainder of section 36.03(a)(1) and the "coercion" definition — including the other two 1989 amendments — has survived without substantive change to this day.

In addition to excluding a category of public-servant threats from section 36.03(a)(1)'s prohibition, subsection (c) also provides further illumination regarding the nature of threats that remain covered. By requiring that an excepted threat be an "official action" of a "governing board" member, subsection (c) implies that section 36.03(a)(1)'s proscription reaches a range of threats that would not be considered official actions of a public servant, whether governing board members or other kinds. This category, as it relates to threats covered by paragraph (F) of the "coercion" definition, would seem to contemplate exceptional instances in which a public servant has utilized threats to take or withhold official action to ends wholly beyond the servant's delegated powers, such as when used to pursue some sort of entirely personal benefit or interest.173 The same feature of subsection (c) also confirms that 36.03(a)(1)'s proscription would reach a range of threats that would be considered official actions of the public servants who are not governing board members. In fact, as the State seems to recognize, this would be the principal effect of the statute.174

If the sparse volume of appellate opinions addressing section 36.03 to date is any indication, the statute has rarely been utilized. However, the evident purpose of criminal statutes of this sort — like the improper-influence statute, the prohibition against bribery of public servants in section 36.02,175 and other provisions within chapter 36 of the Penal Code — is "to reach various means by which the integrity of government can be undermined."176 In that regard, prohibitions like that in section 36.03(a)(1), at least in theory, serve as a "complement" to the bribery offense, in that bribery focuses on the corruptive influence of offering or conferring a "benefit" as consideration for a public servant's actions, whereas section 36.03(a)(1) combats the use of a negative incentive — threats of certain kinds of harm — to induce the public servant's actions.177 The definition of "coercion" incorporated into section 36.03 also has some similarities to the types of threats that would be criminalized under the Model Penal Code through a "residual," stand-alone "coercion" offense when used "with purpose unlawfully to restrict another's freedom of action to his detriment" and not done with "benign purpose."178

But translating such concepts into specific statutes presents numerous line-drawing challenges in regard to a wide range of communications made to, by, or between public servants that can be fairly characterized as "threats," yet are considered to be a commonplace and accepted — if not also constitutionally protected — component of the day-to-day debate, bargaining, and give-and-take that is characteristic of American governmental and political processes.179 In this regard, section 36.03 (aside from subsection (c)), as it incorporates paragraph (F), is most notable in the lines it does not attempt to draw, as revealed by our review of the statutory text:

Absent such limitations, the net result is that section 36.03(a)(1), as it incorporates paragraph (F), would criminalize:

These sorts of implications are the focus of Perry's facial constitutional challenges. In these ways, he urges, section 36.03(a)(1), as it incorporates paragraph (F), criminalizes a "virtually endless" array of threats that are in the nature of "ordinary give and take" between and among the public servants who are not excepted by subsection (c). He posits hypothetical examples that include:

At least to the extent each hypothetical threat would be aimed at bringing about some specific action on the part of another public servant, we agree that Perry's illustrations would indeed come within section 36.03(a)(1)'s proscription, as written. Each example would, in terms of the statutory language, represent a "public servant" (under the Penal Code's broad definition) seeking "by means of coercion" (defined, again, as a "threat" to "take or withhold action as a public servant" that, while "harmful" or detrimental to another, need not be unlawful) to "influence[] or attempt[] to influence" (i.e., bring about) some "specific exercise" or "specific performance" of another public servant's duties or powers (which also need not be unlawful, nor beyond the first public servant's powers to demand or require).

And it is not difficult to think of other similar examples. Among them, it has occurred to the members of this panel that unless appellate court justices can shoehorn themselves into subsection (c)'s exception, section 36.03(a)(1) would seemingly put at risk that time-honored practice whereby one justice will seek changes to another justice's draft majority opinion by threatening to write a dissent exposing flaws in the other's legal reasoning.189 Similar concerns would arise regarding the standard letter our Clerk issues to prompt action whenever briefs or records are late, in which he warns of imminent dismissal if the deficiency is not promptly rectified, at least when the party in interest is a government entity or official.

One's view of section 36.03(a)(1) is further informed by considering its implications as it incorporates a portion of paragraph (F) that Perry has not challenged in this case — a threat "to cause a public servant to take or withhold action."190 Within the literal scope of this prohibition would be ordinary citizens who make credible threats to obtain legislative or judicial remedies as a means of prompting some action by a local government official,191 not to mention members of the media who might have to resort to similar means of persuasion in order to overcome governmental foot-dragging on Public Information Act requests. We mention this not to plant ideas in the minds of local prosecutors across Texas who may be inclined to protect their fellow public servants, but to provide stark demonstration that section 36.03(a)(1) and the "coercion" definition it incorporates were not crafted with the sort of precision necessary to avoid First Amendment problems when, as with these statutes, government purports to criminalize speech.

In rejecting Perry's overbreadth challenge, the district court evidently perceived that section 36.03(a)(1) and paragraph (F) are much more limited in scope than we have concluded they are. And it is this vast breadth that causes section 36.03(a)(1), as it incorporates paragraph (F), to impinge upon substantial amounts of First Amendment-protected speech among Texas's public servants.

Section 36.03(a)(1), as it incorporates paragraph (F), criminalizes First Amendment-protected speech

There is little room for disagreement that section 36.03(a)(1) and paragraph (F) criminalize speech — that much is apparent in the phrase "a threat, however communicated ..." that is a defining characteristic of all types of "coercion" prohibited by section 36.03(a)(1).192 But it is a somewhat closer question whether this speech is a type that the First Amendment actually protects. In contending that the proscription reaches protected speech (and a substantial amount at that), Perry emphasizes the First Amendment's central concern with protecting and fostering discussion and advocacy concerning public issues and political change193 and what the Supreme Court has termed the provision's embodiment of a "profound national commitment" to the ideal that "debate on public issues should be uninhibited, robust, and wide-open."194 This is not simply liberty to engage in "abstract discussion" or "merely to describe facts," as the Supreme Court has emphasized, but the liberty "to persuade to action," as "[t]he First Amendment is a charter for government, not for an institution of learning."195 The court has also said that the First Amendment's protections are at their "zenith" for such "core political speech,"196 and Perry insists that this is precisely the sort of speech section 36.03(a)(1) and paragraph (F) reach — speech by public servants regarding their official actions that is aimed at prompting other public servants to undertake official action.

Perry similarly emphasizes the Supreme Court's longstanding recognition that discussion, debate, and advocacy regarding public issues "is often vituperative, abusive, and inexact," yet is protected by the First Amendment nonetheless.197 In Hanson, the then-Chief Justice of the Tenth Court of Appeals, the late Bob Thomas, made a similar observation regarding speech to and in Texas government at its local levels:

The State counters that section 36.03(a)(1) and paragraph (F) do not implicate any First Amendment interests due to the identity of the speakers who are targeted — public servants — and the content and effect of the speech that is prohibited — what the State characterizes as "coercive" or "extortionate" threats.

Contrary to the State's arguments, public servants have First Amendment rights

The State insists that "[s]tatements made by public officials to other public officials are unprotected" under the First Amendment, at least when uttered as part of an official's job. The State derives this proposition from its understanding of the United States Supreme Court's holdings in Garcetti v. Ceballos.199 Garcetti involved a section 1983 claim brought by a deputy district attorney (Ceballos) against his governmental employer to obtain relief based on alleged violations of Ceballos's First Amendment rights.200 The asserted violations consisted of a job reassignment, transfer, and promotion denial that Ceballos perceived to be in retaliation for a memorandum that he had written — undisputedly as part of his job duties201 — concerning a pending case.202 The Supreme Court framed the dispositive issue as "whether the First Amendment protects a government employee from discipline based on speech made pursuant to the employee's official duties."203 After analysis, the court concluded that "when public employees make statements pursuant to their official duties, the employees are not speaking as citizens for First Amendment purposes, and the Constitution does not insulate their communication from employer discipline."204 And because "the First Amendment does not prohibit managerial discipline based on an employee's expressions made pursuant to official responsibilities," the court held that Ceballos's claim of unconstitutional retaliation "must fail."205

The State reads Garcetti to mean that statements made by public employees while performing their official duties enjoy no First Amendment protection not only as against employer discipline — the issue in Garcetti206 — but for all other purposes, even as against criminal penalties that might be imposed based on that official speech. In that regard, the State places great emphasis on the Garcetti court's phrasing of its holding in terms of public employees "not speaking as citizens for First Amendment purposes" when making statements pursuant to their official duties, while diminishing the significance of the clause that immediately follows, "... and the Constitution does not insulate their communications from employer discipline."207 Perry disputes that Garcetti has any bearing in the present context of a criminal prosecution based on speech, as opposed to a claim or issue involving employee discipline. We agree.

While Garcetti has admittedly generated occasional confusion of this sort among lower courts,208 a close reading of that opinion, especially as informed by antecedent precedents on which the Supreme Court relied in that case, belies the State's expansive reading of it. Garcetti is rooted firmly in a series of cases (sometimes termed the Pickering209-Connick210-Garcetti line of cases) that derives from a long-recognized principle that "a State cannot condition public employment on a basis that infringes the employee's constitutionally protected interest in freedom of expression."211 Pickering and its progeny represent a particularized application of that principle to the termination, discipline, or other adverse actions imposed by governmental employers against their employees based on the employees' expression.212 Accordingly, these cases have consistently emphasized the unique nature of government's relationship to its citizens in that context — government is acting in the role and capacity of employer, as opposed to its usual or general role "as sovereign"213 — and have, correspondingly, consistently involved government's regulation or punishment of speech within the context of that employment relationship, such as through termination or other adverse personnel actions, as opposed to the imposition of criminal penalties or other remedies that government administers in its sovereign capacity.214 In fact, Pickering contrasted the governmental personnel action at issue there (a school board's dismissal of a teacher in response to a letter to the editor that had been critical of the board) with "criminal sanctions and damage awards."215

A central premise of the Pickering-Connick-Garcetti line is that when government acts "as employer," it "has far broader powers than does the government as sovereign" to regulate speech.216 The Supreme Court deems this distinction justified because "[g]overnmental employers, like private employers, need a significant degree of control over their employees' words and actions; [as] without it, there would be little chance for the efficient provision of public services."217 Similarly, the court has voiced a reluctance to "`constitutionalize the employee grievance.'"218 But while a citizen entering government service must thereby "accept certain limitations on his or her freedom,"219 the Supreme Court has also emphasized that "a citizen who works for the government is nonetheless a citizen"220 and "do[es] not surrender all [his] First Amendment rights by reason of [his] employment."221 The point of the Pickering-Connick-Garcetti line has been to delineate the parameters of the First Amendment rights that public employees retain against expression-based personnel actions, and the basic rule is this: a public employee may "speak as a citizen addressing matters of public concern," subject to the governmental employer's power to impose restrictions when there is "adequate justification for treating the employee differently from any other member of the general public."222 Pickering, Connick, and Garcetti each illustrates different points along that continuum — Pickering is an example of a public employee's speech "as a citizen" (a teacher writing a letter to the editor outside of his job duties) addressing matters of "public concern" (school fiscal matters) and where no adequate justification was shown for employee discipline on that basis,223 while Connick and Garcetti are on the opposite side of that line, with the former concerning discharge based on a public employee's speech deemed not to be of "public concern" (various internal office grievances)224 and the latter involving personnel actions allegedly based on a public employee's speech deemed not to have been uttered "as a citizen" (a memo written as part of his job duties).225 None of these cases addresses or would logically extend to the situation presented by the present case — the potential imposition of criminal penalties, as opposed to employee discipline, by the State of Texas acting in its capacity as sovereign rather than as an employer. If the Pickering-Connick-Garcetti line has any bearing here, it would only be to emphasize a baseline rule that public employees possess First Amendment rights just like other citizens do.226

In addition to its misplaced reliance on Garcetti, the State insists more broadly that public servants have no First Amendment protection against criminal punishment for their official speech because, in essence, it is not actually their speech, but that of their governmental employers. The State attempts to tie this notion to the Supreme Court's "governmental speech" jurisprudence, which is to the effect that the First Amendment generally permits the government to communicate its own viewpoints through, e.g., the types of specialty license plates the government decides to issue, the monuments it allows to be placed in public parks, or its promotion of beef consumption, without having a duty to subsidize or promote other viewpoints.227 In that sense and context, the Supreme Court has said that "[t]he Government's own speech ... is exempt from First Amendment scrutiny."228 Seizing on this language, the State reasons that section 36.03(a)(1) implicates no First Amendment interests of public servants, at least with regard to speech made in their official capacities, because it represents Texas's regulation of its "own speech and actions" to advance a governmental "viewpoint" that "official speech should not be coerced speech." And because this State-desired end product of non-coerced official speech is "governmental" speech, the State adds, the Legislature may likewise burden any private speech that might compete with that "viewpoint."

As an initial observation, if, as the State argues, public servants' official speech equals "governmental speech" in the sense of that doctrine, that theory would seem to suggest that the State itself would be criminally responsible for any violations of section 36.03(a)(1) through that official speech.229 And if, as the State suggests, speech inimical to the government's official "viewpoint" can be criminalized for that reason alone, there would seem to be little left of the First Amendment's rights of speech and petition. But more to the point, the Supreme Court has never held (nor has any other court, as far as we can tell) that the "governmental speech" doctrine means that governments can freely criminalize the speech of their citizens (whether spoken in their official or individual capacities) free of any First Amendment constraints. Rather, it is fundamental that the First Amendment limits the sovereign power of government to criminally punish the speech of its officials and agents, including that uttered in their official roles.230 Also belying the State's notion that public servants can be freely punished criminally for their speech in their official capacities, the Supreme Court has emphasized that the First Amendment's application does not turn solely on the identity of the speaker or source of the expression at issue.231

In short, contrary to the State's arguments, "[s]tatements made by public officials to other public officials" on public issues are generally protected by the First Amendment — and neither the "governmental speech" doctrine nor Garcetti holds otherwise. Consequently, if the State is correct that section 36.03(a)(1), as it incorporates paragraph (F), does not impinge upon First Amendment-protected speech, that must be so because of the nature of the speech the statute proscribes.

Nor does any historical exception to the First Amendment's protections apply

The State's arguments concerning the content or effect of the threats proscribed by section 36.03(a)(1) and paragraph (F) implicate the Supreme Court's jurisprudence delineating certain types of speech for which even content-based restrictions have not been considered "to raise any Constitutional problem."232 In recent years, the Supreme Court has rejected any "free-floating test for First Amendment coverage ... [based on] an ad hoc balancing of relative social costs and benefits"233 in favor of a categorical approach that looks instead to certain "historic and traditional categories long familiar to the bar."234 These categories, as identified by the court to date, include the likes of obscenity, defamation, and so-called "fighting words."235 Another refers specifically to a type of threat — what the Supreme Court has termed "true threats."236 But "true threats," according to the court, are only "those statements where the speaker means to communicate a serious expression of intent to commit an act of unlawful violence to a particular individual or group of individuals."237

Although the few reported appellate cases arising from prosecutions under Penal Code section 36.03(a)(1) have tended to involve "true threats,"238 such threats of violence would comprise a relatively small fraction of section 36.03(a)(1)'s potential applications, especially with respect to the threats by public servants to take or withhold official action that would be within the coverage of paragraph (F). Further, as Perry emphasizes, the Supreme Court has never seen fit to extend this unprotected category of threats beyond those parameters, and it has also demarked the category's boundaries strictly in relation to political speech by holding that a wide range of "intemperate" and "vituperative" expression related to public issues — even some referencing or advocating violence — falls short of "true threats" and is protected by the First Amendment.239

Aside from these limitations on proscribable "true threats," Perry points out Supreme Court holdings to the effect that the First Amendment's protections are not rendered inapplicable merely because political speech can be said to be "coercive" in nature. He refers us to Organization for a Better Austin v. Keefe,240 in which the Supreme Court overturned, as an invalid prior restraint, an injunction restraining a pamphleteering campaign by the organization that had been aimed at pressuring a real estate broker to sign an agreement not to engage in certain business practices the group perceived to be detrimental to the community.241 In so doing, the Supreme Court rejected a lower court's view that the campaign was unprotected by the First Amendment because the effort's aim had been to "force" the broker to sign the desired agreement rather than merely informing the public.242 The Supreme Court reasoned:

More recently, in N.A.A.C.P. v. Claiborne Hardware Company,244 on which Perry also relies, the Supreme Court cited Keefe for the proposition that "`offensive' and `coercive' speech was nevertheless protected by the First Amendment."245 The court there also applied similar reasoning in holding that the First Amendment had protected various peaceful measures by which participants in a business boycott, aimed at achieving political and social goals related to racial equality and integration, had endeavored to compel participation through both advocacy and social pressure. While acknowledging that the boycott advocates had "admittedly sought to persuade others through social pressure and the `threat' of social ostracism," the Supreme Court emphasized that "[s]peech does not lose its protected character ... simply because it may embarrass others or coerce them into action."246

The State does not respond directly to Perry's arguments regarding the narrowness of proscribable "true threats" and the implications of Claiborne Hardware and Keefe. Instead, it advances arguments that, while not framed explicitly in these terms, potentially implicate a different "historic and traditional" category of proscribable speech — what the Supreme Court has termed "speech integral to criminal conduct."247 The basic rationale underlying the court's recognition of this historical category of unprotected speech is that the First Amendment does not prohibit government from "mak[ing] a course of conduct illegal merely because the conduct was in part initiated, evidenced, or carried out by means of language, either spoken, written, or printed."248 Within this category, the Supreme Court has placed such activities as picketing having the sole objective of inducing a business to engage in an illegal restraint of trade,249 speech soliciting illegal transactions or other crimes,250 or speech that discloses our Nation's defense secrets in furtherance of treason.251

Applying a similar rationale, a number of courts have held or otherwise expressed the view that speech incident to bribery or extortion is not protected by the First Amendment — and these have included the Court of Criminal Appeals. In Sanchez v. State, the Court of Criminal Appeals addressed overbreadth challenges to the sexual-harassment provision of the Penal Code's official-oppression statute, an offense that it summarized as having the elements of: (1) "an official's use of his official position"; (2) "to intentionally engage in conduct of a sexual nature or attempt to procure sexual favors"; (3) "which he knows is not welcomed by the recipient"; (4) "intending submission to which to be a term or condition of the recipient's or another person's enjoyment of something of value to the person."252 The court anchored its analysis in the premise, derived from some federal intermediate appellate decisions, that "[b]ribery and extortion, while involving `speech,' are not protected by the First Amendment."253 Quoting language from the federal decisions, the Court of Criminal Appeals observed that "[t]hreats and bribes are not protected simply because they are written or spoken; extortion is a crime although it is verbal,"254 and that "`[i]t may categorically be stated that extortionate speech has no more constitutional protection than that uttered by a robber while ordering his victim to hand over the money, which is no protection at all.'"255 To resolve the overbreadth challenge presented in Sanchez, the Court of Criminal Appeals took this analysis a step further by comparing the sexual-harassment offense at issue to extortion as that crime had been known both historically and in modern times. The court observed that the crime of "extortion" had originated under English common law as a proscription against a public official's taking of "money not due him for the performance of official duties," which the court termed "essentially a form of bribery."256 Over time, the court continued, extortion had evolved by statute to encompass the taking of property by private individuals "through force, fear, or threats" — essentially a species of theft.257 Today there is no "extortion" offense by that name in the Texas Penal Code, but vestiges survive within the bribery offense under section 36.02258 and the theft offense created by section 31.03. The theft offense — which was expressly intended to subsume "extortion" and various other theft-related crimes previously known to the law259 — in part proscribes the appropriation of another's property through consent obtained by "coercion," thereby incorporating the same "coercion" element and definition as does section 36.03(a)(1).260

In the view of the Court of Criminal Appeals, the sexual-harassment offense at issue in Sanchez had several material similarities to extortion in both its historical and modern forms, so it followed that any speech the offense proscribed lacked First Amendment protection:

Subsequently, relying on Sanchez, the Fourth Court of Appeals summarily rejected an overbreadth challenge to section 31.03's theft-by-coercion offense in Roberts v. State.262

The gist of the State's position is that we should, similar to the courts in Sanchez and Roberts, regard any threats proscribed by section 36.03(a)(1) and paragraph (F) to be tantamount or analogous to extortion or bribery and, for that reason, unprotected by the First Amendment. To similar effect, the State urges that the statutes reach only "conduct," and cites a Houston Court of Appeals case that rejected an overbreadth challenge to Penal Code section 38.15 (Interference with the Duties of a Public Servant), a statute that explicitly provides a defense where the acts in question "consist[] of speech only."263

In reply, Perry does not dispute that speech in furtherance of extortion or bribery would lack First Amendment protection, but he rejects the State's attempts to equate those crimes or the sexual-harassment offense addressed in Sanchez with section 36.03(a)(1), at least as it incorporates paragraph (F). At most, Perry suggests, any such comparisons could establish only that section 36.03(a)(1), as it incorporates paragraph (F), would have some constitutionally permissible applications, such as if a public servant threatened to take or withhold action as a means of obtaining some sort of personal benefit that would violate the theft-by-coercion or bribery prohibitions. But section 36.03(a)(1) and paragraph (F) extend far beyond this, Perry urges, to impinge First Amendment-protected speech. We ultimately agree with Perry. The threats proscribed by section 36.03(a)(1), as it incorporates paragraph (F), are not limited to those that would fall within any recognized category of speech integral to criminal conduct, like extortion, bribery, or (to follow the Court of Criminal Appeals's lead in Sanchez) any analogues or equivalents to those crimes.

To be sure, section 36.03(a)(1), as it incorporates paragraph (F), would reach some threats aimed at achieving illicit ends that the State could properly proscribe criminally. A public servant's threat to take or withhold action in order to induce another public servant to knowingly violate his or her legal duties would resemble (and in some instances could include) solicitation of a crime, a recognized category of unprotected speech.264 Threats seeking some sort of prohibited personal benefit, a la extortion or bribery, would be another example. But the extortion or bribery analogy does not extend much further — as the Court of Criminal Appeals observed in Sanchez, the crimes have long focused on the use of threats or official power to extract some form of personal benefit to which the actor would otherwise not be entitled.265 About the most that could be extrapolated from the extortion crime is that the State can validly prohibit the use of threats in situations where the parties' relationship does not provide the threatener a legitimate basis for seeking conduct of the recipient (i.e., he had no right to demand or require it) and where the recipient would have no underlying obligation to consider performing the conduct aside from the threat.266 Such use of threats might also be said to seek or effect a sort of transaction in which a public servant interjects a threat to take or withhold official action, then offers freedom from the threatened harm solely to exchange for some otherwise-unrelated desired conduct by another public servant whose sole rationale or justification would be avoidance of the threatened harm. In this respect, such threats would achieve an effect similar to bribery of the recipient.

Perry also grants that if a public servant threatened to take or withhold action in a manner that would in itself be unlawful, that threat would "generally not [be] protected," even if not foretelling violence against persons.267 Such threats would present the victim the choice of engaging in the sought-after conduct in order to regain a right that had belonged to him or her in the first instance — undertaking official action in order to regain freedom from the breach of criminal law or tort duties being threatened. The same basic dynamic is also a feature of robbery — the victim must hand over property in order to regain his or her preexisting right to personal safety — and extortion where the threat foretells unlawful action.

But these analogies do not fit the class of potential applications of section 36.03(1) and paragraph (F) that Perry emphasizes. Where government managers threaten to demote poor-performing employees unless they improve, public defenders threaten to file suppression motions in order to secure better plea deals, or an appellate justice threatens to write a dissent in order to prompt changes in a draft majority opinion, the threatened taking or withholding of action would generally be lawful; consequently, the recipient would have no right to be free of the threatened action in the first instance. The sought-after action would likewise be lawful — the threat would not be akin to solicitation of a crime. And in these sorts of instances, the desired action would generally be within the actor's lawful rights to demand or require of the recipient, and the threat in itself would typically have no independent impact on the recipient's ultimate rights or options relative to the public servant seeking the action. Whatever leverage that was possessed by the public servant making the threat, in other words, would derive not from the threat in itself, but from the underlying legal relationship and relative distribution of powers between public servants that the People have chosen through their Constitution and laws. The threat, as such, would amount merely to an articulated expression of that relative distribution of powers, in the nature of a warning of consequences that the public servant could lawfully impose in response to the recipient's failure to comply with a lawful request or demand.268

These kinds of threats are "speech incident to criminal conduct" only if the basic workings of government are considered criminal conduct, a proposition we cannot sanction. Nor, absent the sorts of unlawful means or ends we have described, would any characterization of these threats as "coercive" suffice to displace them from the First Amendment's protections, as the State would have us conclude; on the contrary, the Supreme Court has instructed in Claiborne Hardware and Keefe that "`offensive' and `coercive' speech [is] nevertheless protected by the First Amendment."269 These threats remain protected instead — indeed, as Perry urges, "[s]tatements so intrinsic to government, particularly when they relate to matters of public policy, lie at the core of First Amendment protection." In the very least, by criminalizing these types of threats, section 36.03(a)(1), as it incorporates paragraph (F), plainly impinges upon the First Amendment-protected speech of Texas's public servants.270

Our holding is hardly a novel one — Hanson reached essentially the same conclusions regarding section 36.03(a)(1) and paragraph (F) a quarter century ago. In that case, the Tenth Court had little difficulty determining that "[w]hen applied to Judge Hanson's alleged conduct" ("threaten[ing] to terminate the county's funding of the salaries of a deputy district clerk and an assistant district attorney in an attempt to coerce the district judge into firing the county auditor and the county attorney into revoking a misdemeanant's probation"), section 36.03(a)(1), as it incorporated paragraph (F), "clearly implicated [Hanson's] First Amendment rights because she was indicted for words allegedly spoken to other public officials" — she was charged with "having made a threat to take action as a public official in an attempt to coerce another public official into performing an official act."271 The court elaborated that "[t]hreats may portend either lawful or unlawful action" and that "First Amendment protection is extended to the former but not the latter."272 And it was likewise true, the court continued, that "[c]oercion of a lawful act by a threat of lawful action is protected free expression."273 In this regard, the court observed, with reference to the statutory powers of all three offices, and assuming the allegations in the charging instruments were true, "Judge Hanson could have lawfully taken the actions threatened, and the district judge and county attorney, had they acted as she desired, would have acted lawfully."274 The State suggests that Hanson also involved a third variable, although one not obviously emphasized by the Tenth Court — as the State explains it, Judge Hanson "could legally request that the county attorney institute proceedings to revoke a misdemeanant's probation [such that] the act requested by the defendant as part of her threat was an act that was within the scope of the lawful authority that she, the defendant, otherwise had." Section 36.03(a)(1), as its incorporates paragraph (F), reaches threats having all of these features.

The Tenth Court addressed these First Amendment concerns under the rubric of vagueness, concluding that the statutes failed to afford Judge Hanson sufficient notice as to whether "the term `threat' encompass[ed] a threat of lawful action or only prohibit[ed] a threat of unlawful action"275 and whether Hanson could seek to "coerce" lawful actions of the other "public servants."276 But the foundation for these holdings was the same conclusion we reach here — section 36.03(a)(1), as it incorporates paragraph (F), impinges upon speech that the First Amendment protects.

Our reasoning, and that of Hanson, is also consistent with decisions from other jurisdictions that have recognized threats to be protected by the First Amendment absent some applicable exclusion. Some of these courts have invalidated criminal prohibitions against threats on overbreadth grounds.277 Still others have resorted to narrowing constructions of threat prohibitions in order to avoid these sorts of free-speech problems.278

Section 36.03(a)(1)'s impingement upon First Amendment-protected speech is not justified

Having concluded that section 36.03(a)(1), as it incorporates paragraph (F), reaches into the realm of First Amendment-protected speech, we must next consider whether that impingement can be justified. Under the Supreme Court's most recent guidance, it now seems clear that section 36.03(a)(1) must be classified as a "content-based" speech prohibition — on its face, it targets speech having a specified content (communicated intent to inflict one of the six categories of harms enumerated in the "coercion" definition) based on the communicative impact the threat has or is intended to have (bringing about certain specified conduct of a public servant), and we are instructed that these facial features of the statute, not any asserted underlying content-neutral rationales, control.279 Where, as here, a statute "seeks to restrict and punish speech based on its content," "the usual presumption of constitutionality is reversed" — the statute is instead presumed invalid, and "the government bears the burden to rebut that presumption."280 In that event, courts apply "strict scrutiny," which requires that a law regulating speech be (1) necessary to serve (2) a compelling state interest and (3) be narrowly drawn (i.e., that it employ the least restrictive means to achieve its goal and there be a close nexus between the government's compelling interest and the restriction).281

The State posits that section 36.03(a)(1), as it incorporates paragraph (F), is justified by a "singular interest in intervening when public officials try to leverage the powers of government for their personal or political whims." It also refers to perceived interests in protecting the "integrity" of Texas's government institutions (similar to the Court of Criminal Appeals's observations in Isassi282) or protecting public servants from "interference" in the performance of their duties. The State also refers us to various court of appeals cases that have referenced interests in protecting participants in the judicial system — such as judges, jurors, or witnesses — against threats of physical violence or threats of exposure to "hatred, contempt, and ridicule" in the course of rejecting overbreadth challenges to section 36.03(a)(1),283 section 36.05's witness-tampering prohibition,284 or section 36.06's retaliation provision.285

Assuming without deciding that each of these asserted interests would rise to the level of "compelling," it cannot be said that section 36.03(a)(1), as it incorporates paragraph (F), is "narrowly drawn" to achieve those objectives. Indeed, it begins to strain even rational-basis scrutiny to conclude that a statute that goes as far as to criminalize a public servant's threat merely to lawfully exercise his delegated powers as a related consequence of another public servant's failure or refusal to comply with a lawful demand for action would protect the "integrity" of government or prevent "interference" with public servants' performance of their delegated functions. If anything, the effect of the statutes in those circumstances would be to undermine these interests by criminalizing much of the ordinary day-to-day workings of government. And to the extent the unique circumstances of the judicial process might justify the proscription as it would apply in that context, we need only observe that the proscription extends far more broadly.286

Perry further questions the extent to which section 36.03(a)(1) and paragraph (F) are "necessary" to effectuate these professed state interests. He points out, for example, that other subsections of the Penal Code's "coercion" definition address threats of violence or other unlawful conduct and that other Code provisions — e.g., the bribery statute, the theft-by-coercion proscription, or the official-oppression prohibition287 — address more directly any instances of public servants wielding their power merely to advance "personal whims." "The only discrete function of [paragraph (F)], when combined with Section 36.03(a)(1)," in Perry's view, "is to prohibit speech protected by the First Amendment." In this regard, Perry draws comparisons to the Court of Criminal Appeals's analysis in Ex parte Lo, in which the court struck down, on overbreadth grounds (and via pretrial habeas), subsection (b) of Penal Code section 33.021, which had prohibited a person from communicating online in a "sexually explicit" manner (defined so broadly as to encompass "many modern movies, television shows, and `young adult' books") with a minor if the person intended to arouse or gratify anyone's sexual desire.288 While acknowledging that the State had a compelling interest — "indeed a solemn duty" — to protect minors from sexual predation, the high court concluded that the statute was not narrowly drawn where "everything that Section 33.021(b) prohibits and punishes is speech and is either already prohibited by other statutes (such as obscenity, distributing harmful materials to minors, or child pornography) or is constitutionally protected."289 We agree that while some applications of section 36.03(a)(1) and paragraph (F) may be permissible or justified under the First Amendment, Lo's rationale would be an additional reason for invalidating the statute as to other applications.

There is no alternative to invalidation

It remains to be addressed whether the potential unconstitutional applications of section 36.03(a)(1), as it incorporates paragraph (F), are so substantial relative to permissible ones that we should apply the "strong medicine" of invalidating the proscription for overbreadth. We conclude that the proscription's overreach into First Amendment-protected territory is to an extent and nature that cannot merely be left to remedy through future case-by-case adjudication. The proscription, as we have seen, can be fairly said to be of "alarming" breadth, reaching even a public servant's declared intention to take or withhold action lawfully, aimed at bringing about another public servant's lawful action that the first public servant could lawfully demand or require, and even where is a close relationship between the foretold action and the desired action.290 Furthermore, there is no requirement that the actor articulate a specific demand for action or quid-pro-quo offer to accompany the threat — it is enough that the public servant makes the threat and a fact-finder might infer the prohibited intent to bring about some sort of official action by the recipient. The statutes would criminalize many utterances that are ordinary and commonplace in the day-to-day workings of government — and that are protected by the First Amendment.

The State downplays the possibility that any perceived overbreadth of section 36.03(a)(1) and paragraph (F) gives rise to any chilling effect on the speech of Texas's public servants, or at least has until now. It emphasizes that the statutes have been on the books "for over four decades" and that there has not been (aside from this case) any evident outcry or concern regarding their potential reach. We must acknowledge being unaware of any widespread incidence of prosecutors pursuing "coercion of a public servant" charges for the likes of government managers who threaten to demote poor-performing employees or public defenders who threaten to file suppression motions, and it seems that the longstanding practice of appellate judges threatening to write dissents to prompt changes in majority opinions remains as vibrant as ever. But then again, section 36.03(a)(1) has apparently been rarely utilized — at least until now. Perhaps more importantly, for the past quarter century one of those few appellate decisions has been Hanson, and its recognition of First Amendment constraints on the enforcement of section 36.03(a)(1) and paragraph (F) would have had its own chilling effect on prosecutorial zeal. Nor have Hanson's limits been seriously questioned since — at least until now.

Nor is it any answer to overbreadth concerns to say that prosecutors have heretofore not exercised their discretion to pursue charges under section 36.03(a)(1) and paragraph (F) as aggressively as they might have, or that we should assume they will display similar reticence in the future. The "First Amendment protects against the Government; it does not leave us at the mercy of noblesse oblige."291 "We would not uphold an unconstitutional statute merely because the Government promised to use it responsibly."292

But before proceeding to invalidate section 36.03(a)(1) as it incorporates paragraph (F), we must consider whether we can stop short of that remedy by adopting some sort of reasonable narrowing construction that would cure the proscription's overreach into First Amendment-protected speech.293 Yet as judges, our tools are limited in this regard — we "may not rewrite a statute that is not readily susceptible to a narrowing construction because such a rewriting constitutes a serious invasion of the legislative domain and would sharply diminish the legislature's incentive to draft a narrowly tailored statute in the first place."294 And short of engaging in this sort of prohibited statutory rewriting, a viable narrowing construction of either section 36.03(a)(1) or paragraph (F) eludes us.

One potential option, followed by courts in some other states when addressing similar First Amendment problems with criminal proscriptions against threats, has been to construe the prohibition so as to apply only to threats of unlawful action.295 While these courts have sometimes relied on textual features not found in the Texas statutes, at least one court relied on an alternative definition of "threat" that appears in Black's Law Dictionary — "[a] declaration of an intention to injure another or his property by some unlawful act."296 But this definition of "threat" was not among those cited with approval by the Court of Criminal Appeals in Olivas — while referring to declared intention to "hurt," "punish," "inflict injury" upon, amount to "danger" to, or "harm" another person, the definitions cited in Olivas did not include the additional specification that such harm or injury rise to the level of unlawful harm.297 Moreover, we cannot conclude, in any event, that the Legislature intended to incorporate an "unlawful" harm limitation into paragraph (F) or section 36.03(a)(1) where it has conspicuously omitted any mention of it in either provision, in contrast to other Penal Code provisions where it has explicitly included such a limitation.298 Further, if we look to legislative history, we discover that the Legislature even specifically removed such a limitation from paragraph (F) in 1993 after having added it in 1989.299 Against this backdrop, we can only conclude that if the Legislature desired to limit section 36.03(a)(1) or paragraph (F) solely to threats by public servants to take or withhold action unlawfully, the statutes would read that way today.

Similar textual constraints preclude us from construing section 36.03(a)(1) or paragraph (F) so as to apply only to threats akin to criminal solicitation — again, section 36.03(a)(1) explicitly prohibits not only threats aimed at bringing about a "violat[ion of] the public servant's known legal duty," but also those seeking "a specific exercise of his official power" or "a specific performance of his official duty" that would be entirely lawful.300 Likewise, we cannot, without baldly inserting new language into the statutes that the Legislature has not seen fit to include, construe section 36.03(a)(1) or paragraph (F) so as to proscribe only threats seeking conduct that the actor could not lawfully demand or require or that lack any nexus to such conduct. Had the Legislature intended such a limitation, section 36.03(a)(1) would have prohibited, e.g., "wrongfully influenc[ing] or attempt[ing] to influence a public servant" by means of "coercion"/threat. However, we acknowledge that courts in some other states have adopted this sort of fix.301

The State suggests no good alternatives, either. While condemning Perry's hypothetical applications of section 36.03(a)(1) and paragraph (F) as "alarmist," the State does not assert either that the statutes are ambiguous or that the Legislature could not possibly have intended them to reach as broadly as they do.302 Nor could it do so persuasively;303 indeed, as previously explained, other states had also enacted similarly broad threat prohibitions with similar implications and outcomes.304 And even if we were to look to legislative history in lieu of the text,305 it tends only to confirm the Legislature's desire that section 36.03(a)(1) and paragraph (F) apply expansively. In addition to the Legislature's 1993 elimination of the "unlawfully" limitation it had added to paragraph (F) in 1989, in the same session it rejected a proposal, initially passed by the House, that would have criminalized only threats made with specific intent to "obtain a benefit" when communicated by members of governing bodies.306

The State does urge that "threat" should be construed in light of the connotation of "coercion" as an overcoming of free will so as to exclude "threats" that would amount merely to "negotiation."307 But where the State would draw this line is unclear. The State posits that "coercive control ... provides the context for the threat," that "[c]oercion implies intimidation to violate any choice or free will of the listener," that "[c]oercion depends on a listener acting against his own interest to avoid a greater harm," and that "[c]oercive threats are in the same category [as] extortion and blackmail [by] cross[ing] a societal line between acceptable and unacceptable areas of negotiation."

Leaving aside whether the State's proposed limitation could salvage the statutes in light of Claiborne Hardware and Keefe, it fails as a matter of statutory construction. Neither the "coercion" definition nor section 36.03(a)(1) provides textual support for imposing qualitative limitations along the lines the State proposes. Instead, the statutes' text, context, and structure embody legislative intent to prescribe a particular meaning for "coercion" under the Penal Code that departs somewhat from — and clarifies — what might otherwise be the dimensions of coercion in common or legal parlance. For example, "coercion" under the Penal Code definition is narrower in some respects than more general notions of "coercion," being limited exclusively to communicated threats regarding particular enumerated subjects; other threats or means of compulsion are excluded. Conversely, neither the "coercion" definition nor section 36.03(a)(1) makes any mention of any actual or potential deprivation of that public servant's free will or exertion of "control" over that public servant.308 At most, the statutes might be said to deem or presume that a threat to inflict one of the harms enumerated within the "coercion" definition should be considered, at least for purposes of the Penal Code, to have coercive effects on a listener, as opposed to requiring proof of an actual deprivation of free will.309 Further confirming this intent is that the "coercion" definition, as we have seen, eschews several other limitations on coercion as known to the common law and other statutes.

To the extent the State is suggesting that the Legislature intended merely to leave the parameters of "coercion" or impermissible-versus-permissible "threats" to the discretion of prosecutors or fact-finders, to determine ad hoc or subjectively whether a given threat "crosses the line" or is deemed "unacceptable" in their own view, it has identified a different constitutional defect — impermissible vagueness.310 We would be required to strike down section 36.03(a)(1), as it incorporates paragraph (F), for the same reasons the Hanson court held it unconstitutional in that case.311

We cannot conceive of any viable means of construing the statutes more narrowly than we have. As such, we are left with the "last resort" of invalidation. However, our analysis may identify some factors the Legislature may wish to consider when and if it chooses to revise the statutes. But in the meantime, we must hold that section 36.03(a)(1), as it incorporates paragraph (F) of the Penal Code's "coercion" definition, is facially invalid under the First Amendment and is thus unenforceable. Because this holding entitles Perry to habeas relief with respect to Count II of the indictment, we need not address his remaining arguments.

CONCLUSION

We reverse the district court's order to the extent it upholds the facial constitutionality of the statute on which Count II of the indictment is based, Penal Code section 36.03(a)(1), as that statute incorporates paragraph (F) of the Code's "coercion" definition. However, because Perry has asserted only "as applied" constitutional challenges regarding Count I of the indictment ("Abuse of Official Capacity"), the binding precedents of the Court of Criminal Appeals preclude any relief as to that charge at this juncture. Accordingly, we affirm the district court's denial of relief as to Count I. We remand the case to the district court for further proceedings consistent with this opinion, including dismissal of Count II.