State Of Texas v. Harper

Justice Brown delivered the opinion of the Court, in which Chief Justice Hecht, Justice Green, Justice Guzman, and Justice Devine joined.Justice Boyd filed a dissenting opinion, in which Justice Johnson and Justice Lehrmann joined.Justice Blacklock did not participate in the decision.The Texas Citizens Participation Act (the TCPA) allows a defendant to obtain expedited dismissal of certain legal actions for which the party bringing the action does not establish prima facie support. we conclude that this suit to remove a county official from elected office under chapter 87 of the Texas Local Government Code (the removal statute) is a legal action under the TCPA. However, the TCPA does not apply when a government attorney brings an enforcement action in the state’s name. We further conclude that only one of the several allegations against the county official constitutes an enforcement action. Finally, as to the allegations that are not enforcement actions, we conclude that the state’s sovereign immunity does not protect it from the county official’s claim for appellate costs. We affirm the court of appeals’ judgment, with a modification.IBackgroundWhen Paul Harper ran for a position on the Somervell County Hospital District Board, he campaigned on pledges to eliminate the tax that supports the district and to replace the district’s administrative employees. He won the election, joined the board, and allegedly tried to make good on his promises. In response, a county resident named George Best sought to remove Harper from the board by filing this suit under the removal statute. See Tex. Loc. Gov’t Code §§ 87.001-.043. Specifically, Best alleged that Harper violated the district’s bylaws at a board meeting by moving to set the district’s tax rate at zero—even though Harper knew that eliminating the tax revenue would bankrupt or otherwise harm the district. Best also alleged that Harper posted a blog that falsely accused the district’s administrative employees of violating the law. Best argued these actions were enough to remove Harper for incompetency under the removal statute. See id. § 87.013(a)(1).The removal statute authorizes any Texas resident who has lived in a county for at least six months to file a petition to remove certain county officers from office (a removal petition). See id. §§ 87.012, .015(b); see also Tex. Const. art. V, § 24 (authorizing “the Judges of the District Courts” to remove “county officers” for “incompetency, official misconduct, habitual drunkenness, and other causes defined by law”).^[1] But it also requires the county attorney to “represent the state” in any removal proceedings that take place. Tex. Loc. Gov’t Code § 87.018(d); see also Garcia v. Laughlin, 285 S.W.2d 191, 194 (Tex. 1955) (orig. proceeding) (“Individual citizens have no private interest distinguishable from the public as a whole and have no right to maintain an ouster suit without being joined by a proper state official.”). Consistent with that requirement, the Somervell county attorney opted to appear in this case as plaintiff on the state’s behalf. The state adopted Best’s allegations, and it added an allegation that Harper engaged in misconduct by violating the Texas Open Meetings Act when he exchanged certain text messages with other board members. See Tex. Gov’t Code §§ 551.001- 146.Harper filed a motion to dismiss the case under the TCPA. See Tex. Civ. Prac. & Rem. Code §§ 27.001-.011. Harper argued that Best filed, and the state joined, the removal petition based on or in response to Harper’s exercise of the right to petition and right of free speech. See id. § 27.003(a) (authorizing a motion to dismiss an action that “is based on, relates to, or is in response to a party’s exercise of the right of free speech, right to petition, or right of association”). Harper contended that the state could not establish a prima facie case for removal because he did not formally move to set the district’s tax rate at zero, did not author or publish the blog, and did not violate the Open Meetings Act by exchanging text messages with other board members. See id. § 27.005(c) (providing that a court may not dismiss a legal action “if the party bringing the legal action establishes by clear and specific evidence a prima facie case for each essential element of the claim”). Harper argued that the trial court should dismiss the removal petition and award him attorney’s fees, costs, and sanctions. See id. § 27.009(a) (instructing that a court dismissing an action shall award attorney’s fees, court costs, and sanctions to a party who obtains dismissal). After conducting an evidentiary hearing, the trial court denied Harper’s motion to dismiss.Harper filed an interlocutory appeal from that order. See id. § 27.008(b) (authorizing an interlocutory appeal from an order denying a dismissal motion). The court of appeals reversed, holding that the TCPA applies to the state’s removal action and that the state failed to establish a prima facie case for removal. See 493 S.W.3d 105, 111, 116 (Tex. App.—Waco 2016). The court remanded the case to the trial court “for rendition of an order granting Harper’s motion to dismiss and for a determination of Harper’s request for court costs, reasonable attorney’s fees, and sanctions.” Id. at 118. The state moved for rehearing in the court of appeals, arguing for the first time that sovereign immunity protects it from any claim for attorney’s fees, court costs, or sanctions under the TCPA. The court of appeals denied the motion. But while the motion was pending, Harper lost his bid for reelection, and as a result he no longer serves on the hospital district’s board. We granted the state’s petition for review.IIMootnessBecause the state’s petition seeks to remove Harper from a position he no longer holds, we must first decide whether this case is moot. A case becomes moot when there ceases to be a justiciable controversy between the parties or when the parties cease to have “a legally cognizable interest in the outcome.” Williams v. Lara, 52 S.W.3d 171, 184 (Tex. 2001) (quoting Murphy v. Hunt, 455 U.S. 478, 481 (1982)). Mootness occurs when events make it impossible for the court to grant the relief requested or otherwise “affect the parties’ rights or interests.” See Heckman v. Williamson Cty., 369 S.W.3d 137, 162 (Tex. 2012). A case can become moot at any time, including on appeal. See id. at 166-67. When a case becomes moot, the court loses jurisdiction and cannot hear the case, because any decision would constitute an advisory opinion that is “outside the jurisdiction conferred by Texas Constitution article II, section 1.” Matthews v. Kountze Indep. Sch. Dist., 484 S.W.3d 416, 418 (Tex. 2016). But a case “is not rendered moot simply because some of the issues become moot during the appellate process.” In re Kellogg Brown & Root, Inc., 166 S.W.3d 732, 737 (Tex. 2005) (orig. proceeding). If only some claims or issues become moot, the case remains “live,” at least as to other claims or issues that are not moot. See id.Both the state and Harper argue that this case remains live. Harper argues we cannot address mootness at all because the trial court’s record contains no evidence that he lost his reelection bid and no longer serves on the board. But we must consider issues affecting our jurisdiction sua sponte. M.O. Dental Lab v. Rape, 139 S.W.3d 671, 673 (Tex. 2004) (per curiam). And we have the power, “on affidavit or otherwise,” to “ascertain the matters of fact that are necessary to the proper exercise of [our] jurisdiction,” even if evidence establishing those facts is not in the trial court’s record. Tex. Gov’t Code § 22.001(d). Here, the state filed a “status report” with the court of appeals that included an election canvass confirming that Harper lost his reelection bid. Harper does not dispute that he lost the election or that he no longer holds the position.The state concedes that Harper’s failed reelection bid renders its removal petition moot. But the state contends that the issue whether the court of appeals properly ordered the trial court to award Harper his costs, attorney’s fees, and sanctions under the TCPA is not moot. We agree that Harper’s request for an award of attorney’s fees and sanctions under the TCPA presents an issue that is separate from the request for removal, and we also agree that this separate issue survives the mootness that would otherwise prevent us from addressing the underlying claim.We have recognized that in some cases—but not all—a claim for attorney’s fees “breathes life” into a suit that has become moot in all other respects. Camarena v. Tex. Emp’t Comm’n, 754 S.W.2d 149, 151 (Tex. 1988) (concluding that when the claimants prevailed in the trial court before their underlying claims became moot, their claim for attorney’s fees as prevailing parties remained live even though the underlying claims were moot); see also Allstate Ins. Co. v. Hallman, 159 S.W.3d 640, 643 (Tex. 2005) (holding that an attorney’s-fees claim based on a statute that allows fee awards to non-prevailing parties remains live even after the underlying claim becomes moot); Speer v. Presbyterian Children’s Home & Serv. Agency, 847 S.W.2d 227, 229 (Tex. 1993) (distinguishing Camarena v. Texas Employment Commission and holding that an attorney’s-fees claim under a prevailing-party statute had become moot along with the underlying claim because the claimant had not prevailed before mootness occurred).Whether an attorney’s-fees claim breathes life into an otherwise moot appeal depends first on whether the claimant seeks the fees under a statute that authorizes fees only for a prevailing party or, alternatively, under a statute that permits fees based on equitable principles regardless of who prevails. If the statute allows a non-prevailing party to recover fees under equitable principles, the claim for fees always breathes life into a case that has otherwise become moot, because the trial court must always consider the relative merits of the parties’ positions (among other factors) when exercising its discretion to award fees to either party. Allstate, 159 S.W.3d at 643.But when the party seeks attorney’s fees under a prevailing-party statute, the determination whether the attorney’s-fees claim is moot depends on whether the party prevailed before the underlying substantive claim became moot. As we recognized in Camarena, if the party prevailed before the substantive claim became moot, the party’s claim for attorney’s fees under a prevailing- party statute remains a live controversy and a court must consider the claim’s merits to determine whether the party properly prevailed. See 754 S.W.2d at 151. But as we recognized in Speer v. Presbyterian Children’s Home & Service Agency, if the party did not prevail before the substantive claim became moot, the party’s claim for attorney’s fees is also moot because the party can never prevail and thus can never be entitled to attorney’s fees. See 847 S.W.2d at 229-30.In this case, Harper seeks fees under the TCPA, which requires the trial court to award court costs, attorney’s fees, expenses, and sanctions to a party who prevails on its motion to dismiss. See Tex. Civ. Prac. & Rem. Code § 27.009(a). The trial court denied Harper’s motion to dismiss, but he prevailed in the court of appeals, which remanded the case and ordered the trial court to grant Harper’s motion. See 493 S.W.3d at 118. Because Harper established his right to dismissal before his failed reelection bid rendered the state’s removal petition moot, we conclude that his claim for attorney’s fees, costs, and sanctions breathes life into this appeal. See Camarena, 754 S.W.2d at 151. Thus, to determine whether Harper properly prevailed and is entitled to assert his claim for attorney’s fees, costs, and sanctions, we must address the merits of the state’s removal petition and of Harper’s dismissal motion, even though both would otherwise be moot. See id.IIIApplicability of the TCPAThe state argues that the court of appeals erred by ordering the trial court to grant Harper’s motion to dismiss under the TCPA. See Tex. Civ. Prac. & Rem. Code § 27.003(a). The reason, the state says, is that the TCPA does not apply to a removal petition. Specifically, the state argues the TCPA is inapplicable because an action under the removal statute is not a “legal action” under the TCPA. See id. §§ 27.001(6), .003(a). The state also argues that a removal action is an “enforcement action” to which the TCPA expressly does not apply. See id. § 27.010(a). We hold that a removal petition is a “legal action” under the TCPA. See id. § 27.003(a). We also conclude that, while one allegation against Harper constitutes a TCPA “enforcement action,” the remaining allegations do not. See id. § 27.010(a).A. TCPA “legal action”The TCPA permits a party to file a motion to dismiss a “legal action” if the action “is based on, relates to, or is in response to a party’s exercise of the right of free speech, right to petition, or right of association.” Id. § 27.003(a). The state argues that a removal petition is not a “legal action” and that, as a result, dismissal under the TCPA is not available. See id. Instead, the state says, the removal statute contains specific provisions for dismissal and costs that prevail over the general provisions in the TCPA. See Tex. Loc. Gov’t Code § 87.016(c). And, according to the state, applying the TCPA’s dismissal and cost provisions to a removal action is inconsistent with the TCPA’s language and purpose. See Tex. Civ. Prac. & Rem. Code § 27.002.The TCPA defines a “[l]egal action” as “a lawsuit, cause of action, petition, complaint, cross-claim, or counterclaim or any other judicial pleading or filing that requests legal or equitable relief.” Id. § 27.001(6). This undeniably “broad” definition appears to encompass any “procedural vehicle for the vindication of a legal claim.” Paulsen v. Yarrell, 537 S.W.3d 224, 233 (Tex. App.— Houston [1st Dist.] 2017, pet. denied) (noting, however, that a TCPA dismissal motion is not itself a TCPA “legal action”); see also In re Elliott, 504 S.W.3d 455, 465 (Tex. App—Austin 2016, orig. proceeding) (holding that the TCPA’s “broad definition” encompasses a petition under Texas Rule of Civil Procedure 202); Better Bus. Bureau of Metro. Dall., Inc. v. Ward, 401 S.W.3d 440, 443 (Tex. App.—Dallas 2013, pet. denied) (holding that the definition “is broad and evidences a legislative intent to treat any claim by any party on an individual and separate basis”).Despite the TCPA’s broad definition, the state argues that a removal petition is not a legal action because it seeks “constitutional” or “political” relief in the form of an order removing an elected official from office rather than “legal or equitable relief” such as damages, an injunction, or declaratory relief. We disagree. A court order requiring the defendant’s removal or ouster from office is undoubtedly a “remedy.” See State ex rel. Dishman v. Gary, 359 S.W.2d 456, 458-59 (Tex. 1962) (orig. proceeding). And “remedy” is another word for “relief.” See Relief, Black’s Law Dictionary (10th ed. 2014) [Black's] (defining “relief” as the “redress or benefit[] . . . that a party asks of a court” and noting that “relief’ is “[a]lso termed remedy”). Here, the remedy the state seeks is only available because Texas law—specifically, article 5, section 24 of the Texas constitution and chapter 87 of the Texas Local Government Code—provides it. See Tex. Const. art. V, § 24; Tex. Loc. Gov’t Code § 87.013. Because a removal petition seeks legal relief in the form of a statutory remedy, the pleading is a “legal action” under the TCPA. See Tex. Civ. Prac. & Rem. Code § 27.003(a).Even so, the state argues, applying the TCPA’s expedited-dismissal provisions to the removal statute “creates a conflict between the two statutory schemes” because the removal statute provides its own protections against meritless petitions. For example, the statute permits a trial court to refuse to issue an order for citation against the county officer. Tex. Loc. Gov’t Code § 87.016(c). If the trial court refuses to issue an order for citation, “the petition shall be dismissed at the cost of the person filing the petition,” and the plaintiff may not appeal. See id. If the trial court does issue an order for citation, it must require the plaintiff to post security for the county officer’s costs. See id. And if the trial court temporarily suspends the officer but later determines that the suspension was improper, the officer may recover damages and costs. See id. § 87.017(b). According to the state, these “specific” provisions govern the dismissal of a removal action, but the TCPA’s “general” dismissal provisions do not. See, e.g., Tex. Gov’t Code § 311.026(b) (providing that when a general provision irreconcilably conflicts with a special or local provision, “the special or local provision prevails . . . unless the general provision is the later enactment and the manifest intent is that the general provision prevail”). So in the state’s view, we cannot apply the TCPA to a removal petition without intruding “into a field governed by a specific law adopted pursuant to a constitutional mandate.”Harper responds that the removal statute’s remedies are not exclusive and that the chapter itself contemplates that a defendant can rely on external defenses such as the TCPA. For example, the chapter states that “the proceedings connected with the trial” of a removal petition “shall be conducted as much as possible in accordance with the rules and practice of the court in other civil cases.” Tex. Loc. Gov’t Code § 87.018(b). Similarly, either party “may appeal the final judgment to the court of appeals in the manner provided for other civil cases.” Id. § 87.019(a). Harper also points to our statement that “[e]xcept where otherwise provided by statute, the rules of practice governing other civil cases control” in a removal proceeding. Dishman, 359 S.W.2d at 458.We agree with Harper. The TCPA’s dismissal provisions complement, rather than contradict, the removal statute. The rule that a specific provision controls over a general provision applies only when the statutes at issue are ambiguous or irreconcilable. See Tex. Gov’t Code § 311.026(a) (“If a general provision conflicts with a special or local provision, the provisions shall be construed, if possible, so that effect is given to both.”); see also Tex. Lottery Comm ‘n v. First State Bank of DeQueen, 325 S.W.3d 628, 639 (Tex. 2010) (“[W]e construe statutes by first looking to the statutory language for the Legislature’s intent, and only if we cannot discern legislative intent in the language of the statute itself do we resort to canons of construction or other aids such as which statute is more specific.”). Here, the state neither identifies nor have we found any provision in either statute that cannot be applied alongside those in the other statute. The removal statute provides for dismissal when the trial court determines that citation should not issue. Tex. Loc. Gov’t Code § 87.016(c). The TCPA’s dismissal provisions provide the defendant the opportunity to argue for dismissal on other grounds—namely, his rights to free speech, to petition, and to associate. Tex. Civ. Prac. & Rem. Code § 27.003(a). These provisions do not conflict.Finally, the state argues we have stated that a removal petition initiates a unique, “quasi criminal” proceeding “such that ordinarily it admits of no cross action.” SeeDishman, 359 S.W.2d at 460. In State ex rel. Dishman v. Gary, the state unilaterally nonsuited its removal petition after the defendant filed a “cross action” asserting, as an affirmative defense, the argument that actions before his election to office could not be the basis for removal. Id. The primary issue in Dishman was whether the defendant’s “cross action” preserved the trial court’s jurisdiction, thus enabling it to enter an order reinstating the state’s removal petition. Id. at 458. We held that it did not, concluding that an ouster suit “admits of no cross action” and that “no statutory provision for a cross action” existed. Id. at 460. We thus found it “difficult to imagine a set of circumstances under which a cross action would properly lie” under the removal statute. Id. at 459. The one scenario where we suggested a “cross action” might “perhaps” be permitted would be where the district attorney had “repeatedly filed and dismissed ouster suits against a defendant for the purpose of harassment,” in which case the trial court would have a common-law basis to dismiss. Id. Otherwise, we concluded, “it would seem that the only issue presented [in a removal action] would be whether or not the defendant w[as] guilty of the charges brought against him,” and the defendant’s “cross action” did “not operate to deprive the district attorney of his control of the statutory ouster suit nor serve as a basis for the reinstatement of a suit in which he had taken a voluntary nonsuit.” Id. at 460.In response to the state’s reliance on Dishman, Harper argues that even if the removal statute prohibits a counterclaim or “cross action” to a removal petition, a TCPA dismissal motion is not a “cross action.” We need not address this argument, however, because even if a TCPA dismissal motion is a counterclaim or cross action, the TCPA could authorize that counterclaim even if the removal statute does not. We decided Dishman in 1962, see id. at 456, long before the legislature enacted the TCPA in 2011, see Citizens Participation Act, 82d Leg., R.S., ch. 341, §§ 1­4, 2011 Tex. Gen. Laws 961 (codified at Tex. Civ. Prac. & Rem. Code §§ 27.001-.011). We noted in Dishman that a court’s common-law power to dismiss a removal petition filed for the purpose of harassment can supplement and coexist with the removal statute. See Dishman, 359 S.W.2d at 460 (suggesting that courts have the common-law power to dismiss “ouster suits [brought] against a defendant for the purpose of harassment”). A court’s TCPA-authorized power to dismiss an action based on the defendant’s exercise of his free-speech and petition rights can surely do the same. Nothing in the constitution or in the removal statute precludes the legislature from granting such power; the question here is whether it did so in the TCPA.We conclude that it did. The TCPA expressly applies to any “legal action,” and—as we have explained—a removal petition is a legal action. See Tex. Civ. Prac. & Rem. Code § 27.001(6). Accordingly, the TCPA applies to a removal petition.B. TCPA “enforcement action”We now turn to whether a removal petition constitutes an “enforcement action” under the TCPA. See id. § 27.010(a) (stating that the TCPA “does not apply to an enforcement action that is brought in the name of this state . . . by . . . a county attorney”). If it does, Harper cannot invoke the TCPA’s protections. See id. The state joined Best’s petition, but that does not necessarily make this suit an enforcement action. Unlike “legal action,” the TCPA’s definitional subsection does not include an entry for the term “enforcement action.” See generally id. § 27.001. Nor have we previously considered its meaning.“Statutory construction is a legal question we review de novo.” See City of Rockwall v. Hughes, 246 S.W.3d 621, 625 (Tex. 2008). “In construing statutes, we ascertain and give effect to the Legislature’s intent as expressed by the language of the statute.” Id. Statutes do not always include express statements of purpose or directions for construction, but the TCPA includes both. See Tex. Civ. Prac. & Rem. Code §§ 27.002, .011. The TCPA’s purpose “is to encourage and safeguard the constitutional rights of persons to petition, speak freely, associate freely, and otherwise participate in government to the maximum extent permitted by law and, at the same time, protect the rights of a person to file meritorious lawsuits for demonstrable injury.” Id. § 27.002 (emphasis added). We must construe the TCPA “liberally to effectuate its purpose and intent fully.” Id. § 27.011(b). Construing the TCPA liberally means construing its exemptions narrowly.One such exemption is that the TCPA “does not apply to an enforcement action that is brought in the name of this state . . . by . . . a county attorney.” Id. § 27.010(a). Because the legislature did not define “enforcement action,” we must determine the term’s “common, ordinary meaning.” City of Richardson v. Oncor Elec. Delivery Co., 539 S.W.3d 252, 261 (Tex. 2018). “To determine a statutory term’s common, ordinary meaning, we typically look first to [its] dictionary definitions . . . .” Tex. State Bd. of Exam’rs of Marriage & Family Therapists v. Tex. Med. Ass’n, 511 S.W.3d 28, 35 (Tex. 2017). The common meaning applies “unless a more precise definition is apparent from the statutory context.” Oncor, 539 S.W.3d at 261; see also Greene v. Farmers Ins. Exch., 446 S.W.3d 761, 765 (Tex. 2014) (“We rely on the plain meaning of the text as expressing legislative intent unless a different meaning is supplied by legislative definition or is apparent from the context, or the plain meaning leads to absurd results.”).An enforcement is “[t]he act or process of compelling compliance with a law, mandate, command, decree, or agreement.” Enforcement, Black’s; see also The American Heritage Dictionary of the English Language 590 (5th ed. 2011) (defining “enforce” as “[t]o compel observance of or obedience to: enforce a law”); Webster’s Third New International Dictionary 751 (1961) (defining “enforcement” as “the compelling of the fulfillment (as of a law or order)”). Applying the Black’s dictionary definition, the court of appeals concluded that “there is nothing in the removal statute with which the State is seeking to compel Harper’s compliance,” because “Harper cannot comply with his duties if he is no longer a board member.” 493 S.W.3d at 111.The dictionary definitions do not capture the full extent of the legislature’s intent in this instance. The reason is that our precedent directing us to the dictionaries, see, e.g., Melden & Hunt, Inc. v. E. Rio Hondo Water Supply Corp., 520 S.W.3d 887, 893 (Tex. 2017), collides with our precedent directing us to “presume the Legislature selected language in a statute with care and that every word or phrase was used with a purpose in mind.” DeQueen, 325 S.W.3d at 635 (emphasis added). Section 27.010 contains the “enforcement action” exemption, but it also contains three others. See Tex. Civ. Prac. & Rem. Code § 27.010. The TCPA does not apply to: “an enforcement action” brought in the name of the state; “a legal action” against certain businesses; “a legal action” in certain personal injury cases; and “a legal action” arising out of certain insurance contracts. Id.The second, third, and fourth exemptions apply to a “legal action,” id. § 27.010(b)-(d), a term the TCPA defines, see id. § 27.001(6). But the first applies to something different: an “enforcement action.” Id. § 27.010(a). We must assume the legislature used a different word because it intended a different meaning. See DeQueen, 325 S.W.3d at 635. Otherwise, the legislature would have said “legal action” all four times.These observations demonstrate that whatever an “enforcement action” is, it must be different from a “legal action.” A “legal action” is “a lawsuit, cause of action, petition, complaint, cross-claim, or counterclaim or any other judicial pleading or filing that requests legal or equitable relief.” Tex. Civ. Prac. & Rem. Code § 27.001(6). It would be difficult to write a more capacious definition, and even more so to conceive one for an undefined term. So the definition of an “enforcement action” cannot exceed the broad definition the TCPA gives for “legal action.” See id. Nor can the two terms share the same meaning. See DeQueen, 325 S.W.3d at 635. Hence, a TCPA “enforcement action” must be a subtype of legal actions in general. That is, the term “enforcement action” must mean something less than “any legal action brought by the state.” Otherwise, “enforcement action” and “legal action” would mean the same thing.We conclude that, within the TCPA, the term “enforcement action” refers to a governmental attempt to enforce a substantive legal prohibition against unlawful conduct. This interpretation is consistent with the TCPA’s text, under which enforcements are but one type of legal action. See id. § 27.001(6). It also accords with the TCPA’s self-expressed purpose, which “is to encourage and safeguard constitutional rights”—not to impede the criminal or civil proceedings that law enforcement or other government agencies initiate. See id. § 27.002. Under this definition, a removal petition is not an “enforcement action” in the abstract. Instead it is a procedural device, and as such a party cannot initiate a removal action to enforce the removal statute itself. There is a range of conduct—some unlawful and some not—for which a public official may properly face removal under the removal statute. And as discussed above, the TCPA is available by default since removal actions are legal actions. However, when a removal action has its basis in unlawful conduct, the “enforcement action” exemption renders the TCPA inapplicable.Accordingly, we must next ask whether the petition against Harper seeks to enforce a substantive legal prohibition against unlawful conduct. Under the removal statute, “[a]n officer may be removed for: (1) incompetency; (2) official misconduct; or (3) intoxication on or off duty caused by drinking an alcoholic beverage.” Tex. Loc. Gov’t Code § 87.013(a).In his original petition, Best sought Harper’s removal based on Best’s allegation that Harper “exhibited incompetency by way of gross ignorance of his official duties and gross carelessness in the discharge of those duties.” See id. § 87.011(2)(A)-(B) (“‘Incompetency’ means: (A) gross ignorance of official duties; (B) gross carelessness in the discharge of those duties[] . . . .”). Incompetency is a basis for removal under the removal statute, but it is not against the law. See id. § 87.013(a)(1). Similarly, while intoxication is also a basis for removal, and while the town drunk might make a lousy official, being the town drunk is not against the law. See id. § 87.013(a)(3).Best’s incompetency claims are a transparent retaliation against Harper’s quixotic political beliefs. Harper opposed the hospital district’s creation. He put his beliefs into action by running for office, and the voters of Somervell County elected him on a clearly stated anti-tax platform. Best’s petition alleges Harper did exactly what he told the voters he would do upon taking office. Harper’s detractors may disagree with his politics, but no law requires elected officials to support the status quo upon arriving in office. Best’s removal petition was a pretext for forcing Harper to cease acting on the beliefs that won him his office in the first place.We are not fooled. We doubt anyone else is. Harper’s refusal to capitulate to Best’s demands does not render him incompetent. Best thought that the hospital district was important, and he sought Harper’s removal because he thought that Harper was setting the hospital up for failure, lacked candor, and sometimes communicated less than cordially. Even if a jury agreed that Harper was unfit for office, he would face no criminal or civil penalty other than removal itself. Efforts like Best’s are attacks on core political speech. But the TCPA “protects citizens who petition or speak on matters of public concern from retaliatory lawsuits that seek to intimidate or silence them.” In re Lipsky, 460 S.W.3d 579, 584 (Tex. 2015) (orig. proceeding) (footnote omitted) (citing Tex. Civ. Prac. & Rem. Code §§ 27.001-.011). Since the incompetency allegations in Best’s petition seek to achieve Best’s political goals rather than to enforce a law, they cannot form the basis of an enforcement action for purposes of the TCPA. See Tex. Civ. Prac. & Rem. Code § 27.010(a).Nor are we swayed by the argument that Best’s incompetency claim included allegations that Harper violated the hospital district’s bylaws. The bylaws require board members to discharge their duties “in good faith, with ordinary care, and in a manner the director reasonably believes to be in the best interest of the District.” Similarly, the bylaws prohibit members from performing “any act with the intention of harming the District or any of its operations” or “that would make it impossible or unnecessarily difficult to carry on the intended or ordinary business of the District.” Best’s petition argues Harper violated the bylaws by attempting to set the district’s tax rate at zero and by accusing the district’s administration of violating the law.The Somervell County Hospital District Board of Directors may promulgate bylaws pursuant to its authority to “adopt rules governing the operation of the hospital and hospital system and the duties, functions, and responsibilities of district staff and employees.” Tex. Health & Safety Code § 286.075. Although violation of an organization’s internal rules concerning “duties, functions, and responsibilities” may expose the violator to liability, the rules are just that— rules. Though they may sometimes overlap with Texas statutes, regulations, or common-law obligations, the bylaws appear nowhere within these bodies of law. Instead, they arise by agreement of the board of directors. Indeed, the bylaws include numerous references to actions required by law or the bylaws. In sum, the bylaws are not legal prohibitions against unlawful conduct, and they do not elevate the allegations in Best’s petition to an “enforcement action.” See Tex. Civ. Prac. & Rem. Code § 27.010(a).In light of our textual conclusion that an enforcement action must be something narrower than a legal action, our general understanding that an enforcement action cannot enforce itself, our specific conclusion that an enforcement action under the TCPA must enforce substantive prohibitions against unlawful conduct, and the legislature’s clear instruction to construe the TCPA liberally to protect citizens’ rights to participate in government, we conclude the that allegations in Best’s petition do not amount to an “enforcement action” under the TCPA. See id.However, the removal statute also allows removal for “official misconduct,” which may include allegations or evidence that a public official has acted unlawfully. Tex. Loc. Gov’t Code § 87.013(2); see also id. § 87.011(3) (defining “[official misconduct" as "intentional, unlawful behavior relating to official duties" including "intentional or corrupt failure, refusal, or neglect of an officer to perform a duty imposed on the officer by law"). An allegation premised on unlawful conduct rather than behavior undesirable in a public official can form the basis of an "enforcement action" for purposes of the TCPA. See Tex. Civ. Prac. & Rem. Code § 27.010(a).Best did not allege official misconduct against Harper, but the state did. After it joined Best's petition, the state added---as an "official misconduct" ground---the allegation that Harper violated the Open Meetings Act. See Tex. Gov't Code § 551.143(a) ("A member . . . of a governmental body commits an offense if the member . . . knowingly conspires to circumvent this chapter by meeting in numbers less than a quorum for the purpose of secret deliberations in violation of this chapter."). This allegation is sufficient to form the basis of an enforcement action. It involves reference to a specific statutory provision that contains a substantive prohibition against certain conduct, and the state alleges Harper violated that prohibition. Whatever the merits of the state's allegation, it amounts to an "enforcement" of a law---the Open Meetings Act---"brought in the name of this state." See Tex. Civ. Prac. & Rem. Code § 27.010(a).A removal petition is not an "enforcement action" unless it seeks to enforce a substantive legal prohibition against unlawful conduct. The removal grounds alleging Harper's incompetency do not meet this definition, which means that the TCPA's "enforcement action" exemption does not apply to them. See id. But under the same definition, the state's additional ground alleging official misconduct based on violations of the Open Meetings Act is an enforcement action. So the enforcement-action exemption renders the TCPA inapplicable to the state's additional ground. See id.We conclude that Harper may benefit from the TCPA's expedited-dismissal provisions for the grounds that Best's initial removal petition raised, but not for the state's additional ground alleging a violation of the Open Meetings Act.IVSovereign ImmunityBecause the TCPA applies, we must consider whether Harper can recover appellate fees and costs from the state. The court of appeals' judgment awarded "judgment against the state of Texas for [Harper's] appellate costs that were paid, if any, by [Harper]; and all unpaid appellate court cost[s], if any . . . against the state of Texas.” Similarly, the court of appeals’ opinion remands the case to the trial court for a “determination of Harper’s request for court costs, reasonable attorney’s fees, and sanctions.” 493 S.W.3d at 118.The state argues the award and remand were improper. Because the trial court has not yet made a “determination of’ Harper’s request for costs, fees, or sanctions, see id., we can do no more than speculate whether Harper will actually obtain a judgment for any of these against Best, the state, both, or neither. See id. So we need not consider the state’s argument that such an award would be improper if entered against the state. Instead, the narrow issue before us is whether sovereign immunity protects the state from the appellate fees and costs that the court of appeals has already awarded to Harper. We conclude it does not.“Sovereign immunity in Texas embodies two concepts: immunity from liability and immunity from suit.” Rusk State Hosp. v. Black, 392 S.W.3d 88, 93 (Tex. 2012). “[I]mmunity from liability is not jurisdictional and protects [only] from judgments.” Harris Cty. Hosp. Dist. v. TomballReg’lHosp., 283 S.W.3d 838, 842 (Tex. 2009). Hence, immunity from liability “must be pleaded, or else it is waived.” Kinnear v. Tex. Comm’n on Human Rights ex rel. Hale, 14 S.W.3d 299, 300 (Tex. 2000) (per curiam); see also Tex. Dep’t of Transp. v. Jones, 8 S.W.3d 636, 638 (Tex. 1999) (per curiam) (“Like other affirmative defenses to liability, [immunity from liability] must be pleaded or else it is waived.”); Davis v. City of San Antonio, 752 S.W.2d 518, 523 (Tex. 1988) (“[T]he City waived any immunity defense by failing to affirmatively plead it.”).By contrast, immunity from suit “implicates a court’s subject-matter jurisdiction” and may “be raised for the first time on appeal.” Engelman Irrigation Dist. v. Shields Bros., Inc., 514 S.W.3d 746, 751, 755 (Tex. 2017); see also Jones, 8 S.W.3d at 638. Absent legislative waiver, and when it applies to begin with, immunity from suit protects the state from suits against it. See Tex. Nat. Res. Conservation Comm’n v. IT-Davy, 74 S.W.3d 849, 855 (Tex. 2002); see also City of Houston v. Williams, 353 S.W.3d 128, 134 (Tex. 2011). Legislative waiver is effective only if it is “clear and unambiguous.” Tooke v. City of Mexia, 197 S.W.3d 325, 328-29 (Tex. 2006) (citing Tex. Gov’t Code § 311.034). Our cases also recognize that the state’s immunity does not apply in certain situations in which the state “join[s] into the litigation process.” Reata Constr. Corp. v. City of Dallas, 197 S.W.3d 371, 376 (Tex. 2006); see also Kinnear, 14 S.W.3d at 300 (“Because the [state] initiated this proceeding . . . and [the defendant] claimed attorney fees as a consequence of that suit, the jurisdictional question in this case was answered when the [state] filed suit[] . . . . Thus the issue is whether the [state] . . . is immune from liability . . . .” (emphasis added)).The state has not argued that it is immune from liability, but only that it is immune from suit. Thus, the state has waived its immunity-from-liability argument. See Kinnear, 14 S.W.3d at 300 (first citing Jones, 8 S.W.3d at 638; and then citing Davis, 752 S.W.2d at 519-20). The dispositive question, then, is whether the state is immune from suit. If it is, then Harper cannot recover his fees, at least not from the state. But if the state is not immune from suit, then he can.The state argues that neither the TCPA nor the removal statute contains a legislative waiver of immunity and that immunity from suit therefore protects it from an award of appellate costs. The removal statute requires “the person filing the petition to post security for costs” and further directs that dismissal shall be “at the cost of the person filing the petition.” Tex. Loc. Gov’t Code § 87.016(c). The state joined Best’s suit, but Best filed the initial petition, so the state argues that the removal statute does not provide a waiver of immunity. Similarly, the TCPA allows a court to award “sanctions against the party who brought the legal action.” Tex. Civ. Prac. & Rem. Code § 27.009(a)(2) (emphasis added). Because Harper initiated this action, the state also argues that the TCPA does not waive immunity. See also id. § 27.011(a) (“[The TCPA] does not abrogate or lessen any other defense, remedy, immunity, or privilege available under other constitutional, statutory, case, or common law or rule provisions.”).Nor, says the state, do any of this Court’s limitations to immunity’s scope apply. Reata Construction Corp. v. City of Dallas details a circumstance in which “a governmental entity does not have immunity from suit for monetary claims against it that are ‘germane to, connected with, and properly defensive to’ affirmative claims made by the entity,” to the extent that the claims against the entity offset the entity’s own claims. City of Dallas v. Albert, 354 S.W.3d 368, 372 (Tex. 2011) (quoting Reata, 197 S.W.3d at 378); see alsoManbeck v. AustinIndep. Sch. Dist., 381 S.W.3d 528, 532-33 (Tex. 2012) (per curiam) (noting that Reata applies in “cases where the governmental entity asserts an affirmative claim for monetary relief against which the opponent’s claims can be offset”).Harper responds that when the state appeared in this suit, it adopted Best’s live pleading stating jurisdiction was proper in the trial court. Moreover, Harper argues, the removal statute contemplates that the temporary replacement officer must pay any “costs” associated with the prosecution of an unsuccessful removal action. See Tex. Loc. Gov’t Code § 87.016(c). Harper argues this officer is an “agent of the state” and therefore the removal statute contemplates payment by a “governmental entity.” So in Harper’s view, the state stepped beyond the sphere of its immunity when it appeared in the suit. Harper also points to the four-factor analysis we set out in Wichita Falls State Hospital v. Taylor for determining whether language less explicit than “‘sovereign immunity to suit is waived’ . . . may nevertheless waive the State’s immunity from suit,” to argue that the removal statute, the TCPA, or both expressly waives immunity. See 106 S.W.3d at 697-98 (footnote omitted).We agree with the state that neither statute waives the state’s immunity from suit. The portions of the removal statute discussing costs treat the temporary replacement officer as an individual, not as an agent of the state. See Tex. Loc. Gov’t Code § 87.017. The officer is an agent of the state for purposes of carrying out his job duties, but not for purposes of paying the costs associated with an unsuccessful removal action. Consequently, the removal statute does not contain a “clear and unambiguous” waiver of the state’s immunity from suit. See Tooke, 197 S.W.3d at 328-29. Nor does the TCPA contain a waiver that meets Tooke’s requirements. See id. Harper argues that the TCPA “d[oes] not expressly exempt any governmental entity from its sphere” and that “[i]f the Legislature intended to exempt the government from the TCPA . . . [it] would have clearly stated so.” That argument states precisely the opposite of the presumption that we apply when determining whether the legislature has waived sovereign immunity. See Taylor, 106 S.W.3d at 701 (“[W]e require the Legislature to express its intent beyond doubt and will construe ambiguities in a manner that retains the State’s immunity.”). The TCPA allows for awards of costs, but it does not contain any “clear and unambiguous” legislative basis for awarding costs against the state. See Tooke, 197 S.W.3d at 329. So neither the removal statute nor the TCPA expressly waives the state’s sovereign immunity.Although neither statute waives the state’s immunity from suit, that conclusion does not answer the question whether immunity applies in the first place. See Engelman, 514 S.W.3d at 753 (“[S]overeign immunity is a common-law creation, and it remains the judiciary’s responsibility to define the boundaries of the doctrine.”); Brown & Gay Eng’g, Inc. v. Olivares, 461 S.W.3d 117, 122 (Tex. 2015) (noting “the doctrine’s judicial origins”); Reata, 197 S.W.3d at 375 (“[I]t remains the judiciary’s responsibility to define the boundaries of the common-law doctrine and to determine under what circumstances sovereign immunity exists in the first instance.”). When sovereign immunity is inapplicable due to judicial modification rather than legislative pronouncement, courts characterize the protection’s absence as arising from abrogation rather than waiver. See Albert, 354 S.W.3d at 375 (“[T]he judiciary has abrogated [governmental entities'] common law immunity from suit as to certain offsetting claims.”); Taylor, 106 S.W.3d at 696 (“[W]e have not absolutely foreclosed the possibility that the judiciary may abrogate immunity by modifying the common law . . . .”).Reata is our foundational case discussing abrogation. See generally 197 S.W.3d at 374­77. Under the Reata rule, immunity from suit does not protect the state from counterclaims that are “germane to, connected with, and properly defensive to” certain of the state’s own claims, but only to the extent that the counterclaims act as a monetary “offset” to the state’s own claim. Id. at 373. In this case, the state has not asserted any claims against which Harper can counterclaim a monetary offset. Because the state seeks Harper’s removal rather than an appropriate monetary recovery, the Reata rule does not apply to abrogate the state’s immunity from Harper’s counterclaims. But that does not mean that no rule does so.Significant as it is, Reata is not our only case discussing abrogation, and it does not purport to map the full boundary separating counterclaims that sovereign immunity bars from those it does not. See, e.g., Nazari v. State,___S.W.3d___,___(“[W]e have never held that the Reata rule always applies when the government seeks any transfer of funds. . . . [N]or have we ever held that Reata applies only to compensatory damages.”).As an example of another type of abrogation, Harper cites our per curiam decision in Kinnear v. Texas Commission on Human Rights ex rel. Hale. See generally 14 S.W.3d 299. There, “[t]he Texas Commission on Human Rights sued [Kinnear] for violating the Texas Fair Housing Act.” Id. at 299. Kinnear prevailed, and he requested attorney’s fees from the commission under the Fair Housing Act, which provides that a court “may award reasonable attorney fees to the prevailing party.” Tex. Prop. Code § 301.156. In analyzing whether the state abandoned its immunity from suit by initiating the litigation, we held that “the jurisdictional question . . . was answered when the [state] filed suit, regardless of whether the [state] can ultimately be liable for fees.” Kinnear, 14 S.W.3d at 300. Put differently, when the state “initiated th[e] proceeding” that spurred the attorney’s-fees claim, it was operating outside the bounds of its immunity from suit. Id. And because, as here, the state had waived immunity from liability by failing to plead it, we “render[ed] judgment awarding Kinnear his attorney fees and costs” under the Fair Housing Act. See id. Harper says his counterclaim is “akin” to the one in Kinnear and that he should prevail on that basis.The state responds by citing another of our per curiam opinions, Manbeck v. Austin Independent School District, for its counterargument that a “governmental entity’s decision to avail itself of a statutory right—without bringing an affirmative claim for monetary damages— does not result in a loss of immunity from a claim for attorney’s fees.” See generally 381 S.W.3d 528. Manbeck addressed a counterclaim for attorney’s fees after the state non-suited a judicial appeal from an administrative proceeding that arose under the Texas Workers’ Compensation Act. See id. at 529 (citing Tex. Lab. Code § 408.021(c)). Finding the Reata rule inapplicable, we “reversed the trial court’s award of attorney fees.” Id. at 533.The parties strive to distinguish Kinnear and Manbeck—each arguing that one case or the other answers the TCPA attorney’s-fees question at issue here. However, neither per curiam opinion addresses the TCPA. And since Manbeck concerned an administrative appeal in the workers’ compensation context, see id. at 529, it does not overrule Kinnear’s conclusion in the fair-housing context, see 14 S.W.3d at 299. But, the state argues, it is not that Manbeck overrules Kinnear—it is that Reata does, or at least it establishes a new paradigm such that Kinnear is no longer good law. We disagree. Far from overruling Kinnear, Reata cited it—along with Anderson, Clayton & Co. v. State ex rel. Allred, 62 S.W.2d 107, 110 (Tex. 1933), and State v. Humble Oil & Refining Co., 169 S.W.2d 707, 708 (Tex. 1943)—as the foundation for the rule it established. See Reata, 197 S.W.3d at 374-77; see also Nazari,___S.W.3d at___(listing Anderson, Humble Oil, and Kinnear as “the three principal cases on which Reata relied”). And we have cited Kinnear since. See Rusk, 392 S.W.3d at 97 & n.4 (listing Kinnear among cases establishing that “immunity deprives courts of subject-matter jurisdiction”); Nazari, ___S.W.3d at___(“[W]hile Kinnear involved an abrogation of immunity, it was not the type of abrogation we announced in Anderson and expounded on in Reata.”); see also Jeffrey S. Boyd, Where Sovereign Immunity and Water Development Issues Collide, 39 Tex. Envtl. L.J. 95, 116 (2009) (“The Texas Supreme Court recently reaffirmed and clarified [Kinnear's] holding in Reata . . . .” (emphasis added)).In Kinnear, sovereign immunity did not protect the state from a claim for attorney’s fees under the Fair Housing Act. See 14 S.W.3d at 299. On the other hand, in Manbeck, sovereign immunity did protect the state from a claim for attorney’s fees under the Workers’ Compensation Act. See 381 S.W.3d at 529. Thus, Kinnear and Manbeck stand together—and at most—for the proposition that sovereign immunity sometimes does and sometimes does not protect the state from counterclaims for attorney’s fees. See Kinnear, 14 S.W.3d at 299; Manbeck, 381 S.W.3d at 528; see also Sharyland Water Supply Corp. v. City of Alton, 354 S.W.3d 407, 424 (Tex. 2011) (reversing attorney’s fees awarded against a governmental entity when the fees were a “subset” of a “breach-of-contract claim, a claim on which [the claimant could not] recover”). But neither case considers the attorney’s-fees question beyond the statute it addresses.Nor does Reata itself apply to all counterclaims for attorney’s fees. Rather, Reata establishes that sovereign immunity never protects the state from certain qualifying counterclaims when the state does assert certain claims for monetary recovery. But it does not establish the inverse. That is, Reata does not establish that sovereign immunity always protects the state anytime the state does not assert a claim for monetary recovery. In short, Reata does not hold that a monetary claim is a necessary condition for abrogation in every instance.So the question whether sovereign immunity protects the state from Harper’s counterclaim for attorney’s fees under the TCPA is one of first impression. Based on the TCPA’s unique status and on the general principles underlying sovereign immunity, we answer that it does not.As part of its purpose to “safeguard the constitutional rights of persons to . . . participate in government,” the TCPA obligates those who fail to prove a “prima facie case” to pay certain litigation costs. See Tex. Civ. Prac. & Rem. Code §§ 27.002, .005(c), .009. Because the state should not be suing to prevent its own citizens from participating in government—especially when it lacks even a prima facie case against them—and because when it does sue, it risks paying only attorney’s fees (rather than damages or some other uncapped sum), abrogating the state’s sovereign immunity in the TCPA context does not present any grave danger to the public fisc. See, e.g., Tex. Dep’t of Transp. v. Sefzik, 355 S.W.3d 618, 621 (Tex. 2011) (per curiam) (“[T]he doctrine of sovereign immunity originated to protect the public fisc from unforeseen expenditures that could hamper governmental functions . . . .”). And because the TCPA already contains an express exemption for enforcement actions, nor does abrogation pose a threat to the state’s ability to sanction wrongdoers. See Tex. Civ. Prac. & Rem. Code § 27.010(a).Abrogation remains the judiciary’s responsibility. See Reata, 197 S.W.3d at 375. We have considered abrogation for attorney’s fees in other contexts—sometimes applying it, sometimes not—but we have never addressed a counterclaim for attorney’s fees under the TCPA. So while cases like Kinnear and Manbeck are instructive, neither is dispositive. Accordingly, given the TCPA’s unique role in protecting the democratic processes that allow our state to function, today we conclude that sovereign immunity does not protect the state from a counterclaim for attorney’s fees under the TCPA.Because the state was not operating within sovereign immunity’s bounds when it joined Best’s suit, the TCPA allows Harper to recover costs against the state pursuant to the TCPA’s terms. See Tex. Civ. Prac. & Rem. Code § 27.009(a). As is relevant here, these are the TCPA’s requirements with regard to fees, costs, and sanctions:(a) If the court orders dismissal of a legal action under this chapter, the court shall award to the moving party: court costs, reasonable attorney’s fees, and other expenses incurred in defending against the legal action as justice and equity may require; and sanctions against the party who brought the legal action as the court determines sufficient to deter the party who brought the legal action from bringing similar actions described in this chapter. Id.The state points to the limit in section 27.009(a)(2) that courts may only award sanctions against the party who “brought” the action. Id. § 27.009(a)(2). But that limitation appears only in section 27.009(a)(2), which addresses sanctions. See id. Section 27.009(a)(2) does not limit section 27.009(a) as a whole or section 27.009(a)(1), which has its own limit—justice and equity. See id. § 27.009(a)(1). Since no sanctions have yet been awarded, we need not—and do not—address whether the state “brought” this action (as opposed to “joined” it or something else). Instead, we hold that because this is a TCPA action, the state’s sovereign immunity from suit does not protect it from the appellate costs that the court of appeals has already awarded.* * *The court of appeals reversed the trial court’s judgment, holding that the TCPA applies to the state’s removal proceeding and that the state failed to establish a prima facie case for Harper’s removal. See 493 S.W.3d at 111, 116. It then remanded the case “to the trial court for rendition of an order granting Harper’s motion to dismiss and for a determination of Harper’s request for court costs, reasonable attorney’s fees, and sanctions.” Id. at 118. We affirm the court of appeals’ judgment with the modification that Harper was not entitled to dismissal of or attorney’s fees for the state’s allegation that he violated the Open Meetings Act. Because Harper is no longer in office, however, our holding bears only on the trial court’s determination of court costs, reasonable attorney’s fees, and sanctions.Jeffrey V. BrownJusticeOPINION DELIVERED: June 29, 2018