the Dallas Morning News, Inc. and Steve Blow v. John Tatum and Mary Ann Tatum

Jeffrey V. Brown, Justice

Words-so innocent and powerless as they are, as standing in a dictionary, how potent for good and evil they become in the hands of one who knows how to combine them.¹

-Nathaniel Hawthorne

*621In this libel-by-implication case, we must determine whether the defamatory meanings the Tatums allege are capable of arising from the words that Steve Blow combined in a column that The Dallas Morning News published. We conclude that the column is reasonably capable of meaning that the Tatums acted deceptively and that the accusation of deception is reasonably capable of defaming the Tatums. However, as we further conclude that the accusation is an opinion, we reverse the court of appeals' judgment and reinstate the trial court's summary judgment for petitioners Steve Blow and The Dallas Morning News.

I

Background

A. Facts

Paul Tatum was the son of John and Mary Ann Tatum.² At seventeen years old, Paul was a smart, popular, and athletic high-school student. By every indication, he was a talented young man with a bright future. One mid-May evening, Paul, driving alone, crashed his parents' vehicle on his way home from a fast-food run. The vehicle's airbag deployed, and the crash was so severe that investigators later discovered Paul's eyelashes and facial tissue at the scene. The crash's cause has never been conclusively established and no evidence suggests that Paul was intoxicated or otherwise under the influence of any substance when the crash occurred.

Paul found his way home on foot. He began drinking and he called a friend. The phone call indicated to the friend that Paul was behaving erratically. The friend, concerned, traveled to Paul's house to see him in person. The friend found Paul at the Tatums' house in a confused state and holding one of the Tatum family's firearms. The friend left the room where Paul was to report Paul's irrational behavior to the friend's parent, who was waiting in a car outside the Tatums' house. Soon after, the friend heard a gunshot. Paul had killed himself.

In the wake of Paul's death, the Tatums discovered medical literature positing a link between traumatic brain injury and suicide. The Tatums concluded that the car accident caused irrational and suicidal ideations in Paul, which in turn led to his death (whether through an irrational failure to appreciate the risks that accompany handling a firearm or through suicidal desires that led to an intentional, suicidal *622action). Paul's mother, a mental-health professional, had never noticed any suicidal tendencies in Paul. By her account, and by all others, Paul was a normal, healthy, and mentally stable young man. For the Tatums, these observations underscored the plausibility of their theory that Paul's car crash generated a brain injury that led to his suicide.

In addition to establishing a scholarship fund in his name, the Tatums sought to memorialize Paul by writing an obituary, which they published by purchasing space in The Dallas Morning News. The obituary stated that Paul died "as a result of injuries sustained in an automobile accident." The Tatums chose this wording to reflect their conviction that Paul's suicide resulted from suicidal ideation arising from a brain injury rather than from any undiagnosed mental illness. The Dallas Morning News published the obituary on May 21, 2010. More than 1,000 people attended Paul's funeral.

Steve Blow is a columnist for The Dallas Morning News.³ On June 20, 2010-Father's Day, and about one month after Paul's suicide-the paper published a column by Blow entitled "Shrouding Suicide Leaves its Danger Unaddressed."⁴

The column characterized suicide as the "one form of death still considered worthy of deception." While it did not refer to the Tatums by name, it quoted from Paul's obituary and referred to it as "a paid obituary in this newspaper." Although those who knew Paul already knew the truth, the column revealed what the obituary left out: Paul's death "turned out to have been a suicide." After providing another example of an undisclosed suicide, the column went on to lament that "we, as a society, allow suicide to remain cloaked in such secrecy, if not outright deception." The reason we should be more open, according to the column, is that "the secrecy surrounding suicide leaves us greatly underestimating the danger there" and that "averting our eyes from the reality of suicide only puts more lives at risk." The reason we are not open about suicide, the column speculated, is that "we don't talk about the illness that often underlies it-mental illness." Despite these perceived risks, the column also suggested that the lack of openness is understandable. Blow wrote that we should not feel embarrassed by suicide and that "the last thing I want to do is put guilt on the family of suicide victims." The column concluded with an exhortation: "Awareness, frank discussion, timely intervention, treatment-those are the things that save lives. Honesty is the first step."

Blow drafted the column without attempting to contact the Tatums and the paper published it without letting the Tatums know that it was going to print. Those who knew the Tatums immediately recognized that the obituary the column referenced was Paul's.

B. Procedural history

The Tatums filed suit. They alleged libel and libel per se against Blow and the paper. In particular, the Tatums alleged the column defamed them by its "gist." They also brought Deceptive Trade Practices Act claims against the paper. The News filed a motion for traditional and no-evidence summary judgment. The News *623asserted several traditional grounds. Among them were that the column was not reasonably capable of a defamatory meaning and that the column was an opinion. Without specifying why, the trial court granted the News's motion.⁵

The Tatums appealed. The court of appeals affirmed as to the deceptive-trade practices claims, but it reversed and remanded the Tatums' claims that were based on libel and libel per se. 493 S.W.3d 646, 653 (Tex. App.-Dallas 2015). As is especially relevant here, the court of appeals began by asking whether there was a "genuine fact issue regarding whether the column was capable of defaming" the Tatums. Id. at 659. Nowhere in this analysis did the court of appeals discuss the column's gist. Yet the court concluded that "a person of ordinary intelligence could construe the column to suggest that Paul suffered from mental illness and his parents failed to confront it honestly and timely, perhaps missing a chance to save his life." Id. at 661. It further concluded that "[t]his meaning is defamatory because it tends to injure the Tatums' reputations and to expose them to public hatred, contempt, or ridicule." Id.

In the next section, the court analyzed "the column's gist regarding the Tatums." Id. at 662-63 (emphasis added). A reasonable reader, the court held, could conclude that "the column's gist is that the Tatums, as authors of Paul's obituary, wrote a deceptive obituary to keep Paul's suicide a secret and to protect themselves from being seen as having missed the chance to intervene and prevent the suicide." Id. (emphasis added). But see id. at 672 ("We assume without deciding that the defamatory publication in this case generally involved a matter of public concern (preventing suicides) ....").

The court's conclusion regarding the column's gist drove the rest of its analysis. It held the column was not an opinion because "the column's gist that the Tatums were deceptive when they wrote Paul's obituary is sufficiently verifiable to be actionable in defamation." Id. at 668. The News's defenses based on fair comment, official proceedings, truth, substantial truth, actual malice, and negligence fared no better. See id. at 666-67. Thus, the court of appeals rejected every possible ground on which the trial court might have based its grant of summary judgment.

The News petitioned this Court for review. It argues that the court of appeals was wrong on four fronts: the column is not reasonably capable of defamatory meaning; it is non-actionable opinion; it is substantially true; and the court of appeals did not properly analyze actual malice.

II

Law

A. Defamation

Defamation is a tort, the threshold requirement for which is the publication of a false statement of fact to a third party. Exxon Mobil Corp. v. Rincones , 520 S.W.3d 572, 579 (Tex. 2017). The fact must be defamatory concerning the plaintiff, and the publisher must make the statement with the requisite degree of fault. Id. And in some cases, the plaintiff must also prove damages. Id. (citing In re Lipsky , 460 S.W.3d 579, 593 (Tex. 2015) ); see also D Magazine Partners, L.P. v. Rosenthal , 529 S.W.3d 429, 434 (Tex. 2017). Defamation may occur through slander or through libel. Slander is a defamatory statement expressed orally. See *624Randall's Food Mkts., Inc. v. Johnson , 891 S.W.2d 640, 646 (Tex. 1995). By contrast, libel is a defamatory statement expressed in written or other graphic form. See TEX. CIV. PRAC. & REM. CODE § 73.001.

Texas recognizes the common-law rule that defamation is either per se or per quod. See Lipsky , 460 S.W.3d at 596. Defamation per se occurs when a statement is so obviously detrimental to one's good name that a jury may presume general damages, such as for loss of reputation or for mental anguish. Hancock v. Variyam , 400 S.W.3d 59, 63-64 (Tex. 2013). Statements that injure a person in her office, profession, or occupation are typically classified as defamatory per se. Id. at 64. Defamation per quod is simply defamation that is not actionable per se. Lipsky , 460 S.W.3d at 596.

In a defamation case, the threshold question is whether the words used "are reasonably capable of a defamatory meaning." Musser v. Smith Protective Servs., Inc. , 723 S.W.2d 653, 655 (Tex. 1987). In answering this question, the "inquiry is objective, not subjective." New Times, Inc. v. Isaacks , 146 S.W.3d 144, 157 (Tex. 2004). But if the court determines the language is ambiguous, the jury should determine the statement's meaning. See Musser , 723 S.W.2d at 655. If a statement is not verifiable as false, it is not defamatory. Neely v. Wilson , 418 S.W.3d 52, 62 (Tex. 2013) (citing Milkovich v. Lorain Journal Co. , 497 U.S. 1, 21-22, 110 S.Ct. 2695, 111 L.Ed.2d 1 (1990) ). Similarly, even when a statement is verifiable as false, it does not give rise to liability if the "entire context in which it was made" discloses that it is merely an opinion masquerading as a fact. See Bentley v. Bunton , 94 S.W.3d 561, 581 (Tex. 2002) ; see also Isaacks , 146 S.W.3d at 156-57.

Both the U.S. Constitution and the Texas Constitution "robustly protect freedom of speech," Rosenthal , 529 S.W.3d at 431, and the Texas Constitution expressly acknowledges a cause of action for defamation. See Tex. Const. art. I, § 8 ("Every person shall be at liberty to speak, write or publish his opinions on any subject, being responsible for the abuse of that privilege ...."); see also Casso v. Brand , 776 S.W.2d 551, 556 (Tex. 1989). These documents also impose substantive limits on defamation law. See Cain v. Hearst Corp. , 878 S.W.2d 577, 584 (Tex. 1994) ("[T]he Texas Constitution [has] independent vitality from the federal constitution, and [it] impose[s] even higher standards on court orders which restrict the right of free speech."). Among these limits, to avoid the threat to free speech that unrestrained defamation liability poses, the U.S. Constitution "imposes a special responsibility on judges whenever it is claimed that a particular communication is [defamatory]." Bose Corp. v. Consumers Union of U.S., Inc. , 466 U.S. 485, 505, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984). For appellate judges, one of these responsibilities is to comply with the "requirement of independent appellate review" as a matter of "federal constitutional law." Bose , 466 U.S. at 510, 104 S.Ct. 1949 ; see also Doubleday & Co., v. Rogers , 674 S.W.2d 751, 755 (Tex. 1984) ("[T]he first amendment requires the appellate court to independently review the evidence.")

B. Standard of review

We review a denial of summary judgment de novo. See Neely , 418 S.W.3d at 59. In the interest of efficiency, "we consider all grounds presented to the trial court and preserved on appeal." Id. "When reviewing a summary judgment, we take as true all evidence favorable to the nonmovant and we indulge every reasonable inference and resolve any doubts in the nonmovant's favor." Rincones , 520 S.W.3d at 579 (citing *625Valence Operating Co. v. Dorsett , 164 S.W.3d 656, 661 (Tex. 2005) ). A trial court properly grants a defendant's traditional motion for summary judgment "if the defendant disproves at least one element of each of the plaintiff's claims or establishes all elements of an affirmative defense to each claim." Am. Tobacco Co. v. Grinnell , 951 S.W.2d 420, 425 (Tex. 1997) (internal citation omitted). Similarly, it is proper for the trial court to grant a defendant's no-evidence motion for summary judgment if the plaintiff has produced no more than a scintilla of evidence on an essential element of the cause of action, that is, if the plaintiff's evidence does not rise "to a level that would enable reasonable and fair-minded people to differ in their conclusions." Ford Motor Co. v. Ridgway , 135 S.W.3d 598, 600-01 (Tex. 2004) (quoting Merrell Dow Pharm. v. Havner , 953 S.W.2d 706, 711 (Tex. 1997) ).

III

Analysis

A. Is the column reasonably capable of a defamatory meaning?

"Meaning is the life of language." Masson v. New Yorker Magazine, Inc. , 501 U.S. 496, 517, 111 S.Ct. 2419, 115 L.Ed.2d 447 (1991). Thus, the first question in a libel action is whether the words used are "reasonably capable of a defamatory meaning." Musser , 723 S.W.2d at 654. Meaning is a question of law. Id. at 654. In answering it, the "inquiry is objective, not subjective." Isaacks , 146 S.W.3d at 157. We note that the question involves two independent steps. The first is to determine whether the meaning the plaintiff alleges is reasonably capable of arising from the text of which the plaintiff complains. See, e.g. , Rosenthal , 529 S.W.3d at 437-41 (first analyzing an article's gist, then discussing whether the gist was defamatory). The second step is to answer whether the meaning-if it is reasonably capable of arising from the text-is reasonably capable of defaming the plaintiff. See id.

1. What does the column mean?

a) Law

In the typical defamation case, the determination of what a publication means involves little beyond browsing the publication's relevant portions in search of the defamatory content of which the plaintiff complains. That is, defamatory meanings are ordinarily transmitted the same way that other meanings are-explicitly. But this is not the typical defamation case. Rather, the Tatums allege that the column defames them by its "gist." This allegation requires us to consider the history of our cases addressing "gist."

(1) Common law

Texas cases recognize a distinction between a statement that is defamatory by its text alone and a statement that is defamatory only by reason of "extrinsic evidence" and "explanatory circumstances." Moore v. Waldrop , 166 S.W.3d 380, 385 (Tex. App.-Waco 2005, no pet.) ; see also Gartman v. Hedgpeth , 138 Tex. 73, 157 S.W.2d 139, 141-43 (1941) (discussing the distinction). The common law employed the term "defamation per se" to refer to the first type of statement-one defamatory by its text alone. See Defamation Per Se , BLACK'S LAW DICTIONARY (10th ed. 2014) (defining as defamatory "per se" a "statement that is defamatory in and of itself"). Similarly, at common law, "defamation per quod" referred to a statement whose defamatory meaning required reference to extrinsic facts. See Defamation Per Quod , BLACK'S LAW DICTIONARY (10th ed. 2014) (defining as defamatory "per quod" a statement whose defamatory meaning is "not apparent but [must be] proved by *626extrinsic evidence showing its defamatory meaning").

However, this distinction "is not the same as that between defamation which is actionable of itself and that which requires proof of special damage." W. PAGE KEETON ET AL. , PROSSER AND KEETON ON TORTS § 111, at 782 (5th ed. 1984). Despite the difference, we have also characterized as "defamation per se" statements that are "so obviously hurtful to a plaintiff's reputation that the jury may presume general damages, including for loss of reputation and mental anguish." Hancock , 400 S.W.3d at 63-64. In this usage, "[a] statement that injures a person in her office, profession, or occupation is typically classified as defamatory per se." Id. at 64. With regard to special damages, "[d]efamation per quod is defamation that is not actionable per se." Lipsky , 460 S.W.3d at 596. Unfortunately, "the terms 'defamation per se' and 'defamation per quod' are used indiscriminately in both senses." KEETON ET AL. supra , § 111, at 782 n.41.

Thus, for clarity, we introduce the following terms. To begin, "textual defamation" refers to the common-law concept of defamation per se, that is, defamation that arises from the statement's text without reference to any extrinsic evidence. On the other hand, "extrinsic defamation" refers to the common-law concept of defamation per quod, which is to say, defamation that does require reference to extrinsic circumstances. Moreover, as we noted in In re Lipsky , "Texas has not abandoned t[he] distinction" between defamation so harmful that the jury may presume general damages and defamation that requires the plaintiff to prove special damages. 460 S.W.3d at 596 n.13. Thus, we ratify the continued usage of (and distinction between) "defamation per se" and "defamation per quod" as used in relation to special damages. See id. ; Hancock , 400 S.W.3d at 63-64. This case concerns, in part, the distinction between textual defamation and extrinsic defamation.

Extrinsic defamation occurs when a statement whose textual meaning is innocent becomes defamatory when considered in light of "other facts and circumstances sufficiently expressed before" or otherwise known to the reader. See Snider v. Leatherwood , 49 S.W.2d 1107, 1109 (Tex. Civ. App.-Eastland 1932, writ dism'd w.o.j.). The requirements for proving an extrinsic-defamation case-including the torts professor's perennial favorites of innuendo, inducement, and colloquium-are somewhat technical. Only two are of interest here. First, it must be remembered that an extrinsically defamatory statement requires extrinsic evidence to be defamatory at all. See id. Second, plaintiffs relying on extrinsic defamation must assert as much in their petitions to present the theory at trial. See Billington v. Hous. Fire & Cas. Ins. , 226 S.W.2d 494, 497 (Tex. Civ. App.-Fort Worth 1950, no writ).

Textual defamation occurs when a statement's defamatory meaning arises from the words of the statement itself, without reference to any extrinsic evidence. See Defamation Per Se , BLACK'S LAW DICTIONARY (10th ed. 2014) (defining as defamatory "per se" a "statement that is defamatory in and of itself").⁶ The ordinary *627textual defamation involves a statement that is explicitly defamatory. Explicit textual-defamation cases share two common attributes. First, none necessarily involve any extrinsic evidence. Thus, none necessarily involve extrinsic defamation. Second, the defamatory statement's literal text and its communicative content align-what the statement says and what the statement communicates are the same. In other words, the defamation is both textual and explicit . As discussed below, our cases also recognize that defamation can be textual and implicit . See generally Turner v. KTRK Television, Inc. , 38 S.W.3d 103 (Tex. 2000). When a publication's text implicitly communicates a defamatory statement, we refer to the plaintiff's theory as "defamation by implication."

(2) Defamation by implication

In a defamation-by-implication case, the defamatory meaning arises from the statement's text, but it does so implicitly. Defamation by implication is not the same thing as textual defamation. Rather, it is a subset of textual defamation. That is, if the defamation is textual, it may be either implicit or explicit. The difference is important because the precepts that apply to construing explicit meanings do not necessarily apply with the same force or in the same manner when construing implicit meanings. And, importantly, nor is implicit textual defamation the same thing as extrinsic defamation, although parties and courts have often confused the two.⁷ Finally, defamation by implication is not the same thing as defamation by innuendo. The dividing line is the same as that between extrinsic defamation and textual defamation generally: the first requires extrinsic evidence, but the second arises solely from a statement's text. The difference is important because plaintiffs relying on extrinsic defamation must say so in their pleadings, whereas plaintiffs relying on textual defamation need not. See Billington , 226 S.W.2d at 497.

Turner v. KTRK Television, Inc. is our foundational case recognizing defamation by implication See generally , 38 S.W.3d at 113. There, we held "that a plaintiff can bring a claim for defamation when discrete facts, literally or substantially true, are published in such a way that they create a substantially false and defamatory impression by omitting material facts or juxtaposing facts in a misleading way." Id. at 115. In particular, Turner focused on the "converse of the substantial truth doctrine." See id. (citing McIlvain v. Jacobs , 794 S.W.2d 14, 16 (Tex.1990) ). The converse of that doctrine, we held, is that a defendant may be liable for a "publication that gets the details right but fails to put them in the proper context and thereby gets the story's 'gist' wrong." See id. Although Turner used the word "gist," commentators were relatively quick to point out that the decision actually addressed libel by implication.⁸

*628The issue in Turner was whether a plaintiff could bring a "gist" claim based on "the entirety of a publication and not merely on individual statements." Id. We answered that question in the affirmative, see id. , and we have maintained the same approach in subsequent cases.⁹ Indeed, just last term we held that "[i]n making the initial determination of whether a publication is capable of a defamatory meaning, we examine its 'gist.' That is, we construe the publication 'as a whole ....' " Rosenthal , 529 S.W.3d at 434 (citations omitted). Thus, Turner and its progeny recognize that a plaintiff can rely on an entire publication to prove that a defendant has implicitly communicated a defamatory statement.

However, and of special importance in this case, there is no reason that implicit meanings must arise only from an entire publication or not at all. Our decision in Rosenthal is illustrative. There, the plaintiff brought a defamation claim based on an article titled "THE PARK CITIES WELFARE QUEEN." Id. at 431. The article was

published under the heading "CRIME" and [was] accompanied by Rosenthal's mug shot from a prior unrelated charge. The article state[d] under the aforementioned "Welfare Queen" title that Rosenthal, described as a "University Park mom," ha[d] "figured out how to get food stamps while living in the lap of luxury." It then invite[d] the reader to see how Rosenthal "pulls it off" despite the assumption that one living in the affluent Park Cities would "never qualify."

Id. at 437. The article's language would not necessarily have been any less defamatory if it had been appended to a larger piece discussing, for example, the biographies of various individual Park Cities homeowners. Of course, the larger context would have been relevant for construing what the article meant. But the language would not have ceased being defamatory solely by being published within a larger article. In recognizing defamation-by-"gist" in Turner , we also recognized the broader category of defamation by implication.

Thus, we acknowledge that in a textual-defamation case, a plaintiff may allege that meaning arises in one of three ways. First, meaning may arise explicitly. See Bentley , 94 S.W.3d at 569 ("[Y]'all are corrupt, y'all are the criminals, [and] y'all are the ones that oughta be in jail."). Second, meaning may arise implicitly as a result of the article's entire gist. See Rosenthal , 529 S.W.3d at 439 ("[E]valuating the article 'as a whole ...' [t]he article's gist is that ...." (citation omitted) ). Third, as in this case, the plaintiff may allege that the defamatory meaning arises implicitly from a distinct portion of the article rather than from the article's as-a-whole gist. As other courts have recognized, the distinction between "as-a-whole" gist and "partial" implication is important. See, e.g. , *629Sassone v. Elder , 626 So.2d 345, 354 (La. 1993) ("[P]laintiffs prove that the alleged implication is the principal inference a reasonable reader or viewer will draw ...."); see also C. Thomas Dienes & Lee Levine, Implied Libel, Defamatory Meaning, and State of Mind: The Promise of New York Times Co. v. Sullivan , 78 IOWA L. REV. 237, 289 (1993) ("The distinction between inferences that may reasonably be drawn from a publication, on the one hand, and the meaning a reasonable reader would ascribe to the publication, on the other, is crucial ....").

Accordingly, we use the following terms. "Gist" refers to a publication or broadcast's main theme, central idea, thesis, or essence. See THE AMERICAN HERITAGE DICTIONARY OF THE ENGLISH LANGUAGE 745 (4th ed. 2000) (defining "gist" as "[t]he central idea; the essence"); WEBSTER'S THIRD NEW INTERNATIONAL DICTIONARY 959 (2002) (defining "gist" as "the main point or material part ... the pith of a matter"); Gist , BLACK'S LAW DICTIONARY (10th ed. 2014) (defining gist as "[t]he main point"). Thus, we use "gist" in its colloquial sense. In this usage, publications and broadcasts typically have a single gist.

"Implication," on the other hand, refers to the inferential, illative, suggestive, or deductive meanings that may emerge from a publication or broadcast's discrete parts. Implication includes necessary logical entailments as well as meanings that are merely suggested. Thus, in the sentence "John took some of the candy," implication includes both the logical entailment that John took at least one piece of the candy as well as the suggestion that John did not take every piece of the candy. "Defamation by implication," as a subtype of textual defamation, covers both "gist" and "implication."

The difference between gist and implication is especially important in two contexts. The first relates to the substantial-truth doctrine. "A broadcast with specific statements that err in the details but that correctly convey the gist of a story is substantially true." Neely , 418 S.W.3d at 63-64. If the plaintiff demonstrates substantial truth, the doctrine "precludes liability for a publication that correctly conveys a story's 'gist' or 'sting' although erring in the details ...." Turner , 38 S.W.3d at 115. We have never held, nor do we today, that a true implication-as opposed to a true gist-can save a defendant from liability for publishing an otherwise factually defamatory statement. Second, the difference between gist and implication matters when considering the requirements that the U.S. Constitution imposes on defamation law.

(3) Construing implications

By nature, defamations by implication require construction. Under Musser v. Smith Protective Services, Inc , the standard for construing defamatory meaning generally is whether the publication is "reasonably capable" of defamatory meaning. 723 S.W.2d at 655. Defamation by implication is simply a subtype of textual defamation, which is itself one way that a publisher can communicate a defamatory meaning. Thus, to determine whether a defamation by implication has occurred, the question is the same as it is for defamatory content generally: is the publication "reasonably capable" of communicating the defamatory statement? But to whose "reason" does "reasonably capable" refer?

Sometimes we have said that "reasonably capable" requires us to construe a publication "based upon how a person of ordinary intelligence would perceive it." Rosenthal , 529 S.W.3d at 434 (emphasis added) (internal quotation omitted); see also Bentley , 94 S.W.3d at 579 ; Turner , 38 S.W.3d at 114. The "would" standard recognizes *630that gist, in particular, is the type of implication that no reasonable reader would fail to notice. But the "would" standard falls short when applied to implications. Not all readers will pick up on all reasonable implications in all publications. In fact, it seems apparent that no reader would internalize every implication from a single article-or even a single sentence.

For example, what implications would a reasonable reader draw from the following sentence, which opens one of Virginia Woolf's best-known novels: "Mrs. Dalloway said she would buy the flowers herself." VIRGINIA WOOLF, MRS. DALLOWAY 3 (1925). The gist is that Mrs. Dalloway plans to buy flowers. But what are the implications? One implication is that someone else was supposed to do the flower-buying. Another implication, apparent from the fact that Mrs. Dalloway "said" she would buy the flowers, is that she is irritated by this other person's failure to purchase the flowers. Although some of these implications may be reasonable, not all reasonable readers would consciously internalize them. Some reasonable readers would notice one implication, while other reasonable readers would notice another. And some reasonable readers would notice no implications. These observations illustrate that the "would" standard, when applied to implications, is overly subjective. The reason is that when applying the "would" standard to implications, the court necessarily must prefer what one reader would discern over what another reader would discern. Since not all reasonable readers "would" perceive all implications, "would" does not capture the entirety of the "reasonably capable" standard.

In other cases, we have said the meaning the plaintiff proposes fails the "reasonably capable" standard only when no "person of ordinary intelligence could conclude" that the publication conveys the meaning alleged. Neely , 418 S.W.3d at 76 (emphasis added); see also Toledo , 492 S.W.3d at 722 (Boyd, J., dissenting) ("[T]he question for us is not whether an ordinary viewer would have understood the broadcasts' gist to be false or defamatory, but whether a 'reasonable jury could have found the broadcast to be false or defamatory.' " (citations omitted) ). The "could" standard recognizes that publications of any length will communicate more than one implication and that not all reasonable readers will notice every one. Thus, the "could" standard avoids one of the problems that the "would" standard creates. But "could" also creates its own problems.

To return to the example above, is Mrs. Dalloway speaking to another person? Is it a servant? Was it the servant's job to get the flowers? Has Mrs. Dalloway implied that the servant is doing her job poorly? Does the servant have a cause of action for slander, or even slander per se, against Mrs. Dalloway? From the nine words that comprise the sentence, any lawyer might construct a chain of implications that required answering each question "yes" and demonstrated that some reader "could" construe the sentence as defamatory. And with only "could" at its disposal, no court would have any choice but to pass the question on to the jury.

Neither "would" nor "could"-to the extent that the two words are distinguishable, which is not always the case-captures the full scope of the "reasonably capable" standard that governs defamation by implication. "Would" applies in gist cases, as we have repeatedly emphasized, and thus it accurately states one condition that, if present, is sufficient for implicit meaning. But in contrast to a publication's single gist, no reasonable reader "would" absorb all implications that a publication puts forth. "Could," on the other hand, *631recognizes that meaning can be transmitted in many ways and that reasonable readers will read some things differently. In this way, "could" states a condition that is necessary for the transmission of implicit meaning. But as the sentence from Mrs. Dalloway illustrates, a reasonable reader "could," without departing from the constraints that pure logic imposes, follow or construct hyper-attenuated inferential chains that stretch beyond the realm of ordinary semantic meaning. Thus, while these standards capture part of the judicial task, they do not capture all of it.

Instead, when the plaintiff claims defamation by implication, the judicial task is to determine whether the meaning the plaintiff alleges arises from an objectively reasonable reading. See Isaacks , 146 S.W.3d at 157 (explaining that "the hypothetical reasonable reader" is the standard by which to judge a publication's meaning (emphasis added) ). "The appropriate inquiry is objective, not subjective." Id. The objectively reasonable reader has made little appearance in our cases discussing gist. The reason, as discussed above, is that publications usually have a single gist that no reasonable reader could fail to notice. Thus, in gist cases, the "would" standard renders the objectively reasonable reader redundant.

But when discrete implications are at issue, the objectively reasonable reader reappears to aid the court in determining what meaning has been communicated. The reason for the sudden reappearance is that an objectively reasonable reading encompasses many more implications than any single reasonable reader necessarily "would" understand a publication to convey. Even reasonable readers do not internalize every single implication that a publication conveys. That is, "[i]ntelligent, well-read people act unreasonably from time to time, whereas the hypothetical reasonable reader, for purposes of defamation law, does not." Id. at 158. Similarly, the objectively reasonable reader notices some-but not all-of the implications that an ordinary reader could draw from a publication's text. So in an implication case, the judicial role is not to map out every single implication that a publication is capable of supporting. Rather, the judge's task is to determine whether the implication the plaintiff alleges is among the implications that the objectively reasonable reader would draw.

Making this determination is a quintessentially judicial task. It involves "a single objective inquiry: whether the [publication] can be reasonably understood as stating" the meaning the plaintiff proposes. Id. The objectively reasonable reader aids in the inquiry, as a "prototype ... who exercises care and prudence, but not omniscience, when evaluating allegedly defamatory communications." Id. at 157. He does not place "overwhelming emphasis on a[ny] single term." See Rosenthal , 529 S.W.3d at 437. Nor does he "focus on individual statements" to the exclusion of the entire publication. See id. The objectively reasonable reader internalizes all of a publication's reasonable implications. When doing so, he considers inferential meaning carefully, but not exhaustively. He performs analysis, but not exegesis.

(4) Meaning's limits

Meanings sometimes terminate in ambiguities. And because defamation involves meaning, ambiguity is often an issue in defamation cases. "Only when the court determines the language is ambiguous or of doubtful import should the jury ... determine the statement's meaning ...." Musser , 723 S.W.2d at 655 ; see also Hancock , 400 S.W.3d at 66 ; Gartman , 157 S.W.2d at 141. And in the very next sentence Musser states that "[t]he threshold *632question then, which is a question of law, is whether [the defendant's] statements are reasonably capable of a defamatory meaning." Musser , 723 S.W.2d at 655. Thus, whether "language is ambiguous" and whether the same language is "reasonably capable of defamatory meaning" are not technically the same question. See, e.g. , Toledo , 492 S.W.3d at 722 (stating both rules); accord Hancock , 400 S.W.3d at 66 ; Carr v. Brasher , 776 S.W.2d 567, 570 (Tex. 1989).

Questions of meaning and ambiguity recur in three different types. First, if a court determines that a statement is capable of defamatory meaning and only defamatory meaning-that it is unambiguous-then the jury plays no role in determining the statement's meaning. See Hancock , 400 S.W.3d at 66 ; see also Brasher , 776 S.W.2d at 570. Second, courts sometimes determine that a statement is capable of at least one defamatory and at least one non-defamatory meaning. When that occurs, "it is for the jury to determine whether the defamatory sense was the one conveyed." KEETON ET AL. , supra , § 111, at 781; see also Hancock , 400 S.W.3d at 66. Third, a court may determine that the statement is not capable of any defamatory meanings. "If the statement is not reasonably capable of a defamatory meaning, the statement is not defamatory as a matter of law and the claim fails." Hancock , 400 S.W.3d at 66. Importantly, when the court makes this determination, the plaintiff cannot present the question of meaning to the jury. This remains true even if the statement is otherwise ambiguous.

Our point in reciting these black-letter applications of our defamation law is to emphasize that the analytical framework for considering ambiguities does not evaporate simply because the plaintiff alleges an implicit meaning. Put differently, a plaintiff who alleges defamation by implication has not thereby alleged an ambiguity. At least, not necessarily. Of course, implications can be ambiguous. They can be ambiguous in what they convey, just like explicit denotative meaning. But unlike explicit meaning, it can also be uncertain whether a certain implication arises from a statement at all. Thus, one question is whether a publication implicitly communicates a certain statement-e.g., that "Bob was at the bank." The second question is what the statement means-was Bob at the river bank? Or was he at the First National Bank? Ambiguity sometimes prevents a court from answering either question. But it does not always prevent an answer. The same rule that allows courts to determine meaning as a matter of law allows them to determine communicative content as a matter of law.

The U.S. and Texas constitutions also limit defamation law. See Bose , 466 U.S. at 510, 104 S.Ct. 1949 (requiring "independent appellate review"); Doubleday , 674 S.W.2d at 751 (recognizing Bose in Texas); see also Rosenthal , 529 S.W.3d at 431 (discussing the constitutional limits); accord Cain , 878 S.W.2d at 584 ; Brand , 776 S.W.2d at 556. Accordingly, answering whether a statement is "reasonably capable of" a certain meaning does not end our inquiry. Instead, upon answering that question in the affirmative, we must further consider whether our answer will lead publishers to curtail protected speech in an attempt to "steer wider of the unlawful zone" of unprotected speech. Time, Inc. v. Hill , 385 U.S. 374, 389, 87 S.Ct. 534, 17 L.Ed.2d 456 (1967). In other words, our decision must not exert too great a "chilling effect" on First Amendment activities.

The potential chilling effect is especially strong in defamation-by-implication cases. Unlike explicit statements, publishers cannot be expected to foresee every implication that may reasonably arise from *633a certain publication. To avoid this chilling effect, the First Amendment "imposes a special responsibility on judges whenever it is claimed that a particular communication is [defamatory]." Bose , 466 U.S. at 505, 104 S.Ct. 1949. For appellate judges, one of these responsibilities is to comply with the "requirement of independent appellate review reiterated" in New York Times v. Sullivan as a matter of "federal constitutional law." Id. at 510, 104 S.Ct. 1949. Although Sullivan emphasized the "actual malice" requirement that applies when the plaintiff, defendant, or subject matter are sufficiently "public," see generally 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), we recognize that its reasoning extends to the First Amendment concerns that defamation by implication raises.

The Constitution requires protection beyond that which the "objectively reasonable reader" standard provides. But what level of protection? And by what means?

One option would be to leave the issue for a jury to decide. However, "[p]roviding triers of fact with a general description of the type of communication whose content is unworthy of protection has not, in and of itself, served sufficiently to narrow the category, nor served to eliminate the danger that decisions by triers of fact may inhibit the expression of protected ideas." Id. at 505, 104 S.Ct. 1949 ; see also Ollman v. Evans , 750 F.2d 970, 997 (D.C. Cir. 1984) (Bork, J., concurring) ("The only solution to the problem libel actions pose would appear to be close judicial scrutiny to ensure that cases about types of speech and writing essential to a vigorous first amendment do not reach the jury."). Since the determination whether a publication is "reasonably capable" of a given meaning involves a textual analysis rather than a credibility determination, displacing the jury does not present any grave danger to due process. Thus, as the U.S. Supreme Court has acknowledged many times, it is consonant with our nation's heritage to recognize a rule requiring judges to answer some of the factual questions that defamation cases present.

For a court to subject a publisher to liability for defamation by implication, the "plaintiff must make an especially rigorous showing" of the publication's defamatory meaning. Chapin v. Knight-Ridder, Inc. , 993 F.2d 1087, 1092-93 (4th Cir. 1993). Under this standard, we must look to the judge rather than the jury to prevent the chilling effect, and the judge's review must be "especially rigorous." Id. But what does that standard entail? In this section's remainder we answer the question, first by examining the methods by which other jurisdictions have implemented a heightened standard of review in defamation-by-implication cases. Next, we lay out our reasons for adopting the rule we do today. Finally, we consider how the rule's application varies within the defamation-by-implication contexts of gist and individual implication.

One way of cabining the dangers that defamation by implication poses would be to subsume the constitutional question within the question of meaning. However, we see no reason for thinking that either the U.S. Constitution or the Texas Constitution has anything to do with what a word in its everyday usage means . Each, of course, has a great deal to say about a statement's legal effect-does it expose the publisher to liability? is it obscene?-but semantic meaning and legal effect are different inquiries. These considerations persuade us that asking what a statement means is a different question from asking whether the law will punish the speaker for saying it. Of course, in practice the two inquiries may take place concurrently. We see no problem with that, but there remains a discernable difference between *634whether a restriction on meaning arises from the particulars of English usage or from the Constitution. We cannot solve the constitutional challenges that the tort of defamation by implication presents simply by heightening our standard of meaning. Doing so would be to swim against the current of our traditional jurisprudence that favors "plain meaning." Consequently, we reject a heightened standard of "meaning" as a workable limit on the chilling effect that defamation by implication poses.

A second category of protection disallows defamation by implication, whether altogether or in certain contexts. Some states have taken this approach. See Sassone , 626 So.2d at 354 ; Diesen v. Hessburg , 455 N.W.2d 446, 451 (Minn. 1990). Our decision in Turner holds that a public figure can "bring a claim for defamation when discrete facts, literally or substantially true, are published in such a way that they create a substantially false and defamatory impression by omitting material facts or juxtaposing facts in a misleading way." Turner , 38 S.W.3d at 115. Our cases allow public figures-and by extension, private figures, see Rosenthal , 529 S.W.3d at 434 -to bring cases alleging defamation by implication. These precedents prevent us from relying on wholesale rejection of defamation by implication to protect the freedoms that the First Amendment enshrines.

Still other courts have taken a third path by suggesting that defamatory implications might presumptively constitute opinion in some contexts. See, e.g. , Janklow v. Newsweek, Inc. , 788 F.2d 1300, 1303 (8th Cir. 1986). We reject the view that implications are opinions, either necessarily or presumptively. Publishers cannot avoid liability for defamatory statements simply by couching their implications within a subjective opinion. See Milkovich , 497 U.S. at 19, 110 S.Ct. 2695. Thus, after the U.S. Supreme Court's landmark decision in Milkovich v. Lorain Journal Co. , the opinion inquiry seeks to ascertain whether a statement is "verifiable," not whether it manifests a personal view. See Neely , 418 S.W.3d at 62. But no court can decide whether a statement is verifiable until the court decides what the statement is -that is, until it conducts an inquiry into the publication's meaning. Of course, implications may frequently turn out to be non-verifiable opinions, but we disagree that implications are presumptively opinion simply by virtue of being implicit. So we see little hope that asking a court to decide from the outset whether a statement is an opinion will limit the number of defamation-by-implication claims that reach a jury.

A fourth and final limit is to rely on or adjust the culpability standards that Sullivan lays out. See 376 U.S. at 280, 84 S.Ct. 710. With regard to public-figure plaintiffs, we note (without adopting) the view in other courts that a defendant cannot act with actual knowledge of or reckless disregard for a statement's falsity if the defendant has no knowledge (either actual or constructive) that the publication communicates the statement at issue. See, e.g. , Newton v. Nat'l Broad. Co. , 930 F.2d 662, 681 (9th Cir. 1990). When the plaintiff is a private figure, the relevant inquiry is whether the defendant acted negligently. See Neely , 418 S.W.3d at 61. But if a statement is defamatory, then it is "virtually inevitable that a jury will return a verdict that the publisher was negligent in not ascertaining the truth of the defamatory character of the statement." Kelley & Zansberg, supra , at 5. We do not think that the defendant's culpability presents the right implement for curtailing the kinds of defamation-by-implication claims that most injure public discourse. A subjective inquiry into whether a defendant *635"knew" or "intended" a certain meaning will unquestionably lead to exactly the kind of lengthy litigation and burdensome discovery that Sullivan and its progeny indicate ought to be avoided. Thus, we decline to recognize "culpability" as a limit on our meaning inquiry.

In place of these tests, we believe the D.C. Circuit was correct when it stated the following limit on the inquiry into meaning:

[I]f a communication, viewed in its entire context, merely conveys materially true facts from which a defamatory inference can reasonably be drawn, the libel is not established. But if the communication, by the particular manner or language in which the true facts are conveyed, supplies additional, affirmative evidence suggesting that the defendant intends or endorses the defamatory inference, the communication will be deemed capable of bearing that meaning.

White v. Fraternal Order of Police , 909 F.2d 512, 520 (D.C. Cir. 1990) ; see also Dodds v. Am. Broad. Co. , 145 F.3d 1053, 1063-64 (9th Cir. 1998) ; Chapin , 993 F.2d at 1093 ; Vinas v. Chubb Corp. , 499 F.Supp.2d 427, 437 (S.D.N.Y. 2007). Thus, a plaintiff who seeks to recover based on a defamatory implication-whether a gist or a discrete implication-must point to "additional, affirmative evidence" within the publication itself that suggests the defendant "intends or endorses the defamatory inference." White , 909 F.2d at 520 (emphasis omitted). A few of the rule's aspects bear emphasizing.

First, the evidence of intent must arise from the publication itself. In considering whether the publication demonstrates such an intent, the court must, as always, "evaluate the publication as a whole rather than focus on individual statements." Rosenthal , 529 S.W.3d at 437. Of the myriad considerations that exist beyond this long-standing guidepost, many are only relevant depending on a publication's case-specific context. But among them are at least the following questions. Does the publication "clearly disclose[ ] the factual bases for" the statements it impliedly asserts? See Biospherics, Inc. v. Forbes, Inc. , 151 F.3d 180, 185 (4th Cir. 1998). Does the allegedly defamatory implication align or conflict with the article's explicit statements? See, e.g. , Wyo. Corp. Servs. v. CNBC, LLC , 32 F.Supp.3d 1177, 1189 (D. Wyo. 2014). Does the publication accuse the plaintiff in a defamatory manner as opposed to simply reciting that others have accused the plaintiff of the same conduct? See, e.g. , McIlvain , 794 S.W.2d at 15. Does the publication report separate "sets of facts," or does it "link[ ] the key statements together"? See, e.g. , Biro v. Conde Nast , 883 F.Supp.2d 441, 467 (S.D.N.Y. 2012). And does the publication "specifically include[ ] facts that negate the implications that [the defendant] conjures up." Deripaska v. Associated Press , 282 F.Supp.3d 133, 148 (D.D.C. 2017), appeal dismissed per stipulation , No. 17-7164, 2017 WL 6553388 (D.C. Cir. Dec. 8, 2017).

Second, in consonance with our precedent and in accord with the judiciary's traditional role when considering plain meaning, the intent or endorsement inquiry "is objective, not subjective." See Isaacks , 146 S.W.3d at 157. Objectivity is one feature that distinguishes this limit from the Sullivan tests that address culpability. See, e.g. , Stepanov v. Dow Jones & Co. , 120 A.D.3d 28, 987 N.Y.S.2d 37, 44 (2014) (noting that actual malice and textually demonstrated intent are "two entirely separate issues"). If the publication itself indicates that the defendant intended to communicate a certain meaning, it is not relevant (at this stage) that the defendant did not in fact intend to communicate that *636meaning. Similarly, the defendant's subjective views about whether a statement is defamatory have no relevance at this stage. By the same token, a defendant will not be subject to liability for subjectively intending to convey a defamatory meaning that the publication's text does not actually support. In either case, the question is whether the publication indicates by its plain language that the publisher intended to convey the meaning that the plaintiff alleges.

Third, the rule may vary in application depending on the type of defamation that the plaintiff alleges. It does not apply in cases of explicit defamation because when the defendant speaks explicitly, the court indulges the presumption that the defendant intended the communicatory content that he conveyed. In a gist case, the court must "construe the publication 'as a whole in light of the surrounding circumstances based upon how a person of ordinary intelligence would perceive it.' " Rosenthal , 529 S.W.3d at 434. Under the "would" standard, courts are usually able to determine a publication's gist as a matter of law. A gist case is similar to an explicit-meaning case in that the very fact of the gist's (or meaning's) existence is presumptive evidence that the publisher intended to convey the relevant meaning. Thus, it will usually be the case that if a meaning is reasonably capable of being communicated from the gist as a whole, the fact that the gist arises will be sufficient textual evidence that the publisher meant to communicate it.

Finally, in a discrete-implication case, it becomes especially relevant for the court to apply the requirement that the publication's text demonstrates the publisher's intent to convey the meaning the plaintiff alleges. In applying the requirement, courts must bear its origin in mind. The especially rigorous review that the requirement implements is merely a reflection of the "underlying principle" that obligates "judges to decide when allowing a case to go to a jury would, in the totality of the circumstances, endanger first amendment freedoms." Ollman , 750 F.2d at 1006 (Bork, J., concurring).

b) Analysis

At the time of summary judgment, the Tatums' live petition alleged that the column defamed them by implicitly communicating the following "gist":

[The Tatums] created a red herring in the obituary by discussing a car crash in order to conceal the fact that Paul's untreated mental illness-ignored by Plaintiffs-resulted in a suicide that Plaintiffs cannot come to terms with. Defendants led their readers to believe it is people like Plaintiffs-and their alleged inability to accept that their loved ones suffer from mental illness-who perpetuate and exacerbate the problems of mental illness, depression, and suicide.

From this paragraph we discern that the Tatums construe the column to mean that:

• The Tatums acted deceptively in publishing the obituary;

• Paul had a mental illness, which the Tatums ignored and which led to Paul's suicide; and

• The Tatums' deception perpetuates and exacerbates the problem of suicide in others.

None of these meanings appear in the column's explicit text. Nor do they depend on any extrinsic evidence. Thus, while the Tatums allege a textual defamation, their claim rests on defamation by implication rather than on explicit meaning.

The column's gist has nothing to do with the Tatums. Rather, the column's gist is that our society ought to be more forthcoming *637about suicide and that by failing to do so, our society is making the problem of suicide worse, not better. So none of the meanings the Tatums allege arise from the column's gist.

As to the first meaning the Tatums allege, we agree that the column's text supports the discrete implication that the Tatums acted deceptively. The standard is whether an objectively reasonable reader would draw the implication that the Tatums allege. Here, the gist of Blow's column is that bereaved families often do society a disservice by failing to explicitly mention when suicide is the cause of death. Blow holds up the Tatums as an example of the very phenomenon that his column seeks to discourage. Blow would have no reason to mention the Tatums' obituary except to support his point that suicide often goes undiscussed. The objectively reasonable reader seeks to place every word and paragraph in context and to understand the relation that a publication's subparts bear to its whole. Here, an objectively reasonable reading must end with the conclusion that Blow points to the Tatums as one illustration of his thesis that suicide is often "shrouded in secrecy." Simply put, he had no other reason for including them in the column. For the same reason, we conclude that the publication's text objectively demonstrates an intent to convey that the Tatums were deceptive.

But we do not agree that the second and third meanings the Tatums allege are implications that an objectively reasonable reader would draw.

The second alleged meaning rests on the premise that the column means that Paul had a mental illness. We do not agree that the column conveys that meaning. Though the column does say that "mental illness" "often" underlies suicide, the column does so immediately after citing the statistic that suicide is "the third-leading cause of death among young people." The author's use of the word "often" means the column does not logically entail that all suicides are the result of mental illness. And we think the space between the discussion of the Tatums and the discussion of mental illness negates the inferential construction that the Tatums allege-especially since the reference to mental illness follows a citation to a population-level statistic rather than the example paragraphs. But even if we agreed that the column implies that Paul had a mental illness, we could not agree that the column communicates the further implication that the Tatums ignored it or that their treatment of Paul is what led to his suicide. Thus, we conclude that the second meaning the Tatums allege does not arise from an objectively reasonable reading of the column.

Nor does their third. The column declares that "the last thing I want to do is put guilt on the family of suicide victims." An objectively reasonable reader must conclude that the column is about our society as a whole, not about the Tatums in particular. Blow wrote the column to affect future conduct, not to direct blame at any particular family (including the Tatums) for past conduct.

2. Is the meaning defamatory?

Because the column is "reasonably capable" of communicating the meaning that the Tatums were deceptive, the next question is whether that meaning is "reasonably capable" of defaming the Tatums. See Musser , 723 S.W.2d at 655. We conclude that it is.

In Texas, a statement is defamatory libel by statute if it "tends to injure a living person's reputation and thereby expose the person to public hatred, contempt or ridicule, or financial injury or to impeach any person's honesty, *638integrity, virtue, or reputation." TEX. CIV. PRAC. & REM. CODE § 73.001. In addition, under our state's common law, a statement is defamatory per se when it is "so obviously harmful that general damages, such as mental anguish and loss of reputation, are presumed." Lipsky , 460 S.W.3d at 596 ; see also Hancock , 400 S.W.3d at 63. For example, "[a]ccusing someone of a crime, of having a foul or loathsome disease, or of engaging in serious sexual misconduct" constitutes defamation per se. Lipsky , 460 S.W.3d at 596 ; see also Moore , 166 S.W.3d at 384. Moreover, "[r]emarks that adversely reflect on a person's fitness to conduct his or her business or trade are also deemed defamatory per se." Lipsky , 460 S.W.3d at 596.

We agree with the Tatums and with the court of appeals that the column's accusation of deception is "reasonably capable" of injuring the Tatums' standing in the community. In accusing the Tatums of deception, the column is reasonably capable of impeaching the Tatums' "honesty[ ] [and] integrity[.]" See TEX. CIV. PRAC. & REM. CODE § 73.001. Thus, the accusation is reasonably capable of being defamatory. "Deception" and "honesty" are antonyms. Blow's statement accusing the Tatums of the first is capable of impeaching their character for the second.

B. Opinion

We conclude that of the defamatory meanings the Tatums allege, the only one capable of arising from Blow's column is the implicit statement that the Tatums acted deceptively. However, "statements that are not verifiable as false" are not defamatory. Neely , 418 S.W.3d at 62 (citing Milkovich , 497 U.S. at 21-22, 110 S.Ct. 2695 ). And even when a statement is verifiable, it cannot give rise to liability if "the entire context in which it was made" discloses that it was not intended to assert a fact. See Bentley , 94 S.W.3d at 581. A statement that fails either test-verifiability or context-is called an opinion.

1. Arguments

The News, of course, denies that it has accused the Tatums of deception. But even if the column explicitly levied that accusation, the News argues that the deception in this case is inherently unverifiable. The Tatums' mental states in the hours following Paul's death simply cannot be factually verified. Unlike in Milkovich , which involved perjury, no "core of objective evidence" exists from which a jury could draw any conclusions about the Tatums' mental states. See 497 U.S. at 21, 110 S.Ct. 2695. The News also argues that the column's context clearly discloses that it contains opinions, and that even if the accusation is capable of verification, it is protected because it is among the opinions that the column contains.

The Tatums contend that the charge of deception is verifiable. The accusation turns on whether the Tatums drafted the obituary with a deceptive mental state. Though the News argues this makes the accusation unverifiable, the law determines mental states all the time. Defamation, the very body of law at issue, has developed a robust process for determining whether a defendant's mental state constitutes actual malice. It cannot be the case, the Tatums argue, that defamation law can ascertain a defendant's mental state but not a plaintiff's. As for context, the Tatums argue that "a reasonable reader ... would conclude that Blow is making objectively verifiable assertions regarding the Tatums and their deliberate misrepresentations of fact in the Obituary." Thus, in the Tatums' view, the statement is both verifiable and contextually stated as a fact.

The court of appeals agreed with the Tatums "that the column's gist that the Tatums were deceptive when they wrote *639Paul's obituary is sufficiently verifiable to be actionable in defamation." See 493 S.W.3d at 668. The court compared the implicit accusation in this case to "[c]alling someone a liar and accusing someone of perjury." Id. The court concluded: "Although the Tatums' mental states when they wrote the obituary may not be susceptible of direct proof, ... they are sufficiently verifiable through circumstantial evidence[ ] ...." Id.

2. Law

"[S]tatements that are not verifiable as false cannot form the basis of a defamation claim." Neely , 418 S.W.3d at 62 (citing Milkovich , 497 U.S. at 21-22, 110 S.Ct. 2695 ). However, Milkovich requires courts to focus not only "on a statement's verifiability," but also on "the entire context in which it was made." Bentley , 94 S.W.3d at 581. And even when a statement is verifiable as false, it does not give rise to liability if the "entire context in which it was made" discloses that it is merely an opinion masquerading as fact. See Bentley , 94 S.W.3d at 581 ; see also Isaacks , 146 S.W.3d at 157 ("[ Milkovich protects] statements that cannot 'reasonably [be] interpreted as stating actual facts' ...." (second alteration in original) (citations omitted) ). Thus, statements that cannot be verified, as well as statements that cannot be understood to convey a verifiable fact, are opinions. Whether a statement is an opinion is a question of law. See Bentley , 94 S.W.3d at 580. Finally, the type of writing at issue, though not dispositive, must never cease to inform the reviewing court's analysis.

3. Analysis

The column's context manifestly discloses that any implied accusation of deception against the Tatums is opinion. Thus, we need not decide whether the accusation is wholly verifiable.

The column does not implicitly accuse the Tatums of being deceptive people in the abstract or by nature. Instead, it accuses them of a single, understandable act of deception, undertaken with motives that should not incite guilt or embarrassment. And it does so using language that conveys a personal viewpoint rather than an objective recitation of fact. The first sentence begins "So I guess," the column uses various versions of "I think" and "I understand," and near the column's close Blow states "the last thing I want to do is put guilt on the family of suicide victims." This first-person, informal style indicates that the format is subjective rather than objective. Nor does the column imply any undisclosed facts. The Tatums list several "exculpatory" facts that they say Blow should have included in the column. But Blow did not imply that he had personal knowledge that any of the facts the Tatums assert were false. Instead, he compared a quotation from the obituary against an account of Paul's suicide. These two accounts diverged, which Blow noted. Any speculation as to why the accounts diverged-if it appears in the column at all-was reasonably based on these disclosed facts. Thus, the column's words indicate that the statement is an opinion. The column's title does the same. The column as a whole, though it includes facts, argues in support of the opinion that the title conveys-society ought to be more frank about suicide. It is an opinion piece through and through.

The court of appeals ignored the column's context, opting instead to focus on de-contextualized words which it-not Blow-emphasized. See 493 S.W.3d at 654-55. In doing so, it disregarded Bentley 's direction that the "entire context in which [a statement] was made" must be analyzed to determine whether a verifiable statement *640of fact is nonetheless an opinion for purposes of defamation. Bentley , 94 S.W.3d at 581 ; see also Isaacks , 146 S.W.3d at 157. We reject that conclusion, and hold instead that if the column is reasonably capable of casting any moral aspersions on the Tatums, it casts them as opinions. See Musser , 723 S.W.2d at 654-55. Thus, under our precedent recognizing Milkovich 's joint tests, the accusation is not actionable. See Bentley , 94 S.W.3d at 581.

C. Truth

Blow's column is an opinion because it does not, in context, defame the Tatums by accusing them of perpetrating a morally blameworthy deception. But to the extent that the column states that the Tatums acted deceptively, it is true. Implicit defamatory meanings-like explicit defamatory statements-are not actionable if they are either true or substantially true. See McIlvain , 794 S.W.2d at 15 ; see also Neely , 418 S.W.3d at 66. The court of appeals held that "a genuine fact issue" existed as to whether the column's implicit accusation of deception was true or substantially true. 493 S.W.3d at 666. We disagree.

The statement at issue, which arises implicitly, is that the Tatums acted deceptively when they published the obituary. "The truth of the statement in the publication on which an action for libel is based is a defense to the action." TEX. CIV. PRAC. & REM. CODE § 73.005(a). A statement is true if it is either literally true or substantially true. See Neely , 418 S.W.3d at 66. A statement is substantially true if it is "[no] more damaging to the plaintiff's reputation than a truthful [statement] would have been." Id. (first citing Turner , 38 S.W.3d at 115 ; and then citing McIlvain , 794 S.W.2d at 16 ). In our view, the statement that the Tatums were deceptive is both literally and substantially true.

The statement is literally true because the Tatums' obituary is deceptive. It leads readers to believe something that is not true. It states that Paul died from injuries arising from a car accident when in fact Paul committed suicide. The Tatums believe that the car accident and the suicide are related, but the obituary does not convey that belief. At oral argument, the Tatums' counsel noted that the public often understands news reporting a death due to mental illness synonymously with death by suicide. The same cannot be said of death due to car accident. The obituary purports to convey that a car accident was both the proximate and immediate cause of Paul's death. The former may be true, but the latter is not. That is enough to render the obituary deceptive, which is enough to render truthful the column's implication that the Tatums acted deceptively in publishing it.

The Tatums respond that they earnestly believed that the obituary was true. But the Tatums' beliefs, however sincere, do not make the obituary's message any less deceptive. Indeed, the Tatums argue that Blow should have included all kinds of background facts about the Tatums' beliefs concerning traumatic brain injuries, cause of death, and other matters. But the Tatums themselves did not include any of this information in Paul's obituary. The Tatums cannot argue both that the obituary was true without this background information and that the column is false for failing to include it.

The Tatums also respond that deception implies intentionality. We agree. But the Tatums plainly and intentionally omitted from the obituary the crucial fact that Paul committed suicide. Their motive with regard to the omission is immaterial to whether the obituary is deceptive. What matters is that they intentionally omitted that Paul committed suicide; in so doing, *641they drafted an obituary that conveyed a deceptive message to the substantial majority of the News's readership. At root, the Tatums' argument regarding intentionality muddles the concepts of intentionality and moral blameworthiness. True, an intentional deception often brings with it the implication of wrongdoing, but that is not always the case. And it is certainly not the case here, where the column's author expressly stated that "the last thing I want to do is put guilt on the family of suicide victims." Accordingly, we conclude that the column is literally true.

And even if the statement is not literally true, it is substantially true because it is no more damaging to the Tatums' reputation than a truthful column would have been. See Neely , 418 S.W.3d at 63. The column does not damage the Tatums' reputation among the cohort of people who knew, before the obituary, that Paul committed suicide. The reason is that these people, assumedly, read the obituary the way the Tatums claim that they intended it to be read-as a tactful way of acknowledging Paul's death without dishonoring his memory. Nor does the column, by omitting the Tatums' belief as to the reason for Paul's suicide, cause additional damage to the Tatums' reputation among the much larger group of people who first learned that Paul committed suicide upon reading the column. These readers, even if they believed the column's implication that the Tatums intended to be deceptive, would heed the column's exhortation that those who shroud suicide in secrecy do not deserve blame.

The column does not accuse the Tatums of being deceptive people in general, but instead of buckling to the current societal pressure to avoid disclosing suicide when it occurs. And to the extent that readers thought less of the Tatums after reading the column, it would be because they concluded on their own that the Tatums acted deceptively, not because they decided to believe the column's implied assertion to that effect. Put differently, the column revealed something that the obituary did not: Paul committed suicide. If readers formed a negative view of the Tatums as a result of that revelation, it was of their own volition, not because the column urged them to. In fact, the column urged precisely the opposite when it said that our society should not place any guilt on families who conceal suicide.

The Tatums respond that a literally truthful column would have included many caveats beyond the fact that the Tatums did not intend to deceive. These facts all relate to whether the Tatums' view of Paul's death was reasonable or scientifically justified. But, of course, the Tatums do not claim to have been defamed by an allegation that they failed to rely on reason or scientific evidence in coming to their conclusion. Instead, they claim the column defames them by omitting their belief-the same belief that they themselves omitted from the obituary. Thus, even accepting the Tatums' contention that the column was less than literally true, a "hypothetically truthful" account would require nothing more than a recitation that the Tatums did not intend to deceive anyone.

Because we agree with the News that a recitation to that effect would not have mitigated the reputational harm that the accusation of deception caused the Tatums, if any, the statement does not fail our standard for substantial truth. Blow's column was callous, certainly, but it was not false.

IV

Conclusion

The publication of Blow's column may have run afoul of certain journalistic, ethical, *642and other standards. But the standards governing the law of defamation are not among them. Accordingly, we reverse the judgment of the court of appeals and reinstate the trial court's summary judgment in favor of petitioners Steve Blow and The Dallas Morning News, Inc.

Justice Boyd filed a concurring opinion, in which Justice Lehrmann and Justice Blacklock joined.

APPENDIX

So I guess we're down to just one form of death still considered worthy of deception.

I'm told there was a time when the word "cancer" was never mentioned. Oddly, it was considered an embarrassing way to die.

It took a while for honesty to come to the AIDS epidemic. Ironically, the first person I knew to die of AIDS was said to have cancer.

We're open these days with just about every form of death except one-suicide.

When art expert Ted Pillsbury died in March, his company said he suffered an apparent heart attack on a country road in Kaufman County.

But what was apparent to every witness on the scene that day was that Pillsbury had walked a few paces from his car and shot himself.

Naturally, with such a well-known figure, the truth quickly came out.

More recently, a paid obituary in this newspaper reported that a popular local high school student died "as a result of injuries sustained in an automobile accident."

When one of my colleagues began to inquire, thinking the death deserved news coverage, it turned out to have been a suicide.

There was a car crash, all right, but death came from a self-inflicted gunshot wound [page break] in a time of remorse afterward.

And for us, there the matter ended. Newspapers don't write about suicides unless they involve a public figure or happen in a very public way.

But is that always best?

I'm troubled that we, as a society, allow suicide to remain cloaked in such secrecy, if not outright deception.

Some obituary readers tell me they feel guilty for having such curiosity about how people died. They're frustrated when obits don't say. "Morbid curiosity," they call it apologetically.

But I don't think we should feel embarrassment at all. I think the need to know is wired deeply in us. I think it's part of our survival mechanism.

Like a cat putting its nose to the wind, that curiosity is part of how we gauge the danger out there for ourselves and our loved ones.

And the secrecy surrounding suicide leaves us greatly underestimating the danger there.

Did you know that almost twice as many people die each year from suicide as from homicide?

Think of how much more attention we pay to the latter. We're nearly obsessed with crime. Yet we're nearly blind to the greater threat of self-inflicted violence.

Suicide is the third-leading cause of death among young people (ages 15 to 24) in this country.

Do you think that might be important for parents to understand?

*643In part, we don't talk about suicide because we don't talk about the illness that often underlies it-mental illness.

I'm a big admirer of Julie Hersh. The Dallas woman first went public with her story of depression and suicide attempts in my column three years ago.

She has since written a book, Struck by Living. Through honesty, she's trying to erase some of the shame and stigma that compounds and prolongs mental illness.

Julie recently wrote a blog item titled "Don't omit from the obit," urging more openness about suicide as a cause of death.

"I understand why people don't include it," she told me. "But it's such a missed opportunity to educate."

And she's so right.

Listen, the last thing I want to do is put guilt on the family of suicide victims. They already face a grief more intense than most of us will ever know.

But averting our eyes from the reality of suicide only puts more lives at risk.

Awareness, frank discussion, timely intervention, treatment-those are the things that save lives.

Honesty is the first step.

See Steve Blow, Shrouding suicide in secrecy leaves its danger unaddressed , THE DALLAS MORNING NEWS (July 12, 1010), https://www.dallasnews.com/news/news/2010/07/12/20100620-Shrouding-suicide-in-secrecy-leaves-its-9618.

1 Nathaniel Hawthorne, American Notebook (1841-52), in THE AMERICAN NOTEBOOKS 73, 122 (R. Stewart ed., Yale Univ. Press 1932).

2 We draw our recitation of the Tatums' factual and legal allegations from their third amended petition, which was their live petition when the trial court granted The Dallas Morning News's motion for summary judgment.

3 Throughout the rest of this opinion, we refer to The Dallas Morning News as "the paper." Similarly, we refer to Blow and the paper together as "the News."

4 The column, which this opinion attaches as an appendix, spanned two pages. The headline on the second page was slightly different from the headline on the first: "Shrouding Suicide in Secrecy Leaves Its Dangers Unaddressed" (emphasis added).

5 The News amended its motion once. The trial court granted summary judgment on the News's Amended Motion for Final Summary Judgment.

6 See, e.g. , Salinas v. Salinas , 365 S.W.3d 318, 319 (Tex. 2012) (per curiam) (discussing a defamation claim in which defendant accused plaintiff of being "a drug dealer and a corrupt politician," who had "stolen and lied and killed"); Bentley , 94 S.W.3d at 569 (discussing public official's defamation action based on plaintiff's statement that "y'all are corrupt, y'all are the criminals, [and] y'all are the ones that oughta be in jail"); Leyendecker & Assocs., Inc. v. Wechter , 683 S.W.2d 369, 374 (Tex. 1984) (holding that a letter's explicit accusation that plaintiff "committ[ed] a criminal act by attempting to conspire ... to file fraudulent insurance claims" was textually defamatory and libelous per se); Cullum v. White , 399 S.W.3d 173, 178 (Tex. App.-San Antonio 2011, pet. denied) (discussing defamation claim in which defendant accused plaintiff of being a "pathological liar" who was "flagged" for "terrorist activity").

7 See, e.g. , Turner , 38 S.W.3d at 113 (mentioning that the plaintiff "strenuously argued that the broadcast's 'gist' was both false and defamatory .... [but] regarded libel by implication as a separate theory"); Leatherwood , 49 S.W.2d at 1109-10 (discussing a letter's "innuendo" concurrently with "all reasonable implications thereof or inferences to be drawn therefrom").

8 See, e.g. , Elizabeth Blanks Hindman, When Is the Truth Not the Truth? Truth Telling and Libel by Implication , 12 Comm. L. & Pol'y 341, 363 (2007) ("[Turner ] took up the issue of libel by implication ...."); see also Thomas B. Kelley & Steven D. Zansberg, Libel by Implication , Comm. Law ., Spring 2002, at 3, 10 (discussing Turner ); John P. Border et al., Recent Developments in Media Law and Defamation Torts , 37 Tort & Ins. L.J. 563, 578-79 (2002) (discussing Turner immediately under the heading "Libel by Implication").

9 See KBMT Operating Co. v. Toledo , 492 S.W.3d 710, 723 (Tex. 2016) ("[T]he question is whether [plaintiff] submitted some evidence that the gist of [defendant's] broadcasts was false." (emphasis omitted) ); Lipsky , 460 S.W.3d at 594 (discussing "the gist of [plaintiff's] statements"); Neely , 418 S.W.3d at 56-57 (reversing summary judgment because plaintiff "raised a genuine issue of material fact" as to broadcast's gist).