HLR Free Speech Symposium: Rebecca Tushnet, More than a Feeling: Emotion and the First Amendment

My paper versus the other topics: One of these things is not like the others.  Regulation of non-press entities is also an important part of modern speech regulation, and I do think my topic is grounded in Sullivan, although not in Sullivan’s protection for the press as such.

One important aspect of Sullivan, especially as later elaborated, is its bar on liability for derogatory but nonfactual speech.  It’s not because such speech can’t do harm.  Everyone who survived an American high school knows that it’s the nonfactual statements that do the most damage.  Instead, the fact/nonfact distinction limits and channels defamation liability in a way that, first, we hope is objective and thus more protective of speech than a subjective standard, and second, we believe limits liability when factfinders would be inclined to protect powerful and respected figures over marginal characters—both when the marginal characters are plaintiffs and when they’re defendants.  You can do as much harm as you like to a person’s reputation as long as you do it with nonfacts—mockery, slurs, and the like.  Under modern defamation law, a particular mechanism—emotional manipulation—is fair game, or at least off limits for suppression by the government.

My paper contrasts two private causes of action and two government disclosure mandates. First, trademark law allows exactly the kind of control over nonfactual, emotional appeals that modern defamation law precludes.  Second, in compelled speech cases, courts have struck down mandatory smoking warnings in visual form, but have approved mandatory abortion disclosures and ultrasound requirements that operate in the same emotional register. 

I suggest that the contradictions of current doctrine could be ameliorated by less distrust of emotion and more acceptance that, where information is being conveyed, emotion will regularly follow.  Our focus then should not be on whether deployment of emotion is “manipulative” but whether it is part of a discriminatory or factually misleading regulation.  When the government regulates speech, the regulation will generally have an emotional component, because human thought is emotional.  Objections to emotion-based regulations should not be based on the obviousness of that component.  Rather, the acceptability of the government’s aim should be the guide, especially when nongovernmental speakers are free to use emotional appeals to press their own cases.  The government may be required to be neutral as between classes of private speakers, as Sullivan requires and as I will argue should be the case with respect to trademark law.  It is not required to be neutered.  When the government can otherwise constitutionally mandate disclosure, the fact that these disclosures have emotional resonance is not an independent constitutional barrier.

Trademark and defamation:

Trademark law presently authorizes the kind of private control over emotional meaning that defamation law’s limitation to facts precludes: The government provides a remedy as between two private parties when one interferes with the emotional valence of the other. 

Courts routinely protect the aura of distinction and uniqueness claimed by a trademark owner.  For example, the Second Circuit enjoined a copyist of Levi’s stitching, ostensibly on confusion grounds, because otherwise Levi’s sales would be “affected adversely by . . . buyers’ ultimate realization that the pattern is no longer exclusive.”   Trademark protection in the absence of consumer confusion at the point of sale is justified, at least in part, as a means of preserving the status of consumers of “true” luxury products, even if no one ever makes a mistaken purchase.  Trademark dilution is another doctrine that allows trademark owners to control emotional meaning: dilution prevents commercial actors from interfering with the singular meaning of a mark even if no one is ever mistaken about any fact at all.  For example, the Second Circuit found that a lawn tractor ad actionably diluted John Deere’s mark by featuring an animated Deere logo running away in terror from a small, yappy dog.

American regulators rarely try to suppress nonfactual commercial speech—adorable spokesbears and body-spray-using boys surrounded by inexplicably attracted women are safe devices to build consumer interest and loyalty.  Within this regime, trademark law creates systematic bias.  If nonfalsifiable speech about someone else’s trademark can be banned because it’s only used to get attention, build image, or amuse consumers, the result is special positional advantages given to now-dominant brands.   This discrimination in favor of the already powerful is what makes trademark law, in its role as emotional regulator, constitutionally problematic.

This power can’t be justified by trademark-specific concepts of economic harm or property, because those concepts can’t be made trademark-specific except by pure fiat.  Inconsistency isn’t itself a vice, but there’s a particular problem when the inconsistency favors the presently powerful and dominant, which is explicitly what trademark law does: it protects strong brands more than new or weak entrants, and dilution in particular only protects famous marks, which is to say household names.

So I’ve identified a tension in the treatment of private torts targeting emotional methods of persuasion.  This tension also exists in other areas.  Because the topic of government speech is so broad and unwieldy, I will focus on compelled speech rather than on claims like “this is your brain on drugs.”  Can the government mandate that a speaker disclose negative facts in a way that is likely to trigger emotional reactions? 

The answer presently is no and yes.  No when the speaker provides cigarettes.  Yes when the speaker provides abortions.

On tobacco, the DC Circuit upheld the invalidation of congressionally required graphic warnings on cigarette packages.  The main problem the R.J. Reynolds majority had was that the warnings were too emotional.   The graphic warnings were not “purely” factual because “they are primarily intended to evoke an emotional response, or, at most, shock the viewer into retaining the information in the text warning.”

The FDA defended the use of images by citing research showing that pictures are easier to remember than words, which meant that the images’ health messages—smoking is bad for you—would be better received. The majority interpreted this as the FDA’s desire to “shock[]” consumers, but it didn’t reject the FDA’s factual claims that memory is aided by emotional cues or that other alternatives had failed to make the risk message stick. The warnings were designed “to shame and repulse smokers and denigrate smoking as an antisocial act,” making the message “ideological and not informational.”

Under the majority’s reasoning, the government is apparently not allowed to mandate a warning that works through an emotional mechanism.  One immediate problem with that conclusion is that “purely” factual words also work that way.  For example, research shows that price signals trigger negative emotional reactions that mitigate consumers’ desires for the advertised products.   They cause consumers to feel bad about buying, mitigating positive emotions triggered by desire for the product.  Yet price disclosures have previously been upheld as acceptable regulations of commercial speech, obviously related to rational consumer decisionmaking.

Nonetheless, the Reynolds majority held that the graphic warnings were unacceptable because they didn’t provide “purely factual and uncontroversial” information.  By contrast, the majority was confident that mandatory price-related disclosures “were both indisputably accurate and not subject to misinterpretation by consumers.”  The majority constructed a false dichotomy between accurate statements and misinterpretation (which is connected to the false dichotomy between reason and emotion).  There is no such thing as a disclosure that is not subject to misinterpretation by consumers.   People are just too variable in their attention, prior beliefs, and other cognitive resources; someone always ends up reading “this claim has not been evaluated by the Food and Drug Administration” as “this claim has been evaluated by the Food and Drug Administration.”

Professor Caroline Mala Corbin’s careful analysis uses the social science evidence to contest this result.  Among other things, she points out that vivid images seem more personal, so that the viewer will imagine herself at risk rather than assuming herself immune.  Relatedly, salient images are more likely to be noticed and comprehended, unlike current textual warnings.  

However, Corbin considers one emotional pathway to be dangerous: mere association of smoking with negative affect—“the reverse process of what advertisers do when they link their product with something that triggers a positive emotional response.”  As long as the negative image is truly connected to smoking, though, she considers this reaction to be a legitimate transfer of affect, whereas Clockwork Orange-style aversive conditioning—exposing consumers to images of maggot-infested meat next to cigarettes, for example—would be illegitimate.   Corbin suggests that most of the FDA’s images were unproblematic in this regard, except for one of a woman “weeping uncontrollably,” which exploits cognitive shortcuts instead of relying on the merits. 

I think this reaction illustrates the difficulties of distinguishing shortcuts from merits.  Is it really tenuous to think that nonsmokers who contract fatal lung disease will be mourned?  Or even that smokers who discover that an intimate has fatal lung disease will feel guilt?  It is these rational but not perfectly correlated associations that government will most often want to make when it imposes disclosure requirements.  Unlike private advertisers using sex and pleasure to sell, government mandates aren’t likely to associate completely unrelated things.

But private advertisers’ willingness to sell products with positive imagery has to be considered as well in any full analysis of the constitutional issues.  Commercial sellers—including tobacco companies—routinely and even predominantly use images and nonfactual matter to make their products attractive—that’s the story of modern trademark law.  If imagery can distort rational judgment, Reynolds makes it impossible for the government to correct tobacco advertisers’ own distortions with a countervailing emotional appeal at the very point where that emotion is likely to have an influence: the time a decision is made to smoke. 

Now for the other side: While Reynolds invalidated visual tobacco warnings, courts have mainly upheld a different set of emotional interventions related to abortion.  I depart from many critics of mandatory abortion-related disclosures in accepting the idea that emotional government appeals are legitimate.  The real concerns—shared by critics of emotional appeals as well—relate to factual misleadingness, burdens on the practical ability to obtain an abortion, and misattribution of the government’s message to doctors.  

So, the Eighth Circuit upheld South Dakota’s requirement that physicians provide their patients with a written statement informing women contemplating an abortion that a “known medical risk[] of the [abortion] procedure” is an “[i]ncreased risk of suicide ideation and suicide,” even though the best available scientific studies suggest that abortion is “psychologically benign.”  The majority concluded that this statement didn’t suggest a causal relationship between abortion and an increased risk of suicide, and was therefore not false or misleading because of the possibility of causation.  By any normal standard, the required disclosure is clearly misleading.   Correlation may not be causation, but reasonable listeners would undeniably receive the message that having an abortion caused the increased risk, using the ordinary rules of implicature.  It would have been just as “true” to mandate a disclosure that people who take Advil have an increased risk of headaches.

Courts have also upheld various other statements about the whole, separate, unique life that will be destroyed by abortion and requirements forcing women to view fetal images and hear fetal heartbeats, by way of mandatory ultrasounds.

While courts have treated these abortion requirements as “truthful” and “nonmisleading,” there is no way that they’d satisfy the standard applied in the tobacco cases.  The aim here is really persuasion, not just information.  The state’s position is that a woman’s future self is sufficiently likely to have a different perspective on the present decision—much like a current smoker and that smoker’s future self—that providing her the information now will enhance her overall decisionmaking. We may disagree about the likelihood that these regrets will materialize, but the two situations can’t be separated on their logic, only on their facts.  And the state absent the constraints of Lochner often has a great deal of freedom to find facts.

Given this government aim, focusing on the emotionality of the government’s emotional appeal may divert us from the more basic question of whether the appeal ought to be allowed.  In this vein, the reasoning behind one common analogy deployed against abortion disclosures demonstrates how our underlying concepts about which choices are rational affect our conclusions about acceptable emotional appeals.  Critics of graphic abortion disclosures often make the point that graphic images of ordinary surgery aren’t considered necessary for informed consent to surgery.  Rationally, the argument goes, the fact that surgery is bloody and gross shouldn’t deter someone who needs it from having it. 

But imagine a state with a substantial population of Jehovah’s Witnesses and Christian Scientists.  Would they agree that the only rational, necessary choice is surgery?  Are we prepared to accept the existence of reasons to reject surgery that acknowledge its potential efficacy but value other considerations more?  To press further: consider that in many cases, surgery is not the only option—there may be other possible treatments, or a patient nearing the end of her life may choose not to attempt to extend it with surgery given the costs in pain and recovery time.  To the extent that graphic images make more salient the physical costs of surgery,  they could be justified in such situations.  Just as an image of a crying woman concretizes and stands for the many costs of tobacco use, or the image of a fetus stands for a human life, surgical images are directly related to considerations that a patient rationally could take into account. 

The surgical analogy does prove a somewhat different point: abortion disclosure mandates discriminate, in a constitutionally significant way, between a choice made uniquely by women and other medical choices, and this discrimination is likely to reflect sexist assumptions about women’s rational capacities.  To the extent that the state is deploying emotional appeals in a biased way—and especially to the extent that its is forcing doctors to represent the state’s emotional opinions as its own —arguments against forced disclosures are still available even if general appeals to emotions are constitutional. 

Indeed, it is now well-recognized that emotion is a crucial component of decisionmaking.   There is no weighing of alternatives, or judgment between them, without emotions guiding choices.   Some decisions don’t look like they’re emotionally influenced, often because the cultural definition of “emotional” decisions excludes those that make certain interactions flow smoothly.   But that just means the emotion is invisible, not that it’s absent.  

Advertising provides the most obvious examples.  People are regularly influenced by aspects of presentation that have no rational relationship to decisions.  In one especially striking study, a photo of a smiling, attractive woman in the bottom right-hand corner of a loan offer produced an increased response rate for men equivalent to dropping the monthly interest rate by 200 basis points, or approximately 25% of the total rate, a substantial percentage.  The recipients didn’t get any increased exposure to the woman by accepting the offer; even if they valued her smile at 200 basis points, taking the offer was irrational.   We’re regularly influenced in ways we’d be embarrassed to admit, if we even acknowledged that we’d been influenced.

So what can be done with this reality?  With respect to mediating private disputes, it’s desirable for the government to refuse to allow liability when a defendant’s nonfactual speech affects audiences’ evaluation of the plaintiff.  This is the rule of Sullivan and should be the rule in trademark law as well.  In that sense, it is possible for the government to be hands-off with respect to certain kinds of privately generated emotion, though emotion will necessarily remain one mechanism by which factually false claims can do harm.

However, the government may also wish to require certain disclosures from commercial speakers, and here disconnection from emotion is impossible.  Some have argued that the proper inquiry in evaluating required disclosures is whether the government has a purpose to change behavior along its preferred lines, or merely to inform.   But the government will always have both purposes when it requires disclosure.   Zauderer, the classic commercial speech disclosure case, found a lawyer’s failure to disclose that consumers would have to pay court costs even if they lost their cases to be deceptive because that liability would matter to a significant number of consumers.  If it would matter, it would change at least some behavior: that’s what it means for information to be material.

If the government does have a legitimate interest in having people become aware of certain information, then emotion has to enter into the calculation of how to deliver that information.  Humans are, as Dan Ariely says, predictably irrational.  We are and will remain poor decisionmakers who rely on salience over statistics.  Salience is produced by concrete, human examples that trigger emotional reactions.  Because emotion and reason are inextricable, emotional appeals should be fair game for the government as well as for private parties, unless the emotion is tied to factual deception.

Explicitly emotional government messages often make us uncomfortable because they seem unduly paternalist and judgmental—they put undue pressure on the citizen.  But, among other things, the government isn’t the only entity addressing the citizen.  If she’s vulnerable to emotional manipulation by the government, then she’s probably vulnerable to other kinds of emotional manipulation.   Other entities are already selecting what messages they want her to hear and pressing them on her.  Even accepting that deliberate appeals to emotion create autonomy problems, her autonomy is profoundly at risk already, and we might want the government to step in to help her restore that autonomy. 

Consider a hypothetical from David Strauss on the question of illegitimate manipulation:  He asks us to “Suppose that the government could manipulate people’s minds directly, by irradiating them in a way that changed their desires. No one would say that the power to ban an activity automatically included the ‘lesser’ power to irradiate people so that they no longer had the desire to engage in that activity.”  But that’s not really what many of these emotional appeals do.  Could the government ban private parties from irradiating people, on the ground that this harmed the victims’ autonomy?   Could it irradiate people to reverse the effects of that private irradiation?  To the extent that government argues that it is operating in an already emotion-saturated field such as that created by tobacco advertising, it becomes harder to identify illegitimate manipulation.

Ultimately, I would allow the government to express the preferences it can constitutionally have, and to express them in emotional terms, as long as it isn’t deceptive.  At least with respect to commercial speech, this should give it significant freedom to require disclosures, even ones that concededly rely on emotional appeals to communicate information.

Caroline Corbin

Agrees that emotion and reason are inextricably linked; compelled disclosures aren’t automatically suspect and Reynolds was wrong.  We also agree about factual errors and misleadingness, which means gov’t shouldn’t be able to force disclosures that abortions cause suicide/etc., since that’s false.  Gov’t may compel cigarette sellers to reveal gruesome facts, and compel doctors to reveal state’s view of abortion.

Limits of this approach: defines deception too narrowly.  Just as state can mislead with facts, it can mislead with emotion.  Plus, if deception is problematic because it fails to respect autonomy, then the goals of the compelled speech need to be considered as well. Also must consider autonomy of person forced to give gov’ts message.

Misleading: emotionally false/misleading—intentionally exploit cognitive shortcuts involving affect.  Make you like something not by virtue of its own merits but by association with something you like.  Sell toaster by draping beautiful woman over a toaster.  Response to tobacco disclosures is due to the merits—you reoil from smoking because of what smoking does, not because smoking has been associated with something else you don’t like.  It is hard to draw the line, but also hard to draw the line about what’s misleading.  Not comfortable having no line at all.  Gov’t might not be likely to associate smoking with maggot infested meat, but is not sure that she trusts the gov’t that much.  If the 1A allows gov’t to intentionally exploit affect heuristics, it can create negative associations with tobacco or with cigarettes.  Weeping regretful women could be used for both.  Intentionally exploiting affect heuristics is deceptive and manipulative. Lead audience to draw factual conclusions as well as emotional conclusions it wouldn’t have done. Decisional autonomy is at the heart of the 1A.

If autonomy is at stake, then goals of disclosure also matter.  Tushnet says autonomy is intact if you can still make a choice. But interference with the decisionmaking process also matters.  Very hard to distinguish between compelled disclosures meant to inform and those meant to persuade, but there is a difference in “smoking is harmful” and “abortion is murder.”  Less an insult to captive audience when the gov’t’s goal is one you’re likely to agree with—noncontroversial, well established fact v. controversial moral position.  Tushnet says everything’s controversial today, but Corbin disagrees.  (Heh.)  Smoking causes cancer—that’s a well-established fact.  Abortion is murder is not a fact, but an issue of deep division.

Autonomy and non-autonomy-enhancing are not clear lines, but can still be done.  Smoking is addictive and interferes with autonomy. Most women don’t regret abortion; availability of abortion enhances women’s autonomy. Abortion disclosure is based on gov’t viewpoint on controverted moral issue, and urge a course of conduct that’s not necessarily autonomy-enhancing.

Based on moral aims: raises free speech rights of compelled speakers. Insult to dignity/autonomy is worse if you’re forced to affirm a viewpoint you don’t share, not facts.

If free speech jurisprudence considers it anathema to force articulation of viewpoints, why is it ok to force audiences to hear it. 

RT: Breadth of definition of deception is a key issue: Difference between you’re making the wrong decision and you’re making an incorrectly informed decision is very difficult.  This is why Lochner is important: what facts can the state find?  This is why consumer protection rhetoric has become so important in abortion disclosures. Abortion is murder: That’s not what the gov’t is saying. Instead, it uses the language of information.  And though abortion is special, there are a bunch of disclosures that we mandate that wouldn’t satisfy these standards, which is why I don’t like it.

West: we do have some defenses to these messages.  Are there audiences without defenses? Children, people in a bad health situation. 

RT: we think we have defenses.  “Persuasion knowledge” varies a lot.  Real q for me is how to deal with private parties also using emotion to influence you.

Q: in an age of mind sciences, is decisional autonomy something that can survive?

RT: Yes; these are all probabilistic.  Even if there is determinism, we can't get at it and must act as if there are degrees of freedom.