Emory Law Journal

Freedom from Thought
Jane Bambauer *Associate Professor of Law, University of Arizona James E. Rogers College of Law. J.D., Yale Law School; B.S., Yale College. Many thanks to Derek Bambauer, Caroline Corbin, Eugene Volokh, and Alexander Tsesis for helpful feedback, and to Ryan Pulley for superb editing. I am also tremendously grateful to Alexander Tsesis, Gerard Bifulco, and the Emory Law Journal for the opportunity to participate in the 2015 Thrower Symposium.

April Salazar was thrilled to be pregnant. Shortly after learning the good news, she created an account with a pregnancy website so that she would receive email updates about the development of her fetus and advice for pregnancy health. 1April Salazar, The Pregnancy Is Gone, but the Promotions Keep Coming, N.Y. Times: Motherlode (Feb. 2, 2014, 8:25 AM), . By the time Salazar made the devastating decision to terminate her pregnancy (the fetus had a fatal defect), the pregnancy website had apparently sold Salazar’s due date and contact information to Enfamil, a producer of infant formula. 2Id. Months later, when Enfamil sent Salazar a box of infant formula marked with the poignant phrase “You’re almost there!,” the experience lurched Salazar back into the personal hell of a lost pregnancy. 3Id.

Salazar’s story is just one poignant example of inept promotions and tactless services reminding consumers about something unpleasant. 4Other women who have experienced miscarriage found themselves in similar circumstances to Salazar with respect to targeted advertising. E.g., Why Targeted Ads Suck, Maybe Baby? (Apr. 27, 2009), ; Emily Wilkins, Life After Miscarriage: Reminders in the Mail, Pea in the Podcast (July 18, 2010), . Some users of Google services resent learning that their true tastes in music or reading material are pedestrian, or even crude. 5James Carmichael, Google Knows You Better than You Know Yourself, Atlantic (Aug. 19, 2014), . Still others have had an unpleasant experience learning something intimate about their family members through promotional materials, such as the (possibly apocryphal 6Tim Harford, Big Data: Are We Making a Big Mistake?, Fin. Times Mag. (Mar. 28, 2014, 11:38 AM), ; Gregory Piatetsky, Did Target Really Predict a Teen’s Pregnancy?: The Inside Story, KDNuggets (May 7, 2014), . ) example of a father who learned that his teenaged daughter was pregnant when Target accurately predicted her pregnancy and sent coupons for baby formula. 7Charles Duhigg, How Companies Learn Your Secrets, N.Y. Times Mag. (Feb. 16, 2012), . These stories help respond to criticism that privacy protects amorphous and unspecific interests. 8E.g., Daniel Castro, A Catalog of Every “Harm” in the White House Big Data Report, Ctr. for Data Innovation (July 15, 2014), . They demonstrate that widespread data tracking and behavioral advertising can cause specific and perfectly understandable distress. And they share very little with a skeptic’s conception of privacy as a shroud for misbehavior and social fraud. 9See Richard Posner, The Right of Privacy, 12 Ga. L. Rev. 393, 395 (1978).

The privacy harm that Salazar experienced does not track the typical course of privacy harm. Salazar’s pain did not stem from the lost control of her personal information, nor did it stem from negative consequences that could result when friends, employers, or the government discover something embarrassing. It did not even stem from the indignity of being treated like a commodity by the pregnancy website (although this was a predicate step). 10For a description of the commodification critique of data collection, see Alexander Furnas, It’s Not All About You: What Privacy Advocates Don’t Get About Data Tracking on the Web, Atlantic (Mar. 15, 2012), . The direct cause of Salazar’s injury was information that she, herself, received. The box of Enfamil reminded Salazar that if she had carried her pregnancy to term, she would be giving birth soon, and this unwelcome message prompted a predictable series of distressing thoughts and memories. 11Salazar, supra note 1.

Salazar’s harm came from self-knowledge, 12Or, in Salazar’s case, self-reminding. and it is not unique. Although self-reflection and understanding usually improves lives, it can disrupt our efforts to define ourselves or manage our internal monologues. 13See generally Susan Nolen-Hoeksema, Blair E. Wisco & Sonja Lyubomirsky, Rethinking Rumination, 3 Persp. on Psychol. Sci. 400 (2008) (discussing potential negative side effects of self-reflection, including depression and anxiety). These psychic disruptions range from trivial revelations about workplace productivity to profound revelations of paternity. Sometimes new information disrupts conscious attempts to avoid thought, as when a cancer patient avoids learning details about a prognosis, 14See generally Andrew Steptoe et al., Satisfaction with Communication, Medical Knowledge, and Coping Style in Patients with Metastatic Cancer, 32 Soc. Sci. & Med. 627 (1991) (describing oncology patients’ coping strategies as either information-seeking or avoidance). or when the new divorcée avoids learning information about his ex. 15Chandra Steele, How to Avoid Your Ex on Social Media, PC Mag. (Feb. 13, 2014), . Other times, unwanted information can seep into consciousness without any preexisting plan to avoid it, as when evidence of unanticipated health risks are returned to research participants, 16See generally Gail P. Jarvik et al., Return of Genomic Results to Research Participants: The Floor, the Ceiling, and the Choices in Between, 94 Am. J. Human Genetics 818 (2014) (discussing whether results with significant health implications should be reported to participants undergoing genetic studies). or when once-buried family secrets rise from the grave. 17Daniel Engber, Who’s Your Daddy?: The Perils of Personal Genomics, Slate (May 21, 2013, 5:45 AM), . In all cases, though, the new information conflicts with the recipient’s self-image and causes distress. As a result, the information cannot be ignored or easily dislodged from memory by the listener.

This short Essay explores when ignorance can be supported or even coerced by law, and when it cannot. Freedom from thought is a particularly intriguing topic for me for two distinct reasons.

First, freedom from thought may be a promising form of privacy and dignity rights. Although self-definition is frequently listed in the long list of interests encompassed by the concept of privacy, 18See, e.g., Julie E. Cohen, What Privacy Is For, 126 Harv. L. Rev. 1904, 1910–11 (2013); Ruth Gavison, Privacy and the Limits of Law, 89 Yale L.J. 421, 455 (1980) (moral autonomy); Daniel J. Solove, A Taxonomy of Privacy, 154 U. Pa. L. Rev. 477, 553 (2006) (decisional privacy). the difficult task of managing knowledge about ourselves has not received sufficient appreciation. 19One exception is new work by Kiel Brennan-Marquez theorizing that one of the under appreciated harms from surveillance comes from the anxiety and attention demanded by a subject who becomes aware that he is being watched and judged, and therefore judges himself. Brennan-Marquez questions whether the optimal amount of information about ourselves in order to maximize human flourishing is something less than what modern technology can offer. See generally Kiel Brennan-Marquez, The Freedom Not to Think (unpublished manuscript), . Privacy scholars and lawmakers focus principally on a subject’s loss of control when information about him is discovered by second parties or disclosed to third parties. The trouble with these forms of privacy, however, is that they stay in inescapable tension with the interests of those other parties who have something to gain by learning more about the subject. But the tension with second and third party interests does not necessarily afflict the topic of this Essay—the interest in self-ignorance. If a person sets out to avoid learning something about herself, another individual’s decision to tell that person what she sought to avoid may be entirely wasteful. In at least some cases, the listener is put in a worse position and the speaker gains nothing (discounting whatever utility the speaker may enjoy from bursting the listener’s bubble). Therefore, law may be able to incisively find and treat areas of particular vulnerability to create zones where a person should be left alone without running into significant free speech problems.

The second reason freedom from thought holds special intrigue for me is that it presents a challenge to the assumptions I typically carry from project to project—that more knowledge is better. 20See, e.g., Jane Bambauer, All Life Is an Experiment. (Sometimes It’s a Controlled Experiment.), 47 Loyola U. Chi. L.J. (forthcoming 2015); Jane Bambauer, Is Data Speech?, 66 Stan. L. Rev. 57, 63 (2014). I suspect all readers can imagine or remember circumstances in which learning something related to their lives would cause more harm than good. These circumstances contradict the premise of much First Amendment theorizing that prizes the messy quest for truth. Thus, when the government does act to preserve ignorance, the free speech implications are particularly interesting.

This Essay will borrow some basic free speech principles from the work of Eugene Volokh and James Grimmelmann. Part I will describe the analytical model used throughout the Essay, which depends on the willingness of the speaker to provide a message and the listener to receive it. Parts II through IV will apply the model to three categories of potential freedom-from-thought lawmaking. The first category involves a willing speaker and an unwilling listener and will explore what the government can do to support the listener’s preferences. The second category considers circumstances in which the government may want to keep a listener in ignorance even though both the speaker and listener prefer to transfer information. The third category considers circumstances in which neither the speaker nor listener wish to transfer information, but the government nevertheless compels information transfer. Together, these three explorations cover the field of government responses when either the listener or the government itself perceives a need for ignorance.

In the end, although freedom from thought has much to offer to the development of privacy and dignitary rights, interests in self-ignorance are better handled through norms than through law. Like other forms of privacy, First Amendment commitments are likely to frustrate legal efforts to support or coerce self-ignorance. If a speaker wishes to disclose information, the government is unlikely to be able to interfere with that disclosure unless the speaker’s interests are demonstrably weak. 21See infra Parts II & III. However, when both the speaker and the listener prefer silence, government compulsion of information disclosure will offend privacy and First Amendment principles alike. 22See infra Part IV.

I. The Speaker–Listener Hierarchy in First Amendment Law

Before jumping into the clash between free speech rights and freedom from thought, I would like to outline the First Amendment rules I will use to predict how courts will respond. The two most common approaches to First Amendment analysis are woefully inadequate for discrete projects like this one. At one extreme lives the general proposition that all content-based restrictions on speech are constitutionally suspect and will undergo exacting strict scrutiny. 23Reed v. Town of Gilbert, 135 S. Ct. 2218, 2226 (2015); Consol. Edison Co. v. Pub. Serv. Comm’n, 447 U.S. 530, 536–37 (1980). This principle is accurate in a very general sense, but it is far too simplistic to be useful here. It cannot account for a good deal of nuance that exists both in free speech case law and in the wide range of laws that are generally accepted without First Amendment challengers. 24See Frederick Schauer, The Boundaries of the First Amendment: A Preliminary Exploration of Constitutional Salience, 117 Harv. L. Rev. 1765, 1768 (2004). At the other extreme are analyses that start from scratch and ask how first principles should guide courts. 25Martin H. Redish, The Value of Free Speech, 130 U. Pa. L. Rev. 591, 591–96 (1982); Alexander Tsesis, Free Speech Constitutionalism, 2015 U. Ill. L. Rev. 1015, 1018–27. This approach has too much nuance to provide practical answers for this short study.

So, I will use two relevant speech theories put forward by Eugene Volokh and James Grimmelmann that inhabit the middle range between clear but misleading concrete rules and nuanced-but-impractical abstract principles. Both theories start with the presumption that communications between willing speakers and willing listeners have the greatest constitutional value, and each offers a means of addressing the legal ambiguities when willing speakers wish to communicate with unwilling listeners.

From Volokh, I borrow the distinction between one-to-one and one to many communications. 26See generally Eugene Volokh, One-to-One Speech vs. One-to-Many Speech, Criminal Harassment Laws, and “Cyberstalking, 107 Nw. U. L. Rev. 731 (2013) (discussing this distinction). Speech addressed to an unwilling listener does not have as strong a claim to constitutional protection if the unwilling speaker is the only recipient of the message. 27Id. at 742–43. This is a corollary to the basic presumption that the willing-speaker, willing-listener combination has the greatest value and receives the most constitutional protection. If a speaker communicates to many people at the same time, some of whom may be willing listeners, that communication will receive more protection than a communication directed to only one, unwilling listener. 28Id. at 751. To be sure, one to one communications with unwilling listeners must receive some constitutional protection. A political candidate will have an interest in calling or leaving literature with listeners who ex ante might identify themselves as uninterested in the message, and preserving this opportunity for the speaker to persuade is an important First Amendment goal. 29See Rowan v. U.S. Post Office Dep’t, 397 U.S. 728, 741 (1970) (Brennan, J., concurring) (highlighting the interest that children may have in receiving political, religious, or other materials); Mainstream Mktg. Servs. v. FTC, 358 F.3d 1228, 1229, 1233–34, 1238 (10th Cir. 2004). But the government will have more leeway, and attract less scrutiny, if it attempts to protect an unwilling listener from messages that are directed exclusively at him since the speaker’s autonomy clashes with the listener’s. 30See Volokh, supra note 26, at 750 (describing careful restrictions of unwanted one-to-one communication as “something of an exception to First Amendment protection”); see also Hill v. Colorado, 530 U.S. 703, 716–17 (2000) (“The unwilling listener’s interest in avoiding unwanted communication has been repeatedly identified in our cases. It is an aspect of the broader ‘right to be let alone’ that one of our wisest Justices characterized as ‘the most comprehensive of rights and the right most valued by civilized men.’” (quoting Olmstead v. United States, 277 U.S. 438, 478 (1928) (Brandeis, J., dissenting))).

Grimmelmann has begun very exciting new work to extend Volokh’s insights about unwilling listeners. 31James Grimmelmann, Active Listening (unpublished manuscript), . For Grimmelmann, the distinction between one-to-one and one-to-many communications is really a difference in “separation costs.” If it is costless (or nearly so) for a speaker to sort willing listeners from unwilling listeners and distribute a message to the former, then it would be trivially easy to convert a one-to-many communication to several one-to-one communications. 32Id. at 20. Moreover, if these sorting costs are very low, the government arguably should have some flexibility to require speakers to do that sorting in order to protect the interests of unwilling listeners. 33Id. Grimmelmann also accounts for the self-separation costs of listeners so that his model will not afford undue deference to unwilling listeners who can very easily avoid and ignore unwanted communications, but this Essay addresses communications that cannot easily be disregarded by the listener. Grimmelmann’s friendly amendments to Volokh’s theories are useful for this project because the government can more legitimately serve the interests of an unwilling listener if it first helps the speakers identify unwilling listeners. 34Id. at 14, 20–21 (adding the “many-to-one” classification of listeners, as well as the “separation costs” that sort willing from unwilling listeners).

I also make use of the compelled speech doctrine. In a nutshell, that doctrine applies something like intermediate scrutiny when the government compels a speaker to communicate a message that it would rather not communicate. 35Zauderer v. Office of Disciplinary Counsel of the Supreme Court of Ohio, 471 U.S. 626, 651 (1985). It is, of course, much more complex than this. For a thorough account of the factors that seem to push toward and away from constitutional intervention for compelled disclosures, see Jennifer M. Keighley, Can You Handle the Truth?: Compelled Commercial Speech and the First Amendment, 15 J. Const. L. 539 (2012); Toni M. Massaro, Tread on Me!, 17 U. Pa. J. Const. L. 365, 416–17 (2014). Compelled speech is all the more troubling for the authentic expression of speakers and the autonomy of listeners when neither speakers nor listeners wish for a communication to take place, and would presumably merit greater scrutiny if listeners were easily identified as unwilling participants in the compelled communication. 36I am not aware of case law that clearly articulates this position, but the unwilling-speaker, unwilling-listener is the compelled speech doctrine’s analogue to the willing-speaker, willing-listener combination when considering prohibitions on speech.

Combining these principles, I adopt and apply the following rules: (a) When a regulation concerns a willing speaker and willing listener, government interference will be heavily scrutinized; (b) When a regulation concerns a willing speaker and an unwilling listener, government interference will undergo some, but less, scrutiny; (c) When a regulation concerns a willing speaker and a listener of unknown status, government interference of the message will be heavily scrutinized unless the speaker can easily learn the status of the listener and avoid communicating to unwilling listeners; (d) When the government compels a communication, its justifications will be scrutinized especially carefully if the speaker and listener are both unwilling participants.

This set of rules will help frame and guide the discussion for the three categories of lawmaking we will explore. We start with lawmaking that protect a would-be listener’s interest in staying ignorant about themselves.

II. Willing Speaker, Unwilling Listener

A small group of American young adults share the harrowing experience of watching a parent suffer and die from Huntington’s disease—a rare and fatal disease that causes degeneration of brain cells. 37Robert Lentner, Huntington’s Disease, 2002 Mich. St. L. Rev. 921, 921 (discussing Huntington’s disease). At the same time, these young adults often must decide whether to get a genetic test to determine whether they, too, will suffer the same fate. For children of Huntington’s victims, the chances of carrying the gene and contracting the disease is 50%—frighteningly high. 38Richard H. Myers, Huntington’s Disease Genetics, 1 NeuroRx 255, 255 (2004). Thus, for the young individuals in these circumstances, the testing decision is a difficult philosophical choice: Do I want to know if I face certain, premature death?

This is an extreme example of the interest a person may have to stay ignorant about himself, but for that reason it creates a useful starting place to think through whether, and how, the government could support a freedom from knowledge with the force of law. Historically, lawmakers would have had little need to come to the aid of people at risk of having Huntington’s disease. Just a few years ago, the only would-be speakers who could know whether a person carries the gene were doctors and medical researchers, and neither would return this type of result to a patient unless the patient asked for it explicitly. But as the DNA home testing industry matures and attracts more entrants to the market, a person with casual interest in their health or ancestry could wind up learning something quite unsettling. Genomic researchers and genetic test producers are accustomed to warning subjects that they may turn out to be among the 4% of the population whose father is not their biological parent. 39Mark A. Bellis et al., Measuring Paternal Discrepancy and Its Public Health Consequences, 59 J. Epidemiology & Community Health 749, 751–52 (2005). So it would not be farfetched for a legislature to craft narrow or not-so-narrow rules to protect consumers who do not want to receive information that others may very well have.

Some privacy and health ethics laws already serve the purpose of shielding people from certain types of information about themselves. For example, California used to have a very strict prohibition on HIV testing. 40Cal. Health & Safety Code § 120990 (West 1995) (amended 2008). The HIV statutes prohibit testing rather than the reporting of HIV test results. But the purpose of the limitation on testing is to prohibit the creation of information that can harm the test subject. So I do not believe it is entirely misleading to characterize the HIV testing prohibitions as speech restrictions. In addition to nondisclosure rules, health providers and researchers were prohibited from routinely testing for HIV in patients without written consent. 41Id. In 2008, the legislature changed the law so that health providers could disclose their intent to screen for HIV and give the patient an opportunity to opt out, 42Cal. Health & Safety Code § 120990 (West 2012). HIV testing outside the medical field still requires written consent from the subject. but the statute continues to protect a patient from learning he has HIV if he prefers to avoid testing. This strong protection from knowledge is also separately and explicitly afforded to pregnant woman when they are in labor. 43Id. § 125090. This is a particularly strong legal commitment to the patient’s freedom from knowledge; modern HIV testing and treatment can decrease the chance HIV is transmitted to the infant, and yet even a woman showing symptoms of AIDS can avoid the test. 44HIV Testing in Health Care Settings—Legal Background, Cal. Dep’t Pub. Health, (last updated May 13, 2009). But the protection is not absolute. A person convicted for prostitution in California loses the freedom from knowledge and will be compelled to undergo HIV testing. 45Cal. Pen. Code § 1202.6 (West 2015). See generally Center for Infectious Diseases, Cal. Dep’t Pub. Health, California HIV/AIDS Laws, 2009, at 5–6 (Jan. 2010), .

New York has a curious approach to HIV. New York requires health professionals to get written consent before performing an HIV screen, and to respect a patient’s wishes if he does not consent to a test. 46N.Y. Comp. Codes R. & Regs. tit. 10, § 63.3(a) (2015). Like California, New York law protects a patient who does not want to know his HIV status. However, if a patient does consent, a doctor whose patient has tested positive for HIV is legally obligated to request the identities of the patient’s former partners and to make efforts to tell them that they are at risk. 47 HIV Reporting and Partner Notification Questions and Answers, N.Y. Dep’t of Health, (last updated Nov. 2013). And that message is delivered whether those former partners would like to know the information or not.

This schizophrenic treatment of a person’s interest in staying ignorant about HIV status (or risk) reflects the fact that one person’s freedom from knowledge can rarely be separated from the interests of other people. Family bonds and other intimate social connections complicate the value of ignorance, even when ignorance is in the interest of the listener. Even the Huntington’s disease example, which looked so promising for the ignorance rights argument, is deceptively complex. If a law prohibited disclosures about the Huntington’s disease gene to all subjects who opted out of (or failed to opt into) knowing their status, a censored speaker might argue that they have an interest in delivering the information despite the listener’s reluctance to receive it. The listener’s family may be better off if the listener is able to plan for his impending illness. And the listener himself may be better off in every possible way if the speaker is able to assure him that he does not carry the gene. (Note, though, that if this practice were widespread enough, it would compromise things for others who would rationally interpret silence as bad news.)

How would a law supporting a person’s freedom from knowledge fare under the First Amendment? Based on the principles outlined in the last Part, much will depend on how the law is set up.

A prohibition on communications that include multiple listeners, some of whom are willing listeners, will have to be justified by compelling state interests in order to survive scrutiny. For example, if a woman reveals that her son was the product of incest and rape on a daytime television program, whatever interests her son might have in self-ignorance will be overpowered by the speaker’s expressive interests and the television audience’s listener rights. 48This hypothetical is roughly based on Anonsen v. Donahue, 857 S.W.2d 700 (1993). This one-to-many communication will receive the greatest level of constitutional protection.

But let’s return to April Salazar, the woman who lost her pregnancy and then received the package from Enfantil reminding her about her would-be due date. 49Salazar, supra note 1. Although the individual piece of mail looks at first blush like a one to one communication, this would not accurately describe the communication from Enfantil’s point of view. Enfantil mechanically sent out mail with baby formula samples based on a limited amount of information about its recipients. (In her New York Times op-ed, Salazar argued that Enfantil had too much information about her, but Enfantil could argue that the problem was that they had too little.) Other than altering the timing of the message, Enfantil’s broadcast of formula samples is similar to the broadcast of a commercial message on television. Enfantil crafted one message and delivered it to many listeners, most of whom were willing listeners who may have benefited from the formula sample and the celebratory reminder that they are “almost there!” Unless Enfantil can easily separate Salazar and other unwilling listeners from the willing ones, their mailing is better understood to be a one-to-many communication. To prohibit the mailing, the government would have to be prepared to meet significant scrutiny (though perhaps not strict scrutiny since the mailing should qualify as commercial speech 50Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 564 (1980) (“If the communication is neither misleading nor related to unlawful activity, the government’s power is more circumscribed. The State must assert a substantial interest to be achieved by restrictions on commercial speech. Moreover, the regulatory technique must be in proportion to that interest. . . . Compliance with this requirement may be measured by two criteria. First, the restriction must directly advance the state interest involved . . . . Second, if the governmental interest could be served as well by a more limited restriction on commercial speech, the excessive restrictions cannot survive.”). ).

However, some legal solutions that stop short of blanket prohibition will be subjected to less First Amendment scrutiny. The key move for a government eager to empower unwilling listeners is to facilitate the separation between unwilling and willing listeners. If lawmakers were determined to protect women in Salazar’s position from receiving messages about pregnancy, they could create a notice scheme for unwilling listeners. A weak scheme would facilitate notice from unwilling listeners to each speaker so that the speaker knows the listener is not interested in pregnancy-related material. A middle scheme would permit unwilling listeners to provide notice to one organization and force that organization to alert all affiliates and business partners who received the listener’s contact information. And a stronger (unrealistic) scheme would set up a government-run “No Pregnancy” list that listeners could opt into in order to provide notice to all relevant commercial speakers that they do not want pregnancy-related messages. Whether the government forces speakers to respect the listener preferences or simply facilitates notice, each of the options transforms one-to-many communications into one-to-one communications.

It bears mentioning that these legal options impose some costs on the speakers. Although the notice schemes are not as restrictive as an outright ban, not every speaker will believe that they are better off identifying and removing unwanted listeners from a mass communication. But if we believe the communication to the unwilling speaker has a low value (and this is a big “if,” which I will return to), the separation costs will sometimes be cheaper for a speaker than the listener, and should for that reason be borne by them. 51Grimmelmann, supra note 31, at 20.

This determination of who should bear the cost of separation, though, are medium- and technology-specific. 52Id. Today, our means of receiving messages from the mail, the radio, and cable and broadcast television give listeners limited opportunity to filter information before learning about its content. The mailbox, the AM/FM radio, and (to some extent) the television are “stupid” receiving devices. 53Yochai Benkler, Some Economics of Wireless Communications, 16 Harv. J.L. & Tech. 25, 38 (2002). Cable services have an expanding array of options for subscribers who want to block certain types of content for family or religious reasons, but these options make very crude distinctions between content. Messages in their media are expensive for the listener to sort through without learning something about the content. Contrast this with email, which gives listeners a chance to set up filters to sort out unwanted content without the listener’s awareness and direct involvement. 54These are not perfect screens, of course, so the listener bears some costs in setting up the filters and losing content that they didn’t actually want screened out. Salazar, for example, would have had a hard time setting up a filter that effectively screened out unwanted pregnancy-related messages without also screening out personal emails that reference her pregnancy.

Let’s step back for a moment to understand why First Amendment law might have less to say about a “No Pregnancy” notice scheme than a ban on Enfamil’s mailings (or, for that matter, a ban on the disclosure of personal information from the pregnancy website to Enfamil). With the speech bans, the regulation interferes with willing speakers and some willing listeners. Since these communications involve willing speakers and listeners who have found each other, the government’s reasons for blocking the message are viewed skeptically. 55Volokh, supra note 26, at 743. The notice scheme, on the other hand, allows the unwilling listeners to self-identify. 56Compare Lamont v. Postmaster Gen., 381 U.S. 301, 307 (1965) (burden of filling out a reply card to receive Communist propaganda by mail was an unconstitutional burden on willing listeners), with Rowan v. U.S. Post Office Dep’t, 397 U.S. 728, 740 (1970) (statute permitting unwilling listeners to opt out of receiving mail from advertisers was a constitutional limitation on speakers). These willing-speaker, unwilling-listener matchups can then be minimized with more public law or through private ordering.

A consent scheme (like the HIV laws) that require a speaker to check with the listener about their willingness to receive a message will fall somewhere between a notice scheme and an outright ban. Like the notice schemes, a consent scheme facilitates the separation of unwilling listeners. But the process of identifying unwilling listeners is much more burdensome since the listeners are not expected to do the work of identifying themselves. Moreover, the process of asking for consent sends a meta-message that the content the speaker wishes to disclose could be dangerous. So, while the regulation of messages that the speaker knowingly directs at unwilling listeners presumably receives lower constitutional scrutiny (by supposition from the model I laid out in Part I), this more relaxed scrutiny should be heightened again if the government imposes an elaborate process on the speaker in order to discover that the listener is unwilling.

Let me now return briefly to that “if” I left emphasized but unexplained above. I began this project with a great amount of sympathy for the plight of the unwilling listener who attempts to live his life in an autonomous and self determined way. These sympathies are both intellectual and personal; I can think of a number of occasions in my life when I decided to avoid certain types of information relevant to my life and made efforts to ensure my ignorance. But after thinking about the freedom from knowledge (particularly self-knowledge) for this project, I have concluded that the First Amendment should provide more than casual protection to a speaker, even when he wishes to communicate to an unwilling listener, and even when he wishes to tell the listener about the listener himself.

I came to this conclusion because I could find only two instances in which a speaker has negligible interest in making the communication: when the speaker is harassing the listener (the subject of Volokh’s work on which I rely), or when the speaker is mistaken about the listener’s preferences. 57Volokh, supra note 26. The latter is best tackled by law, if it is tackled at all, through a notice statute alone without any accompanying prohibition on communications. And the former is best treated as a separate narrow category—perhaps even one that could survive heightened scrutiny anyways because of the vindictive and uninspiring interests of the speaker. In all other instances—that is, in the vast majority of noncommercial cases—the speaker believes that the message is in the best interest of the listener, in the best interest of third parties, in the speaker’s own interest, or some combination thereof. Like the politician who hopes to persuade a skeptic, speakers with a message to and about a skeptical listener have First Amendment value that cannot be discounted through listener preferences alone. Thus, except for exceptional cases, the law is better off facilitating notice of a listener’s lack of interest in certain types of messages. Otherwise we should leave it to social norms and the decency of the speakers to refrain from contacting unwilling listeners.

The next two Parts explore (in less detail than this one did) lawmaking that interferes with listener preferences rather than supporting them.

III. Willing Speaker, Willing Listener

Next we consider the circumstances in which the speaker wants to tell the listener something about herself, the listener wants to listen, and the government decides to stifle the communication. This Part will be brief because the theory outlined in Part I and existing case law suggest that this setup should be exceedingly rare and exceedingly difficult for the government to survive scrutiny. After all, the willing-speaker, willing-listener is the favored combination in free speech law, and this favoritism should only grow when the message offers the chance for the listener to become enlightened about himself.

Despite the obvious application of strong First Amendment protection, it is not difficult to find examples of this sort of speech restriction. Two recent examples suffice to make the point.

First is the Food and Drug Administration’s action (or threat of action) against 23andMe for reporting health information to its customers. 58Letter from Alberto Gutierrez, Dir., Office of In Vitro Diagnostics & Radiological Health, U.S. Food & Drug Admin., to Ann Wojcicki, CEO, 23andMe, Inc. (Nov. 22, 2013), . 23andMe sells DNA kits for under $100, and for several years customers who used the service received both ancestry and health information. 5923andMe, (last visited Sept. 16, 2015); Status of Our Health-Related Genetic Reports, 23andMe, (last visited Oct. 17, 2015). The health information reported the results of genome-wide association studies (GWASs) for the genes that the customer actually had. 60Staff Reporter, 23andMe Publishes Web-Based GWAS Using Self-Reported Trait Data, GenomeWeb (June 25, 2010), . To be clear, 23andMe did not (typically) produce these studies. Instead, it curated the results based on the customer’s genotype. 61Id. The FDA claims to have jurisdiction over 23andMe by classifying the at-home DNA sample kit as a “medical device.” 62Letter from Alberto Gutierrez, supra note 64. This is a little bit of regulatory sleight of hand because the many at-home DNA kits currently available without reporting the GWASs are not “medical,” and the provision of GWASs without a home kit is not a “device.” It was only the combination of the kit and the studies, by the same company, that gave the FDA an opening. 6321 U.S.C. § 321(h) (West 2013) (providing the definition of “device”). As a medical device, 23andMe is expected to submit evidence of efficacy for each GWAS before reporting them to customers (even though the same GWASs are relied on routinely by doctors in the course of care). 64Letter from Alberto Gutierrez, supra note 58.

Of all the theories explaining the FDA decision to shut down 23andMe’s health reporting, the only non-cynical one is a bit paternalistic. 65The others include a public-choice theory that doctors influenced the FDA so that patients would have to come to them to have their genomic profile interpreted and an even less generous theory that the FDA is engaged in a regulatory power grab. Apparently one of the FDA’s concerns was that 23andMe customers would overreact to the information they received. 66See George J. Annas & Sherman Elias, 23andME and the FDA, 370 New Eng. J. Med. 985, 985 (2014). If a GWAS suggested the customer had an elevated risk of some sort, the customer may demand more invasive medical testing that wastes resources and exposes him to more health risks.

A second example comes from the trenches of the abortion debates. Indiana’s legislature is considering a bill that would outlaw aborting a fetus because of its sex or because of a prenatal diagnosis of disease. 67S.B. 334, 119th Gen. Assemb., Reg. Sess. (Ind. 2015). Although the law does not forbid prenatal screening for diseases like Down syndrome, enforcement of the law may require as a practical matter that expecting parents stay ignorant about certain congenital diseases. If sex-selective abortion were common in the United States, even the knowledge of an aborted fetus’s sex could ensnare the parents in suspicion (if the state’s ban on sex- or disease selective abortions is itself constitutional).

Both of these communications—the reporting of genetic health risks of 23andMe customers or the reporting of fetal sex and health risks to expecting couples—could be banned by the government if it had a compelling justification. That is, if the government had sufficient evidence that the listeners were actually using the information in a way that harmed themselves or others, the state may be able to withstand the First Amendment scrutiny. But without clear and convincing evidence of harm the government should not be permitted to cut off a willing speaker’s message from its willing listener. 68This is especially true with the current speech-protective Supreme Court that scrutinizes even the regulation of lies. See United States v. Alvarez, 132 S. Ct. 2537, 2548–51 (2012).

IV. Unwilling Speaker, Unwilling Listener

The final category of lawmaking that affects the freedom from knowledge involves legal rules that coerce knowledge. Laws in this category mandate a speaker to tell the listener something about himself even if the speaker does not want to say it, and the listener does not want to hear it.

Of the three categories of lawmaking examined in this Essay, this category is by far the most prevalent because it contains a wide range of compulsory disclosures that organizations and professionals must make to their customers and clients, some of whom are uninterested or worse. Most mandatory disclosures by regulated entities are messages related to a product or service the listener is about to use. For example, the nutrition label on the packaging of nearly every processed food ensures that a person consuming the food knows some basic nutritional information about the food.

Nutrition labels are a ubiquitous example, but they might not be the best since an unwilling listener could avoid the information by taking modest precautions. That is, a man on vacation who does not want to know how many calories are in the Twinkie he is about to consume can simply avoid turning the packaging around or staring too closely at the label. This depends on listeners to pay attention to convention and self-manage their attention, but this is a reasonable price for the social benefit offered to all the consumers who do want easy access to nutritional information. 69Whether this sort of listener self-management is possible or not is the critical question determined by the application of the “captive audience doctrine” within compelled speech analysis. See Caroline Mala Corbin, The First Amendment Right Against Compelled Listening, 89 B.U. L. Rev. 939, 944 (2009). However, these sorts of precautions are not available to New Yorkers dining at fast food restaurants. In New York City and California, calorie counts must be reported on the menus and signs at all chain restaurants, and often the calories are reported right next to prices. 70N.Y.C., N.Y., Health Code § 81.50 (2011). These requirements make avoiding the information impractical or impossible for most consumers, much to the chagrin of the New Yorkers who resented being nudged off their Big Macs and Frappuccinos. The Affordable Care Act will soon require all restaurants with twenty or more locations to report calorie counts on all signs and menus (including advertisements). 71Jason Stverak, Obamacare’s Restaurant Calorie-Label Mandate Is a Complete Mess, Forbes (Feb. 7, 2014, 2:46 PM), . But see N.Y.U. Langone Med. Ctr., Mandatory Calorie Postings at Fast-Food Chains Often Ignored or Unseen, Does Not Influence Food Choice, EurekAlert! (Nov. 15, 2013), .

These mandatory disclosures serve an obvious public health purpose and would presumably survive First Amendment scrutiny under the compelled speech doctrine since some (perhaps many) of the listeners appreciate and use the calorie information. Using the admittedly simple rules for compelled speech described in Part I, courts could apply a less rigorous, intermediate form of free speech scrutiny in cases involving compelled disclosures by commercial actors, even if a subset of listeners dislike the message, as long as some of the audience are willing listeners.

Next consider a more ideologically divisive example. Texas and North Carolina each have statutes requiring abortion providers to perform a sonogram and display the video image to patients before carrying out an abortion procedure. 72The Texas statute, which the Fifth Circuit upheld, allowed abortion patients to opt out of seeing the sonogram images but required doctors to describe and explain what he saw on the sonogram even if the patient did not observe the images. Tex. Med. Providers Performing Abortion Servs. v. Lakey, 667 F.3d 570, 583, 577–78 (5th Cir. 2012). These abortion disclosure rules differ from others that have been adopted and debated requiring the disclosures about the alleged increased risk of suicidal ideation and suicide caused by abortions. 73See Planned Parenthood of Minn., N.D., S.D. v. Rounds, 686 F.3d 889, 905–06 (8th Cir. 2012) (en banc) (Rounds II); Planned Parenthood Minn. v. Rounds, 653 F.3d 662, 672–73 (8th Cir. 2011) (Rounds I). The suicidal ideation disclosures are considered by most doctors and public health experts to be incompatible with existing evidence, and the objections to those disclosures have therefore focused on the inaccuracy and disinformation in the disclosures. 74See Robert Post, Informed Consent to Abortion: A First Amendment Analysis of Compelled Physician Speech, 2007 U. Ill. L. Rev. 939, 970–71. The compelled sonograms, by contrast, cannot be accused of providing false information even if they prey on the emotional vulnerability of the listener. 75The “ick factor” of the message does play a role in the compelled speech analysis. Courts will perceive messages and images that are designed to shock a viewer or otherwise overbear their cognitive reasoning as ideological persuasion and will attract more judicial skepticism. See R.J. Reynolds Tobacco Co. v. Food & Drug Admin., 696 F.3d 1205, 1217, 1221–22 (D.C. Cir. 2012) (striking down a requirement to place graphic warning labels on cigarette packaging). Since the state coerces the viewing of sonogram images for its persuasive content rather than its informational content, this aspect of the compelled speech doctrine should apply to strike down compelled sonograms.

Even setting aside the question of accuracy, the abortion-related disclosures are distinguishable from food nutrition labeling because of the legal and ethical duties that attach to doctor–patient relationships and other relationships of trust. When the state uses a trusted fiduciary as a mouthpiece, the doctrine may not apply the same way as it would in a commercial context. 76Especially since access to abortion has its own constitutional guarantees. See Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 855–56 (1992). And there is another difference that may be more enlightening and applicable to a wider range of scenarios for our purposes: these communications are one-to-one rather than one-to-many. Before an unwilling speaker (the doctor) displays the sonogram images to an unwilling listener (the patient), the patient can identify herself as uninterested. By alerting the doctor that she does not consent to the viewing of the sonogram images, the patient has the power to convert the law as applied to one that coerces a communication between a certain unwilling speaker and unwilling listener. If challenged under these conditions, then, Texas and North Carolina should not be able to rely on the possibility of interested listeners, and should have to justify the application of the law under heightened scrutiny.

In other words, if the unwilling-speaker, unwilling-listener connection is the one at the center of the compelled speech doctrine (much as the willing–willing link is at the core of basic free speech freedoms), then compelled disclosures that require one-on-one conversations should receive the highest form of scrutiny.

Up to this point, I have assumed that commercial disclosures like nutrition labels and calorie counts will be analyzed indefinitely as one-to-many communications, but this is not necessarily so. The severability of one-to-many communications into one-to-one communications is just as important to the future of the compelled speech doctrine as it is for speech bans. The nutritional and calorie disclosures were analyzed above using the presumptions that are appropriate for one-to-many communications—namely, that some listeners will be interested in the message even if others are not. But with technological advances, it may be increasingly easy to permit unwilling listeners to opt out of disclosures if they and the speakers are not interested in communicating. Indeed, if it is very easy to convert one-to-many messages into customized one-to-one messages, courts may have to increase scrutiny when the government needlessly compels a one-to-many message (thereby exposing unwilling listeners).

The compelled abortion disclosures that take place in doctor’s offices are one context where the one-to-one nature of the communication easily permits a patient to opt out of disclosures that the government by default would require. But oral conversations are not the only ones amenable to this type of customizing. Websites, too, can be coded to allow for listener opt out if both the owner and reader of the web material find the compelled disclosure conflicts with their personal beliefs or preferences. Thus, restaurants like Krispy Kreme and the Heart Attack Grill may have no choice but to include calories on their signage and print media marketing materials, but could (under this proposed scheme) permit web visitors to click on an a button that removes the calorie information and spares them from the comprehension of their impending unhealthy food choices.

Conclusions

I have used a few simple rules-of-thumb about the outcomes of free speech cases to anticipate whether the First Amendment can tolerate an interest in self-ignorance—that is, in a freedom from thought. These rules are gracelessly but efficiently illustrated here with graphics.

First, communications between a willing speaker and a willing listener are at the core of First Amendment protections while communications from a willing speaker to an unwilling listener are less protected, and more easily regulated.

j-bambauer-fig1

Next, communications from a willing speaker that are disseminated to multiple listeners are presumptively protected because these one-to-many communications contain that key willing-speaker, willing-listener pair.

j-bambauer-fig2

If one-to-many communications can be separated into a parallel series of one-to-one conversations without burdening the speaker, then the government may be in a better position to justify speech restrictions on the communications to unwilling listeners.

j-bambauer-fig3

This separation task is not trivial, so laws that compel speakers to determine the status of their listeners must be scrutinized to ensure that the burden really is better placed on the speaker to separate out unwilling listeners (rather than on the listener to separate out himself). However, if the speaker is in a better position to identify unwilling listeners, a law requiring the inquiry can be justified. Moreover, a law may have a better chance of withstanding free speech scrutiny if it requires speakers to inquire about the interest of its listeners without imposing a ban. After all, outside of harassment, speakers who wish to communicate a message even though they know their listeners are uninterested have some reason to think that the logic or morality of their message should change the listener’s mind, and again outside of harassment, the First Amendment will generally permit them to try. 77Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting) (introducing the concept of the marketplace of ideas). Let’s revisit the doctor–patient relationship to see what I mean. When the law forbids a doctor from communicating to a patient about his HIV status 78See supra notes 40– 46 and accompanying text. or about whether he owns guns, 79Wollschlaeger v. Governor of Florida, No. 12-14009, 2015 WL 4530452, at *6–7 (11th Cir. July 28, 2015) (upholding a ban on doctor inquiries about patient gun ownership). a doctor would have ample reason to carry out the communication despite his patient’s preference to stay ignorant. The information could be more useful to the patient than the patient realizes—for example, if the patient does not know about effective new HIV treatments or about relatively simple measures to keep children away from guns. Moreover, even if the conversation is uncomfortable for the patient, the doctor may still feel compelled to pass along information in case it proves beneficial to third parties, by causing the patient to engage in safe sex or safe gun practices.

All of these caveats make the larger implications of this scheme less than promising for privacy, even for the decidedly narrow category of interests related to managing self-knowledge. However, the privacy and dignity interests for an unwilling listener are much more profound in the area of compelled speech. When the government forces a speaker to deliver a particular message, the First Amendment priorities are inverted. A compelled disclosure to many people can more easily survive scrutiny since an unwilling speaker may be providing information that a listener finds valuable. Thus, regulations requiring one-to-many disclosures are better able to withstand judicial scrutiny. But a one-on-one conversation between an unwilling speaker and an unwilling listener should be drafted with much more caution by legislatures, and should be scrutinized with much more skepticism by the courts.

For these communications, courts cannot infer from the speaker some countervailing reason to permit the message despite the preferences of the listener, for the speaker is also an unwilling participant. Thus, disclosure regimes can put free speech and privacy interests on the same side of the ledger, particularly when the communication has no other potentially interested listeners. 80This marriage between privacy and the compelled speech doctrine is not new. See Corbin, supra note 69, at 952–55. The modest insight offered here is that the validity of these arguments increase when the message is converted from a one-to-many format to a one-to-one conversation. If they do not give listeners a way to opt out, they simultaneously burden speaker and autonomy interests.

j-bambauer-fig4

Footnotes

*Associate Professor of Law, University of Arizona James E. Rogers College of Law. J.D., Yale Law School; B.S., Yale College. Many thanks to Derek Bambauer, Caroline Corbin, Eugene Volokh, and Alexander Tsesis for helpful feedback, and to Ryan Pulley for superb editing. I am also tremendously grateful to Alexander Tsesis, Gerard Bifulco, and the Emory Law Journal for the opportunity to participate in the 2015 Thrower Symposium.

1April Salazar, The Pregnancy Is Gone, but the Promotions Keep Coming, N.Y. Times: Motherlode (Feb. 2, 2014, 8:25 AM), .

2Id.

3Id.

4Other women who have experienced miscarriage found themselves in similar circumstances to Salazar with respect to targeted advertising. E.g., Why Targeted Ads Suck, Maybe Baby? (Apr. 27, 2009), ; Emily Wilkins, Life After Miscarriage: Reminders in the Mail, Pea in the Podcast (July 18, 2010), .

5James Carmichael, Google Knows You Better than You Know Yourself, Atlantic (Aug. 19, 2014), .

6Tim Harford, Big Data: Are We Making a Big Mistake?, Fin. Times Mag. (Mar. 28, 2014, 11:38 AM), ; Gregory Piatetsky, Did Target Really Predict a Teen’s Pregnancy?: The Inside Story, KDNuggets (May 7, 2014), .

7Charles Duhigg, How Companies Learn Your Secrets, N.Y. Times Mag. (Feb. 16, 2012), .

8E.g., Daniel Castro, A Catalog of Every “Harm” in the White House Big Data Report, Ctr. for Data Innovation (July 15, 2014), .

9See Richard Posner, The Right of Privacy, 12 Ga. L. Rev. 393, 395 (1978).

10For a description of the commodification critique of data collection, see Alexander Furnas, It’s Not All About You: What Privacy Advocates Don’t Get About Data Tracking on the Web, Atlantic (Mar. 15, 2012), .

11Salazar, supra note 1.

12Or, in Salazar’s case, self-reminding.

13See generally Susan Nolen-Hoeksema, Blair E. Wisco & Sonja Lyubomirsky, Rethinking Rumination, 3 Persp. on Psychol. Sci. 400 (2008) (discussing potential negative side effects of self-reflection, including depression and anxiety).

14See generally Andrew Steptoe et al., Satisfaction with Communication, Medical Knowledge, and Coping Style in Patients with Metastatic Cancer, 32 Soc. Sci. & Med. 627 (1991) (describing oncology patients’ coping strategies as either information-seeking or avoidance).

15Chandra Steele, How to Avoid Your Ex on Social Media, PC Mag. (Feb. 13, 2014), .

16See generally Gail P. Jarvik et al., Return of Genomic Results to Research Participants: The Floor, the Ceiling, and the Choices in Between, 94 Am. J. Human Genetics 818 (2014) (discussing whether results with significant health implications should be reported to participants undergoing genetic studies).

17Daniel Engber, Who’s Your Daddy?: The Perils of Personal Genomics, Slate (May 21, 2013, 5:45 AM), .

18See, e.g., Julie E. Cohen, What Privacy Is For, 126 Harv. L. Rev. 1904, 1910–11 (2013); Ruth Gavison, Privacy and the Limits of Law, 89 Yale L.J. 421, 455 (1980) (moral autonomy); Daniel J. Solove, A Taxonomy of Privacy, 154 U. Pa. L. Rev. 477, 553 (2006) (decisional privacy).

19One exception is new work by Kiel Brennan-Marquez theorizing that one of the under appreciated harms from surveillance comes from the anxiety and attention demanded by a subject who becomes aware that he is being watched and judged, and therefore judges himself. Brennan-Marquez questions whether the optimal amount of information about ourselves in order to maximize human flourishing is something less than what modern technology can offer. See generally Kiel Brennan-Marquez, The Freedom Not to Think (unpublished manuscript), .

20See, e.g., Jane Bambauer, All Life Is an Experiment. (Sometimes It’s a Controlled Experiment.), 47 Loyola U. Chi. L.J. (forthcoming 2015); Jane Bambauer, Is Data Speech?, 66 Stan. L. Rev. 57, 63 (2014).

21See infra Parts II & III.

22See infra Part IV.

23Reed v. Town of Gilbert, 135 S. Ct. 2218, 2226 (2015); Consol. Edison Co. v. Pub. Serv. Comm’n, 447 U.S. 530, 536–37 (1980).

24See Frederick Schauer, The Boundaries of the First Amendment: A Preliminary Exploration of Constitutional Salience, 117 Harv. L. Rev. 1765, 1768 (2004).

25Martin H. Redish, The Value of Free Speech, 130 U. Pa. L. Rev. 591, 591–96 (1982); Alexander Tsesis, Free Speech Constitutionalism, 2015 U. Ill. L. Rev. 1015, 1018–27.

26See generally Eugene Volokh, One-to-One Speech vs. One-to-Many Speech, Criminal Harassment Laws, and “Cyberstalking, 107 Nw. U. L. Rev. 731 (2013) (discussing this distinction).

27Id. at 742–43.

28Id. at 751.

29See Rowan v. U.S. Post Office Dep’t, 397 U.S. 728, 741 (1970) (Brennan, J., concurring) (highlighting the interest that children may have in receiving political, religious, or other materials); Mainstream Mktg. Servs. v. FTC, 358 F.3d 1228, 1229, 1233–34, 1238 (10th Cir. 2004).

30See Volokh, supra note 26, at 750 (describing careful restrictions of unwanted one-to-one communication as “something of an exception to First Amendment protection”); see also Hill v. Colorado, 530 U.S. 703, 716–17 (2000) (“The unwilling listener’s interest in avoiding unwanted communication has been repeatedly identified in our cases. It is an aspect of the broader ‘right to be let alone’ that one of our wisest Justices characterized as ‘the most comprehensive of rights and the right most valued by civilized men.’” (quoting Olmstead v. United States, 277 U.S. 438, 478 (1928) (Brandeis, J., dissenting))).

31James Grimmelmann, Active Listening (unpublished manuscript), .

32Id. at 20.

33Id. Grimmelmann also accounts for the self-separation costs of listeners so that his model will not afford undue deference to unwilling listeners who can very easily avoid and ignore unwanted communications, but this Essay addresses communications that cannot easily be disregarded by the listener.

34Id. at 14, 20–21 (adding the “many-to-one” classification of listeners, as well as the “separation costs” that sort willing from unwilling listeners).

35Zauderer v. Office of Disciplinary Counsel of the Supreme Court of Ohio, 471 U.S. 626, 651 (1985). It is, of course, much more complex than this. For a thorough account of the factors that seem to push toward and away from constitutional intervention for compelled disclosures, see Jennifer M. Keighley, Can You Handle the Truth?: Compelled Commercial Speech and the First Amendment, 15 J. Const. L. 539 (2012); Toni M. Massaro, Tread on Me!, 17 U. Pa. J. Const. L. 365, 416–17 (2014).

36I am not aware of case law that clearly articulates this position, but the unwilling-speaker, unwilling-listener is the compelled speech doctrine’s analogue to the willing-speaker, willing-listener combination when considering prohibitions on speech.

37Robert Lentner, Huntington’s Disease, 2002 Mich. St. L. Rev. 921, 921 (discussing Huntington’s disease).

38Richard H. Myers, Huntington’s Disease Genetics, 1 NeuroRx 255, 255 (2004).

39Mark A. Bellis et al., Measuring Paternal Discrepancy and Its Public Health Consequences, 59 J. Epidemiology & Community Health 749, 751–52 (2005).

40Cal. Health & Safety Code § 120990 (West 1995) (amended 2008). The HIV statutes prohibit testing rather than the reporting of HIV test results. But the purpose of the limitation on testing is to prohibit the creation of information that can harm the test subject. So I do not believe it is entirely misleading to characterize the HIV testing prohibitions as speech restrictions.

41Id.

42Cal. Health & Safety Code § 120990 (West 2012). HIV testing outside the medical field still requires written consent from the subject.

43Id. § 125090.

44HIV Testing in Health Care Settings—Legal Background, Cal. Dep’t Pub. Health, (last updated May 13, 2009).

45Cal. Pen. Code § 1202.6 (West 2015). See generally Center for Infectious Diseases, Cal. Dep’t Pub. Health, California HIV/AIDS Laws, 2009, at 5–6 (Jan. 2010), .

46N.Y. Comp. Codes R. & Regs. tit. 10, § 63.3(a) (2015).

47 HIV Reporting and Partner Notification Questions and Answers, N.Y. Dep’t of Health, (last updated Nov. 2013).

48This hypothetical is roughly based on Anonsen v. Donahue, 857 S.W.2d 700 (1993).

49Salazar, supra note 1.

50Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 564 (1980) (“If the communication is neither misleading nor related to unlawful activity, the government’s power is more circumscribed. The State must assert a substantial interest to be achieved by restrictions on commercial speech. Moreover, the regulatory technique must be in proportion to that interest. . . . Compliance with this requirement may be measured by two criteria. First, the restriction must directly advance the state interest involved . . . . Second, if the governmental interest could be served as well by a more limited restriction on commercial speech, the excessive restrictions cannot survive.”).

51Grimmelmann, supra note 31, at 20.

52Id.

53Yochai Benkler, Some Economics of Wireless Communications, 16 Harv. J.L. & Tech. 25, 38 (2002). Cable services have an expanding array of options for subscribers who want to block certain types of content for family or religious reasons, but these options make very crude distinctions between content.

54These are not perfect screens, of course, so the listener bears some costs in setting up the filters and losing content that they didn’t actually want screened out. Salazar, for example, would have had a hard time setting up a filter that effectively screened out unwanted pregnancy-related messages without also screening out personal emails that reference her pregnancy.

55Volokh, supra note 26, at 743.

56Compare Lamont v. Postmaster Gen., 381 U.S. 301, 307 (1965) (burden of filling out a reply card to receive Communist propaganda by mail was an unconstitutional burden on willing listeners), with Rowan v. U.S. Post Office Dep’t, 397 U.S. 728, 740 (1970) (statute permitting unwilling listeners to opt out of receiving mail from advertisers was a constitutional limitation on speakers).

57Volokh, supra note 26.

58Letter from Alberto Gutierrez, Dir., Office of In Vitro Diagnostics & Radiological Health, U.S. Food & Drug Admin., to Ann Wojcicki, CEO, 23andMe, Inc. (Nov. 22, 2013), .

5923andMe, (last visited Sept. 16, 2015); Status of Our Health-Related Genetic Reports, 23andMe, (last visited Oct. 17, 2015).

60Staff Reporter, 23andMe Publishes Web-Based GWAS Using Self-Reported Trait Data, GenomeWeb (June 25, 2010), .

61Id.

62Letter from Alberto Gutierrez, supra note 64.

6321 U.S.C. § 321(h) (West 2013) (providing the definition of “device”).

64Letter from Alberto Gutierrez, supra note 58.

65The others include a public-choice theory that doctors influenced the FDA so that patients would have to come to them to have their genomic profile interpreted and an even less generous theory that the FDA is engaged in a regulatory power grab.

66See George J. Annas & Sherman Elias, 23andME and the FDA, 370 New Eng. J. Med. 985, 985 (2014).

67S.B. 334, 119th Gen. Assemb., Reg. Sess. (Ind. 2015).

68This is especially true with the current speech-protective Supreme Court that scrutinizes even the regulation of lies. See United States v. Alvarez, 132 S. Ct. 2537, 2548–51 (2012).

69Whether this sort of listener self-management is possible or not is the critical question determined by the application of the “captive audience doctrine” within compelled speech analysis. See Caroline Mala Corbin, The First Amendment Right Against Compelled Listening, 89 B.U. L. Rev. 939, 944 (2009).

70N.Y.C., N.Y., Health Code § 81.50 (2011).

71Jason Stverak, Obamacare’s Restaurant Calorie-Label Mandate Is a Complete Mess, Forbes (Feb. 7, 2014, 2:46 PM), . But see N.Y.U. Langone Med. Ctr., Mandatory Calorie Postings at Fast-Food Chains Often Ignored or Unseen, Does Not Influence Food Choice, EurekAlert! (Nov. 15, 2013), .

72The Texas statute, which the Fifth Circuit upheld, allowed abortion patients to opt out of seeing the sonogram images but required doctors to describe and explain what he saw on the sonogram even if the patient did not observe the images. Tex. Med. Providers Performing Abortion Servs. v. Lakey, 667 F.3d 570, 583, 577–78 (5th Cir. 2012).

73See Planned Parenthood of Minn., N.D., S.D. v. Rounds, 686 F.3d 889, 905–06 (8th Cir. 2012) (en banc) (Rounds II); Planned Parenthood Minn. v. Rounds, 653 F.3d 662, 672–73 (8th Cir. 2011) (Rounds I).

74See Robert Post, Informed Consent to Abortion: A First Amendment Analysis of Compelled Physician Speech, 2007 U. Ill. L. Rev. 939, 970–71.

75The “ick factor” of the message does play a role in the compelled speech analysis. Courts will perceive messages and images that are designed to shock a viewer or otherwise overbear their cognitive reasoning as ideological persuasion and will attract more judicial skepticism. See R.J. Reynolds Tobacco Co. v. Food & Drug Admin., 696 F.3d 1205, 1217, 1221–22 (D.C. Cir. 2012) (striking down a requirement to place graphic warning labels on cigarette packaging). Since the state coerces the viewing of sonogram images for its persuasive content rather than its informational content, this aspect of the compelled speech doctrine should apply to strike down compelled sonograms.

76Especially since access to abortion has its own constitutional guarantees. See Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 855–56 (1992).

77Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting) (introducing the concept of the marketplace of ideas).

78See supra notes 40– 46 and accompanying text.

79Wollschlaeger v. Governor of Florida, No. 12-14009, 2015 WL 4530452, at *6–7 (11th Cir. July 28, 2015) (upholding a ban on doctor inquiries about patient gun ownership).

80This marriage between privacy and the compelled speech doctrine is not new. See Corbin, supra note 69, at 952–55. The modest insight offered here is that the validity of these arguments increase when the message is converted from a one-to-many format to a one-to-one conversation.