¹ “If there is any one proposition that commands general agreement among theorists and practitioners of the penal law, it is that judicial punishment ought not to be inflicted for private thoughts … ” (Alan Brudner, Punishment and Freedom: A Liberal Theory of Penal Justice [Oxford: Oxford University Press, 2009], 108, as quoted in Gabriel Mendow, “Why Is It Wrong to Punish Thought?” Yale Law Journal 127, no. 8 [2018]: 2342–86, at 2345, with Mendlow observing that “It’s a venerable maxim of criminal jurisprudence that the state must never punish people for their mere thoughts …”). And see also Harold Laski, Authority in the Modern State (New Haven, CT: Yale University Press, 1919), 22, announcing that “[f]reedom of thought, then, the modern state must regard as absolute; … ,” and George Sher, “A Wild West of the Mind,” Australasian Journal of Philosophy 97, no. 3 (2019): 483–96, concluding (484) that “the realm of the purely mental is best regarded as a morality-free zone. Within that realm, no thoughts or attitudes are either forbidden or required.”

² See, most prominently, Article 18 of the 1948 Universal Declaration of Human Rights, which states that “Everyone shall have the right to freedom of thought, conscience, and religion.” Article 9 of the European Convention on Human Rights repeats the same language. And Article 13 of the American Convention on Human Rights, adopted by the Inter-American Specialized Conference on Human Rights in 1969, guarantees the “freedom of thought and expression.” For influential endorsements of the idea of freedom of thought as a moral goal, see Seana Valentine Shiffrin, Speech Matters: On Lying, Morality, and the Law (Princeton, NJ: Princeton University Press, 2014), 79–82; Gabriel S. Mendlow, “Why Is It Wrong to Punish Thought?” Meir Dan-Cohen, “Harmful Thoughts,” Law and Philosophy 18, no. 4 (1999): 379–405, at 379, notes the “inviolability” of thoughts, but the label is agnostic as between the descriptive claim that thoughts are unreachable as a matter of physiology and the normative claim, which is Dan-Cohen’s focus, that thoughts ought to be treated as immune from state restriction.

³ This conclusion is similar to that in Mendlow, “Why Is It Wrong to Punish Thought?” Mendlow, although focusing largely on the reach of the criminal law as a matter of (moral) criminal law theory, and not, as I do here, on freedom of thought as a putative political, moral, legal, or constitutional right, recognizes the difficulties with the standard platitudes about freedom of thought, but then defends, contrary to my argument here, a distinct thought-based moral principle marking one of the boundaries of the criminal law.

⁴ And, of course, from observing those actions of others that we seek to avoid.

⁵ Mendlow, “Why Is It Wrong to Punish Thought?” does not set them aside, and bases his defense of a categorical moral prohibition on criminalizing pure intent on the notion of mental integrity, the extreme form of which is the use of “a mind-altering drug in order to disrupt your criminal intentions” (ibid., 2368). For Mendlow, punishing (but not necessarily otherwise restricting [ibid., 2379]) for intentions (or other thoughts) is analogous to using pharmacological or technological means to change one’s thoughts, but I resist the notion that there is—the theory of the criminal law aside, and the theory of punishment qua punishment aside—some profound difference between highly effective means of persuasion, especially where the target understands that she is being persuaded, and sanctioning someone for their thoughts, or making it more difficult or costly to have some thoughts rather than others.

⁶ This was very much the enduring message of George Orwell’s 1984, but it is important to consider just why the alarmist message of 1984 was alarming, and just what features of the superstate—Oceania—featured in the book produced that alarm.

⁷ Some years ago, there was a flurry of interest in the claim that government speech, which it was argued had the potential to drown out private speakers, ought to be constrained by the free speech principles embodied in the First Amendment. See Mark G. Yudof, When Government Speaks: Politics, Law, and Government Expression in America (Berkeley, CA: University of California Press, 1983); Steven Shiffrin, “Government Speech,” UCLA Law Review 27, no. 3 (1980): 565–655. But for a skeptical view, see Frederick Schauer, “Is Government Speech a Problem?” Stanford Law Review 35, no. 2 (1983): 373–86. As a matter of existing positive law, the speech of government has never been considered to raise free speech problems, a conclusion embodied in such Supreme Court cases as Meese v. Keene, 481 U.S. 465 (1987), and Walker v. Texas Division, Sons of Confederate Veterans, Inc., 135 S. Ct. 2239 (2015). When presidents condemn the Ku Klux Klan or praise the American automobile industry, for example, we do not hear criticism of such government speech, even while recognizing that some citizens will disagree with the message. But when government messages endorse a particular religion or religious point of view, the Establishment Clause of the First Amendment comes into play, prohibiting at least some forms of government speech that would be constitutionally permissible with respect to any topic or point of view other than religion. See Town of Greece v. Galloway, 134 S. Ct. 1811 (2014); Jesse H. Choper, “The Endorsement Test: Its Status and Desirability,” Journal of Law and Politics 18, no. 2 (2002): 499–536; Shari Seidman Diamond and Andrew Koppelman, “Measured Endorsement,” Maryland Law Review 60, no. 3 (2001): 713–60.

⁸ Among the most enduring justifications for a principle of freedom of speech is the idea, inherited from John Milton (Areopagitica [1644]) and then John Stuart Mill (On Liberty [1859]) and then Oliver Wendell Holmes (Abrams v. United States, 250 U.S. 616, 630 [1919] [dissenting opinion]), that a regime of freedom of speech—the marketplace of ideas—is facilitative of the identification of truth, the exposure of falsehood, and the increase of human knowledge. See Joseph Blocher, “Institutions in the Marketplace of Ideas,” Duke Law Journal 57, no. 4 (2008): 821–89; Jill Gordon, “John Stuart Mill and the Marketplace of Ideas,” Social Theory and Practice 23, no. 2 (1997): 235–49. Implicit in this theme is the view that the truth of a proposition has substantial explanatory force in determining, for a population, which propositions will be accepted and which rejected. But the lesson of a great deal of modern research in cognitive and social psychology is that various other attributes of a proposition—the charisma or authority of the agent uttering the proposition; the frequency with which the proposition is uttered; the extent of rhetorical or technological enhancement of the proposition; the extent to which a proposition is consistent or inconsistent with an audience’s prior beliefs; the extent to which a proposition reinforces an audience’s normative commitments; and much more—compete with and often dominate the truth of a proposition in determining acceptance or rejection of a proposition. For summaries of the relevant research, and application to legal and philosophical questions about freedom of speech, see Daniel Ho and Frederick Schauer, “Testing the Marketplace of Ideas,” New York University Law Review 90, no. 4 (2015): 1160–1228; Frederick Schauer, “Free Speech, the Search for Truth, and the Problem of Collective Knowledge,” SMU Law Review 70, no. 2 (2017): 231–52.

⁹ The common phrase, “Don’t tell me what to do,” is typically used against those whose tone is one of giving orders rather than of giving advice. But is there a difference among “Don’t smoke,” “You shouldn’t smoke,” “I advise you not to smoke,” “I urge you not to smoke,” and “If I were you, I wouldn’t smoke”? Perhaps there are differences among these, differences that might be measured in terms of how much the particular locution recognizes (or not) that the final decision is to be made by the recipient. But if “I were you, I wouldn’t smoke” represents the extreme of such acknowledgment, then the difference between that and “Don’t smoke” may only be one of tone, or only about the potential sanctions that might be imposed upon the subject for noncompliance.

¹⁰ Thus, Webster’s New Universal Unabridged Dictionary (New York: Simon and Schuster, 2d ed. 1983) gives as one of the definitions of “propaganda,” “any organization or movement working for the propagation of particular ideas, doctrines, practices, etc.,” and another as “any systematic, widespread, deliberate indoctrination or plan for such indoctrination: now often used in a derogatory sense, connoting deception or distortion.” Perhaps the latter comes closer to now-common usage.

¹¹ See Bliss, Thomas Albert, “Subliminal Projection: History and Analysis”, Hastings Communications and Entertainment Law Journal 5, no. 2 (1983): 419–41Google Scholar.

¹² For a well-known critique of the use of various psychological practices to influence consumer preferences without the knowledge of those consumers, see Vance Packard, The Hidden Persuaders (New York: R. McKay Co., 1957).

¹³ See, prominently, Richard H. Thaler and Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (New York: Penguin Books, 2009). What makes this widely-discussed book an example of the phenomenon described in the text is the extent to which it is less about helping people improve their own decisions as it is about assisting individuals and institutions in influencing the decisions of others.

¹⁴ And that is why the view that punishment for thoughts is objectionable (see note 5, and Mendlow, “Why Is It Wrong to Punish Thought?”) must be based on a theory of just what punishment is, why to punish someone for engaging in some behavior is different from other forms of sanction for engaging in that behavior, and why punishment is different from other ways of coercion. This essay is not the place to engage with a theory of punishment, but it is worth noting that at least some of the argument here resists the widespread notion that punishment is importantly different from other forms of behavioral control.

¹⁵ Influenced by long-standing British usage and by the language of the American First Amendment (“Congress shall make no law … abridging the freedom of speech, …”), the traditional label for the right under discussion is freedom of speech. But because that right is commonly understood in most liberal democracies to encompass art, music, parades, demonstrations, rallies, flag waving, flag burning, armband wearing, picket sign carrying, and silent protests, among others, none of which would be considered “speech” in ordinary language, the more common contemporary locution, seen in most modern human rights documents and many modern constitutions as well, is freedom of expression. Just as “freedom of speech” may be underinclusive, however, “freedom of speech” may be overinclusive, including almost any way in which we might non-communicatively express ourselves, including how we dress, how we wear our hair, where we live, what professions or avocations we pursue, and with whom we choose to associate. See Frederick Schauer, “Must Speech Be Special?” Northwestern University Law Review 78 no. 5 (1984): 1284–1306. If “freedom of expression” is not therefore to be a synonym for or to collapse into a principle of general liberty, it seems best to refer to freedom of communication, even while recognizing that not all forms of communication —think about the communications between two parties to produce a legal contract, or the communication of a testator in a last will and testament—are even covered by a principle of freedom of communication.

¹⁶ Thomas Scanlon, “A Theory of Freedom of Expression,” Philosophy and Public Affairs 1, no. 2 (1972): 204–26.

¹⁷ See Scanlon, T. M., “Comment on Baker’s Autonomy and Freedom of Speech,” Constitutional Commentary 27, no. 2 (2011): 319–25Google Scholar; Scanlon, Jr. T. M., “Freedom of Expression and Categories of Expression,” University of Pittsburgh Law Review 40, no. 4 (1979): 519–50Google Scholar.

¹⁸ In addition to Scanlon, see David A. Strauss, “Persuasion, Autonomy, and Freedom of Expression,” Columbia Law Review 91, no. 2 (1991): 335–71. Seana Shiffrin argues that listener-based theories, while valuably explaining some aspects of, and foundations for, freedom of speech, fail to explain too large a range of intuitions and accepted legal doctrine to serve as the exclusive or even principal basis for a free speech principle. Shiffrin, Speech Matters, 82–85. But without taking a position on whether listener-based accounts should figure exclusively, somewhat, or not at all in justifying a free speech principle, I nevertheless leave open the possibility that much of legal doctrine, and many of the intuitions that may be based on that doctrine, are mistaken, as is, possibly, the idea of a justifiably distinct principle of free speech at all.

¹⁹ Seana Valentine Shiffrin, “A Thinker-Based Approach to Freedom of Speech,” Constitutional Commentary 27, no. 2 (2011): 283–307, and in revised form in Shiffrin, Speech Matters, at 79–115.

²⁰ Other arguments for freedom-of-thought-based justifications for a principle of freedom of speech include Charles Fried, Modern Liberty: And the Limits of Government (New York: W. W. Norton, 2007), 95–123; Dana Remus Irwin, “Freedom of Thought: The First Amendment and Scientific Method,” Wisconsin Law Review 2005, no. 6 (2005): 1479–1534; Christina E. Wells, “Reinvigorating Autonomy: Freedom and Responsibility in the Supreme Court’s First Amendment Jurisprudence,” Harvard Civil Rights—Civil Liberties Law Review 32, no. 1 (1997): 159–96. Arguments for the relationship between freedom of thought and freedom of speech can be traced to Benedict de Spinoza’s Tractatus Theologico-Politicus, ch. XX (R. H. M. Elwes trans., New York: Dover, 1951 [1670]), which, anticipating “Die Gedanken sind Frei” by centuries, observed that “If men’s minds were as easily controlled as their tongues, every king would sit safely on his throne.” And on Spinoza’s conjunction of freedom of thought and freedom of speech, see Martha Womack Haun, “Spinoza on Freedom of Thought and Speech,” Free Speech Yearbook 16, no. 1 (1977): 47–53; Edward I. Pitts, “Spinoza on Freedom of Expression,” Journal of the History of Ideas 47, no. 1 (1986): 21–35. It is noteworthy that Spinoza’s concerns with the freedoms of thought and conscience were devoted almost exclusively to questions of religious belief. But whether religious belief is in some way different from, say, moral or social or political belief, is itself a religious question, and not one I feel competent to address.Note that the claim that some idea of freedom of thought or freedom of the mind justifies a right to freedom of speech or expression is different from the claim that freedom of thought and freedom of speech are largely the same thing, and different from the claim that the two are naturally conjoined in a way that is deeper than one being instrumental to the other. See, for example, Adam J. Kolber, “Two Views of First Amendment Thought Privacy,” University of Pennsylvania Journal of Constitutional Law 18, no. 5 (2016): 1381–1423.

²¹ For related ideas about the collective aspects of a free speech principle, see Joshua Cohen, “Freedom of Expression,” Philosophy and Public Affairs 22, no. 2 (1993): 207–63; Robert C. Post, “Participatory Democracy and Free Speech,” Virginia Law Review 97, no. 3 (2011): 477–89; James Weinstein, “Participatory Democracy as the Central Value of American Free Speech Doctrine,” Virginia Law Review 97, no. 3 (2011): 491–514.

²² Thus, Shiffrin says that “the foundation of free speech protection is that freedom of speech is necessary for the development and maintenance of the self qua thinker, for freedom of thought, and for discharging other aspects of our moral relations.” Shiffrin, Speech Matters, 80. But although it is difficult to deny that thinking is intrinsically a part of being, we need more explanation of how that aspect of being might be threatened, or how that aspect of being is relevantly different from those aspects of being that are largely concerned with doing.

²³ Scanlon notes that heightened or special protection is a necessary condition for recognizing a right to freedom of speech (“A Theory of Freedom of Expression,” 204): “The doctrine of freedom of expression is generally thought to single out a class of ‘protected acts’ which it holds to be immune from restrictions to which other acts are subject.” This structural aspect of what it is for there to be a right to free speech (and, en passant, what it is for there to be a right of any kind) is developed in Frederick Schauer, Free Speech: A Philosophical Enquiry (Cambridge: Cambridge University Press, 1982), and, more recently and at greater length and depth, in Frederick Schauer, “Free Speech on Tuesdays,” Law and Philosophy 34, no. 2 (2015): 119–40.

²⁴ The phrase “thought police” comes originally from Orwell’s 1984, but Orwell’s Thought Police (Thinkpol) were mainly concerned with what was said, as are those who now wield Orwell’s phrase in political and social debate. See, for example, Bruce, Tammy, The New Thought Police: Inside the Left’s Assault on Free Speech and Free Minds (New York: Random House, 2001)Google Scholar.

²⁵ See note 21.

²⁶ See note 2.

²⁷ I recognize that some people may commit bank robberies under duress (threats or other forms of coercion), or because of desperate financial need, and so on. But even such people have positive thoughts, however produced, about bank robberies before they commit them. I thus use “involuntary” in a non-moralized sense. There are involuntary muscular contractions, but there are not involuntary bank robberies.

²⁸ Imagine that Jones is on trial for committing a bank robbery. The prosecution offers as evidence the testimony of one of Jones’s friends that she heard Jones say, “I think bank robbery is a good way to make a lot of money with not much effort.” Defense counsel objects, saying that this statement does not show that Jones committed the robbery. Under the existing law of most common law jurisdictions, the evidence will be admitted, not because it establishes by itself that Jones committed the bank robbery with which he is charged, but because, in probabilistic or Bayesian fashion, the likelihood that Jones committed the bank robbery is greater with this evidence than without. See Federal Rules of Evidence, Rule 401.

²⁹ By “probabilistic indicator,” I mean only that the indicator makes some conclusion more likely than without the indication, even if that conclusion remains highly unlikely.

³⁰ Often the inferential chain is complex. For example, people who possess child pornography are often prosecuted (constitutionally permissibly—see New York v. Ferber, 458 U.S. 747 [1982])—because such possession is thought to be either indicative of, or causal of, child molestation by the possessor. Assuming that possession of child pornography is good evidence of thinking about child pornography, and assuming (controversially) that thinking about child pornography is some (but not necessarily good) evidence of thinking about child molestation, it remains true that those who think about child molestation will not necessarily, and perhaps even not probably, engage in actual child molestation. And that is why prosecutions for possessing child pornography commonly attract the charge of thought control. See Clay Calvert, “Freedom of Thought, Offensive Fantasies, and the Fundamental Human Right to Hold Deviant Ideas: Why the Seventh Circuit Got It Wrong in Doe v. City of Lafayette, Indiana,” Pierce Law Review 3, no. 2 (2005): 125–59; Eric M. Freedman, “Digitized Pornography Meets the First Amendment,” Cardozo Law Review 23, no. 6 (2002): 2011–17; CBC News, “Canada’s Porn Law Close to ‘Thought Control,’” April 27, 1999, available at www.cbc.ca. See generally Carissa Byrne Hessick, ed., Refining Child Pornography Law: Crime, Language, and Social Consequences (Ann Arbor: University of Michigan Press, 2016). Still, in the language of evidence law, possessing child pornography appears to be probative of child molestation.

³¹ See R. A. Duff, Criminal Attempts (Oxford: Oxford University Press, 1996), 390. This is the general tenor of, for example, Husak, Douglas N., Overcriminalization: The Limits of the Criminal Law (Oxford: Oxford University Press, 2008)Google Scholar; Husak, Douglas N., “Drug Proscriptions as Proxy Crimes,” Law and Philosophy 36, no. 3 (2017): 345–66CrossRefGoogle Scholar.

³² See, for example, New York Penal Law, §140.35 (McKinney 2010). It appears that every state except one has or has had such laws. See Paul A. Clark, “Do Statutes Criminalizing Possession of Burglary Tools Reduce Crime?” Capital University Law Review 42, no. 4 (2014): 803–60. Indeed, the crime of possession of burglary tools is commonly a part of larger discussions about incomplete crimes and inchoate crimes generally, a topic that includes questions about the criminalization of intent and motive and the criminalization of attempts. See Kimberly Kessler Ferzan, “Inchoate Crimes and the Prevention/Punishment Divide,” San Diego Law Review 48 (2011): 1273–96; Douglas N. Husak, “The Nature and Justification of Nonconsummate Offenses,” Arizona Law Review 37, no. 1 (1995): 151–83.

³³ See Florida Statutes 893.135 (2018), which provides that the threshold for the trafficking offense, for cocaine, is 28 grams.

³⁴ For examples and analysis, see Stefanie Bock and Findlay Stark, “Preparatory Offences,” University of Cambridge Faculty of Law Legal Studies Research Paper Series, no. 64 (October 2018).

³⁵ Preparatory crimes should be distinguished from indicative or proxy crimes, in which people are punished for having committed some act on the theory that committing that act—an act that is often not itself unlawful—probabilistically indicates that they have committed some other act, one that is in fact illegal. For example, mothers of deceased babies who had not reported the birth were once prosecuted for infanticide, sellers of books and magazines without covers were prosecuted for defrauding distributors, and those who travel abroad with substantial quantities of unreported cash are penalized as if they were money launderers, even though in such cases the relationship between the proxy crime and the wrong for which it is a proxy is only probabilistic. See Pyotr Bystranowski, “Retributivism, Consequentialism, and the Risk of Punishing the Innocent: The Troublesome Case of Proxy Crimes,” Diametros 53, no. 1 (2017): 26–49; Frederick Schauer, “Bentham on Presumed Offenses,” Utilitas 23, no. 4 (2011): 363–79. In many contexts, the probabilistic nature of both preparatory and proxy offenses invites us to treat them similarly (see Frederick Schauer, Profiles, Probabilities, and Stereotypes (Cambridge, MA: Harvard University Press, 2003), but in the context of the specific question of freedom of thought the difference between probabilistic predictions of future behavior and probabilistic determinations of past acts is crucial.

³⁶ It is common in criminal law debates about punishing attempts and intentions to resist this assumption, and to argue that there is a class of thoughts that is largely inconsequential. See R. A. Duff, Answering for Crime: Responsibility and Liability in the Criminal Law (Oxford: Hart, 2007), 102–4. Duff’s conclusions about the inertness of a category of thoughts (but not all thoughts), conclusions that are challenged by the probabilistic account I offer here, are also challenged in Mendlow, “Why Is It Wrong to Punish Thought?” And see also, paralleling Duff’s resistance to the efficacy (and thus the culpability) of supposedly “mere” thoughts, Larry Alexander and Kimberly Kessler Ferzan (with Stephen Morse), Crime and Culpability: A Theory of Criminal Law (New York: Cambridge University Press, 2009), 197–216; Federico Picinali, “A Retributive Justification for Not Punishing Our Intentions Or: On the Moral Relevance of the ‘Now-Belief,’” Law and Philosophy 32, no. 4 (2013): 385–403.

³⁷ See Seana Valentine Shiffrin, “Methodology in Free Speech Theory,” Virginia Law Review 97, no. 3 (2011): 554.

³⁸ Mendlow, “Why Is It Wrong to Punish Thought?”

³⁹ See Dubber, Markus Dirk, “Toward a Constitutional Law of Crime and Punishment,” Hastings Law Journal 55, no. 3 (2004): 509–72Google Scholar.

⁴⁰ See Frederick Schauer, “On the Relationship Between Chapters One and Two of John Stuart Mill’s On Liberty,” Capital University Law Review 39, no. 3 (2011): 572–92.

⁴¹ American libel law prohibits public officials and (broadly defined) public figures from recovering in libel actions in which they can prove with “convincing clarity” not only that what was said about them was false, but that it was said with actual knowledge or actual suspicion of its falsity. This approach, more speaker and press protective than that present in any other country (see Frederick Schauer, “The Exceptional First Amendment,” in American Exceptionalism and Human Rights, ed. Michael Ignatieff [Princeton, NJ: Princeton University Press, 2005], 29–56), owes its origins to New York Times Co. v. Sullivan, 376 U.S. 254 (1964). But its bite can be seen most clearly when applied to protect plainly false and plainly harmful negligently published accusations. See Ocala Star Banner Co. v. Damron, 401 U.S. 295 (1971).

⁴² See Snyder v. Phelps, 562 U.S. 443 (2011).

⁴³ See Brandenburg v. Ohio, 395 U.S. 444 (1969). And so too with racial insult and intimidation. See R.A.V. v. City of St. Paul, 505 U.S. 377 (1992).

⁴⁴ For example, the increasing conflict between demands to be free from discrimination on the basis of sexual orientation and the claims that offering services to same-sex couples (as, for example, a wedding cake for a same-sex couple, see Masterpiece Cakeshop. Ltd. v. Colorado Civil Rights Commission, 138 S. Ct. 1719 [2018]) would infringe the religious liberty rights of the provider.

⁴⁵ A prominent and powerful rejoinder to much of the contemporary celebration of autonomy is Sarah Conly, Against Autonomy: Justifying Coercive Paternalism (Cambridge, UK: Cambridge University Press, 2012).