I

I said: “[A]ll speech is conduct, involving an agent, and all conduct, being intentional action, is expressive (of the motivating intention).”⁹

In response to this claim, Collier notes that although

I agree that this is the Court's view, but I do not agree that it is correct, and certainly not for that reason alone. As others, for example Robert Post, have argued, it has proven to be notoriously difficult—some would say impossible—for the Court to distinguish speech from other forms of intentional action.¹¹

Collier cites Spence v. Washington, a flag-desecration case in which the Court's analysis began with an inquiry into whether the defendant's “activity was sufficiently imbued with elements of communication to fall within the scope of the First and Fourteenth Amendments.”¹² The test used by the Court, which has come to be known as the Spence test, was whether “an intent to convey a particularized message was present and [whether] the likelihood was great that the message would be understood by those who viewed it.”¹³ The Spence test, however, does not succeed in distinguishing speech from nonspeech conduct, as Robert Post has persuasively argued:

Although Collier adds, quoting Peter Meijes Tiersma, that “‘[t]o be functionally equivalent to speech, conduct must be the means of communication, rather than simply providing evidence or stimuli from which an observer might draw inferences or in some other way derive information,’”¹⁵

this qualification does not help to distinguish speech (and its functional equivalents) from other forms of intentional action. Post's public-property defacement example meets this criterion as well, as indeed does any example of clearly criminal activity (including arson, rape, and murder) that is intended as a means for conveying a message. As Post notes, “[i]f the Spence test were to describe actual judicial practice, we would expect criminals routinely to attempt to immunize their crimes by endowing them with particular messages.”¹⁶

But Collier has more to say. To illustrate his view of what counts and what does not count as speech or its functional equivalent, he presents four scenarios for our consideration. Each takes place in a train compartment, and in each case, one occupant, X, smokes a cigar, prompting another occupant, Y, to cough.

Example 1: Y is asleep when X lights up a cigar. The smoke causes Y to cough, which wakes him up. In this case, as Collier notes, the cough is not “conduct in the sense of intentional action.”¹⁷

So far, so good. The cough, I agree, is not an action, since it is not intentional under any description. It is like snoring or anything else one “does” while asleep or otherwise unconscious.

Example 2: Y is now awake, and the smoke is affecting his throat, but, being exceedingly polite, Y does not wish to cause a fuss. Eventually, though, “it becomes too much” and Y decides to clear his throat “with a brief, quiet cough.”¹⁸

Collier stresses that, in this case, Y was not forced to cough. (He could have held his breath or rushed out of the compartment.) Instead, Y “freely decided that, all things considered, coughing was the best thing to do.” Y “meant (intended) to cough,” but “‘meant’ nothing else by it.” “So here,” Collier states, “there is intention and conduct but not (I shall argue) communication or speech.”¹⁹

The assumption here seems to be that in order to count as communication or speech, intentional conduct must be intended to be perceived by someone else as intentionally conveying a particular (that is, the intended) message. But no argument is given for this assumption, and there are good reasons for rejecting it. In deciding not to suppress the cough or leave the room, Y acted in a way that he knew would communicate something to X. The very fact that he thought about the alternatives to coughing and decided against them (“all things considered”), indicates that conveying a message to Y was at least a foreseeable consequence of coughing.

Now, one familiar with the doctrine of double effect could argue that even if conveying a message to Y was a foreseeable consequence of coughing, it need not have been an intended consequence (and in this scenario, we shall grant that it was not intended). Y's coughing in this case is similar to one's yawning (but not loudly or ostentatiously) when one has decided not to resist the urge to yawn during a talk that has gone on for too long, knowing (but not intending) that the yawn will signal the speaker to wind up the talk. This contrasts not only with the case in which one's stomach growls because the talk has gone on past lunchtime (perhaps making the speaker aware that the talk has gone on too long) but also with the case in which one yawns because of a desire (that is, an intention) that the speaker stop talking. However, the yawn in both cases communicates to the speaker that the talk has gone on too long, whether or not the yawner intends this to be the message communicated.

Example 3: Y is “getting a little peeved” by X's continued smoking but, being exceedingly polite, wants to get the smoker to put out his cigar but “without seeming to complain.” This time Y's cough (which “sounds exactly like the previous ones”) is intended to get the smoker to make the inference that the smoke is bothersome. “If [Y is] successful, the smoker will get the message without realizing [Y] was sending a message.”²⁰

This cough, according to Collier, still does not amount to communication or speech, because Y “is simply trying to induce changes in [X's] behavior as one might in training an animal.”²¹

Y's coughing in this case is, to use an example related to hate speech, analogous to someone's burning a cross on (or right next to) a black family's lawn in an attempt to induce the family to move away, without intending that the black family view it as an act of terror by an identifiable white neighbor. Regardless of the cross-burner's intentions, the cross-burning nonetheless communicates a message of terror to the black family.

Example 4: Y is losing patience and so coughs loudly and ostentatiously, with the intention of conveying to X: “Your smoking is bothering me, I want you to know that, I want you to know that I want you to know that, and I want you to stop.”²²

I agree with Collier that this cough counts as communication but not with his claim that this series of examples shows that communication “involves not only the conveying of an idea (Example 3) but also the realization [presumably on the part of the audience] that a message is intentionally being sent (Example 4).”²³ If this were the case, communicating anything that one does not intend to communicate would be impossible, making the everyday phenomenon of miscommunication impossible.

Collier assumes that one cannot send a message (or communicate anything) without at the same time sending a message that one is sending a message. But in the cross-burning example above, the cross-burner terrorizes the black family in spite of the fact that he does not want the black family to know that he is sending a message of terror.²⁴

Collier also claims that “[w]ith a true act of communication, I share my belief with you and also implicitly commend it to you for your own adoption.”²⁵ If this were the case, however, the only true (genuine) acts of communication or speech (Collier seems to use the terms interchangeably) would be clear, unambiguous, and truthful expressions of doxastic propositional attitudes that are understood by the listener exactly as intended. But speech does many things apart from (accurately) expressing our beliefs about things. It alarms, commands, degrades, honors, disgraces, consoles, cuts, seduces, and terrifies, among other things. It can be as soothing as a caress or as stunning as a slap. It quickens the pulse, puts you to sleep, makes you blush, and takes your breath away.

Collier's theory of language is further revealed later in his article, when he tries to discern the message conveyed by the cross-burning in R.A.V. v. St. Paul.²⁶

Citing Ayer and Popper, Collier states that “we may classify statements as true, false, or meaningless,”²⁷ and it seems that for Collier this classification is exhaustive. The “translations” Collier gives of the burning cross are all statements that could be true or false. But there are also many translations that are not assertions having a truth value (e.g., “Get out of the neighborhood,” “Leave town or else,” “Whites only here,” “Death to blacks,” “N—go home”). But suppose, per impossibile, that Collier is right—that unless hate speech is meaningless, it is true or false. What follows? According to Collier, it is that the “contemptible and repugnant ‘message’ is what should be challenged and opposed, not the right to express it,”²⁸ although no argument is given to support this exhortation. What might the argument be? What might motivate the long discussion of hate speech as the sort of thing that can (and apparently must) be either true or false (since if it is meaningless, it does not convey any message and so, a fortiori, does not convey a hateful message)? Perhaps it is that if speech is the sort of thing that can (indeed, must) be true or false if meaningful, then, since actions are not true or false (they just are what they are), they cannot be meaningful unless they are accompanied by a verbal statement or some conventional sign that functions in some equivalent way. But speech is not the sort of thing that must be true or false in order to be meaningful, so this way of attempting to draw the distinction between speech and action never gets off the ground.²⁹

Collier concludes his discussion of my first point by arguing that verbal hate speech is speech—more obviously than what he calls “symbolic speech” (e.g., cross-burning). The argumentative force of this is unclear until he reveals that he interprets me as saying that hate speech should be denied constitutional protection because it is not really speech. But nowhere do I say that (nor do I believe it). Nor am I committed to it by anything I do say. (Does this mean that my article was not “a true act of communication” since I did not succeed in “sharing my belief” about this with at least one reader?)

II

In response to my claim that “speech does not differ from other conduct in being context-dependent and subject to interpretation,”³⁰

Collier contrasts someone shouting “Blurp!” at him with someone chopping off his finger and suggests that no matter how he interprets someone shouting “Blurp!” at him, it is worse for him to have his finger chopped off. I agree, but this is not because “Blurp!” is subject to interpretation and finger-chopping is not but, rather, because the finger-chopping causes lifelong, irreversible harm (the extent of which will vary from individual to individual,³¹ but will still be significant). Take a slightly different example: Which is more harmful: being constantly subjected to distracting, humiliating, brain-disordering racist speech in one's workplace, or having one's finger broken (in a way that will swiftly heal) because of a malfunctioning piece of equipment? The former can easily be seen to be a more serious setback to one's significant interests, especially if it makes functioning well on the job impossible. But my point was not that verbal assaults are (always or even often) more harmful than physical assaults, but simply that sometimes they can be.

I also argued that the context of a physical assault can (again, sometimes, not always) affect the extent to which it is harmful. Collier responds by comparing the case in which an assailant hacks off his finger to the case in which his finger succumbed to “some form of organic disease” and had to be removed by a surgeon. “The fact that this context is a blind force of nature rather than a violent criminal attack,” he observes, “does not matter to my finger.”³²

But the physical damage to one's finger is not the sole arbiter of the degree to which one has been harmed. Compare having one's finger broken in an accident with having one's finger broken (with exactly the same physical damage) by a Mafia don who says, convincingly, “next time I'll kill you.” The contextual difference does not matter to your finger, but it makes all the difference to you.

By the end of his article, Collier concedes that “[i]t is true that one can usefully ‘interpret’ both speech and conduct. And it is true that conduct can seem more or less ‘harmful,’ depending on the context. But these analytical truths,” he concludes, “do not somehow affect the preexisting and obviously experienced differences between speech and conduct.”³³

As flattering as it is to see these two claims of mine elevated to the status of “analytic truths,” on no account of analyticity that I am familiar with can these statements be considered analytically true. Nor could experienced differences (between speech and conduct) be considered to “preexist” analytic truths.

III

I said: “[S]peech is a physical phenomenon, having physical effects on its listeners, effects that can be caused by the content of the speech.”³⁴

Collier replies: “Saying that speech is a physical phenomenon is like saying that playing chess is moving pieces of wood.”³⁵ This is intended to be a fatal objection, a reductio ad absurdum of my position, since “someone who thought that chess was nothing but moving pieces of wood would have deeply misunderstood the game of chess.”³⁶ But to say that chess is moving pieces of wood is not to say that it is nothing but moving pieces of wood, and to say that speech is a physical phenomenon is not to say that it is nothing but a physical phenomenon (and that it cannot be described in any other way). Obviously speech is not simply vibrating one's vocal chords and moving one's lips. It is not even simply vibrating one's vocal chords and moving one's lips in a very precisely determinate way (otherwise homonyms would be impossible and “Empedocles leaped” would mean the same thing as “Empedokles liebt”).

It is a mystery why Collier spends so much time attacking the view that playing chess is nothing but moving pieces of wood (and by implication that speech is nothing but a physical phenomenon), since I can think of no one who holds this view. As Collier rightly points out, in order for behavior to count as human action, it must be intentional under some description, and the context in which the behavior occurs can make all the difference. Under “special circumstances,” Collier writes, “the physical phenomenon (correctly) described as ‘moving pieces of wood’ has or takes on the meaning (or ‘symbolic content’) of being a chess move.”³⁷

That is a perfect illustration of my point that the meanings of physical actions are context-dependent (and thus subject to interpretation).

Collier still wants to distinguish speech (including expressive actions) from actions that do not (and cannot) count as speech, however, and so he states: “Meaningful actions are, unlike actions in general, necessarily intentional; there is no such thing as an unintentional greeting, or chess move, or auction bid.”³⁸

But, given Collier's claim that a behavior must be intentional under some description to count as an action, there is no such thing as an action that is not intentional; indeed, all actions are, by his own definition, intentional under some description.

Collier assumes, without examination, the premises that (1) if speech is a physical phenomenon, it cannot also be meaningful; and (2) if something is a cause, it cannot also be a reason. (It took me quite a while to realize that this was why he thinks I must accept “a theory of communication based on classical conditioning.”³⁹

) But these premises presuppose the very mind-body substance dualism that I claim the traditional speech-action distinction is grounded in (a claim Collier seems to think is absurd). For only a substance dualist would (or could) hold that if x is a physical phenomenon, it cannot be meaningful (i.e., intentional), and that if x is meaningful (intentional), it cannot be physical. These claims presuppose the truth of a Cartesian-type dualism according to which anything physical, by definition, has no mental characteristics (such as intentionality), and anything mental, by definition, has no physical characteristics. But this ontological view has been thoroughly discredited by work in philosophy of mind. Even if Collier wants to cling to this outmoded solution to the mind-body problem, it would seem to be not only a philosophical mistake but also a pragmatic error to attempt to ground an account of free speech in a metaphysical theory that is so widely rejected.

IV

In response to my saying that “verbal assaults can be harmful in the same ways that assaults involving direct bodily contact are,”⁴⁰

Collier claims: “There is a definite biological basis for the criminalization of assault, but speech is not normally part of an assault.”⁴¹ On his view, only the allegedly noncommunicative aspects of physical acts contribute to their being assaults. As an example of an assault, he mentions the act of someone who “walks into my office, says ‘I’m going to kill you,' and briefly opens his jacket to show me that he has a gun.” What makes this an assault, according to Collier, is “the physical menace of displaying the gun.”⁴² He calls this entire episode a “noncommunicative” act.⁴³ If no threat is communicated in this scenario, however, there is no reason to feel threatened and hence no assault. If you were a police officer and your partner walked into your office before you headed out on assignment, briefly opening his jacket to show you that he had a gun, you would take this to be the opposite of a threat. It would be an assurance that he would be able and willing to protect you from assault. His act of showing you his gun has meaning only within a given context, and that context is affected by the speech involved in the act. The message conveyed by your partner's act of showing you that he had a gun would change dramatically if accompanied by his saying “I'm going to kill you.”⁴⁴

Collier further asserts that to hold, as I do, that speech can not only contribute to the harmfulness of a physical attack but also be a form of assault in itself presupposes a “therapeutic model,” the first step of which “is to change the focus from actual physical injuries (‘hurts’) to ‘harms,’ understood as setbacks to important interests.”⁴⁵

This “step” is not an original move on my part,⁴⁶ however, nor does it have anything to do with therapy. The law already takes into account harms understood in this way. Even in the case of physical injury, it is not merely the extent of the hurt (the physical pain or damage) that is taken into account by the law, but rather the harm, construed so as to include long-term financial and emotional damage. Collier claims that my view pathologizes targets of hate speech, turning them into victims by attributing to them harms that they, absent the theory, would not be considered to have suffered. This is no more plausible than the claim that tort law (or antidiscrimination law, including sexual harassment law) turns complainants into victims by attributing to them harms that they would not otherwise be considered to have suffered.

Granted, before sexual harassment was recognized as a legal category, the conduct constituting it was not (legally) recognized as harmful. Likewise, had case law developed differently from the way it in fact has, certain harms currently recognized as such in tort law would not be (legally) recognized as harms. Even clearly physical assaults considered to be crimes have not always been recognized as harms to those physically assaulted. Rape was once considered not as a harm to the raped woman but rather as a harm to her husband or father. (And the reason we treat rape so seriously now is not because of the lasting physical harm to the victim but rather because of the psychological damage wreaked by the soul-destroying message conveyed by the assailant.) Our assessments of harm have shifted over time, shaping the law and in turn being shaped by it.

Collier ends his article by waxing nostalgic for the days when things were “so much simpler” and “the ‘liberals’ were those who stood for free speech”⁴⁷

and—it is very tempting to add—men were men, and women knew their place, and blacks had their hands full getting rid of Jim Crow, and gays were clearly criminals, and harm was defined by those with the power and privilege to dismiss conduct they could not imagine being harmed by as being simply a matter of the construction the alleged victims chose to put on it. But I would not want Collier to suppose that I have suggested that he might think anything like this, and, fortunately for me, on his view of speech, I have not.