A reply to Posner.

In The First Amendment's Purpose, (1) I criticized the cost-benefit approach to free speech, of which Richard Posner has been a leading advocate. (2) On the cost-benefit view (or at least Posner's view of that view), speech can be prohibited when "in American society its harmful consequences are thought to outweigh its expressive value." (3) Or, in another formulation: "[S]peech should be allowed if but only if its benefits equal or exceed its costs." (4)

In a sense, stating these formulations should be sufficient to refute them. After all, from the cost-benefit viewpoint, all activity "should be allowed if but only if its benefits equal or exceed its costs." In other words, the cost-benefit approach to free speech holds that individuals have no greater freedom to speak than to engage in any other activity. Which is to say, in an important sense, that there is no freedom of speech at all--no special, constitutional freedom to speak that is different from and greater than the freedom to do anything else.

Nevertheless, cost-benefit notions, together with their cousin, the "balancing test," are legion in First Amendment law today. (Example: Fort Lauderdale can ban begging on its beach and boardwalk because this "speech," although "entitled to First Amendment protection," "adversely impacts tourism" and interferes with "a safe, pleasant environment." (5)) This way of thinking seems natural to a lot of people, even unavoidable. Don't we all know that the freedom of speech is not "absolute," that it must be "weighed" against other "interests"?

In fact, cost-benefit "reasoning" in free speech law is unnecessary and unacceptable--or so I tried to show. It turns judges into legislators, evaluating pure policy matters under the guise of constitutional review. Worse, it betrays fundamental First Amendment commitments.

Whenever a majority wants to abridge the freedom of speech--let's say, hypothetically, by banning the teaching of Islam--we can be pretty sure that "in American society the harmful consequences" of the speech "are thought to outweigh its expressive value." Moreover, perhaps the costs of the speech do outweigh its benefits. The teaching of Islam today might well offend and alarm a hundred million people; by contrast, it might benefit only a relatively small minority. If such costs and benefits can intelligibly be "weighed" at all, the former may exceed the latter; at a minimum, we can stipulate hypothetically that the costs do in fact exceed the benefits. In such circumstances, cost-benefit analysis tells us that the speech is properly prohibited.

In place of such balancing, I argued that judges ought to ask and historically have asked, much more categorically, whether the government has sought to punish the expression of opinion or to further other objectives impermissible under the First Amendment. Responding to my article, Posner defends the cost-benefit approach to free speech and takes a few shots at my alternative, purpose-based account. This reply replies to Posner's reply.

A confession: I am a die-hard Posner admirer. His concurrence, for example, in the Barnes litigation, however wrong I regard it on certain important points, is in my judgment one of the finer First Amendment opinions written since the Second World War. (6) Posner's academic work, however, is another story. It is increasingly filled with hair-raising propositions that even those who find Posner basically congenial would ridicule if other judges asserted them.

Here are three examples, taken from his response to my article:

1. Government may prohibit dissenting political opinions when the country "feels" (the word is Posner's) that such opinions are dangerous. (7)

2. There is no such thing as "`reality'" (the internal quotation marks are Posner's) in the sense that we commonsensically suppose. When we evaluate any proposition--for example, "O.J. Simpson killed his wife"--we should not think of ourselves as trying to get "`reality'" right or wrong. Instead we should adopt the belief that produces the most desirable consequences. "[A] proposition should be tested not by its correspondence with `reality' but by the consequences of believing or disbelieving it." (8)

3. An American judge has no duty to abide by Supreme Court precedent or even the text of the Constitution. Readers will think this cannot be Posner's view; they will think I must be distorting his text. Here is what he says: "The point is not that the judge has some kind of moral or even political duty to abide by constitutional or statutory text, or by precedent; that would be formalism." (9)

A judicial nominee who confessed to these views at a confirmation hearing would, I hope and expect, be rejected. A "liberal" judge who took position (2) or (3)--for example, to support rights of abortion--would be ridiculed by "conservative" academics, newspapers and talk show hosts around the country. In fact, if any sitting judge other than Richard Posner took these positions in print, there would be genuine reason to fear for his judicial competence.

Fortunately, in Posner's case, we can be confident that the judge is too smart actually to believe what he's saying, or at any rate too pragmatic to act on it. In what follows, I will try to show how positions (1) through (3), as well as other astonishing arguments that appear in Posner's response, confirm everything that is wrong with the cost-benefit approach. I will also say a few words in defense of my purposivist account of free speech.

I. FEELING DANGEROUS

Proponents of First Amendment balancing begin with the proposition that speech can be prohibited, as current doctrine puts it, if it threatens compelling state interests. The usual example offered to prove this point is incitement law. Preventing violence is a compelling state interest; that is why (the argument goes) incitement can be made a criminal offense, even if the inciter is doing no more than making fiery statements of a political opinion. Incitement law is thus supposed to demonstrate that the freedom of speech is not "absolute," but must give way when its costs are high enough.

Taking this logic seriously, however, produces a First Amendment with very little spine when it comes to protecting "seditious" or "subversive" speech. Sedition and subversion laws are typically enacted in turbulent times when such speech really might have a high probability of leading to violence. In addition, in just these circumstances, seditious and subversive speech will seem very "low in value," while also highly offensive and alarming to a great many people--costs that a consistent cost-benefit judge would have to take into account as well. As a result, nothing could be easier to justify, from a cost-benefit view, than a prosecution for maligning the President in 1798, opposing war in 1918, demonstrating for civil rights in 1968, or teaching Islam today.

Many proponents of "balancing" would try to dodge this conclusion. Posner, with characteristic candor and consistency, does not. Will the cost-benefit judge say that mere offense is not a legally cognizable harm? Not Posner: "People are often deeply offended by hearing their religious, moral, political, or even aesthetic beliefs challenged; and offense is a cost." (10) Will the cost-benefit judge say that what counts is the value of the speech "category" (e.g., "political speech"), not the particular speech acts, in order to circumvent the problem that some seditious or subversive political opinions may be "low in value"? Not Posner: "There are indeed valueless, noxious, and dangerous opinions," (11) and the attempt to avoid this problem by focusing on a speech category "fails, because it confuses total with marginal benefits." (12) Might a consistent cost-benefit judge be obliged to uphold the suppression of religious or political dissent just because "the country feels endangered" by such dissent? Most balancers would surely say no. Not Posner: "When the country feels very safe," judges and professors can "plume themselves on their fearless devotion to freedom of speech," but

   they are likely to change their tune when next the country feels
   endangered. The word 'feels' is important here. The country may have
   exaggerated the danger that Communism posed [in the 1950s]. But the fear of
   Communism was a brute fact that judges who wanted to preserve their power
   had to consider. (13)

As a purely descriptive claim, this passage is trivial. Of course some judges and professors "may" turn coward when they have to put up or shut up. Of course "the fear of Communism was a brute fact" in the 1950s, as was McCarthyism, which judges "who wanted to preserve their power had to consider." But Posner does not mean only to state facts we already know. He may not want to say so unequivocally, but he favors First Amendment cowardice.

Posner observes without qualm that cost-benefit thinking "endorses what may seem an unbecomingly timid judicial response" whenever there is "public concern with offensive or dangerous speech." (14) Why? Because judges knuckling under to McCarthyism is not merely a fact of life for Posner. Such judges may be engaging in proper cost-benefit analysis. Remember: Expression is prohibitable when "in American society its harmful consequences are thought to outweigh its benefits." (15)

Consider that formulation carefully. In the 1960s, the costs of anti-war and civil rights protest might well have been "thought" by "society" to "outweigh its benefits." I raised this problem in my article; Posner conspicuously fails to address it. He does say this: "Nazi" speech "ought to be allowed" today, but only because "Nazi ideology poses no threat to American institutions or decencies in existing circumstances." (16) Again, consider this formulation carefully. The clear implication is that when offensive speech does pose a "threat to American institutions or decencies," it can be banned.

What constitutes such a "threat"? Posner doesn't say. But there is nothing in his essay to suggest that judges could not have found the relevant "threat to American institutions" in the anti-war or civil-rights demonstrations of a few decades ago. His silence on this issue, given the specific question I raised about it in my article, can perhaps be read to concede the point.

If Posner chose to disown all these propositions as descriptive (how judges actually behave), rather than prescriptive (how judges should behave), I would welcome it. The nation could profit from Posner's leadership if he reminded judges of their obligation to stand firm with the Constitution in the face of popular pressure. But I don't think this reminder will be forthcoming. As noted, for Posner, judges have no "moral or even political duty to abide by" the Constitution. (17) Instead, the primary duty for Posner's judges is to "preserve their [own] power" (18) so long as they think that staying in power will be best for the nation's welfare--and what self-respecting judge would think otherwise?

These questions are not academic. The country "feels endangered" once again today. We are endangered today. If, out of this fear and this endangerment, "teaching Islam" or "espousing Islamic fundamentalism" were made a criminal offense, who among our judges would stand with the First Amendment? Who would say, with Brandeis and Holmes, that "[t]hose who won our independence by revolution were not cowards"? (19) Apparently not Posner: If the "harmful consequences" of "Islamic speech" are "thought" to exceed its "expressive value," (20) if such speech is "felt" to "pose a threat to American institutions or decencies," (21) out it goes.

With this caveat: Posner's cost-benefit judge would presumably have to consider whether banning Islam or Islamic fundamentalism was good policy. Perhaps this ban would only create sympathy for Islamicists. Perhaps it would be counter-productive to security at home and to our anti-terrorism efforts abroad, while producing relatively little positive yields on the "benefits" side. A cost-benefit judge would take this kind of calculus very seriously and might well strike down the prohibition on this basis. Thus on top of constitutional cowardice, cost-benefit analysis perversely adds aggressive judicial review of the policy merits of governmental action.

A purpose-based freedom of speech rejects judicial superlegislative policy review. It rejects the whole rhetoric of judicial "balancing" and of judges assessing the "costs" and "benefits" of speech. As a result it can capture, as cost-benefit thinking never can, the categorical nature of the First Amendment's protection of political dissent. The state cannot punish individuals for expressing their opinion--period.

Some will object to this "absolute" formulation. Can't a person be punished if he expresses his political opinions by car-bombing the White House?

Of course, but this kind of case is no counter-example to a purpose-based First Amendment. A car-bomber may be expressing opinions through his conduct, but he will not be punished for expressing those opinions. He will not be punished for speaking at all. He will be prosecuted for attempted murder (or reckless endangerment, assault, property damage, etc.), regardless of any opinions he might have been expressing. The purpose of our homicide and property laws is fully explicable without any reference to the communicativeness of the conduct they prohibit.

True, incitement cases are different. Individuals imprisoned for inciting others to commit crimes are indeed punished for their speech. But incitement, when defined with the intent and imminence criteria that our law requires (22) (and which cost-benefit thinking cannot explain (23)), is punishable for the same reason that uttering the words "the police are coming," to assist in a bank robbery, can make the speaker liable as an accomplice. In both cases, the speaker participates directly and intentionally in bringing about particular criminal conduct itself prohibited, once again, regardless of any opinions its participants may express through it. The incitement cases do not stand for a cost-benefit First Amendment. They stand for the following simple principle: If you knowingly and directly participate in criminal conduct (where the criminal conduct is itself defined speech-neutrally), you are liable to punishment, whether your participation consisted of words or actions.

The preceding paragraph states merely the conclusions of arguments, not the arguments themselves. But I treat these points at great length in my article, and Posner offers no quarrel with them. Accordingly, I will say no more here.

II. IS FREEDOM COST-EFFECTIVE?

In a sense, cost-benefit thinking is at odds with the freedom of speech as such. As I said in my article, "no one can pretend to know whether the freedom of speech itself is worth its costs." (24)

Posner makes some biting comments about this assertion. It is understandable that he feels a need to rebut it. Posner has to claim that the cost-effectiveness of the freedom of speech is obvious, that it is not subject to reasonable dispute. For if reasonable minds could differ on whether the First Amendment was cost-effective, then (on his view) reasonable judges might properly decline to enforce it at all--a position unappealing, it seems, even to Posner.

Now, it is very hard to see how we could test the proposition that the freedom of speech is cost-effective. Statistical studies are not likely to do the trick here. Nor is philosophy, because Posner's "pragmatism" (more on this later) purports to rebel at philosophical demonstration. When out of arguments, people sometimes turn up the vitriol. Here is what Posner says:

   The academic tactic of studied obtuseness ... is ... on display when
   Rubenfeld remarks "that no one can pretend to know whether the freedom of
   speech itself is worth the costs".... [P]hilosophy is not needed to show
   that a democratic political system, a scientific and technological culture,
   college and university education, electronic media, a diverse religious
   culture, and a diverse popular and elite artistic culture cannot prosper
   without freedom of inquiry and expression.... If one doesn't know that
   much, one doesn't know enough to write an article about the First
   Amendment. But of course Rubenfeld does know that much; he's just
   pretending not to. (25)

Well, I'm obliged to say I don't know that much. Singapore, I am told, has "a scientific and technological culture, college and university education, [and] electronic media," even though it lacks the freedom of expression. According to some, Singapore also has "a diverse religious culture, and a diverse popular and elite artistic culture" too. To be sure, Singapore does not have a "democratic political system." But England does.

Yes, England, which has long possessed every one of the desiderata Posner enumerates, all without benefit of any judicial protection of free speech against parliamentary transgression. Could Posner say that England nonetheless has a vibrant "freedom of inquiry and expression," even without American-style constitutional judicial review? He could say so, and it would be true, but if he answered this way, Posner would have conceded the point he is trying to rebut.

It is, precisely, the constitutional, judicial protection of the freedom of speech, as against legislative transgressions, that Posner must defend on cost-benefit terms. The whole question is whether the institution we call First Amendment law--meaning the body of law authorizing judicial invalidation of governmental action that (according to judges) violates the freedom of speech--can be justified from a cost-benefit perspective. If Posner's passage block-quoted above is to be a cost-benefit defense of the First Amendment, his claim has to be that democracy, science, higher education, and religious and artistic diversity cannot be had without a judicially enforced freedom of speech. Unfortunately, Posner fails to make an argument for this point; he only makes an assertion, and his assertion is wrong.

For myself, I hope and believe that the freedom of speech is worth its costs. But we do the Constitution disservice if we say we adhere to it because we know that it maximizes our welfare. We don't know that. Indeed, when it comes to certain constitutional guarantees, such as equal protection, it may be logically impossible to say whether this guarantee ultimately increases or decreases overall social welfare. For whatever "social welfare" is thought to be, it cannot be defined without begging the questions to which the Fourteenth Amendment's principles offer an answer--whose welfare counts, and whether each person's welfare counts equally.

The cost-benefit thinker cannot understand deontological commitments. Such commitments are, however, at the heart of constitutional law. As a result, there will always be an unbridgeable gap between constitutional law and cost-benefit analysis.

III. FAINTLY RIDICULOUS

Why do our courts spend so much time, under the First Amendment, worrying about laws regulating offensive speech or street demonstrations? Posner finds it baffling. "[M]ost of the cases that have vindicated freedom of speech in recent years have been faintly ridiculous, or at least distinctly marginal, involving as they have pornographic art and entertainment, old-fashioned street demonstrations ... and indecent Web sites." (26)

Posner is right to be baffled. A consistent cost-benefit judge would find our free speech law almost incomprehensible, with respect to what judges find problematic and with respect to what judges find unproblematic. I tried to bring this out in my article by showing that from a cost-benefit viewpoint, routine traffic violations could in theory be turned into serious First Amendment cases.

For example, a person arrested for speeding would simply have to say that through speeding he was "expressing himself" in one way or another--which might be true. Then the cost-benefit judge, I argued, would have to measure the "expressive value" of speeding against the competing state interests. This in turn would require the judge to consider any evidence the speeder introduced showing that the current speed limit was bad policy.

Posner responds by saying that I am "making the easy seem difficult." (27) He says that "comparing costs and benefits" here is "an easy one." Carving out an exemption for expressive speeders, he says, "would emasculate speed limits while doing only a very little to promote the expression of useful ideas." (28)

Piece of cake, isn't it? The only problem is that Posner is not responding to the hypothetical I posed. Posner is talking about exempting expressive speeders. But my hypothetical speeder was not claiming an exemption; he was trying to get a judge to strike down the current speed limit. My speeder claims to have empirical data and expert evidence showing that the 55-mile-per-hour speed limit (which he violated) fails to advance any legitimate state interests. He offers to prove that a higher speed limit or no limit would actually reduce fatalities, increase fuel efficiency, and so on. Accordingly, the speeder asks the court to strike down the 55-mile-per-hour speed limit--perhaps in favor of a 65-mile-per-hour limit, perhaps altogether, perhaps only on certain highways, etc. A cost-benefit judge, I said, would be obliged to hold a full-dress trial to consider this kind of claim, because if the judge were persuaded (1) that driving over 55 miles per hour has some "expressive value" (which Posner does not dispute), and (2) that the 55-mile-per-hour limit is actually a bad thing for highway safety, fuel efficiency, and so on, then the judge would indeed have to strike the speed limit down.

Posner knows as well as anyone else the difference between exemption cases and cases that challenge a law more thoroughly. By eliding the distinction here, however, Posner dodges the real issue. The point of the banal speeding hypothetical is to observe the kind of analysis to which cost-benefit judges are committed, an analysis under which the policy merits of a 55-mile-per-hour speed limit can indeed become a First Amendment question.

For all I know, the no-speed-limit rule prevailing on certain European highways is better policy than a 55-mile-per-hour speed limit. Although I assume that the latter saves lives, I suspect that the question is debatable. I imagine that it is the subject of numerous studies. I'm confident, however, that this is not a constitutional question, and I'm certain it is not a First Amendment question. But under cost-benefit analysis, it is a First Amendment question. The cost-benefit approach allows a routine traffic violation to be converted into a First Amendment case, with full-dress judicial review of utterly legislative policy questions.

The root problem here is that a cost-benefit judge cannot reach the proper response to the speeder's argument: dismissal for failure to state a claim. A purpose-based First Amendment explains this result without difficulty. Empirical data tending to show that a 55-mile-per-hour speed limit is bad policy in no way suggest a state purpose to punish anyone for speaking. A person arrested in the ordinary course for speeding is arrested for what he was doing, not for anything he was communicating. As a result, the defendant has no First Amendment claim--period. Looking at the case this way makes the policy merits of the speed limit exactly what they should be: constitutionally irrelevant.

I don't want to leave the impression that the issue of exemptions would be easy if we had a cost-benefit First Amendment. They could be extremely difficult, and there would be a flood of them. Everyone engaged in First Amendment activity would demand exemptions from onerous laws, especially taxes. Why, for example, on a cost-benefit view, shouldn't newspapers have a serious claim for an exemption from sales taxes, which presumably decrease the quantity of news and ideas to which we are exposed? Wouldn't cost-benefit judges have to balance the loss of tax revenue against the speech gains? And mightn't those gains be sufficient in numerous cases to demand an exemption? Posner's answer:

   The government has to tax, and to force it ... to exempt all activities
   that involve the production or dissemination of ideas and opinions would
   constitute an enormous subsidy to those activities. It would be absurd, for
   example, to exempt authors and TV anchors from federal income tax on the
   ground that the exemption would lead to an increase in expressive activity
   (which it would). The direct regulations of speech often have a smaller
   effect on the speech market, but the bad consequences of prohibiting such
   regulations are immensely smaller. (29)

Read this passage carefully. Try to follow its argument. Question: Why wouldn't the courts confront a flood of very plausible, very difficult exemption cases on a cost-benefit view of free speech? Answer: Because such cases would be "absurd." "It would be absurd, for example, to exempt authors and TV anchors from federal income tax on the ground that the exemption would lead to an increase in expressive activity...." But why, exactly, would this be "absurd"? Because the exemption would not increase expressive activity? No: Posner specifically admits that "it would." The "absurdity," it turns out, lies in an asserted "subsidization": The exemption "would constitute an enormous subsidy to those activities." This "subsidy" is the large "bad consequence" referred to in the final sentence.

"Subsidy" is a nasty word in the economist's lexicon, but even assuming that a tax exemption is a subsidy in the relevant sense, what exactly is wrong with the "subsidy" here? Is the "bad consequence" of this "subsidy" that it will, precisely, increase the subsidized activity? If so, then Posner's argument runs as follows: "It would be absurd to exempt authors and TV anchors from federal income tax on the ground that the exemption would lead to an increase in expressive activity, because the exemption would lead to an increase in expressive activity." There is an absurdity here, but not the one Posner intends.

The "bad consequence" of subsidization seems to be that subsidizing "authors and TV anchors" would lead to excessive authorship and television news programming. But what does "excessive" mean here--inefficient? How does the cost-benefit judge know how much authorship and news programming is enough? No explanation is offered.

The truth is that our law almost never grants free speech exemptions from laws of general applicability. This is not because courts do a cost-benefit analysis and find the scales always tipping against exemption. It is because the freedom of speech, in American law, recognizes no right to an exemption, even when the generally applicable law is burdensome. Cost-benefit analysis simply cannot handle this doctrinal fact. Newspapers just do not have a free speech claim to an exemption from taxes, from minimum wage laws, from labor relations laws, or from any of the other generally applicable laws that impinge on them. "It is beyond dispute that [government] can subject newspapers to generally applicable economic regulations without creating constitutional problems." (30) On the other hand, newspapers cannot be deliberately singled out for worse legal treatment, including worse tax treatment, without triggering the most exacting judicial review. (31)

This result, consistent with a purpose-based First Amendment, is incomprehensible to a cost-benefit judge. Taxing newspapers at a slightly higher tax than other products would have only a tiny marginal adverse impact on newspapers, while the generally applicable tax rates might have very large adverse consequences, which current doctrine treats as unproblematic. This incomprehension slips out in Posner's essay: "[M]easures that have a really big impact on speech are regularly ignored by the courts, which, however, pounce on tiny ones." (32)

Now we see why the great run of First Amendment case law seems faintly ridiculous, or at any rate distinctly marginal, to the cost-benefit judge. Pornography and street demonstration cases seem "ridiculous" to Posner for two reasons: (1) A cost-benefit First Amendment would not be terribly protective of offensive or indecent speech ("offense is a cost"; (33) speech is not protected if it threatens "American ... decencies" (34)) or even of "street demonstrations" (35) (which "feel" dangerous (36)); and (2) a cost-benefit First Amendment would be concerned much more with laws of general applicability (labor relations laws, perhaps, as applied to newspapers) that may have a "big impact" on speech.

The message is clear: The cost-benefit First Amendment is not our First Amendment. There is a profound gap between a cost-benefit First Amendment and the First Amendment as American courts have understood it. If one doesn't know that much, one doesn't know enough to write a First Amendment opinion for a federal appellate court. But of course Posner does know that much; he just pretends not to, in order to further the impression that cost-benefit analysis is the true key to the freedom of speech.

IV. THE COST-BENEFIT APPROACH TO TRUTH

Posner traces his cost-benefit take on free speech to "pragmatism," which he calls a "complex philosophical movement" that "challenge[s] ... the preoccupation of the central philosophical tradition of the West." (37) This essay is not the place for a serious analysis of pragmatism or of its implications, if any, for jurisprudence. But I will say a few words about Posner's pragmatism, because he places such weight on it in his reply.

   At the core of this "pragmatism" is the following claim about knowledge:
   The test for knowledge should not be whether it puts us in touch with an
   ultimate reality ... but whether it is useful in helping us to achieve our
   ends.... The human mind developed not to build a pipeline to the truth but
   to cope with the physical environment in which human beings evolved, and so
   a proposition should be tested not by its correspondence with "reality" but
   by the consequences of believing or disbelieving it. (38)

Observe that we are dealing here with a supposed "test for knowledge," as opposed to some kind of usefulness test to be imposed on knowledge, independently defined. The claim, in other words, is not that knowledge is worthwhile only when it is "useful." The claim is that belief in a proposition is knowledge only if it is "useful." A factual proposition should not be accepted as true because of its "correspondence with `reality.'" Instead, we should adopt whatever beliefs produce the best "consequences."

It is hard to imagine greater nonsense than this.

Consider this proposition: "The human mind developed not to build a pipeline to the truth...." (39) I am unsure what this means, but it certainly sounds like an assertion about "reality," doesn't it? Should we believe it? You might think that the test for knowledge here would be backward-looking: Did or didn't the human mind develop, eons ago, to "build a pipeline to the truth"? You might think that figuring out what the human mind "developed to do" would involve an inquiry requiring difficult excavations of the faraway past. But if you thought all this, you would be way off track. The "test for knowledge" is entirely forward-looking. If believing this proposition produces desirable consequences, go for it.

It is not enough for Posner to reduce constitutional freedom to cost-benefit analysis. He must do the same to knowledge. For Posner, the cost-benefit approach to the First Amendment is supposed to follow from a cost-benefit approach to truth.

Posner's "pragmatist" thinks we should believe that dropped objects will fall toward the earth because it is useful to believe it. There are several ways of bringing out the peculiarity of this way of thinking. One is as follows. Posner's "pragmatist" will have a hard time explaining why it is so useful to believe that dropped objects will fall toward the earth. He will miss the clearest reason why: namely, because dropped objects do fall toward the earth--because, in other words, the proposition is true.

If God himself appeared before you and said that he would make the whole world good and peaceful if only you acknowledged that the events of September 11 had never happened, you still could not acknowledge it. Some things are true. And they remain tree even when the consequences of knowing them are, on the whole, less desirable rather than more.

Consider the following proposition: "The human mind developed ... to cope with" a "physical environment ... and so a proposition should be tested not by its correspondence with `reality' but by the consequences of believing or disbelieving it." (40) The "and so" in this sentence sounds like a logical connector, doesn't it? This "and so" might make you think that the second clause is supposed to follow logically from the first. In fact, it doesn't follow at all. There is no intelligible logic whatsoever connecting these two clauses. (You can see this by observing that the "and so" would go equally well, perhaps better, if the word "not" were omitted from the second clause: "The human mind developed ... to cope with ... a physical environment ... and so a proposition should be tested by its correspondence with reality.") But why should the absence of logic be fatal to a "pragmatic" argument? What does a "pragmatist" care about illogic, so long as his argument ends with a proposition useful to believe? Posner specifically rejects "formalism"; he specifically rejects the idea of holding rules of law to any "high-level abstract validating principle." (41) Well, logic is precisely a formal, high-level abstract validating principle, so we should hardly be surprised if Posner feels no need to conform to it.

In reality, Posner does believe in logic and reality, not to mention the "high-level abstract validating principle" that he calls "pragmatism." For just this reason, however, Posner's "pragmatism" is almost spectacularly self-refuting. Posner knows that readers will want an answer explaining why they should believe his claim that the correct "test for knowledge" is a cost-benefit test of the consequences of belief. And so he offers an answer or two: Because "validating our beliefs as objective" is "impossible," he says, and because "the human mind developed" "to cope" with a "physical environment." (42) But these arguments are not claims about the consequences of believing the "pragmatic" account of truth. They are non-consequentialist appeals to "reality" or to "some other high-level abstract validating principle." Thus Posner tries to defend non-consequentially a claim about troth according to which every other math claim must be tested through consequential cost-benefit analysis.

Why doesn't Posner simply say, when explaining why we should accept the "pragmatic" account of knowledge, that believing it will produce better consequences--which is, after all, supposed to be the true "test for knowledge"? There is a good reason: This answer would be ridiculous. He might as well say, "The troth of a proposition does not depend on its correspondence with `reality,' but on whether believing it will make me, Richard Posner, happy," and then, when we ask him why we should accept this claim, answer, "Because believing it will make me, Richard Posner, happy."

Instead, to support his own claim about truth, Posner's "pragmatist" recurs, like anybody else, to claims about reality or about the possibility of objective validation. A "pragmatism" defended on these grounds refutes itself. It is not self-refuting in the sense of flatly self-contradictory (Posner would be flatly self-contradictory only if he said that holding the pragmatic view of truth will produce bad consequences, but one should hold it nevertheless because it is true), but self-refuting in the sense that it acknowledges criteria of knowledge other than the cost-benefit analysis that is supposed to furnish the only proper criterion.

The truth, however, is that Posner's "pragmatism" is largely beside the point. His cost-benefit jurisprudence in no way depends on his cost-benefit approach to truth. In fact, Posner championed the cost-benefit approach to law before he championed pragmatism. Although he prides himself today on his "anti-theoretical," "bottom-up" pragmatism," Posner used to be a high theorist of law and economics. He argued that judges should engage in cost-benefit, economic analysis of law on the ground that wealth-maximization was a kind of preeminent, legitimizing principle of political and legal authority. When the presuppositions underlying this claim were shot down, he took up his current "anti-theoretical" stance, according to which all moral and constitutional theory should be avoided. "Anti-theory" has to be in quotation marks here, because "anti-theory" so palpably requires a theory; the oddness of someone denouncing "high-level abstract validating principles" in one breath, while resting his position on a highly disputed, "complex philosophical movement" on the other, will not be lost on anyone.

But the great thing is that this lurch to "bottom-up" has had no effect on the bottom-line. By the most felicitous of coincidences, Posner's anti-theoretical "pragmatism" yields the same prescription as did his high theory of wealth-maximization. Judges are still told to engage in cost-benefit, economic analysis of law. (43)

The irony is that Posner's chief complaint against "theory" is that theory serves as window-dressing for beliefs held or results desired independently. Sad to say, that is usually true. The mistake lies in thinking that "anti-theory" is any different.

V. NO OBLIGATIONS

To me, the finest sentence in Posner's response is the one denying that judges have "some kind of moral or even political duty to abide by constitutional or statutory text, or by precedent." (44) That, he says, "would be formalism." (45) All of Posner's many virtues are on display here: candor, confidence, willingness to take his premises to their conclusions, readiness to flout expectations and scout the conventional wisdom.

On the other hand, I would find it understandable if Congress took the view that a judge who denies a judicial duty to follow the clear commands of the Constitution or of federal statutes has committed a serious offense. Unless Posner intends a distinction between abiding by the Constitution and abiding by "constitutional text" (and I don't think he intends this distinction), Posner's statement could be said to amount to an express repudiation of his oath of office. In fact, Posner's view seems to make the oath a kind of lark. The whole point of an oath is to create a moral or political duty. But for Posner, an oath cannot create such a duty; that would be formalism. Judges should do what they have sworn to do only to the extent that it will maximize "social goods" and "preserve their power." (46)

Technically speaking, a judge who feels this way even as he takes his oath of office may well be perjuring himself. And we all know that perjury is a very serious offense.

But put aside this formalistic talk of oaths and perjury. Imagine, more concretely, that Posner is jailed under a federal statute making it an offense for a federal appellate judge to express the view that judges have no duty to support the Constitution. On Posner's view, this law might well be constitutional. After all, a federal appellate judge's declaration that there is no judicial duty to follow the Constitution, statutes or precedents could easily be thought to "pose a threat to American institutions." (47)

Let Posner, then, assess the costs and benefits of this hypothetical law. While he does so, the rest of us should remember that this law would violate one of the First Amendment's core principles: that a person cannot be punished for expressing his moral or political opinions. Posner denies this principle, and he has every right to do so. But it is this principle that gives him the right to deny it. Since Posner will not defend his right to speak freely, others evidently will have to do so for him. Some of us will have to remember to honor and abide by the Constitution's great guarantees, even if our "pragmatic" judges no longer feel any duty to do so.

(1.) Jed Rubenfeld, The First Amendment's Purpose, 53 STAN. L. REV. 767 (2001).

(2.) See, e.g., RICHARD POSNER, The Speech Market, in FRONTIERS OF LEGAL THEORY 62-94 (2001).

(3.) Miller v. Civil City of S. Bend, 904 F.2d 1081, 1097 (7th Cir. 1990) (en banc) (Posner, J., concurring), rev'd sub nom., Barnes v. Glen Theatre, Inc., 501 U.S. 560 (1991).

(4.) POSNER, supra note 2, at 67.

(5.) Smith v. City of Fort Lauderdale, 177 F.3d 954, 956-57 (11th Cir.), cert. denied, 528 U.S. 966 (1999).

(6.) This is the same opinion cited above, which suggests that otherwise protected speech becomes prohibitable when "in American society its harmful consequences are thought to outweigh its benefits." Miller, 904 F.2d at 1097 (Posner, J., concurring). The opinion contains, among other things, an extremely lucid discussion of the place of art in First Amendment law. Id. at 1093-99.

(7.) Richard A. Posner, Pragmatism Versus Purposivism in First Amendment Analysis, 54 STAN. L. REV. 737, 741 (2001) (emphasis added).

(8.) Id. at 738.

(9.) Id. at 747.

(10.) Id. at 750.

(11.) Id. at 745.

(12.) POSNER, supra note 2, at 75.

(13.) Posner, supra note 7, at 741.

(14.) Id. at 743.

(15.) Miller v. Civil City of S. Bend, 904 F.2d 1081, 1097 (7th Cir. 1990) (en banc) (Posner, J., concurring), rev'd sub nom., Barnes v. Glen Theatre, Inc., 501 U.S. 560 (1991).

(16.) Posner, supra note 7, at 748.

(17.) Id. at 739.

(18.) Id. at 741.

(19.) Whitney v. California, 274 U.S. 357, 377 (1927) (Brandeis, J., concurring).

(20.) Miller, 904 F.2d at 1097.

(21.) Posner, supra note 7, at 741, 748.

(22.) See Brandenburg v. Ohio, 395 U.S. 444, 447 (1969).

(23.) In theory, a cost-benefit analysis should be concerned, on the cost side, only with the magnitude of the harm (with which the Brandenburg doctrine is not concerned at all) and its probability. Non-imminence could figure trivially in calculating the present discounted value of the harm (certainly not as a factor barring prosecution altogether), and the speaker's intention to bring about the harm would be largely irrelevant.

(24.) Rubenfeld, supra note 1, at 793.

(25.) Posner, supra note 7, at 743.

(26.) Id. at 743-44.

(27.) Id. at 743.

(28.) Id. at 742.

(29.) Id. at 744.

(30.) Minneapolis Star & Tribune Co. v. Minn. Comm'r of Revenue, 460 U.S. 575, 581 (1983).

(31.) See id. at 582-83 (holding that strict scrutiny applies when government "has singled out the press for special treatment").

(32.) Posner, supra note 7, at 744.

(33.) Id.

(34.) Id. at 748.

(35.) Id. at 744.

(36.) Id. at 741.

(37.) Id. at 738.

(38.) Id.

(39.) Id.

(40.) Id. (emphasis added).

(41.) Id.

(42.) Id.

(43.) See Anthony T. Kronman, Response: The Value of Moral Philosophy, 111 HARV. L. REV. 1751, 1752 (1998) ("Instead of elevating economics as he sought to do before, Posner seeks here to depose philosophy, but the intended result is the same....").

(44.) Posner, supra note 7, at 739.

(45.) Id.

(46.) Id. at 741.

(47.) Id. at 748.

Jed Rubenfeld, Slaughter Professor of Law, Yale Law School.