Weighing the Value of Messages Against the Value of Content-Neutral Regulations

The above argument is not one directed exclusively at courts. It applies to all coercive impositions of such valuations of information and thus suggests that legislative evaluation of information gains and losses is equally problematic. It suggests that legislatures should perhaps measure only the strength of the con­stituents’ preferences in deciding, for instance, between open space suitable for speech and alternative land uses. And it suggests that taxing one group because of another group’s preference for speech activities - for example, in building a town meeting hall - is illegitimate to the extent its justification depends upon the good the public will derive from the speech. (Building a meeting hall, an auditorium, or a library is nonproblematic to the extent it reflects the major­ity’s private-good preference for information of a certain type rather than, say, more tennis courts.) Finally, when government itself speaks with taxpayers’ resources, it is evaluating specific information - the most problematic of all government activities from a freedom of expression standpoint. Public schools and universities, public grants and subsidies for research, public broadcasting, public financing of election campaigns, and a variety of other activities, rather than being extensions of the animating spirit of freedom of expression, are in tension with it.

I will discuss each of these points at greater length in this and the following chapters, particularly Chapter Five.

A. Balancing Speech Interests Against Non-Speech Values

Served by Incidental Regulations

The entire corpus juris, from the general common law of contracts, property, and torts to the most particular tax regulations, affects what gets said, by whom, to whom, and to what effect. Speech and listening are costly activities. They use resources such as space, newsprint, radio frequencies, presses, and police protection, and impose other costs - noise, litter, and clutter. Schneiderv. State,²¹ for example, in which the United States Supreme Court struck down a municipal ordinance banning pamphleteering on city streets and sidewalks in order to reduce litter, resulted in the imposition on the public of the costs of litter or, alternatively, the costs in excess of the state’s next best alternative for eliminating litter. (If a less restrictive alternative is on the order of a Pareto superior move - the alternative does all the good at no greater cost and without affecting speech - then less restrictive alternatives are unlikely to exist; all alternatives will have greater costs in some respects.)

Thus, the Court’s decision in Schneider constitutionally mandated what can be viewed as a subsidy of pamphleteers. But why such a subsidy of pamphleteers?^[40]

Consider Jane, who complains about the high costs of The New Republic, cable television, books from Oxford Press, and a college education. Those costs result from laws - laws regarding property rights, laws conserving trees, laws affecting labor costs, laws regarding tax liability, and many other laws. Jane’s receipt of speech - which is, after all, what the First Amendment is arguably about - is adversely affected by those laws. Why should her attempt to receive this speech not be subsidized? (Alternatively, if one resists the notion that listeners’ rights are central to the First Amendment, why should The New Republic and Oxford Press, for example, not receive subsidies or relief from various laws in order to communicate with a wider audience?)

Next, consider John, who wishes to demonstrate on Main Street, which will tie up traffic and require police presence.

Next, consider Joan, who is denied several outlets for her message that, given her limited resources, would be the most effective: putting graffiti on the side of city hall, using a loudspeaker at night in a residential neighborhood, or putting up a pamphlet stand on land that, due to various zoning laws, is currently unaffordable for her. Why should Schneider but not Joan get a First Amendment subsidy here?

Finally, consider Jason, who wants the city to build an auditorium suitable for public lectures and rallies, but who is opposed by Jean, who would like the city to build more tennis courts because she and others prefer playing and discussing tennis to attending public lectures, and Jerry, who wants lower taxes so that he can afford to go to night school.

The O’Brien line of cases raises the same questions. If a medium is being regulated for noncensorial reasons, then striking down the regulation under the First Amendment imposes a costly freedom of expression easement on others. If Jake wishes to protest inflation by burning an inexpensive work of art in the lobby of the Treasury Building, presumably a court will not protect him, even though he has cost each taxpayer an insignificant amount. Yet, how is Jake different from any other user of a tangible medium that the government wishes to regulate to avoid costs to a variety of values? This question is especially relevant because those values can always be translated into amounts of money that the taxpayers would pay to preserve them.

Track Two also includes such obviously important-to-speech governmental decisions as what resources should be dedicated to the police and courts, and as a component of that decision, to protecting speakers from hostile audiences beyond those dedicated to protecting persons from assault and battery generally.

Without a theory of proper message effects, non-content-related regulations cannot be evaluated under freedom of expression, except in an arbitrary manner. Schneider did win, of course, as did the Hare Krishnas in International Society for Krishna Consciousness, Inc. v. Lee² But why they and not O’Brien,^{[41] [42]} the sleep-in protesters in Clark v. Community for Creative Non-Violence,^[43] or the many other actual losers? And if Schneider and the Krishnas, why not my hypothetical Jane, John, Joan, and Jason?

As discussed previously, assigning the speech value a constant weight in the calculus - for instance, equal to a significant governmental interest - does not help. Without a theory regarding message effects - what gets said, by whom, to whom, and with what effect - assigning any weight will be arbitrary. More importantly, because speech interests are affected regardless of what set of Track Two laws is chosen, the speech “constant” appears on both sides of the equation and does not produce a winner.

A possible alternative would be straightforward balancing rather than as­signing speech an arbitrary constant value. Under this approach, all of the information at stake under all alternative sets of laws would be examined, as well as all the nonspeech values, and a determination would be made as to which set of laws is superior.

There are obvious practical and institutional objections to such a balancing proposal. Given that this approach would involve nothing less than a comparison of all possible entire sets of laws, both for their message effects and for their effects on all the nonspeech values, the proposal is a practical impossibility for a legislature and surely for a court.

This practical objection should by itself be sufficient to undermine all Track Two judicial decisions and to dictate complete judicial withdrawal from Track Two. There remains, however, a theoretical objection to the enterprise as well. In principle, we cannot evaluate the message effects of Track Two laws: Either we assume the viewpoint of one who does not know what the information at issue will turn out to be, in which case we cannot evaluate it at all, or we assume the viewpoint of one who does know what the information will turn out to be, in which case we can evaluate it, but only from a partisan perspective inconsistent with freedom of expression itself. I shall now discuss these two points at same length.

1. assume ignorance of what information will turn out to be. Information is in some respects a commodity just like toothbrushes, auto­mobiles, and bananas. We buy it and sell it. When we buy it, we place a value on it in the sense that we decide to pay a particular price for it. Nevertheless, information is unique as a commodity in this crucial respect: The value of infor­mation, and therefore the price we should pay for it, is generally unascertainable until it has been purchased and received.^[44]

For example, the fifty-dollar medical book at the bookstore may contain information that will save my life or my loved ones’ lives. Then again, it may not. Should I pay the fifty dollars and find out? Similarly, in deciding whether to buy a seventy-five-dollar electric razor, should I purchase the issue of Consumer Reports that evaluates it? The magazine may save me some money, perhaps an amount greater than the magazine’s cost. But I will not know that until I pay the price to get the magazine.

Interestingly, economic theory is mute here. It cannot tell me whether infor­mation is worth its price, except when I have good actuarial information about the information itself, such as that there is a one-in-three chance that Consumer Reports will save me ten dollars on a seventy-five-dollar purchase, which makes it worth its $2.00 asking price. Economic theory itself normally assumes full or adequate information, or at least actuarial information about information, which makes economic theory largely irrelevant to purchases of information itself.^[45]

Economic theory does tell us, however, that the enterprise of producing information for sale - for example, what newspapers do - suffers from a public goods problem. Because information, once purchased, can be disseminated by the original purchaser for free^[46] or at nominal charge, the producer cannot capture the full public benefit of the information in the price charged. As a result, if the cost of production is greater than the price purchasers will pay, though less than the total public benefit of the information, the information will not be produced even though it would have been socially beneficial to produce it.^[47]

In view of this inability of producers to recapture the full public value of the information in the price charged and the resulting underproduction of the information, one might argue that we should subsidize the production of in­formation for distribution.^[48] There are, however, several difficulties with this conclusion.

First, there are an indefinite number of activities that might produce - or might produce if subsidized - information of public benefit. Without knowing what information would be produced that is not produced now, how can we determine which information’s production costs should be subsidized?^[49]

There are several layers to this problem. Because we cannot know what infor­mation will be produced by any information producer if we subsidize its produc­tion, we cannot know whether the information’s total social value - the public’s willingness to pay for the information if free-riding were precluded - will equal its social cost. In turn, this means we cannot determine which of the many producers to subsidize or at what level.^[50] Moreover, there are other pub­lic goods apart from information that require protection or subsidization, some of them in direct opposition to information production. In addition, both the content of information and its production also produce negative externalities not reflected in the costs because of collective action problems. For example, those who do not like the noise and congestion of demonstrations, the scandal­mongering of the tabloids, or the eyesores of campaign posters might, but for collective action problems, pay off the information producers in question to eliminate the negative externalities.

The second general problem with this public goods argument for subsidiz­ing information production is that its own logic renders it impotent as a Track Two tool. It highlights a problem information producers face - they cannot recapture in their price the public value of the information they produce. But defining who is a “producer of information” and what is a “subsidy” requires analytically privileging a certain set of entitlements and background laws, for it is only against that background that we identify who is producing information and what is a subsidy. A Track Two challenge, however, is a challenge to precisely that set of background entitlements and laws on freedom of expres­sion grounds. With a different set of background entitlements and laws, there would be different information producers producing - or potentially producing - different information and facing the recapture (of costs of production) prob­lem.^[51] Paradoxically, the public goods argument works as a potential Track Two freedom of expression argument only if we first decide on freedom of ex­pression grounds which set of entitlements and background laws to privilege; but once that is decided, everything is decided, leaving nothing for the public goods argument to do. In short, the public goods argument has no force as a Track Two freedom of expression argument because it provides no grounds for criticizing the background entitlements and laws that determine who the information producers will be.

2. assume knowledge of what information will turn out to be. The previous section raised the difficulties that arise for any Track Two freedom of expression enterprise if we do not know what information will be gained and what information will be lost under alternative sets of laws. The difficulties stem from two fundamental points. First, we cannot evaluate information and whether it is worth the costs required to obtain it until we actually have it; therefore, because different information will be gained and lost under every alternative set of Track Two laws, we have no way of assessing which of those sets of law is preferable. Second, the fact that we will tend to underproduce information as a commodity because of the opportunity to free­ride is no grounds for making a Track Two attack, because each alternative set of Track Two laws produces its own distinct set of information that is underproduced.

If now we make the heroic assumption that we know what specific infor­mation will be gained and lost under each alternative set of entitlements and background laws, are we in a better position to evaluate the information effects of alternative sets of laws?

Yes and no. Yes in the obvious sense that we can and do evaluate informa­tion’s importance once we have it. No, however, in the most crucial sense for freedom of expression analysis. That is because on at least its orthodox under- tanding, freedom of expression has as its core value that government not preempt individuals’ evaluations of information.^[52] Put differently, freedom of expres­sion - again, on its orthodox understanding - protects a process of citizens’ evaluating information and forbids preemption of that process by government’s privileging certain evaluations.

Al believes that knowing intimate facts about Madonna’s life is more im­portant than knowing how Bush’s tax policy will affect the economy. Barbara believes the opposite. And Charles believes that knowing how many blades of grass there are in his lawn is more important than either Madonna’s private life or Bush’s tax policy. There may be a point of view from which it is possible to say whether Al or the others are correct or incorrect in their evaluations, but whose point of view is it?^[53] Each person - you, Al, Barbara, Charles, and I - will believe his or her point of view to be the correct one; that is why he or she holds that point of view.

The Hare Krishnas in Lee no doubt believed the information about religion that they wished to distribute to airport patrons was more important than the information conveyed by an airport uncongested and uncluttered by solicitors and proselytizers and their litter or the information that would be produced by the resources used to police and clean up after the Krishnas. Is it? What if the religious teachings are false or sinister? What if an uncluttered environment heightens valuable aesthetic sensibility?

Jed Rubenfeld puts the point succintly:

The intelligibility of balancing in First Amendment law is hardly perspicuous. Nothing can be balanced against anything else without a common unit of measure. What is the unit of measure when First Amendment rights are “weighed” against governmental interests? No court has ever said.⁴²

Rubenfeld goes on to illustrate the point by imagining that a movie about race relations induces a few people to commit acts of violence. Under current First Amendment doctrine, the movie could not be banned. But, asks Rubenfeld, if that result is because the “balance” comes out in favor of the movie, what is the metric that supports such a balance?⁴³ Indeed, Rubenfeld notes, “no one can pretend to know whether the freedom of speech itself is worth its costs, which might conceivably be, someday, a complete breakdown in social order.⁴⁴ Moreover, balancing runs afoul of the core freedom of expression principle of evaluative neutrality.⁴⁵

The point is not that an individual cannot have a point of view about the values of the competing types of information, or that one point of view cannot be correct. The point is, rather, as I stated in Chapter One, that freedom of expression on any of its orthodox understandings requires government to treat any point of view on these matters as just one point of view among many.⁴⁶ The only point of view it privileges is its own - namely, that no particular point of view shall be privileged.⁴⁷

Of course, this view of freedom of expression’s central value is contestable and contested at its margins (though, as I have argued, not at its core). For example, some argue that freedom of expression requires only that no “politi­cal” position be privileged, and that only “political” speech and information are

⁴² Rubenfeld, supra note 24, at 788.

⁴³ Id. at 789.

⁴⁴ Id. at 793 (emphasis in original).

⁴⁵ Id. at 818

⁴⁶ See WestVirginiaStateBd. of Educ. v. Barnette, 319 U.S. 624, 642 (1943) (“If there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in... matters of opinion.... ”); Susan M. Gilles, AU Truths Are Equal, But Are Some Truths More Equal Than Others?, 41 Case W Res. L. Rev. 725, 726, 740-1 (1991); Rubenfeld, supra note 24, at 818; cf. David A. Strauss, Persuasion, Autonomy, and Freedom of Expression, 91 Colum. L. Rev. 334, 370 (1991).

Indeed, the liberal ideal observer construct I have described may not even be coherent. Real individuals have more than a bare desire to decide correctly what candidates they should vote for, or what views to hold about others in society, or what occupation they should pursue, or what products they should buy. They have concrete desires and views about products and candidates and occupations. It is possible that unless one knows what those concrete desires are, one cannot make sense of the question: what information would this person want in circulation? The bare desire to reach a correct decision may leave that question unanswerable in principle.

Id.

⁴⁷ This is paradoxical. Indeed, the First Amendment has been held to protect anti-First Amendment views, which is also paradoxical. See Carl A. Auerbach, The Communist Control Act of 1954: A Proposed Legal-Political Theory ofFree Speech, 23 U. Chi. L. Rev. 173 (1956); see also Post, Racist Speech, supra note 25, at 303-4 (discussing the paradox of public discourse exemplified by tolerating the intolerant). And see Chapter Nine infra. protected.^[54] This position, however, is, for related reasons, both unhelpful for Track Two analysis and also wrong. Regardless of what alternative sets of laws we compare in assessing Track Two challenges, political information will be lost and gained under each.^[55] Furthermore, the evaluation of the political infor­mation at stake surely will be political in the sense that requires governmental nonpartisanship. More fundamentally, the question of what information has po­litical relevance is itself political.^[56] If that is true, the position that freedom of expression protects only political speech devours itself.

Would Track Two analysis avoid the “point of view” problem if it sought “balance” or “representation” or “adequate airing” of points of view? This tack is equally unavailing. First, how many issues are there, and how many points of view are there on each? Second, whose position “represents” a point of view? Third, when has that position been “adequately aired”? The first question, if it is a meaningful one, is not a metaphysical one but one itself referable to points of view. Moreover, the number of points of view per issue is not determinable, since any answer depends on how the “issue” is described and what information and state of the argument is posited.^[57] As for representation and adequacy of airing, no one can represent my point of view except me; and that point of view has not been adequately aired until it is universally accepted. To illustrate, if people disagree with me, they obviously have not listened or understood, or they are intellectually deficient. How else can I explain their disagreement if I still hold my point of view to be correct?

This discussion reveals a deep paradox in freedom of expression theory. There are those, such as Owen Fiss and Jtirgen Habermas, who want the realm of expression to be a realm of pure process, the substantive results of which are correct because of the purity of the process from which they emerge. Yet, to make the process “fair,” they would build in substantive constraints. Fiss wants more diversity, more balance; Habermas wants to build in the various conditions that define his “ideal speech situation.”^{[58] [59] [60] Anyconception of diversity and balance, however, will be the product of, and will privilege, particular points of view. Balance and diversity can only be ascertained relative to an opinion of what points of view are plausible or sound.53 Fiss’s version of a fair process will be partisan, and the partisanship deprives any outcome of pure procedural legitimacy.54 Similarly, Habermas’s conditions for the ideal speech situation build partisan positions on all the major substantive issues into his procedural conditions. What comes out of the process will be predetermined by what went into setting it up.[61]}

In short, a pure freedom of expression process is a vacuous concept. There are just different background conditions for expression, each of which will lead to different substantive outcomes.

(1986). It turns out that when one builds in the substantive conditions required to make the “speech situation” “ideal,” there is little if anything left to debate. See Tom Campbell, Justice 221 (2001); LarryAlexander, Liberalism, Religion, andthe UnityofEpistemology, 30 SanDiego L. Rev. 763, 782 n.43.

⁵³ See Post, Managing Deliberation, supra note 26; Kenneth L. Karst, Equality as a Central Principle in the First Amendment, 43 U. Chi. L. Rev. 20, 40 (1975); cf. Frank I. Michelman, Property and the Politics of Distrust: Liberties, Fair Values, and Constitutional Method, 59 U. Chi. L. Rev. 91, 103-4 n.47 (1992):

There are other arguable cases of unfair impairment by some people’s speech of the value of other people’s speech that a committedly liberal constitutional-legal order may have little choice but to disregard. It may happen that less meritorious arguments backed by an individual speaker’s superior personal endowment of wit, chutzpah, eloquence, or charisma gain undue advantage in the speech market over more meritorious insights that a slower-witted, duller-spoken person has trouble articulating. However, personal handicapping in such circumstances seems not a liberally entertainable possibility. Again, it may happen that what an audience experiences as comparative cogency and soundness of argument is just a reflex of the comparative familiarity or conventionality of the ideas being urged. If this is at all a frequent occurrence, then public­forum doctrines of content-neutral order-maintenance and equal access may be a recipe for ensuring that currently prevalent views and perspectives will continue to prevail regardless of their responsiveness to the interests and values of the audience. But by what standard can liberals deal out ‘deviance’ or ‘dissidence’ subsidies?

Id. (citation omitted).

⁵⁴ See Charles Fried, TheNew First AmendmentJurisprudence: A Threat to Liberty, 59 U. Chi. L. Rev. 225, 251 (1992).

To be sure, if government controlled all the resources, then very little would be left of the value of free speech. There are two responses, however, to this concern. First, as classic liberals have always known but socialists have forgotten, a limited government and a vigorous private sector firmly beyond government’s reach are crucial to freedom of the spirit as well as to economic liberty. This is the basis of the old saying that liberty is indivisible. Since community control over resources is the light that beckons on the left, the left - to the extent it cares about freedom of the spirit - must seek out devices that will discipline the government’s monopoly in the realm of ideas. But any such device must consist of an official arbiter (that is, a government arbiter) to attempt to distribute access to the public forum. And that device cannot be content-neutral. It must decide which views have been heard too much, which not enough, and which should not be heard at all. The only content-neutral device turns out to be a society in which a significant portion of the resources are in private hands and beyond the reach of government altogether.

Post understands the legitimacy difficulty faced by those who, like Fiss and Habermas, seek to manage freedom of expression. On the other hand, Post fully appreciates the other side of this paradoxical predicament.^[62] For Fiss, Habermas, and others with similar views are correct that the set of background entitlements and laws will affect the outcomes of the freedom of expression process.^[63] Yet, one such set must always be in place. If we cannot evaluate that set and compare it with other possible sets in terms of message effects, we cannot evaluate the most important determinants of outcomes of freedom of expression, at least not on freedom of expression grounds.

This is the paradox. Track Two laws are extremely important in determining the quality of public debate. Indeed, Track Two laws almost assuredly have a much greater effect on expression than laws covered by principle (5), where government’s purpose is to affect messages, if relative effects can be mean­ingfully measured. Nevertheless, the core freedom of expression value that no partisan point of view be privileged precludes government from evaluating in­formation gains and losses, which, in turn, means that Track Two analysis is precluded by freedom of expression itself.

Again, the point is not just an institutional one about the limits of courts. It applies to any institution subject to the obligations imposed by the putative right of freedom of expression. Legislative evaluations of information, though democratic (unlike judicial evaluations), are themselves antithetical to the core freedom of expression value.^[64]

Finally, one might argue that adopting a comprehensive normative theory - such as Rawlsianism, utilitarianism, or libertarianism - as the vantage point for freedom of expression analysis provides a basis for evaluating information gains and losses and thus for evaluating Track Two laws. The rejoinder to this position is that although comprehensive normative theories are in one sense the proper bases for evaluating Track Two laws, they are the proper bases only insofar as they are not distinctly related to freedom of expression. Rawlsianism, utilitari­anism, and libertarianism are frameworks for assessing the entire corpus juris - to see if it maximizes liberties and wealth, maximizes aggregate or average welfare, or reflects libertarian rights - and as such frameworks, they are impor­tant to Track Two analysis. As comprehensive frameworks, however, they have nothing to do with expression or relative information value as distinct issues apart from liberties, welfare, and libertarian rights. Usually a comprehensive normative theory assumes full information of whatever type the theory makes relevant and then directs that that information be deployed to produce the results the theory dictates. And even if we assume the lack of full information, infor­mation effects can only be relevant to the choice of Track Two laws in a very restricted way. We would have to assume that the choice among Track Two laws were entirely neutral with regard to all values made relevant by the theory in question, and that relative information effects were the only ground for choice. The theory would then dictate that we choose that set of laws most likely to lead to the information upon which the theory itself places the higher value.

Yet, even in this extremely restricted way in which comprehensive normative theories would be relevant to Track Two analysis, choosing laws because their message effects are endorsed by a specific normative theory poses a freedom of expression problem. Presumably freedom of expression, even if its ultimate justification rests on a comprehensive normative theory, forbids imposing such a comprehensive normative theory on citizens’ evaluative decisions. Further, since imposition of a single comprehensive normative theory is the strongest of all governmental intrusions on the citizens’ evaluative processes, it is arguably the most antithetical to freedom of expression.

The upshot of this is that comprehensive normative theories are of no help when it comes to Track Two analysis, at least with respect to how freedom of expression bears on Track Two laws. If utilitarianism or Rawlsian theory favors a particular set of laws regulating property, contracts, torts, crimes, taxa­tion, the environment, and so on, then, assuming the comprehensive normative theory is correct, those are the laws we should have irrespective of their mes­sage effects. Indeed, the message effects of those laws are not only irrelevant, but they may be perverse. In a world of Rawlsian laws, anti-Rawlsian views may be generated more pervasively and with more vigor than in a world of non- Rawlsian laws. Rawlsian theory cannot favor Rawlsian Track Two laws because of their message effects, but must do so because those laws maximize equal basic liberties and the primary goods of the least advantaged.^[65] Similarly, util­itarianism cannot favor welfare-maximizing Track Two laws because of their message effects - which may be anti-utilitarian - but must do so because they are welfare-maximizing. Comprehensive normative theories are not irrelevant to the choice of Track Two laws - they may be all that is relevant ultimately - but they are irrelevant to that choice on freedom of expression grounds.

3. summary. If we assume the position of ignorance regarding message effects - what would be said, to whom, by whom, and to what effect - under alternative sets of Track Two laws, then we can make no evaluation of those sets in terms of information gains and losses. At most we can assume that under each such set, the deliberate production of information will face a recapture problem, a point that is normatively impotent because it applies to all alterna­tive sets of laws.

59

On the other hand, if we assume the position of knowing the information effects under each alternative set of Track Two laws, we then face the problem that all evaluations of such information are positional and partisan and should not be officially endorsed by government. Even a comprehensive normative theory underlying freedom of expression will be too strong to be a Track Two tool because it will usually endorse a set of Track Two laws on grounds inde­pendent of information gain and loss. Moreover, comprehensive theories are nothing more than partisan positions that the government should not endorse in attempting to structure public debate, even if, paradoxically, they themselves provide the justification for freedom of expression and its proscription of their own endorsement.

B. The Implication of the Failure OfBalancing

Track Two freedom of expression analysis requires that we balance information gains and losses and all other values furthered by the entire set of laws in question against the information gains and losses and all other values furthered by all the alternative entire sets of laws. We cannot do this as a practical matter. Perhaps more importantly, we cannot do this as a theoretical matter. Even if we knew what information each alternative set of laws would generate, freedom of expression itself forbids government to act on an evaluation of that information. Track Two analysis is not an extension of freedom of expression but a violation of it. Track Two laws can be evaluated on many grounds - but not on freedom of expression ones.

There are two major implications for the freedom of expression to be drawn from this conclusion. Most obviously, the courts should no longer hear Track Two challenges to laws.

The second implication, however, is directed at legislation, not adjudication, and is quite farreaching. If my analysis is correct, the legislature violates free­dom of expression if it adopts and imposes through law any partisan evaluation of information, which is to say, any evaluation. Essentially, although the gov­ernment may enact laws based on a variety of values, it may not rely on the value of information gained or lost in choosing which laws to enact or repeal. Or at least it may not do so without violating the evaluative neutrality at the core of freedom of expression.

This, in turn, suggests that the government’s decisions whether to build an auditorium to provide a forum for public debates, to fund election campaigns, to expand public education, to fund research, and a multitude of other decisions resting wholly or in part on government’s evaluation of information rather than its aggregation of private preferences, are problematic.

This implication, however, appears to offend common sense. Do we not need more fora for public debate, more public funding of elections, more broadcasting options, more education, and more research? Should government not support such endeavors?^[66] Yet, how can one answer those who disagree without presup­posing what information such endeavors would produce and without evaluating that information and its opportunity costs through the prism of a controverted, partisan point of view?

Ultimately, Track Two analysis reveals not only its own freedom of expres­sion illicitness, but also the illicitness of all information-conscious legislative decisions regarding the quality and quantity of public discourse. Consider a legislative debate, such as might have occurred within the New York City Port Authority prior to Lee, over whether to allow the sale and distribution of lit­erature inside the airport terminals and thereby incur the costs of congestion, annoyance, extra policing, clean-up, and so forth. The legislative body will be aware that if sale and distribution are not allowed, some passengers will not re­ceive some messages, usually religious or political, and usually not mainstream. On the other hand, a clean, uncongested airport, aside from being valuable in itself, also communicates various messages to its patrons. Furthermore, the money saved on clean-up can be used to provide other services - including per­haps schools and libraries - or can be put in the pockets of taxpayers who might spend it on information. How is the choice to be made? The legislative body might disregard remote, uncertain, or amorphous message effects and focus only on the most direct, certain, and specific effects, such as the effects on the Hare Krishnas or the LaRouchites. Even so, what should it do? Should it count these effects as outweighing the nonspeech costs? Why? Because it sympa­thizes with the Hare Krishnas and LaRouchites and believes others should hear their messages at the cost of other values and messages? Alternatively, should it consider these effects as outweighed by the costs? Again, why? Legislative evaluation of the Hare Krishnas' and LaRouchites' messages seems inevitable if message effects are to be taken into account, but such evaluation on closer examination appears quite illicit.

Alternatively, consider a debate within a legislature or administrative agency over whether the broadcast media are sufficiently “diverse” or “balanced” in the array of information they provide the listeners and viewers in their markets. This debate can be meaningful only if there are criteria for identifying degrees of importance of information,^[67] criteria for diversity,^[68] and criteria for balance.^[69] The choice of such criteria is deeply evaluative and partisan.

Current United States Supreme Court Track Two jurisprudence seems most consistent with the following position: The legislative bodies may enact or mod­ify Track Two laws explicitly in order to facilitate speech activities or to promote balance and diversity as long as the legislature is presumed to have had no clear impression of what information would be favored and disfavored, or no partic­ular sympathy for the messages likely to be favored. Governmental decisions to allow demonstrations, to open facilities for pamphleteering, to build public auditoria and other communicative fora, to break-up media monopolies, and to subsidize broad categories of information media will uniformly be upheld, even if those decisions were premised on some estimate of the relative value of information lost and gained.^[70] Because the content of the information gained in these cases cannot be known with any certainty when the decisions are made, the decisions are no more arbitrary than had they been made without regard to information gains and losses. As long as the legislative body is operating in the dark, the Court will not care that concern for message effects was what determined the outcome of the legislative balance.

This analysis explains as well why the Court often strikes down obviously content-based regulations of time, place, and manner. If, for example, the leg­islative body has taken a partisan position in favor of speech related to labor disputes and against all other speech,^[71] or in favor of non-religious speech and against religious speech,^[72] the Court will strike down the time, place, and manner regulation.

Thus, when the government makes a Track Two decision based on message effects, the Court will sustain the decision if it believes that the legislature could not have evaluated information effects in a partisan manner. Where the legislative decision appears to be partisan, however, the Court will invalidate it.

When government’s Track Two decision is not based on message effects - when it is purely content-neutral - Track Two First Amendment jurisprudence is destined to fail. That is so because no Track Two theory is available that is consistent with the core value of freedom of expression - citizen autonomy regarding the evaluation of information. The courts can strike down and, as in Schneider and Lee, have struck down laws in such Track Two cases. Nonethe­less, Track Two judicial decisions can only be either arbitrary or illicit.

We must therefore reject principle (4) as definitive of the scope of freedom of expression. Principle (4) would deem to implicate freedom of expression all governmental acts that affect quantitatively or qualitatively the receipt of mes­sages that someone intended to communicate. But all governmental acts, and all alternatives to those acts, do so, and there is no neutral evaluative perspective from which to review the message effects of every possible corpus juris.

Jed Rubenfeld also rejects principle (4) and endorses principle (5), which is concerned solely with whether government’s purpose is to affect the content of messages. Rubenfeld,responding to the proposal by Judge Richard Posner^[73] that all laws - even laws that are content-neutral and not enacted for the purpose of affecting messages - be assessed by balancing their benefits against their costs, including their effects pro and con on expression, points out the absurdity of the proposal by considering someone arrested for violating the 55-mile-per-hour speed limit who claims that he was speeding to express some idea.

A cost-benefit judge... 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....

For all I know, the no-speed-limit rule prevailing on certain European highways is a 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 suggests 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.^[74]

Rubenfeld points out that under principle (4) and thebalancing test that would attend it, everyone engaged in expressive conduct could claim a freedom-of- expression exemption from whatever content-neutral law otherwise restricted them. Newspapers, for example, might claim an exemption from sales taxes, which decrease sales and ultimately access to information.^[75] Posner denies that such an exemption, or a similar exemption of authors and television anchors from the income tax, would be a plausible result Ofbalancing because it would amount to subsidizing such activities.^[76] Rubenfeld replies:

“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.^[77]

As Rubenfeld points out, this result - the absence of freedom-of-expression exemptions from Track Two laws - is incomprehensible to a balancer.^[78] Track Two laws probably have a much larger impact on what gets said, to whom, and with what effect than the Track One, content-based laws that fall under principle (5). For Rubenfeld, and for me, “[t]he message is clear: The cost-benefit First Amendment - that is, principle (4)-defined freedom of expression - is not our First Amendment.”^[79]