Harvard Law Review

November, 1998

The Supreme Court 1997 Term


(Cite as: 112 Harv. L. Rev. 84)


Frederick Schauer [FNa1]

Copyright © 1998 Harvard Law Review Association; Frederick Schauer

American free speech doctrine has never been comfortable distinguishing among institutions. Throughout its history, the doctrine has been persistently reluctant to develop its principles in an institution-specific manner, and thus to take account of the cultural, political, and economic differences among the differentiated institutions that together comprise a society. Despite the textual mandate of the Press Clause, cases dealing with press access and reporters' privileges have rejected the notion that the Press Clause confers on the press distinct rights not conferred by the Speech Clause on all speakers and all writers. [FN1] With few exceptions, constitutionalized defamation law applies the same principles to a libel in the New York Times as it does to a slander over the back fence. [FN2] Universities have generally been unsuccessful *85 in gaining special rights under the rubric of academic freedom. [FN3] And in their recurring encounters with attempts to restrict speech that is "indecent" but not legally obscene, the Justices have been hesitant, save for the case of broadcasting, to distinguish among the contexts in which the supposedly indecent words and images might appear. [FN4]

Historically, there have been good reasons for the Supreme Court's unwillingness to draw institutional distinctions in the free speech context. Much of the very idea of free speech depends on ignoring what might otherwise appear to be politically and morally relevant features of speakers and speeches, [FN5] so it should come as no surprise that free *86 speech doctrine has resisted the pressure to distinguish among media and among institutions. [FN6] Yet however sensible this strategy has been in the context of threatened governmental restrictions on the activities of private speakers and publishers, it is increasingly clear that the refusal to draw doctrinal distinctions among culturally distinct institutions is simply unworkable in the context of the vast and increasing domain of free speech claims about government land, government funds, and government employees. The necessity of drawing institutional distinctions in the context of free speech questions arising within the precincts of government has never been more apparent than in two cases decided in the 1997 Term, Arkansas Educational Television Commission v. Forbes [FN7] and National Endowment for the Arts v. Finley. [FN8] In these cases, the Court toyed with the idea that the autonomy of journalistic and artistic decisionmaking, respectively, might be worth protecting precisely because of the political and social situations of the journalistic and artistic cultures, thus allowing First Amendment outcomes to turn on the particular characteristics of the specific institutions. Although in the final analysis the Court in both cases clung to the wreckage of doctrines designed for the demonstrably different situations of earlier First Amendment controversies, the Court's explicit reliance on institution-specific ideas moved it closer to a workable approach to managing the free speech issues that arise within the government's own enterprises.

The importance of Forbes and Finley thus lies partly in the path not fully taken, yet suggested in ways that have significant implications. Moreover, the interest in these cases lies also in what the Court's no-more- than-tentative willingness to take institutions seriously says about an intractable tension between free speech theory and *87 judicial methodology. If freedom of speech, especially but not only for government enterprise cases, is largely centered on the policy question of institutional autonomy, but the Court's own understanding of its role requires it to stay on the principle side of the policy/principle divide, then the increasingly obvious phenomenon of institutional differentiation will prove progressively more injurious to the Court's efforts to confront the full range of free speech issues.

I. Forbes and the State as Journalist

Arkansas Educational Television Commission v. Forbes presented the familiar problem of how best to allocate scarce communicative resources as between the competing demands of diversity of participation and appeal to the actual preferences of actual audiences. Among the most common manifestations of this dilemma is the political debate, in which the inclusion of minor party and other "fringe" candidates increases the diversity of voices at the risk of losing much of the audience, [FN9] but in which limiting participation to the major party or otherwise most viable candidates increases the audience and sharpens the debate at the risk of further entrenching mainstream views against challenge from the outside. [FN10] In Forbes, the Arkansas Educational Television Commission, "[a] state-owned public television broadcaster," [FN11] sponsored a debate featuring candidates in an upcoming congressional election. The Commission, which operates five public television stations in Arkansas, invited the Republican and Democratic candidates to participate in the debate, but did not invite Ralph Forbes, an independent candidate who had qualified for a place on the *88 ballot, but who appeared to have at best an infinitesimal possibility of actually winning the election. [FN12]

Not surprisingly, Forbes challenged his exclusion, and because the public broadcaster that excluded him is an arm of the state, he framed his challenge in the First Amendment rhetoric of state exclusion from a public forum. He failed in the district court, both initially and on remand after an Eighth Circuit reversal en banc, [FN13] but prevailed on the merits in the Eighth Circuit, [FN14] which held that the Arkansas Educational Television Commission had created a public forum for all legally qualified candidates--a public forum from which it could not exclude Forbes without a much stronger justification than Forbes's lack of political viability. [FN15]

In reversing the Eighth Circuit and upholding the prerogative of the broadcaster to exclude Forbes from the debate, the Supreme Court nominally attempted to shoehorn the case into its pre-existing categories and sub- categories of public forum doctrine. [FN16] Justice Kennedy, writing for the majority, reiterated the now-familiar points that the universe of traditional public forums, to which access is compelled by direct constitutional mandate, [FN17] is limited to the streets and parks, and that the category of public forums by designation, the existence of which entails similar state obligations to allow access, requires an explicit state decision to open the forum to all members of some relevant category. [FN18] At this point, however, the analytic structure of Forbes became *89 more complex, as Justice Kennedy offered a distinction between nonpublic forums, as to which access is not required but viewpoint neutrality is, and government enterprises that are not forums at all, as to which the commands of the First Amendment, including the command to avoid viewpoint discrimination, are irrelevant. [FN19] And then the Court's doctrinal structure developed into something even more intricate. Justice Kennedy concluded that public broadcasting in general is not a forum at all, and is thus permitted to engage in viewpoint-based exclusions, except when public broadcasting sponsors candidate debates, which are nonpublic forums and thus subject to the prohibitions on viewpoint discrimination. [FN20] Having set up this rather complex framework, the Court held that even though political debates are nonpublic forums from which speakers may not be excluded on the basis of viewpoint, Forbes could still be excluded from the Arkansas debate because the decision to exclude him on the basis of his lack of political viability did not amount to a viewpoint-based exclusion. [FN21]

This precis tracks the abstract and complex structure of the Court's analysis, yet still fails to capture the essential flavor of Justice Kennedy's opinion. For only from one aspect is that opinion about forums, whether traditional public forums, designated public forums, nonpublic forums, or non- forums. From another aspect, and a more important one, the opinion is about state journalism as journalism, as opposed to state journalism as an enterprise of the state. In the introductory paragraph of the opinion, Justice Kennedy characterizes Forbes's adversary not as "the State" but as "the broadcaster," [FN22] and characterizes this state broadcaster's action as an "exercise of journalistic discretion." [FN23] The theme continues, and saturates the opinion. Justice Kennedy talks of a "professional staff" and "editorial discretion" at Arkansas Educational Television, [FN24] stresses the links and similarities between public broadcasting and private broadcasting, [FN25] places great weight on the way in which the debates at issue were designed and organized in close collaboration with the Arkansas Bureau Chief for the Associated Press, [FN26] and repeatedly describes as "journalistic judgment," [FN27] "journalistic discretion," [FN28] "journalistic purpose," [FN29] *90 and "journalistic freedom" [FN30] what others would describe as "state action."

Indeed, there is a striking contrast here between Justice Kennedy's majority opinion and the dissenting opinion of Justice Stevens. [FN31] In arguing that Forbes's exclusion violated the First Amendment, Justice Stevens spent little time characterizing the excluder as a broadcaster or a journalist. For him, Arkansas Educational Television was, at its core, the state, and the constitutional flaw lay in the unconstrained exercise of state discretion. That the unconstrained state decisionmaker happened in this case to be a broadcaster and not, say, an airport [FN32] or a post office, [FN33] was as irrelevant to Justice Stevens as it was, arguably, dispositive to Justice Kennedy and the majority. [FN34]

A comparison of Forbes to two hypothetical cases suggests that the journalistic character of Arkansas Educational Television may have been more determinative than is indicated by the structure of the majority opinion. Consider a hypothetical exclusion, on the same grounds as in Forbes, from an airport [FN35] or from the advertising space in a city bus. [FN36] Existing doctrine suggests that neither of these forums is a traditional or designated public forum, [FN37] but that both are nonpublic forums nevertheless subject to the constraints of viewpoint neutrality. [FN38] Yet if an airport or a transit authority were to restrict its speaking or advertising space to speakers who had previously demonstrated their appeal to a significant portion of the potential audience, and had imposed *91 the restriction on the same grounds of preventing cacophony and focusing the public debate around existing widely- held positions, I suspect that the Court would strike down the exclusion without dissent. And if in that context the state agency were to have argued that the viability test was viewpoint-neutral, I again suspect that the claim would have been treated with far more skepticism than was the same claim in Forbes. [FN39]

Consequently, it appears that all of the discussion of broadcasting, journalism, and editorial discretion in Forbes is not the typical makeweight we often find in First Amendment opinions, [FN40] in which the Court says flattering things about the institution whose representative it is protecting even though the institutional affiliation of that representative makes no difference to the result. Rather, in Forbes it exercises a crucial influence on the outcome. Although the doctrinal structure of the majority opinion in Forbes is focused on public forum doctrine, and to some extent on the idea of viewpoint discrimination, in the end it is the institutional character of public broadcasting as broadcasting, heightened here by the involvement of broadcasting professionals in the very decision under attack, that appears to have determined the outcome of the case.

Institutional specificity appears not only to have determined the outcome in Forbes, but also to have played a crucial role in the doctrinal structure around which that outcome was erected. In concluding that candidate debates play such a special role in public life that they deserve to be treated differently from other forms of public *92 broadcasting, [FN41] Justice Kennedy again drew on the specific characteristics of a relatively discrete social institution. The Court might have seen the institution of the candidate debate simply as a fungible part of public (or, in other contexts, private) broadcasting. Alternatively, the Court might have seen the institution simply as a fungible part of the overall landscape of political discussion. By rejecting both of these assimilations, and instead identifying a special doctrinal status for candidate debates on the basis of the special purpose that candidate debates serve, [FN42] Justice Kennedy reinforced the institutional specificity with which his opinion is soaked.

II. Finley and the State as Patron

Intriguingly, institutional status is equally important, although less explicitly so, in the other free speech case of the 1997 Term, National Endowment of the Arts v. Finley. [FN43] In Finley, the Court was presented with a challenge to the "general standards of decency" factor that Congress had inserted into the 1990 legislation reauthorizing and controlling the grants awarded by the National Endowment for the Arts. [FN44] The performance artist Karen Finley and three other artists were denied funding by the NEA Council despite initial approval in the NEA's peer review process. The artists claimed that the "standards of decency" provision, pursuant to which their funding was denied, constituted impermissible discrimination on the basis of both artistic and political viewpoint. They prevailed in the District Court [FN45] and then again in the Ninth Circuit. [FN46] Reversing the Ninth Circuit, however, the Supreme Court upheld the provision and affirmed the ability of the NEA to use factors in making awards that would be impermissible in other contexts.

Just as with Forbes, Finley was not a "classic" free speech case in which a speaker, wishing to speak with only the use of her own [FN47] resources, *93 is impeded by state-imposed criminal or civil sanctions on engaging in that activity. [FN48] No state actor sought to prevent Ralph Forbes from using his own voice or his own resources to urge his own claim to represent the electorate of the Third Congressional District in Congress. And no state actor sought to prevent Karen Finley from removing her clothes and coating herself with chocolate in order to make political statements in the public forum or on any property whose non-state owner would permit the activity. Rather, in both cases, as with an increasingly large percentage of First Amendment adjudication, the disputes arose in the context of the government's control over its own speech-related enterprises--schools, universities, libraries, museums, arts funding, public broadcasting, state publications--and the central issue was whether the state can impose content- based criteria in the use and allocation of its intrinsically content-based enterprises.

The content-based criterion challenged in Finley involved the section of the National Foundation of the Arts and Humanities Act requiring the Chairperson of the National Endowment for the Arts to ensure that "artistic excellence and artistic merit are the criteria by which [grant] applications are judged, taking into consideration general standards of decency and respect for the diverse beliefs and values of the American public . . . ." [FN49] The history of the "general standards of decency" provision is well-known, [FN50] with public reaction to NEA *94 funding of both Andres Serrano's Piss Christ [FN51] and the homoerotic photographs of Robert Mapplethorpe prompting Senator Jesse Helms and others to call for elimination of the Endowment itself, or at the very least for elimination of its sponsorship of so-called "indecent" art. [FN52] This reaction prompted a public and congressional counterreaction, and the compromise that emerged from Congress contained a prohibition on NEA funding of the "obscene," [FN53] as well as the constitutionally more problematic "decency and respect" provision, the facial challenge [FN54] to which provided the key question in Finley.

As with Justice Kennedy's majority opinion in Forbes, Justice O'Connor's majority opinion in Finley used an implausible doctrinal structure to retreat, at least in part, from an explicitly institution-specific approach to determining which governmental institutions can employ which kinds of content- based, and indeed viewpoint-based, criteria. Most implausible of all, as both Justices Scalia [FN55] and Souter [FN56] made clear in their otherwise quite divergent opinions, is the view, superficially determinative of the result in Finley, that the decency standard is permissible because it is only a factor and not a categorical rule, [FN57] such that some indecent art might still be funded even though its indecency is taken into account by the Endowment in making its funding decisions. For Justice O'Connor and the majority, this absence of categorical effect played a large part in saving what would otherwise have been an unconstitutional standard.

*95 The majority's reliance on the factor/rule distinction is surprising, for the distinction's inability to support the result in Finley is apparent from even the briefest survey of both real and hypothetical cases. It is hard to imagine that the result in Texas v. Johnson [FN58] would have been different had the degree of respect shown for the American flag been merely a "factor" to be considered in deciding when unofficial uses of the flag would be permitted. Similarly, it is highly unlikely that the outcome would have been different in a civil action such as Hustler Magazine v. Falwell [FN59] if the impermissible rule of liability had been merely a "factor" for jury consideration. And even in the context of government facilities or government enterprise cases, if a legislature directed an airport commission to take into consideration a group's patriotism or Republicanism in deciding whether to allow it to solicit for donations at the airport, it seems equally unlikely that the distinction between "take into consideration" and a categorical prohibition would save an otherwise clear case of impermissible viewpoint discrimination. [FN60]

Because the factor/rule distinction cannot do the work that Justice O'Connor's opinion expects it to do, [FN61] then perhaps the laboring oar in Finley is pulled by the fact that this is a government funding case and not a government restriction case. One possibility (and one that appears on the surface of the opinion to have been important to the outcome) [FN62] is therefore that otherwise impermissible viewpoint-based distinctions, as with the distinctions between decency and indecency and between respect for the diverse beliefs and values of the American public and scorn for those beliefs and values, [FN63] are impermissible when *96 they are part of a direct prohibition, but not when they are merely among a larger number of criteria used to determine what government will support and what it will not. But again the cases belie this possibility. As Justice O'Connor stressed in Finley itself, only those funding cases that can be narrowly understood as the government's own speech [FN64] are immune from the prohibition on viewpoint discrimination. When government is operating in its subsidizing mode and not in its speaking mode, the existing caselaw supports the view that viewpoint- based distinctions are impermissible, [FN65] even though content-based distinctions that are not viewpoint-based are both inevitable and constitutional.

So if neither the factor/rule nor the restriction/subsidy distinction can account for the Court's deference to and allowance of a viewpoint-based funding criterion, what might explain the result, and what might justify it? The answer comes partly from the opinion, [FN66] but even more from the oral argument of the Solicitor General, who asked the Court to accept the proposition that "there is something unique . . . about the Government funding of the arts for First Amendment purposes." [FN67] Although Justice O'Connor did not explicitly rely on this argument, its overtones pervade the majority opinion. In contrast to Justice Souter's dissent, [FN68] which repeatedly referred to the NEA as "the government," Justice O'Connor described the NEA grant process in great detail, referring to the key role of "experts in the relevant field of the arts," "advisory panels," "the nature of arts funding," and the necessity of "aesthetic judgments." [FN69]

As in Forbes, the aspects of this institution, especially those aspects that link the viewpoint-discriminatory process to a larger and largely non- governmental social institution (journalism in Forbes, the arts in Finley), appear to do more of the work than the structure of the majority opinion indicates. In light of other relatively contemporary *97 cases, the Court's most explicitly stressed features of a factor/rule distinction and a subsidy/prohibition distinction do not, even when taken together, appear to explain the result. But if the nature of art, the culture of art, and the practice of arts funding are in important ways special or distinct, and if the nominally governmental structure of NEA decisionmaking connects with this culture and its institutions, then the nature of art may have done the work in Finley that the nature of journalism did in Forbes. [FN70] It is this possibility--that both of these cases, and potentially more cases in the future, can best be explained and justified in terms of institutionally specific rules and principles--that I propose to explore in the pages to follow.

III. The False Paths

These brief overviews of Forbes and Finley are sufficient to signal and ground the theme of institutional specificity that I wish to develop. Before doing so, however, it is necessary to look slightly further down some of the paths that the Court has taken, both here and earlier, in trying to deal with government enterprise free speech cases. Only by understanding why none of these more "principled" approaches has been successful can we explain why the seemingly less principled policy-oriented and institution-specific approach has begun to surface.

A. Forums, Public and Otherwise

Of all of the paths down which the Court might go in dealing with government enterprise cases, the so-called "forum doctrine" appears least satisfactory. Originally designed for the perfectly sensible purpose of mandating a speaker's easement on certain forms of government property, [FN71] public forum doctrine began to run into trouble once *98 confronted with the argument that other pieces of government property, in both the literal and figurative senses, were the constructive equivalent of the classic public forums of streets, parks, and sidewalks. Thus, so the argument goes, these sites should likewise be subject to a speaker's easement. [FN72] Yet for obvious reasons, the Court could not create such an easement for all government property. For example, any place at a public school or state university where students congregated, or the lobbies outside courtrooms and executive offices, might be thought by some to be constructive equivalents to streets, parks, and sidewalks, because these are locations in which people may find themselves talking about the issues of the day. The category of such places, though, seems limitless. Consequently, the Court shut off the "constructive equivalent" argument, [FN73] with the one narrow exception of "designated public forum" doctrine, [FN74] which enforces against the state the state's own prior decisions to allow a speaker's easement. [FN75]

Beyond designated public forums, however, it is hard to see the point of forum analysis in government enterprise cases. In the typical *99 case, the complaint is not about access, but about discriminatory treatment. [FN76] And at the heart of this issue is the seemingly banal but quite important point that content-based discriminatory treatment is appropriate in some contexts, but not in others. Yet once we recognize this idea, the point of combining the determination of which contexts permit content discrimination and which do not with public forum analysis is elusive. If access is mandatory, then the focus on content discrimination is redundant. But if access is not mandatory, then the existence (or not) of a public forum is superfluous. What is not superfluous is the question whether this is one of the government enterprises which may control for content or viewpoint, and as to this question public forum doctrine offers no assistance. As long as the Court is not prepared to expand the category of public forums, and there is no indication that it is, then there seems no reason to think that the idea of a public forum will be of any assistance in determining which non-access government enterprises may be content-based and which must be content-neutral.

The irrelevance of public forum analysis is apparent from Justice Kennedy's opinion in Forbes. Justice Kennedy correctly identified the televised debate as neither a traditional nor a designated public forum. [FN77] But this idea could hardly have seriously been maintained in the first place. Forbes was not claiming access to other programming; instead, he was seeking the right not to be discriminated against on account of content from a forum the very content of which created the nature of his claim. So the question is whether a state- created political debate may exclude some candidates on the basis of their non- viability. This is, in a literal sense, a question of access, but so is every government discrimination case. Karen Finley was claiming access to NEA funding, just as the NAACP Legal Defense and Education Fund was claiming access to a fund-raising campaign [FN78] and the Hare Krishnas were claiming access to airports. [FN79] But given that in all of these cases there was no serious claim for mandatory access--for an easement even against the state's claim to exclude all speakers--it is hard to see how an analysis that starts with the fact that the state must set aside some of its property for all speakers will get us very far in solving the problem. That forum analysis plays no role at all in Finley, and that the conclusory distinction between a nonpublic forum and a *100 not-forum does all of the work in Forbes, serves only to underscore the point.

B. The State as Speaker

If the false path of public forum analysis points us inexorably toward the question of content, then one approach would be to start with the idea that the First Amendment does not apply to the state as speaker. [FN80] Although suggestions to the contrary have occasionally surfaced in the literature, [FN81] even these do not purport, for example, to limit the President's content-based determination of what content will come out of his mouth in the State of the Union Address, or anywhere else. This is, of course, the extreme case, and things admittedly become more difficult once we recognize that the state cannot literally speak, but can speak only through the voices of others, others who have their own First Amendment rights in many contexts. Indeed, this very difficulty might be thought to provide the appropriate analytical framework, with the courts charged with the task of sorting out which forms of content regulation within government enterprises should count as constitutionally immune government speech and which should be thought of as constitutionally constrained limitations on the free speech rights of those members of the citizenry who happen to be employed by the government.

*101 Yet when we try to determine when the First Amendment rights of government employees [FN82] or contractors [FN83] are sufficient to block the state's right to compel them to serve as the state's voice, we again get little help from the abstract principle that the state can speak. The state might decide, for example, that one way of speaking is through the books it places in its libraries, or the words it requires its university teachers to utter. It might require its librarians to select books expressing only the state's viewpoint or order its university teachers to present only the state's viewpoint on matters of social policy. To many people these would be highly problematic cases, but the claim that these actions are not the speech of the state is simply false. Indeed, if there is a difference between the state hiring a librarian to select only books expressing the state's viewpoint or hiring a university professor to argue the state's viewpoint, on the one hand, and the state hiring an artist to paint a mural on a state office building including George Washington but not Vladimir Lenin, [FN84] on the other, that difference cannot be explained by characterizing the former as private speech and the latter as government speech, for such a characterization is merely the announcement of a result. Rather, the task is to identify when the state can use its power over its employees or contractors to conscript them into service as the state's voice. This is a real and important question, but the claim that the state can speak does not help us to answer it. Instead, the occasions on which the state may require its employees or contractors to utter or refrain from uttering certain words [FN85] can then be characterized as government speech cases, [FN86] but this label for the conclusion does not provide us with any of the resources that will help in reaching that conclusion.


C. Unconstitutional Conditions

Much the same can be said about the decreasingly useful doctrine of unconstitutional conditions. [FN87] It is, of course, unconstitutional for the state to condition tax exemptions, [FN88] welfare benefits, [FN89] and some forms of non-policy public employment [FN90] on refraining from engaging in otherwise protected speech unrelated to the purpose of the governmental program. But what looks like an unconstitutional condition from one angle may look strikingly like government speech, or government support of its own activities and policies, from another. [FN91] Numerous commentators have offered solutions to the problem of unconstitutional conditions, [FN92] all in an attempt to negotiate the difficult terrain created by the fact that government speaks only through its employees and its policies. Requiring an employee or contractor to speak the government's message will therefore look like an unconstitutional condition insofar as it conditions employment, for example, on speaking words with a certain content, but will look like government speech insofar as it embodies the government's prerogative of sending out its own message. The problem is not in deciding which one of *103 these features is present in a particular case, for both of them are present in almost all cases. And as with most controversies thus subject to equally accurate but mutually exclusive characterizations, the fact that both characterizations are technically accurate means that the decision to prefer one characterization over another must be driven by factors external to the bare idea of unconstitutional conditions itself.

Because these external factors tend to vary considerably, the resulting doctrine looks quite messy, a state of affairs decried by commentators offering a better standard or test to distinguish unconstitutional conditions from those conditions on participation in government enterprises that must of necessity be accepted as permissible. [FN93] Yet while suggestions for improving unconstitutional conditions doctrine proliferate, the doctrine itself has all but disappeared from the Supreme Court's arsenal. [FN94] Just when the Court's free speech docket has become increasingly dominated by government enterprise cases, the Court has jettisoned the classic tool for handling such cases. At times, as in Rust v. Sullivan, the Court mentions the doctrine for the purpose of rejecting its applicability, [FN95] but when some condition is actually invalidated, the work is almost invariably done not by the doctrine of unconstitutional conditions, but by something else entirely.

This is not surprising. Although the doctrine of unconstitutional conditions serves well enough for conditions unrelated to the program at issue-- conditioning employment in the sanitation department on a willingness to refrain from criticizing American foreign policy--such conditions are rare in a post-McCarthy world. Rather, most contemporary controversies involve conditions plainly related to the very program in which they are imposed. This connection does not necessarily make the conditions constitutional, but it does change the character of the dispute, especially because almost all relevant conditions will involve some sort of content discrimination.

Finley illustrates this point nicely. If Karen Finley's eligibility for an NEA grant had been contingent upon agreeing to speak, or not to speak, outside of the context of the very art for which she sought support, we would have seen a classic unconstitutional condition. If, for example, she had been told not only that she could not get funding for *104 her own form of performance art, but also that she could not get funding for anything, unless she refrained from performing in chocolate anywhere, the unconstitutional conditions doctrine would compel invalidation. But that was not the case here. Instead she was told that her eligibility for this grant was contingent upon some characteristic of this art, and thus her freedom to produce whatever kind of art she wished--including bad art [FN96]--was curtailed only by her desire to obtain the grant, and only to the extent of her grant-seeking conduct. Nothing within unconstitutional conditions doctrine provides the resources for permitting content-based conditions of producing good art, or contemporary art, or original art, while striking down a condition that only art supporting the President would be subsidized. [FN97] For such a distinction we must look elsewhere, and we thus see why the cases that a generation ago would have attracted unconstitutional conditions language are now commonly focused on the distinction between different forms of content regulation. That the principle of unconstitutional conditions is not so much as mentioned in any of the five different opinions in Forbes and Finley is less an oversight than an epitaph.

D. Viewpoint Discrimination

Neither Finley nor Forbes says anything about unconstitutional conditions, but both say a great deal about viewpoint discrimination, a principle that is becoming increasingly dominant in government enterprise cases. [FN98] Again, this is not surprising. Almost all government enterprise controversies involve content restrictions, even if in an undeniably benign form. And this is especially true with respect to those government enterprises that are intrinsically in the business of offering speech. Those government employees who work in or on museums, libraries, television and radio broadcasting, primary and secondary schools, colleges and universities, government-produced films, and state promotional publications are all in the business of offering or contracting for non-random speech content, and will thus necessarily *105 be making content-based decisions and imposing content-based restrictions in the very process of doing their jobs. [FN99]

Yet it is one of the important advances of (relatively) recent free speech doctrine that it has recognized and stressed the distinction among various forms of content regulation, in particular the distinction between subject matter discrimination and viewpoint discrimination. [FN100] However necessary it might be for a state university law school to condition continuing employment for its constitutional law faculty on using class time to discuss constitutional law and not the fortunes of the New York Yankees, or even torts for that matter, it seems objectionable to condition such employment on the possession or articulation of one rather than another viewpoint about the Constitution. If libertarian approaches to constitutional law were permissible but Marxist approaches were not, the viewpoint discrimination would appear to create problems not created by the subject matter discrimination standing alone.

Even this will not do, however. First, it is hardly clear that the line between viewpoint and other forms of content discrimination can be sustained, except possibly in extreme cases. Many subject matter restrictions (or standards of quality) will mask or reflect viewpoint distinctions, and this is especially likely to be true in literature, in academic inquiry, and in the arts. To support painting but not installations might not strike everyone as being about point of view, [FN101] but my strong suspicion is that many contemporary artists would disagree. [FN102]

*106 More significantly, undeniable and uncontroversial viewpoint discrimination is part and parcel of many state enterprises, [FN103] as Chief Justice Rehnquist's example of the National Endowment for Democracy in Rust v. Sullivan makes clear. [FN104] And even those who disagree with the outcome in Rust would hardly disagree with the example, or with the ability of a librarian to select books accepting that the Holocaust happened to the exclusion of books denying its occurrence. As with the other false paths, the path of viewpoint discrimination ultimately leads to the point at which we must acknowledge that some forms of viewpoint discrimination by government enterprises are permissible and some forms are not, with the bare idea of viewpoint discrimination of little assistance in separating the one from the other.

IV. Institutional Autonomy and the Legal Point of View

None of the foregoing is meant to suggest that the various doctrines I have glanced at are without value. Public forum doctrine is essential in staking out an important area of mandatory access, especially for those speakers who might otherwise be unable to secure some forum for their speech. The analysis of government speech reminds us both of the special power of the government's voice in the marketplace of ideas and of the implausibility of expecting that First Amendment doctrine and judicial action can check that power. The doctrine of unconstitutional conditions is necessary to prevent government provision of essential services from being used every bit as coercively as a criminal punishment, for instance by conditioning welfare benefits on a recipient's agreement to refrain from criticizing the government. Finally, a presumption against viewpoint discrimination enforces a basic principle of free speech theory by permitting opposing viewpoints to do battle on their own merits without strong state interference on one side or the other.

But although all of these doctrines have their functions, none appears to fit well with very many, if any, of the modern government enterprise controversies, and the fit with those government enterprises that are themselves in the content and viewpoint business is especially poor. In most of the modern government enterprise cases, these devices appear to be little more than conclusions masquerading as analytical tools. Yet it is in the nature of law to look to the past, and try to take from its storehouse of precedents, analogies, rules, doctrines, *107 and principles, all of them designed for yesterday's controversies, the instruments that will deal with today's cases and tomorrow's problems. That the fit is frequently a bad one should come as no surprise.

In this case, however, the bad fit occurs not only at the doctrinal level, but at one level higher as well. Even apart from the particular doctrines and principles selected, the very process of selecting doctrines and principles in First Amendment jurisprudence has traditionally embodied a kind of institutional agnosticism. It is a process--some might call it the meta- doctrine, and others might call it second-order doctrine--that presupposes the undesirability of having a rule, principle, or doctrine for one institution that is not applicable to another. A First Amendment with separate principles for the institutional press, schools, colleges and universities, libraries, museums, the performing arts, traditional public forums, elections, and the Internet would find itself at odds with central features of the American free speech tradition or at least, as the examples I mentioned at the outset of this Comment indicate, [FN105] the doctrinal and precedential embodiment of that tradition. It is no surprise, therefore, that none of the proposed solutions to the question of free speech in government-sponsored institutions is willing to embrace an institutionally specific approach. [FN106]

Thus, although it may appear that recognizing the special characteristics and functions of culturally differentiated institutions might provide some much-needed assistance in resolving government enterprise free speech cases, we should not be surprised that this recognition, prior to Forbes and Finley, was strikingly absent from the cases. Although the salience of institutionally specific factors is apparent in most such cases, the Court's refusal to make institution-specific decisions is supported not only by most of existing First Amendment doctrine, but also, and more importantly, by a battery of extraordinarily *108 well-entrenched views about the nature and function of law itself, views that are especially concentrated in the First Amendment context.

Consider first the law's preference for juridical categories, by which I mean categories that are created by the law and imposed on a world whose categories and divisions often pre-exist and cut across those of the law. As a starting point, recall Justice Holmes's anecdote about the justice of the peace and the churn in The Path of the Law: [FN107]

There is a story of a Vermont justice of the peace before whom a suit was brought by one farmer against another for breaking a churn. The justice took time to consider, and then said that he had looked through the statutes and could find nothing about churns, and gave judgment for the defendant. [FN108] This example serves many purposes for Holmes, [FN109] but the most relevant one here is to stress that only a non-legally trained bumpkin would imagine that "churn" was a relevant category for legal analysis. Holmes decried the fact that "[a]pplications of rudimentary rules of contract or tort are tucked away under the head of Railroads or Telegraphs . . . ." [FN110] For him, one who is a "master" of the law "look[s] straight through all the dramatic incidents" [FN111] and makes decisions according to proper legal categories. [FN112]

Holmes's ideas about the primacy of legal categories both reflected and influenced legal thought. As he himself believed, and as he recognized most legally trained people believed, law did, and should, divide the world into legal categories like contract, negligence, estoppel, strict liability, equity, and possession, and not into pre-legal or extra-legal categories like railroads, telegraphs, churns, and Cleveland. Only by identifying the proper legal category, Holmes appears to have thought, could the purposes of law be achieved and legal outcomes be correctly predicted. Any legal category that was specific to a non-legal institution, Holmes and countless others have believed, would demonstrate a lack of understanding of just what law is all about. [FN113]

*109 It is important to recognize that Holmes was not making a point about the virtues of abstraction, asserting a preference for generality over particularity, or suggesting the appropriate size of legal categories. Juridical categories are not necessarily larger than the categories of the world. There are, after all, far more insects than there are torts. Yet to Holmes and many others "torts" is a proper legal category and "insects" is not, for "insects" is a category of the world, but "torts" is a category of the law. As Holmes correctly identified, legal analysis tends [FN114] towards the latter, and those who aspire to Holmes's posthumous praise would strive for principles, rules, standards, maxims, and tests couched in the language of juridical categories.

Holmes was concerned not about the size of categories but about their character. Others, however, have seen generality as itself an essential feature of the legal point of view. Law's preference for generality and for categories simpliciter thus constitutes a second feature of the legal point of view, different from, but standing alongside, law's preference for juridical categories. Although recent scholarship [FN115] and occasional judicial statements [FN116] have questioned law's traditional preference for generality, that preference still pervades a great deal of legal consciousness: Lon Fuller saw the generality wrought by rules as an essential feature of the idea of law; [FN117] so-called "neutral" principles were an important feature of much of legal process thinking; [FN118] Edwin *110 Patterson praised generality as an "ethically neutral characteristic of law"; [FN119] Justice Scalia has identified the rule of law with the law of rules; [FN120] legislation about specific companies has been disfavored; [FN121] and the great bulk of what goes on in the typical first year law school Socratic interrogation is designed to induce law students to internalize a process by which they think in terms more general than the terms of the particular case under discussion. The very idea that there is more to law than reaching the right result in this case is the enduring lesson of the traditional form of legal education, and thus the traditional understanding of what it is to "think like a lawyer."

The preference for generality as part of the legal craft, or as a necessary condition for legality, is under some attack these days, [FN122] but continues to thrive within the precincts of the First Amendment, where particularity, usually under the name of "balancing," or even worse, "ad hoc balancing," still carries a bad odor. [FN123] This distaste is in part a legacy of the debates of the 1950s and 1960s, [FN124] in which "balancing," especially as championed by Justice Frankfurter, was associated with a tendency to take the substance of governmental justifications for restricting speech quite seriously and with a tendency to defer to the government's own determinations of the weight of those justifications. [FN125] As a consequence of this alignment of balancing with scantier free speech protection, the opponents of balancing, who often *111 rode under the banner of "absolutism," [FN126] came to be associated with a strong commitment to free speech, and the rule-based idea of absolutism came to be understood as a necessary condition for a robust commitment to freedom of speech. [FN127]

Even apart from these debates of an earlier era, official discretion to determine the value of speech content has long been understood to be incompatible with the principle of free speech itself, one of whose central themes is distrust of government. [FN128] That vagueness is a special First Amendment vice, [FN129] that overbreadth doctrine [FN130] and the idea of the chilling effect [FN131] are so central to First Amendment thinking, and that where-do-you-draw-the-line, who-decides, thin-edge-of-the-wedge, camel's-nose- is-in-the-tent, foot-in-the-door, slippery-slope rhetoric features so prominently in the First Amendment culture, [FN132] all exemplify an attitude that distrusts particularity and insists that rules that are simultaneously broad and concrete are the essential conditions of a strong First Amendment. [FN133] In light of this attitude, it is not surprising that "general principles" hold sway, and that a medium- or institution-specific First Amendment jurisprudence is thought inconsistent with the First Amendment itself. [FN134]

*112 Finally, and perhaps most importantly, law, and especially constitutional law, have long been understood to be matters of principle and not of policy. Although the distinction between principle and policy owes its contemporary jurisprudential defense to Ronald Dworkin, [FN135] the idea that Dworkin has identified and defended--that matters of policy are for legislatures and matters of principle are for the courts--resonates with numerous long-standing features of judicial methodology. Courts are uncomfortable with controversies in which the number of interests is greater than two; they are uncomfortable with splitting the difference; they are uncomfortable with quantitative assessments of the optimal solution; and they are especially uncomfortable with obtaining the kind of empirical information that is necessary to policy analysis. [FN136] Although Dworkin undoubtedly exaggerates the descriptive aspect of his claim that courts do principle and not policy, [FN137] the basic idea that courts should look for principle-based solutions to the problems they confront is ingrained in the legal, and especially the judicial, mindset. [FN138]

*113 These three features of the legal point of view--the preference for large categories, the preference that those categories be juridical, and the preference for principle over policy--all stand together, and are together arrayed against the intrinsically particular, non-juridical, and policy-based determination that the rules that apply to one social institution ought not necessarily apply to another. In the First Amendment context, the anathema [FN139] of having one rule for schools, another for the arts, still another for libraries, and so on, reflects the central features of the First Amendment tradition, but at a deeper level it also reflects how judges see their job, and how many of the rest of us see the idea of law. In light of this attitude, it should come as no surprise that institution-specific solutions to First Amendment problems, even if they otherwise have much to recommend them, come to the judiciary under a cloud, having to confront a strong presumption against designing a principle for one institution that could not be applied in essentially the same terms to another.

From this perspective, the institutional specificity that seemed pervasive and controlling in Justice Kennedy's opinion in Forbes, and that seemed at least important in Justice O'Connor's opinion in Finley, challenges much of conventional First Amendment thinking, and much of the legal mindset as well. The history of failed attempts to deal with the issue of free speech in government enterprises reveals the attraction that institutional specificity may have had to Justices Kennedy and O'Connor. And the history of First Amendment thought, and even more the history of legal thought, show us at the same time why institutional specificity, for all of its attractions, was at the end of the day subjugated in both Forbes and Finley.

V. Free Speech and Institutional Autonomy

The fact that the American legal and free speech traditions hold such a deep-seated and widespread antipathy toward institutional specificity, however, should not lead us to ignore the empirical reality of that institutional specificity. [FN140] Institutional specificity and institutional *114 differentiation are a reality of modern life, and this reality is reflected, as elsewhere, in the institutions relevant to free speech adjudication. For example, as the use of NEA arts panels makes clear, in the area of commissioning, selecting, and subsidizing art, governments tend not to use their own full-time employees, but instead rely on an assortment of arrangements--panels, juries, competitions, and the like--such that the government's content-based, and often viewpoint-based, decisions are made by those whose primary allegiances may remain in the private or non-profit sector.

By contrast, public libraries, school libraries, and university libraries, for example, all make their necessarily content-based and frequently viewpoint- based acquisition decisions by use of their own employees. [FN141] Indeed, when there is a suggestion that non-governmental officials should make these decisions, we occasionally hear the argument that the privatization decision itself has free speech implications. [FN142] In the case of libraries, but not in the case of arts funding, the expertise and the professionalism lie within the government, or at least lie within the domain of state employment.

This contrast between the decisions about art funding and art commissioning, on the one hand, and library book selection (and occasional de-selection as well), on the other, suggests that it would be a mistake to draw a distinction on the basis of whether the decision is made by a government employee, for the two institutions draw on professional expertise in quite different ways. Moreover, it would be a mistake as well to draw an abstract distinction between selection and de-selection, because de-selection is a fact of life in the world of libraries, but hardly a necessary process in the world of arts funding.

*115 Numerous other differences between the two domains exist as well, but surveying them here would serve little purpose. My only point is that these two processes employ professionals in different ways, and involve different kinds of decisions. A set of library-specific principles, for example, might focus on the nature of the content-based and viewpoint-based decisions typically and historically made in the selection and de-selection process. These library-specific principles might conclude that this process, when made by a certain cadre of professionals, is constitutionally permissible, but might conclude as well that external influence in this process by non-professionals raises First Amendment problems. [FN143] Consideration of the role of library professionals in the selection and de-selection processes might explain the source of the First Amendment worry that the Court noted, but left unresolved, in Board of Education v. Pico, [FN144] and might also help the Court develop a set of principles for an area that has generated a great deal of litigation in the lower courts [FN145] but virtually no Supreme Court guidance.

Similar principles of professional deference might also be developed in the case of arts funding, although the content of those principles might be different because of the difference in the nature of the institutions. Because art, under at least one view, may be defined in terms of the institutions that create, support, and evaluate it, [FN146] and because artistic quality may consequently exist as an institution-dependent idea, [FN147] the constitutional permissibility of content and viewpoint control in arts funding, as well as in related areas such as the selection and de-selection of public art, might depend on the involvement of the institutions of art in the decision under attack, in much the same way that the involvement of the institutions of journalism appears to have been the saving factor in Forbes. In Finley, therefore, the use of peer panels of artists might have created a presumption in favor of even viewpoint discrimination by those panels or other decisionmaking structures similarly connected to the world of the arts. However, the *116 external (to the institutions of the arts) and political source of the particular standard under attack might have been sufficient to overcome the presumption. [FN148]

The key point here is that such an approach would be plainly implausible for all content-based government enterprises or even for all content-based government subsidies, as the National Endowment for Democracy example makes clear. But even if this approach of respecting institutional autonomy would not work for all government enterprise or government subsidy cases, it might very well work for some. If so, government subsidy cases involving institutions that have a certain First Amendment aura--the arts, libraries, universities, and the institutional press, for example--would serve as leading candidates. But to make this decision would require giving certain institutions special First Amendment status, [FN149] and to make that decision would require the Court to inquire much more deeply into the specific character of the institution, and the functions it serves, than it has been willing to do in any of its past confrontations with these issues. [FN150]

When the state agency in a case such as Forbes reaches beyond the state and uses an Associated Press official to make its decisions, we are not surprised by judicial unwillingness to see this as a stereotypical "government censorship" case, because the journalist/non-journalist distinction carries far more cultural salience, to the Court as well as to the public, than does the public/private distinction. Some of the icons of state action doctrine would simply have said that an Associated Press employee becomes the state in these circumstances, [FN151] but Justice *117 Kennedy's emphasis on public broadcasting as broadcasting rather than public resonates with a considerable amount of empirical reality, just as it grates on an equally considerable amount of existing doctrine.

The empirical reality of institutional specificity manifests itself in other places as well. For example, the school teacher and the university professor are both involved in education, but likely have very different training, very different understandings of their roles, and travel in essentially non-overlapping professional circles. Moreover, the institutions themselves serve functions that have more differences than similarities. Although not all, and not even most, educational institutions are research universities, a wide sociological, cultural, and functional gulf exists between the primary or secondary school and the research university. The Court's suggestion in Hazelwood School District v. Kuhlmeier [FN152] that its holding allowing school control over the content of a high school newspaper might not apply at the university level [FN153] serves as an acknowledgment of an institutional difference, but simultaneously reinforces the idea that the presumption would be against drawing a doctrinal distinction.

The teachers and school administrators in cases like Hazelwood and Bethel School District No. 403 v. Fraser [FN154] were admittedly state employees. It seems strange, however, to suppose that these state employees think of their role in the same way as do all other state employees, or that for most state employees the fact of state employment is more salient than what they are and what they do for the state. Just as the state journalists in Forbes seemed to the Court more like journalists than state employees, so might it seem to a court that state teachers were teachers first and state employees second, and so too for state librarians, state curators, and many others whose state employment may be less relevant to what they do on the job than the professional identification they bring to that job.

This analysis, of course, does not answer very many of the difficult questions that arise in the context of free speech within government enterprises. It does suggest, however, that the quick equation of these professionals making content-based and even viewpoint-based distinctions *118 to political bosses punishing those who criticize them [FN155] is somewhat too quick. If First Amendment doctrine were subdivided along institutional lines, [FN156] it is possible that the doctrine would be better poised not only to capture important institutional differences, but also to recognize the potentially distinct First Amendment status that the arts, universities, libraries, and journalism, and possibly other institutions such as elections, possess. And if, as in both Forbes and Finley, institutionally specific factors are already doing more of the work than the Court is willing to admit, then increased transparency in acknowledging the role that policy considerations of institutional autonomy play may enable the Court to evaluate those policy considerations not only with better analysis, but with better information as well. Finally, the cultural institutions antecedent to First Amendment doctrine have a salience and visibility to officials and to the public to which the niceties of legal principle can never even hope to aspire. As long as a prime goal of constitutionalism is the goal of affecting the behavior of officials, then doctrine couched in the categories of cultural rather than legal understanding is more likely suited to this task. This function has often been thought to be especially important to the First Amendment, whose values are so counter-intuitive to so many people, [FN157] and thus the greatest advantage of judicial openness about institutional specificity may be an increased likelihood of official compliance with, and public assimilation of, the mandates of the First Amendment.

VI. Conclusion

Holmes was wrong. Whatever may be the prescriptive merits of designing law around juridical categories, the path of the law has been otherwise. It has, in many important respects, followed Karl Llewellyn, *119 who persistently urged that law should be designed around the categories of the pre-legal world and not around the categories of legal abstraction. [FN158] Although Holmes scorned those who thought that law might have separate principles for railroads or telegraphs, a century later we see a legal environment that has separate statutes and, increasingly, separate common law principles for sports, for health, for consumers, for insurance contracts, for securities, and for a host of other non-juridical categories.

Against this background, it should only have been expected that the same path of the law that led from contract law simpliciter to (somewhat) separate bodies of contract law for consumer contracts, for insurance contracts, for securities contracts, and for sports contracts might lead from free speech law simpliciter to (somewhat) separate bodies of free speech law for libraries, for schools, for elections, for broadcasting, for the fine arts, and so on. Justice O'Connor peered down this path in Finley, and Justice Kennedy peered down it even further in Forbes. By focusing on the specific empirical characteristics and affiliations of the institutions involved in these two cases, the Court at the very least suggested its receptiveness to arguments from institutional specificity in future First Amendment cases, and suggested as well its receptiveness to the kind of empirical institutional evidence that would support the creation of institution-specific First Amendment principles.

Yet these were, in the final analysis, only suggestions from a Court which in Finley avoided the hardest questions by drawing on implausible claims about the constitutional difference between a factor and a categorical rule and which in Forbes refused to relinquish its grip on an unhelpful public forum doctrine that provided no assistance in the case at hand. The Court peered down the path of institutional specificity in both cases, but when all was said and done it refused to do much more than peer.

This was to be expected. Not only does a strong and often justified First Amendment ideology resist this kind of institutional subdivision, but central features of an even more firmly entrenched legal and judicial ideology work in the same direction. As long as courts think it suspect to depart from juridical categories, to decide according to narrow rather than broad categories, and to make decisions of policy rather than of principle, institutionally specific court-made rules and principles will always remain the exception rather than the rule.

For many areas of law, and for most areas of the First Amendment, these presumptions against institutional specificity may do more good *120 than harm, even though, like any rule-based approach, [FN159] they achieve important systemic goals at the expense of tolerating some number of sub-optimal results in particular cases. But when the issue is how to apply free speech ideas to government enterprises that are themselves in the business of supplying speech because of its content, numerous attempts to resolve the issues by the use of institutionally agnostic general principles have all come up empty. In Forbes and Finley the Court hinted that it might consider abandoning its history of institutional agnosticism in content-oriented government enterprise cases in order to move beyond the failed attempts of the past. But whether the Court will build on this hint will depend on whether it is willing to challenge centuries of legal theory and a generation of free speech theory. If the Court becomes increasingly willing to make institution-specific decisions, [FN160] and increasingly willing to look at the kind of information that will help it to make policy decisions, [FN161] there is a chance that the hints of an institutionally focused approach to government enterprise free speech cases will in the future become more than hints. But for now it is only a chance.


1a.   Frank Stanton Professor of the First Amendment and Academic Dean, John F. Kennedy School of Government, Harvard University. Research for this Comment was supported by the Joan Shorenstein Center on the Press, Politics and Public Policy. Larry Alexander, Michael Dorf, Richard Fallon, and Richard Pildes provided comments, the promptness of which was exceeded only by their helpfulness. I am also pleased to acknowledge the congenial facilities of the Research School of Social Sciences at the Australian National University, where this Comment was written and where many of its ideas were fruitfully discussed.

  1. See Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555, 575-78 (1980) (rejecting special access rights to criminal trials for the press); Gannett Co. v. DePasquale, 443 U.S. 368, 391-93 (1979) (same); Zurcher v. Stanford Daily, 436 U.S. 547, 565 (1978) (rejecting special immunity for the press against search warrants); Pell v. Procunier, 417 U.S. 817, 833-35 (1974) (rejecting special access rights to prisons for the press); Saxbe v. Washington Post Co., 417 U.S. 843, 850 (1974) (same); Branzburg v. Hayes, 408 U.S. 665, 691-92 (1972) (rejecting the argument that the Press Clause generates a distinct reporters' privilege to refuse to disclose sources). The most extensive discussion of the Court's rejection, in the above cases and others, of the claim that the Press Clause justifies treating the press as a separate institution with separate rights is in Chief Justice Burger's concurring opinion in First National Bank of Boston v. Bellotti, 435 U.S. 765, 798-802 (1978). See also David A. Anderson, The Origins of the Press Clause, 30 UCLA L. Rev. 455, 533 (1983) (arguing generally for special protection for the press); Randall P. Bezanson, The New Free Press Guarantee, 63 Va. L. Rev. 731, 731-36 (1977) (same); David Lange, The Speech and Press Clauses, 23 UCLA L. Rev. 77, 118-19 (1975) (opposing special protection).

  2. The statement in the text draws on the Court's willingness to apply the standards of New York Times Co. v. Sullivan, 376 U.S. 254, 286 (1964), to non-media slander. Courts have applied New York Times v. Sullivan to non- media defendants in a number of subsequent cases. See Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 753 (1985) (rejecting the distinction made by the Vermont Supreme Court between the institutional media and other communicators); Gertz v. Robert Welch, Inc., 418 U.S. 323, 347 (1974) (requiring a negligence standard for all defamation about public issues); Garrison v. Louisiana, 379 U.S. 64, 76 (1964) (applying New York Times v. Sullivan to oral statements); Davis v. Schuchat, 510 F.2d 731, 734 n.3 (D.C. Cir. 1975) (affirming the application of the New York Times standard in an action for slander); Sas Jaworsky v. Padfield, 211 So. 2d 122, 126 (La. Ct. App. 1968) (applying the New York Times standard to oral face-to-face allegations of communist affiliation); Fox v. Kahn, 221 A.2d 181, 183-84 (Pa. 1966) (applying the New York Times standard to statements made by an individual at a political rally). See generally Joel D. Eaton, The American Law of Defamation Through Gertz v. Robert Welch, Inc. and Beyond: An Analytical Primer, 61 Va. L. Rev. 1349, 1406-07 (1975) (noting that the actual malice rule is applicable to oral statements by non-media defendants); Steven Shiffrin, Defamatory Non-Media Speech and First Amendment Methodology, 25 UCLA L. Rev. 915, 931-35 (1978) (arguing against a distinction between media and non-media defendants).

  3. See University of Pa. v. EEOC, 493 U.S. 182, 199 (1990) (rejecting an academic freedom claim of secrecy for records of tenure and promotion). In almost all of the existing cases using the words "academic freedom," the free speech rights recognized have been no different from the rights that would likely have been recognized for speakers unaffiliated with educational institutions. See, e.g., Keyishian v. Board of Regents, 385 U.S. 589, 603 (1967); Sweezy v. New Hampshire, 354 U.S. 234, 250-51 (1957); Adler v. Board of Educ., 342 U.S. 485, 511 (1952) (Douglas, J., dissenting). The best discussion of the distinction between academic freedom as an instantiation of a general free speech right and academic freedom as a distinct right comes in William W. Van Alstyne, The Specific Theory of Academic Freedom and the General Issue of Civil Liberties, 404 Annals Am. Acad. Pol. & Soc. Sci. 140, 142-43 (1972).

  4. See Reno v. ACLU, 117 S. Ct. 2329, 2334 (1997) (rejecting indecency regulation for computer communications); Denver Area Educ. Telecomms. Consortium v. FCC, 518 U.S. 727, 733 (1996) (same, with qualifications, for cable television); Sable Communications, Inc. v. FCC, 492 U.S. 115, 126-28 (1989) (same for telephonic communications); cf. Erznoznik v. City of Jacksonville, 422 U.S. 205, 210-12 (1975) (same for screens of outdoor theaters, although not with specific "indecency" language). The only exceptions are for broadcasting, see FCC v. Pacifica Found., 438 U.S. 726, 748-51 (1978), and non-prohibitory zoning regulation, see Young v. American Mini Theatres, Inc., 427 U.S. 50, 62-63 (1976). Sable, Reno, Denver Area, and Erznoznik follow the earlier rejection of the view that motion pictures were to be treated differently from print media for First Amendment purposes. See Joseph Burstyn, Inc. v. Wilson, 343 U.S. 495, 501-02 (1952), overruling Mutual Film Corp. v. Industrial Comm'n, 236 U.S. 230 (1915).

  5. Noteworthy examples are the distinctions between Nazis and other people, see National Socialist Party v. Village of Skokie, 432 U.S. 43, 44 (1977) (per curiam); Collin v. Smith, 578 F.2d 1197, 1207 (7th Cir. 1978), stay denied, 436 U.S. 953 (1978), and cert. denied, 439 U.S. 916 (1978) (refusing to prohibit a march by a self-described Nazi Party in Skokie, Illinois), between advocating racial tolerance and opposing it, see R.A.V. v. City of St. Paul, 505 U.S. 377, 393-96 (1992), and between endorsing sexual violence and condemning it, see American Booksellers Ass'n, Inc. v. Hudnut, 771 F.2d 323, 325 (7th Cir. 1985), aff'd, 475 U.S. 1001 (1986). Much of my own writing about the First Amendment has stressed the centrality to the Amendment of refusing to draw otherwise politically or morally compelled distinctions. See, e.g., Frederick Schauer, Categories and the First Amendment: A Play in Three Acts, 34 Vand. L. Rev. 265, 282-96 (1981) [hereinafter Schauer, Categories]; Frederick Schauer, Codifying the First Amendment: New York v. Ferber, 1982 Sup. Ct. Rev. 285, 286-87; Frederick Schauer, Harry Kalven and the Perils of Particularism, 56 U. Chi. L. Rev. 397, 405-09 (1989) (book review); Frederick Schauer, Justice Stevens and the Size of Constitutional Decisions, 27 Rutgers L.J. 543, 551-52 (1996); Frederick Schauer, The Second-Best First Amendment, 31 Wm. & Mary L. Rev. 1, 2-3 (1989).

  6. This pressure often arises in the context of new technologies, as seen in the discussions in Denver Area Educational Telecommunications Consortium, 518 U.S. at 732-36, and Red Lion Broadcasting Co. v. FCC, 395 U.S. 367, 386 (1969). See also Ithiel de Sola Pool, Technologies of Freedom 189-225 (1983); Thomas G. Krattenmaker & Lucas A. Powe, Jr., Regulating Broadcast Programming 203-36 (1994); Christopher M. Kelly, "The Spectre of a "Wired' Nation": Denver Area Educational Telecommunications Consortium v. FCC and First Amendment Analysis in Cyberspace, 10 Harv. J.L. & Tech. 559, 643-46 (1997); Glen O. Robinson, The Electronic First Amendment: An Essay for the New Age, 47 Duke L.J. 899, 908-12 (1998). However, it is important to recognize that the issue of institutional differentiation is different from that of media differentiation. A single cultural institution may use multiple media, and a single type of medium may find itself used by numerous cultural institutions. More broadly, there is clearly some form of categorization within First Amendment doctrine. See Richard H. Fallon, Jr., Sexual Harassment, Content Neutrality, and the First Amendment Dog that Didn't Bark, 1994 Sup. Ct. Rev. 1, 23-28. But, as I shall stress throughout this Comment, the existing categorical structure of the First Amendment insists on drawing lines based on abstract legal principles and not on the pre-legal contours of the world.

  7. 118 S. Ct. 1633 (1998).

  8. 118 S. Ct. 2168 (1998).

  9. A diversity of off-center voices is characteristic of public access television, as it is of that historically prototypical public forum, Speaker's Corner in Hyde Park, where in the late twentieth century the speakers tend to come from society's fringes and the listeners tend to be foreign tourists taking pictures.

  10. Virtually identical issues arise in the context of press access under conditions in which access opportunities are scarce and must accordingly be allocated. When The Nation, The Progressive, Mother Jones, The Village Voice, and other non-centrist publications were excluded from the 186-member Gulf War press pool, see Frederick Schauer, Parsing the Pentagon Papers, Research Paper R-3 (Joan Shorenstein Barone Ctr., John F. Kennedy Sch. Gov't, Harv. Univ.), May 1991, at 6, on the grounds that scarce space had to be reserved for publications with larger circulations (such as The New York Times, US News and World Report, and CBS News), they sued, claiming viewpoint discrimination against non-mainstream publications. The defense, in which the Defense Department and the mainstream publications intriguingly shared a common interest, was framed in terms of maximizing the news and information that could be delivered to the largest audience. The cessation of hostilities foreclosed a decision on the merits. See Nation Magazine v. United States Dep't of Defense, 762 F. Supp. 1558, 1569-75 (S.D.N.Y. 1991). On the "reluctance" of the mainstream pool members to join the litigation, see Marcia Chambers, Cutting Off War Access to the Press, Nat'l L.J., Feb. 25, 1991, at 13.

  11. Forbes, 118 S. Ct. at 1637.

  12. Although not directly related to my larger themes in this Comment, it is worth noting that a "possibility of winning" standard slights numerous ways in which third parties may influence political debate. See Samuel Issacharoff & Richard H. Pildes, Politics as Markets: Partisan Lockups of the Democratic Process, 50 Stan. L. Rev. 643, 681-90 (1998). If we assume that pre-election debates have a certain visibility not possessed by non-electoral policy and political discussion, then limiting participation in the most salient of electoral discourse has the effect of narrowing political discourse generally, a consequence seemingly at odds with the larger goals of a free speech system. The same tension arises, of course, with respect to the two- party system generally, as is apparent from the clash of opinion in Timmons v. Twin Cities Area New Party, 117 S. Ct. 1364 (1997), which upheld Minnesota regulations banning "fusion" candidates.

  13. See Forbes v. Arkansas Educ. Television Communication Network Found., 22 F.3d 1423 (8th Cir. 1994).

  14. See Forbes v. Arkansas Educ. Television Comm'n, 93 F.3d 497 (8th Cir. 1996).

  15. See id. at 500.

  16. See infra pp. 97-100.

  17. When a public forum is found to exist, that fact is sufficient to support a claim of access to that forum as a matter of constitutional law. See Harry Kalven, Jr., The Concept of the Public Forum: Cox v. Louisiana, 1965 Sup. Ct. Rev. 1, 29-30. This access by direct operation of the Constitution is usefully distinguished from instances in which state or federal statutory or common law creates the right of access, subject to potential countervailing constitutional claims by the property owner. See Pacific Gas & Elec. Co. v. Public Util. Comm'n, 475 U.S. 1, 20-21 (1986) (striking down an attempt by a state utilities commission to compel a utility to distribute materials of a private advocacy group in the utility's mailings); Pruneyard Shopping Ctr. v. Robins, 447 U.S. 74, 80-81 (1980) (upholding state-mandated access against a property owner's free speech claims).

  18. See Forbes, 118 S. Ct. at 1641-43.

  19. See id. at 1640-41.

  20. See id. at 1640-42.

  21. See id. at 1644.

  22. Id. at 1637.

  23. Id.

  24. Id.

  25. See id. at 1639-40.

  26. See id. at 1637.

  27. Id. at 1638-39.

  28. Id. at 1637, 1639-40, 1644.

  29. Id. at 1639.

  30. Id.

  31. See id. at 1644 (Stevens, J., dissenting). Justice Stevens's dissent was joined by Justices Souter and Ginsburg.

  32. See International Soc'y for Krishna Consciousness, Inc. v. Lee, 505 U.S. 672, 679, 685 (1992) (sustaining a ban on the distribution of literature in an airport); Board of Airport Comm'rs v. Jews for Jesus, Inc., 482 U.S. 569, 577 (1987) (striking down a statute restricting any First Amendment activities in an airport).

  33. See United States v. Kokinda, 497 U.S. 720, 736-37 (1989) (plurality opinion) (allowing manner-based but not viewpoint-based restrictions for nonpublic forums).

  34. As an aside, it is ironic that the abstraction from "state journalism" to "state" comes from Justice Stevens, because it is Justice Stevens who has been at the forefront of First Amendment particularism, objecting with some frequency to obsessions with categories and to the failure of First Amendment doctrine to draw on, rather than to elide, various socially salient distinctions. See The Honorable John Paul Stevens, The Freedom of Speech, 102 Yale L.J. 1293, 1301-04 (1993). For my own commentary on this aspect of Justice Stevens's First Amendment methodology, see Schauer, Harry Kalven and the Perils of Particularism, cited above in note 5, at 398-99 n.3, and Schauer, Justice Stevens and the Size of Constitutional Decisions, cited above in note 5, at 544-52.

  35. Cf. International Soc'y for Krishna Consciousness, 505 U.S. at 680-83 (rejecting the argument that an airport terminal was either a traditional public forum or a public forum by designation).

  36. Cf. Lehman v. City of Shaker Heights, 418 U.S. 298, 304 (1974) (upholding a prohibition on political advertising in city buses).

  37. See International Soc'y for Krishna Consciousness, 505 U.S. at 680- 83; Lee v. International Soc'y for Krishna Consciousness, Inc., 505 U.S. 830 (1992) (per curiam).

  38. See Cornelius v. NAACP Legal Defense & Educ. Fund, Inc., 473 U.S. 788, 806 (1985) (holding viewpoint discrimination to be impermissible even in nonpublic forums).

  39. The Court's quick dismissal of the claim that a viability standard is not viewpoint-neutral is troubling. Although it is true that a viability standard leaves open the possibility that any viewpoint may be found viable by the decisionmaker, such as the public broadcaster in Forbes, and although in some contexts viability may be a function of factors other than point of view, a viability standard still might be better seen as an across-the-board prohibition on unpopular viewpoints, or in this context, on people holding unpopular viewpoints. Given that a central part of the free speech tradition is about protecting challenges to received opinion, concluding that a preference for received opinions over other opinions (which is essentially what a viability standard would do) is not a form of viewpoint discrimination seems surprising. See John Stuart Mill, On Liberty 6 (David Spitz ed., W.W. Norton & Co. 1975) (1859); 2 Karl Popper, The Open Society and Its Enemies, at 267-68 (5th ed. 1966); Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting) ("[T]ime has upset many fighting faiths ...."); see also supra p. 85 and note 5. And if we view the First Amendment more broadly as a mechanism for facilitating change of viewpoint, see Vincent Blasi, Reading Holmes Through the Lens of Schauer: The Abrams Dissent, 72 Notre Dame L. Rev. 1343, 1345-47 (1997), concluding that a viability standard is viewpoint-neutral is even more surprising.

  40. For example, most of the Court's references to academic freedom, as in Keyishian v. Board of Regents, 385 U.S. 589, 603 (1967), and Sweezy v. New Hampshire, 354 U.S. 234, 250 (1957), arise in contexts in which the Court was equally willing to grant the same protection to nonacademic state employees. See, e.g., Gardner v. Broderick, 392 U.S. 273, 277-79 (1968); Garrity v. New Jersey, 385 U.S. 493, 500 (1967). The same phenomenon appears in Miami Herald Publishing Co. v. Tornillo, 418 U.S. 241, 258 (1974), in which the Court says lots of nice things about newspaper editors, but plainly would have reached the same result in a totally nonjournalistic context.

  41. See Forbes, 118 S. Ct. at 1640-41.

  42. This identification of a special status for candidate debates suggests the possibility that the electoral sphere, as such, could be treated as a distinct domain for First Amendment purposes. On this possibility, see C. Edwin Baker, Campaign Expenditures and Free Speech, 33 Harv. C.R.-C.L. L. Rev. 1, 24-28 (1998).

  43. 118 S. Ct. 2168 (1998).

  44. In this legislation, Congress required that the Chairperson of the National Endowment for the Arts ensure that "artistic excellence and artistic merit are the criteria by which applications are judged, taking into consideration general standards of decency and respect for the diverse beliefs and values of the American public." 20 U.S.C. 954(d)(1) (1994).

  45. See Finley v. National Endowment for the Arts, 795 F. Supp. 1457, 1476 (C.D. Cal. 1992).

  46. See Finley v. National Endowment for the Arts, 100 F.3d 671, 674 (9th Cir. 1996), reh'g en banc denied, 112 F.3d 1015 (9th Cir. 1997).

  47. I recognize that "own" is tendentious in this context, since a wave of attacks on the public/private distinction in the free speech context would challenge the distinction I draw in the text on the basis of the inability to identify a speaker's "own" resources independent of the state action that makes those resources hers in the first instance. See, e.g., Cass R. Sunstein, Democracy and the Problem of Free Speech 36-38 (1993); Owen Fiss, State Activism and State Censorship, 100 Yale L.J. 2087, 2101 (1991). But as long as the distinction between state action and private action continues to make a constitutional difference, as in Hurley v. Irish-American Gay, Lesbian and Bisexual Group, 515 U.S. 557, 566 (1995), and Edmonson v. Leesville Concrete Co., 500 U.S. 614, 619 (1991), it would be silly to ignore that distinction when thinking within the broad framework of existing doctrine.

  48. This characterization fits all of the icons of the modern American free speech tradition, including New York Times Co. v. United States, 403 U.S. 713 (1971) (per curiam), Cohen v. California, 403 U.S. 15 (1971), Brandenburg v. Ohio, 395 U.S. 444 (1969) (per curiam), and New York Times Co. v. Sullivan, 376 U.S. 254 (1964).

  49. 20 U.S.C. 945(d)(1), quoted in Finley, 118 S. Ct. at 2173.

  50. The voluminous literature on the topic includes Marjorie Heins, Sex, Sin, and Blasphemy: A Guide to America's Censorship Wars (1993), Alice Goldfarb Marquis, Art Lessons: Learning from the Rise and Fall of Public Arts Funding (1995), John Wesley Ziegler, Arts in Crisis: The National Endowment for the Arts Versus America (1994), Amy Adler, What's Left?: Hate Speech, Pornography, and the Problem for Artistic Expression, 84 Cal. L. Rev. 1499 (1996), John H. Garvey, Black and White Images, Law & Contemp. Probs., Autumn 1993, at 189, Thomas P. Leff, The Arts: A Traditional Sphere of Free Expression? First Amendment Implications of Government Funding to the Arts in the Aftermath of Rust v. Sullivan, 45 Am. U. L. Rev. 353 (1995), Note, Standards for Federal Funding of the Arts: Free Expression and Political Control, 103 Harv. L. Rev. 1969 (1990), and Pamela Weinstock, Note, The National Endowment for the Arts Funding Controversy and the Miller Test: A Plea for the Reunification of Art and Society, 72 B.U. L. Rev. 803 (1992). See also Bella Lewitzky Dance Found. v. Frohnmayer, 754 F. Supp. 774, 781-85 (C.D. Cal. 1991) (invalidating requirement that NEA grant recipients pledge not to use grant to produce obscene material).

  51. Piss Christ was a photograph of a crucifix submerged in a jar of urine.

  52. It is worth noting that the political alignments in the NEA dispute might under quite imaginable circumstances have been different, might still be different in the future, and in analogous circumstances are quite different now. Had the NEA awarded federal funds to Damien Hirst, whose art features, inter alia, sliced cows in formaldehyde, see Hirst Protesters Fined, Independent (London), Oct. 14, 1997, at 10, or Hermann Nitsch, whose performance art involves the slaughter of animals in front of a live paying audience, see Kate Connolly, You'll Never Eat Again, Guardian (London), Aug. 6, 1998, at 12, it is quite likely that there would have been protests, albeit from a different political quarter. See also Lois Walker, Gale Teaster & Gloria Kelley, Employee Rights Versus the First Amendment: Is It Censorship for a Library to Remove an Art Exhibit Containing a Dead Kitten?, Am. Libr., May 1998, at 60.

  53. See 20 U.S.C. 954(d)(2).

  54. The Court made it clear that a facial challenge, like the one in this case, creates an especially heavy burden for the challenger. See Finley, 118 S. Ct. at 2175.

  55. See id. at 2180-81 (Scalia, J., concurring in the judgment). Justice Scalia's opinion was joined by Justice Thomas.

  56. See id. at 2189-90 (Souter, J., dissenting).

  57. Justice O'Connor wrote:

[T]he text of 954(d)(1) imposes no categorical requirement ....

.... That 954(d)(1) admonishes the NEA merely to take "decency and respect" into consideration, and that the legislation was aimed at reforming procedures rather than precluding speech, undercut respondents' argument that the provision inevitably will be utilized as a tool for invidious viewpoint discrimination.

Finley, 118 S. Ct. at 2176 (emphasis added).

  1. 491 U.S. 397, 399 (1989) (striking down a Texas statute prohibiting flag desecration because of its viewpoint non-neutrality); see also United States v. Eichman, 496 U.S. 310, 312 (1990) (striking down the federal flag desecration statute).

  2. 485 U.S. 46, 57 (1988) (reversing a damage award for infliction of emotional distress in the parody of a public figure).

  3. The impermissibility of viewpoint discrimination in this context is clear from International Society for Krishna Consciousness, Inc. v. Lee, 505 U.S. 672, 683-85 (1992), and the argument that a decency restriction is a restriction on style and not on viewpoint does not survive Reno v. ACLU, 117 S. Ct. 2329, 2344-46 (1997), or Sable Communications, Inc. v. FCC, 492 U.S. 115, 126-31 (1989), even though such a distinction was earlier suggested by Justice Stevens in Bolger v. Youngs Drug Products Corp., 463 U.S. 60, 83-84 (1983) (Stevens, J., concurring in the judgment).

  4. At the very least, one might expect that the presence of an impermissible factor, even if not dispositive, would serve to change the burden of proof, as when the existence of an impermissible factor in an administrative determination places the burden on the agency to establish that it would have reached the same result in the absence of that factor. See Mt. Healthy Bd. of Educ. v. Doyle, 429 U.S. 274, 285-87 (1977).

  5. See Finley, 118 S. Ct. at 2177-79.

  6. It is not clear whether the "respect for the diverse beliefs and values" standard is parallel to the "decency" standard. Under one interpretation, and one supported by the ejusdem generis canon of statutory interpretation, both apply to individual works, such that a lack of decency or a lack of respect for diversity are negative factors in the funding determination in particular cases. Under another interpretation, however, the "respect for the diverse beliefs" standard applies only to the full array of support, exhorting the Endowment to pursue diversity across all of its grants. Under the latter interpretation, the potential First Amendment problems of the "respect for the diverse beliefs" standard are much less problematic than those of the "decency" standard.

  7. The leading case is Rust v. Sullivan, 500 U.S. 173, 192-95 (1991), in which the Court conceptualized the so-called "gag rule" (prohibiting recipients of federal family-planning funding from offering counseling on the use of abortion as a method of family planning) on federally funded health care clinics as a determination by the government about what it chose to say and what it chose not to say.

  8. See, e.g., Rosenberger v. Rector of the Univ. of Va., 515 U.S. 819, 828-32 (1995); FCC v. League of Women Voters, 468 U.S. 364, 381-84 (1984).

  9. See Finley, 118 S. Ct. at 2177 ("[C]ontent-based considerations ... are a consequence of the nature of arts funding.").

  10. Transcript of Oral Argument, Finley, 118 S. Ct. 2168 (1998) (No. 97- 371), available in 1998 WL 156955, at *27 (Mar. 31, 1998). This is noted in Justice Souter's dissent. See Finley, 118 S. Ct. at 2192-93 (Souter, J., dissenting). For a similar suggestion, see Amy Sabrin, Thinking About Content: Can It Play An Appropriate Role in Government Funding of the Arts?, 102 Yale L.J. 1209 (1993).

  11. See Finley, 118 S. Ct. at 2185 (Souter, J., dissenting).

  12. Finley, 118 S. Ct. at 2172, 2173, 2177, 2178.

  13. The institutional factor, however, may justify the outcome in Forbes more than it justifies the outcome in Finley. If we accept that both the viability restriction in Forbes and the decency restriction in Finley are viewpoint-based, then we might justify Forbes in terms of deference to the viewpoint distinctions drawn within the institution of journalism, just as we might justify many of the viewpoint distinctions that the arts panels drew in recommending NEA funding. But when a viewpoint distinction is imposed on the arts panels by Congress, the same rationale of institutional autonomy that might uphold viewpoint discrimination by those panels themselves would be less tolerant of similar discrimination when imposed by non-institutional actors, as in Finley. See infra pp. 115-16.

  14. The origins of mandatory access lie in Hague v. CIO, 307 U.S. 496 (1939), in which Justice Roberts offered the view that the streets and parks, regardless of formal title, "have immemorially been held in trust for the use of the public ... for purposes of assembly, communicating thoughts between citizens, and discussing public questions." Id. at 515. It is now clear that sidewalks are treated the same as streets and parks. See United States v. Grace, 461 U.S. 171, 183 (1983). Although many of the streets, parks, and sidewalks cases have gone down the different path of worrying about the risks of excess official discretion, see, e.g., City of Lakewood v. Plain Dealer Publ'g Co., 486 U.S. 750, 757-59 (1988); Hynes v. Mayor of Oradell, 425 U.S. 610, 617 (1976); Saia v. New York, 334 U.S. 558, 560-62 (1948), the principle of mandatory access for streets, parks, and sidewalks remains firm, see, e.g., United States v. Grace, 461 U.S. 171, 179-80 (1983); Jamison v. Texas, 318 U.S. 413, 416 (1943); Schneider v. Town of Irvington, 308 U.S. 147, 160, 164 (1939). The classic analysis of the mandatory access public forum is Harry Kalven, Jr., The Concept of the Public Forum: Cox v. Louisiana, 1965 Sup. Ct. Rev. 1. For more recent commentary, see Lillian R. BeVier, Rehabilitating Public Forum Doctrine: In Defense of Categories, 1992 Sup. Ct. Rev. 79, Ronald A. Cass, First Amendment Access to Government Facilities, 65 Va. L. Rev. 1287 (1979), Robert C. Post, Between Governance and Management: The History and Theory of the Public Forum, 34 UCLA L. Rev. 1713 (1987), and Geoffrey R. Stone, Fora Americana: Speech in Public Places, 1974 Sup. Ct. Rev. 233.

  15. See International Soc'y for Krishna Consciousness, Inc. v. Lee, 505 U.S. 672, 674 (1992) (airport terminal); United States v. Kokinda, 497 U.S. 720, 723 (1990) (walkway to post office); Cornelius v. NAACP Legal Defense & Educ. Fund, 473 U.S. 788, 790 (1985) (pooled charitable giving campaign in federal workplaces); City Council v. Taxpayers for Vincent, 466 U.S. 789, 793 (1984) (telephone poles); United States Postal Serv. v. Council of Greenburgh Civic Ass'ns, 453 U.S. 114, 116 (1981) (residential mailboxes); Greer v. Spock, 424 U.S. 828, 830-31 (1976) (military bases); Hudgens v. NLRB, 424 U.S. 507, 508 (1976) (privately owned shopping center); Lehman v. City of Shaker Heights, 418 U.S. 298, 299-300 (1974) (city buses); Grayned v. City of Rockford, 408 U.S. 104, 105 (1972) (schools); Adderley v. Florida, 385 U.S. 39, 40 (1966) (jails).

  16. In each of the cases cited in note 72 above, the public forum mandatory access claim was rejected.

  17. Although the Court's approach allows the possibility that the state might by its words or deeds "designate" a public forum that will then carry the same obligations as the traditional public forum, see, e.g., Kokinda, 497 U.S. at 726-27; Cornelius, 473 U.S. at 800; Perry Educ. Ass'n. v. Perry Local Educators' Ass'n., 460 U.S. 37, 45 (1983), it is telling that the Court has never found such a designated public forum to exist.

  18. In light of my main theme in this Comment, it is worth noting here that the Court's arguably "unprincipled" decision to mandate access to streets, parks, and sidewalks, but not to government property that is like streets, parks, and sidewalks, is the kind of quantitative, about-this-much, policy- based decision that may be necessary in dealing with many constitutional problems, but which nevertheless grates against the view that judicial decisions should be based on principles not requiring this kind of analysis. The classic defense of such an approach in the First Amendment context is Lee C. Bollinger, Freedom of the Press and Public Access: Toward a Theory of Partial Regulation of the Mass Media, 75 Mich. L. Rev. 1 (1976). And the classic attack on such a "checkerboard" approach to adjudication is Ronald Dworkin, Law's Empire (1986).

  19. See Laurence H. Tribe, Constitutional Choices 204-07 (1985) (noting that the problems of viewpoint discrimination cannot be dealt with under public forum analysis); Daniel A. Farber & John E. Nowak, The Misleading Nature of Public Forum Analysis: Content and Context in First Amendment Adjudication, 70 Va. L. Rev. 1219, 1224 (1984) (suggesting that public forum analysis distracts from central issues of content regulation).

  20. See Forbes, 118 S. Ct. at 1641-42.

  21. See Cornelius, 473 U.S. at 793.

  22. See Lee v. International Soc'y for Krishna Consciousness, Inc., 505 U.S. 830, 831 (1992); International Soc'y for Krishna Consciousness, Inc. v. Lee, 505 U.S. 672, 674-76 (1992).

  23. See Rosenberger v. Rector of the Univ. of Va., 515 U.S. 819, 833 (1995) (holding that the government is "entitled to say what it wishes" with its own funds); Meese v. Keene, 481 U.S. 465, 484 (1987) (noting that Congress may label material as political propaganda); Columbia Broad. Sys. v. Democratic Nat'l Comm., 412 U.S. 94, 139 n.7 (1973) (Stewart, J., concurring) ("[The] [g]overnment is not restrained by the First Amendment from controlling its own expression."); Block v. Meese, 793 F.2d 1303, 1313 (D.C. Cir. 1986) (noting that the government need not be evenhanded in criticizing behavior). For explanations of the limited relevance of the First Amendment to government speech, see Marjorie Heins, Viewpoint Discrimination, 24 Hastings Const. L.Q. 99, 150-59 (1996), Laurence H. Tribe, Toward a Metatheory of Free Speech, 10 Sw. U. L. Rev. 237, 244-45 (1978), and Frederick Schauer, Is Government Speech a Problem?, 35 Stan. L. Rev. 373, 383-86 (1983) (book review).

  24. See Mark G. Yudof, When Government Speaks: Politics, Law, and Government Expression in America 200-07 (1983); Robert D. Kamenshine, The First Amendment's Implied Political Establishment Clause, 67 Cal. L. Rev. 1104, 1106 (1979); Steven Shiffrin, Government Speech, 27 UCLA L. Rev. 565, 588 (1980); Note, Contemplating the Dilemma of Government as Speaker: Judicially Identified Limits on Government Speech in the Context of Carter v. City of Las Cruces, 27 N.M. L. Rev. 517, 533-38 (1997).

  25. The employee free speech cases include some involving what might loosely be called "off the job" speech, in which the state seeks to penalize an employee for engaging in speech that is arguably unrelated to the job itself. See United States v. National Treasury Employees Union, 513 U.S. 454, 470 (1995); Branti v. Finkel, 445 U.S. 507, 517-20 (1980); Pickering v. Board of Educ., 391 U.S. 563, 574-75 (1968). The Court in such cases has been generally sympathetic to the employee's free speech rights. Other cases involve what might equally loosely be called "on the job" speech, in which the Court has been generally unsympathetic to free speech claims. See Waters v. Churchill, 511 U.S. 661, 679-82 (1994); Connick v. Myers, 461 U.S. 138, 150- 54 (1983). There have, however, been some noteworthy exceptions, including the "public concern" sliver of Connick, as well as Rankin v. McPherson, 483 U.S. 378, 388-92 (1987).

  26. See Board of County Comm'rs v. Umbehr, 518 U.S. 668, 673 (1996); O'Hare Truck Serv., Inc. v. City of Northlake, 518 U.S. 712, 720-26 (1996).

  27. This example is a fictional variant on the 1933 contretemps in which the socialist muralist Diego Rivera, hired by Nelson Rockefeller to paint a fresco for the newly constructed Rockefeller Center in New York, inserted a likeness of Lenin's head into the fresco. Rockefeller insisted that the head be replaced or removed, and when Rivera refused the fresco was destroyed. The event is described in detail in Laurance P. Hurlburt, The Mexican Muralists in the United States 159-74 (1989), and is captured with great wit in E.B. White's poem, I Paint What I See [A Ballad of Artistic Integrity], New Yorker, May 20, 1933, at 29.

  28. Perhaps the most notable example is Rust v. Sullivan, 500 U.S. 173 (1991).

  29. Rust followed this approach. See id. at 192-200.

  30. At the extreme of decreased utility is the view that the unconstitutional conditions doctrine should be abandoned. See Cass R. Sunstein, Is There an Unconstitutional Conditions Doctrine?, 26 San Diego L. Rev. 337, 339-42 (1989); Cass R. Sunstein, Why the Unconstitutional Conditions Doctrine is an Anachronism (With Particular Reference to Religion, Speech, and Abortion), 70 B.U. L. Rev. 593, 595-608 (1990); cf. Owen M. Fiss, State Activism and State Censorship, 100 Yale L.J. 2087, 2096 (1991) (arguing that there is no important difference between a non-subsidy and a criminal sanction). More restrained skepticism, foreshadowing one of the themes of this Comment, can be found in Frederick Schauer, Too Hard: Unconstitutional Conditions and the Chimera of Constitutional Consistency, 72 Denv. U. L. Rev. 989, 990 (1995). For the classic exposition of unconstitutional conditions doctrine at the height of its acceptance, see William W. Van Alstyne, The Demise of the Right-Privilege Distinction in Constitutional Law, 81 Harv. L. Rev. 1439, 1445- 49 (1968).

  31. See Speiser v. Randall, 357 U.S. 513, 518-20 (1958).

  32. See Robert M. O'Neil, Unconstitutional Conditions: Welfare Benefits with Strings Attached, 54 Cal. L. Rev. 443, 443-44 (1966).

  33. Many of the post-McCarthy era freedom of association and loyalty oath cases fit this characterization, see Keyishian v. Board of Regents, 385 U.S. 589, 591 (1967); Elfbrandt v. Russell, 384 U.S. 11, 13 (1966); Baggett v. Bullitt, 377 U.S. 360, 361 (1964), as do some of the Fifth Amendment self- incrimination cases of the same era, see Spevack v. Klein, 385 U.S. 511, 512 (1967); Garrity v. New Jersey, 385 U.S. 493, 494 (1967).

  34. See Rust v. Sullivan, 500 U.S. 173, 201-03 (1991); Regan v. Taxation With Representation, 461 U.S. 540, 545-46 (1983).

  35. See, e.g., Lynn A. Baker, The Prices of Rights: Toward a Positive Theory of Unconstitutional Conditions, 75 Cornell L. Rev. 1185, 1216-20 (1990); Roberto L. Corrada, Justifying a Search for a Unifying Theory of Unconstitutional Conditions, 72 Denv. U. L. Rev. 1011, 1030 (1995); Richard A. Epstein, The Supreme Court, 1987 Term--Foreword: Unconstitutional Conditions, State Power, and the Limits of Consent, 102 Harv. L. Rev. 4, 26- 28 (1988); Seth F. Kreimer, Allocational Sanctions: The Problem of Negative Rights in a Positive State, 132 U. Pa. L. Rev. 1293, 1378-95 (1984); Julie A. Nice, Making Conditions Constitutional by Attaching Them to Welfare: The Dangers of Selective Contextual Ignorance of the Unconstitutional Conditions Doctrine, 72 Denv. U. L. Rev. 971, 986-87 (1995); Kathleen M. Sullivan, Unconstitutional Conditions, 102 Harv. L. Rev. 1413, 1499-1505 (1989).

  36. See sources cited supra note 92.

  37. The Court has occasionally mentioned the doctrine of unconstitutional conditions for the purpose of pointing out its inapplicability to the case at hand. See Ohio Adult Parole Auth. v. Woodard, 118 S. Ct. 1244, 1252 (1998); Rust, 500 U.S. at 197. Apart from that, the doctrine has done some work in cases of direct retaliation. See Crawford-El v. Britton, 118 S. Ct. 1584, 1592 n.10 (1998); Perry v. Sindermann, 408 U.S. 593, 597 (1972). It also continues to play a genuine part in cases dealing with penalties on political affiliation, the area for which the doctrine was originally designed. See, e.g., O'Hare Truck Serv., Inc. v. City of Northlake, 518 U.S. 712, 714 (1996); Board of County Comm'rs v. Umbehr, 518 U.S. 668, 685 (1996); Rutan v. Republican Party, 497 U.S. 62, 76-79 (1990); Branti v. Finkel, 445 U.S. 507, 516-17 (1980).

  38. See Rust, 500 U.S. at 197.

  39. To put the same point differently, consider whether conditioning NEA grants on producing good art is an unconstitutional condition on the undoubted First Amendment right to produce pastels of little children with big eyes, drawings of dogs playing poker, or paintings of Elvis Presley on velvet backgrounds.

  40. Kathleen Sullivan's heroic attempt to rehabilitate and reinvigorate unconstitutional conditions doctrine, for example, argues that one of the goals of a reinvigorated doctrine would be to "[e]nsur[e] [g]overnment [e] venhandedness," and that "[s]peech is the paradigm example" of a constitutional right that entails an obligation of government evenhandedness. Sullivan, supra note 92, at 1496. But when a librarian in a public library chooses to acquire books accepting the existence of the Holocaust, and not to acquire books denying the Holocaust, he is hardly being evenhanded in allocating the government funds entrusted to his control, and is hardly being neutral in the treatment of book publishers who have equivalent constitutional rights to express the opposing views.

  41. See sources cited infra note 100.

  42. See Advocates for the Arts v. Thomson, 532 F.2d 792, 795-96 (1st Cir. 1976) (noting that content regulation is a necessary part of arts subsidies); Karen Faaborg, Some Constitutional Implications of Denying NEA Subsidies to Arts Projects Under the Yates Compromise, 12 Hastings Comm. & Ent. L.J. 397, 398-99 (1996) (same).

  43. Among the more prominent recent cases focusing on viewpoint discrimination as such are Rosenberger v. Rector of the University of Virginia, 515 U.S. 819, 828 (1995), Turner Broadcasting System v. FCC, 512 U.S. 622, 643 (1994), R.A.V. v. City of St. Paul, 505 U.S. 377, 383-84 (1992), Simon & Schuster, Inc. v. Members of the New York State Crime Board, 502 U.S. 105, 115 (1991), and Texas v. Johnson, 491 U.S. 397, 412 (1989). For general discussion and analysis, see Daniel A. Farber, Content Regulation and the First Amendment: A Revisionist View, 68 Geo. L.J. 727, 749- 62 (1980), Marjorie Heins, Viewpoint Discrimination, 24 Hastings Const. L.Q. 99, 105-22 (1996), Paul B. Stephan III, The First Amendment and Content Discrimination, 68 Va. L. Rev. 203, 215-18 (1982), Geoffrey R. Stone, Content Regulation and the First Amendment, 25 Wm. & Mary L. Rev. 189, 233- 51 (1983), Geoffrey R. Stone, Restrictions of Speech Because of its Content: The Peculiar Case of Subject-Matter Restrictions, 46 U. Chi. L. Rev. 81, 100-14 (1978), and Susan H. Williams, Content Discrimination and the First Amendment, 139 U. Pa. L. Rev. 615, 636-55 (1991).

  44. Thus Justice Souter finds that "there is nothing inherently viewpoint discriminatory about" the "merit-based criteria" of "'artistic excellence and artistic merit."' Finley, 118 S. Ct. at 2192 n.9 (Souter, J., dissenting).

  45. An intriguing example is the Finley majority's statement that "one could hardly anticipate how "decency' or "respect' would bear on grant applications in categories such as funding for symphony orchestras." Finley, 118 S. Ct. at 2177. I suspect that Dmitri Shostakovich, much of whose work was intended to make an anti-Stalin political statement and was understood by Stalin to have done so, would disagree.

  46. See Robert C. Post, Subsidized Speech, 106 Yale L.J. 151, 166 (1996) (recognizing that viewpoint discrimination is inevitable in many state enterprises); see also Erwin Chemerinsky, The First Amendment: When the Government Must Make Content-Based Choices, 42 Clev. St. L. Rev. 199, 204-07 (1994).

  47. See Rust v. Sullivan, 500 U.S. 173, 194 (1991).

  48. See supra pp. 84-85.

  49. See, e.g., Martin H. Redish and Daryl I. Kessler, Government Subsidies and Free Expression, 80 Minn. L. Rev. 543, 545-48 (1996). Some commentators seem more accepting than Redish and Kessler of the possibility that different principles might apply in different domains. See, e.g., David Cole, Beyond Unconstitutional Conditions: Charting Spheres of Neutrality in Government-Funded Speech, 67 N.Y.U. L. Rev. 675, 681-82 (1992); Robert C. Post, Subsidized Speech, 106 Yale L.J. 151, 194-95 (1996). Cole and Post, however, remain unwilling to describe those principles in institutionally specific terms; they implicitly insist that the basis for permitting or prohibiting content regulation in government enterprises must lie in the realm of supra-institutional principle. Post, for example, sees the subsidy question as part of a larger view of the First Amendment that does recognize institutional differences, but for him the institutional differences are instantiations of a principle that distinguishes dialogic communicative environments, which implicate the First Amendment, from other sorts of communicative environments, which do not. See Robert C. Post, Recuperating First Amendment Doctrine, 47 Stan. L. Rev. 1249, 1254-55 (1995). In this respect, Post's sympathies plainly lie with Holmes, see infra pp. 108-09, and with the suspicion of any doctrine that cannot be described in supra- or trans- institutional terms. As I explain below, see infra pp. 113-20, this is a suspicion I do not share, and which the modern history of legal thought does not support.

  50. Oliver Wendell Holmes, Jr., The Path of the Law, 10 Harv. L. Rev. 457 (1897).

  51. Id. at 474-75.

  52. One of which was making fun of Vermonters. For some of the others, see Frederick Schauer, Prediction and Particularity, 78 B.U. L. Rev. 773 (1998).

  53. Holmes, supra note 107, at 475.

  54. Id.

  55. See Justice Oliver Wendell Holmes, His Book Notices and Uncollected Letters and Papers 59-62 (Harry C. Shriver ed., Da Capo Press 1973) (1936) (discussing "proper" legal categories).

  56. See, e.g., Ronald Dworkin, In Praise of Theory, 29 Ariz. St. L.J. 353, 376 (1997). Dworkin argues: "We must strive, so far as we can, not to apply one theory of liability to pharmaceutical companies and a different one to motorists, not to embrace one theory of free speech when we are worried about pornography and another when we are worried about flag burning." Id. Like Holmes, Dworkin appears to resist the idea that the tangible categories of the world can produce legally (or morally) relevant distinctions, but it is not apparent that, for example, the distinction between pharmaceutical companies and motorists is of any less consequence than the distinction between negligence and strict liability.

  57. This observation is an empirical claim about the behavior of legal actors influenced by the jurisprudential ideology that Holmes identified, and not a claim either about normative desirability or natural necessity.

  58. See, e.g., Katharine T. Bartlett, Feminist Legal Methods, 103 Harv. L. Rev. 829, 849-63 (1990); Martha Minow, The Supreme Court, 1986 Term-- Foreword: Justice Engendered, 101 Harv. L. Rev. 10, 26-31 (1987); Cass R. Sunstein, Incompletely Theorized Agreements, 108 Harv. L. Rev. 1733, 1771- 72 (1995); Cass R. Sunstein, Problems With Rules, 83 Cal. L. Rev. 953, 978- 96 (1995).

  59. See, e.g., Stevens, supra note 34, at 1300-08.

  60. See Lon L. Fuller, The Morality of Law 46-49 (rev. ed. 1969). For an extensive analysis, based on Fuller, of the relationship between legality and generality, see Kenneth I. Winston, Toward a Liberal Conception of Legislation, in Liberal Democracy: NOMOS XXV 313 (J. Ronald Pennock & John W. Chapman eds., 1983).

  61. See Herbert Wechsler, Toward Neutral Principles of Constitutional Law, 73 Harv. L. Rev. 1 (1959). I say "so-called" because the concept of neutrality is in this context superfluous. The core of the "neutral principles" idea is that principles of considerable generality can and should be articulated in a form in which they can be followed in future cases, and then should in fact be followed when those future cases arise. See M.P. Golding, Principled Decision-Making and the Supreme Court, 63 Colum. L. Rev. 35, 41 (1963); Kent Greenawalt, The Enduring Significance of Neutral Principles, 78 Colum. L. Rev. 982, 990 (1978). Whatever else might be said about this idea, nothing in it entails or even suggests the proposition that the principles can be or should be in any way "neutral." Yet shorn of the distracting idea of neutrality, the pull towards principle, and thus away from anything that might look like inconsistency, even if the inconsistency could be justified, remains strong. See Deborah Hellman, The Importance of Appearing Principled, 37 Ariz. L. Rev. 1107, 1108 (1995); Eric L. Muller, The Hobgoblin of Little Minds? Our Foolish Law of Inconsistent Verdicts, 111 Harv. L. Rev. 771, 801 (1998). As quoted by Muller, see id., Justice Breyer has noted that "the uniform general treatment of similarly situated persons ... is the essence of law itself." BMW of N. Am., Inc. v. Gore, 517 U.S. 559, 587 (1996) (Breyer, J., concurring). In one sense, the qualification of "similarly situated persons" makes the statement uninteresting, for it suggests only the inadvisability of drawing a distinction without a difference. But in another sense, the statement does reflect the law's historical preference for "general treatment"--a preference reflecting the extent to which rules in general, and the rules of law in particular, often suppress relevant differences in the service of other values. See Frederick Schauer, Playing By the Rules: A Philosophical Examination of Rule-Based Decision-Making in Law and in Life 21- 22 (1991).

  62. Edwin W. Patterson, Jurisprudence: Men and Ideas of the Law 98 (1953).

  63. See Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175 (1989).

  64. See Morey v. Doud, 354 U.S. 457, 469 (1957), overruled by New Orleans v. Dukes, 427 U.S. 297, 306 (1976).

  65. See sources cited supra note 115; see also Lawrence B. Solum, Equity and the Rule of Law, in The Rule of Law: NOMOS XXXVI 120, 145 (Ian Shapiro ed., 1994) ("The rule of law is not a law of rules.").

  66. See, e.g., Eugene Volokh, Freedom of Speech, Shielding Children, and Transcending Balancing, 1997 Sup. Ct. Rev. 141, 167-68.

  67. Compare Laurent B. Frantz, Is the First Amendment Law? A Reply to Professor Mendelson, 51 Cal. L. Rev. 729, 732 (1963), Laurent B. Frantz, The First Amendment in the Balance, 71 Yale L.J. 1424, 1449 (1962), and Alexander Meiklejohn, The First Amendment is an Absolute, 1961 Sup. Ct. Rev. 245, 255-57, with Wallace Mendelson, The First Amendment and the Judicial Process: A Reply to Mr. Frantz, 17 Vand. L. Rev. 479, 484-85 (1964).

  68. See Beauharnais v. Illinois, 343 U.S. 250, 257-64 (1952) (Frankfurter, J.); Dennis v. United States, 341 U.S. 494, 517 (1951) (Frankfurter, J., concurring). In light of the theme of this Comment, it is also interesting to note that in Kovacs v. Cooper, 336 U.S. 77, 96-97 (1949) (Frankfurter, J., concurring), Frankfurter argued that there ought to be separate principles for sound trucks, an argument no more successful with the Court than the Vermont justice's implicit argument about separate principles for churns was with Holmes.

  69. See sources cited supra note 124.

  70. See Thomas I. Emerson, First Amendment Doctrine and the Burger Court, 68 Cal. L. Rev. 422, 429 (1980); Schauer, Categories, supra note 5, at 274-82.

  71. See Forsyth County v. Nationalist Movement, 505 U.S. 123, 130-31 (1992) (holding that excess discretion in granting and conditioning parade permits could produce content discrimination); City of Lakewood v. Plain Dealer Publ'g Co., 486 U.S. 750, 772 (1988) (striking down a newspaper box regulation for granting excessive discretion to government officials).

  72. See Smith v. Goguen, 415 U.S. 566, 572-76 (1974); Coates v. Cincinnati, 402 U.S. 611, 614-15 (1971).

  73. See Gooding v. Wilson, 405 U.S. 518, 528 (1972); Richard H. Fallon, Jr., Making Sense of Overbreadth, 100 Yale L.J. 853, 863-64 (1991); Henry Paul Monaghan, Overbreadth, 1981 Sup. Ct. Rev. 1, 1-3; Martin H. Redish, The Warren Court, the Burger Court and the First Amendment Overbreadth Doctrine, 78 Nw. U. L. Rev. 1031, 1032 (1983).

  74. See New York Times Co. v. Sullivan, 376 U.S. 254, 277-79 (1964); Frederick Schauer, Fear, Risk, and the First Amendment: Unraveling the Chilling Effect, 58 B.U. L. Rev. 685, 685-87 (1978).

  75. See Frederick Schauer, Slippery Slopes, 99 Harv. L. Rev. 361, 361-63 (1985). It is important to recognize that an institution-specific principle may be no less a principle because of its institutional specificity and no less morally or politically sound because of that specificity. A principle that turned, for example, on the difference between books and newspapers would be no less a principle than one that turned on the difference between subject-matter and viewpoint discrimination, even though the former was institutionally specific in a way that the latter was not. Under some circumstances the former might be more susceptible to abuse than the latter (and vice versa), but that should not be confused with a defect in the principle itself. See Gerald Dworkin, Non-Neutral Principles, in Reading Rawls: Critical Studies of A Theory of Justice 124 (Norman Daniels ed., 1975).

  76. See City of Ladue v. Gilleo, 512 U.S. 43, 59-60 (1994) (O'Connor, J., concurring).

  77. Thomas G. Krattenmaker & L.A. Powe, Jr., Converging First Amendment Principles for Converging Communications Media, 104 Yale L.J. 1719, 1740 (1995). For a longer version of Krattenmaker and Powe's argument, applied to a wide range of media and institutions, see Thomas G. Krattenmaker & Lucas A. Powe, Jr., Regulating Broadcast Programming (1994).

  78. See Ronald Dworkin, Law's Empire 178-84 (1986); Ronald Dworkin, A Matter of Principle 72-103 (1985); Ronald Dworkin, Taking Rights Seriously 22-28, 71- 80, 90-100 (1977).

  79. For commentary on this problem in the First Amendment context, see Note, Deference to Legislative Fact Determinations in First Amendment Cases After Turner Broadcasting, 111 Harv. L. Rev. 2312 (1998). This theme also arises in Michael C. Dorf, The Supreme Court, 1997 Term--Foreword: The Limits of Socratic Deliberation, 112 Harv. L. Rev. 4 (1998).

  80. See Melvin A. Eisenberg, The Nature of the Common Law 14-42 (1988) (arguing that common law courts set aside the indications of legal rules in the service of policy as well as in the service of principle); Larry Alexander, The Banality of Legal Reasoning, 73 Notre Dame L. Rev. 517, 519 (1998) (arguing that the application of principles often involves questions of policy). For the argument that many rights claims are best understood in policy terms, see Richard H. Pildes, Why Rights Are Not Trumps: Social Meanings, Expressive Harms, and Constitutionalism, 27 J. Legal Stud. 725, 761 (1998).

  81. One version of the view, largely Dworkin's, is that the policy/principle divide is a question of legitimacy. If questions of policy are largely questions of utilitarian welfare maximization, as Dworkin appears to believe, then this is an area in which the people or their representatives, and not the courts, ought to have a special priority, as Jeremy Bentham, to take a particularly prominent example, believed. See James Steintrager, Bentham 57- 58, 78-96 (1977). Only in areas in which there are rights-based (and therefore principle-based) trumps on utilitarian welfare maximization do the non- majoritarian courts, Dworkin claims, have a role to play. See Dworkin, Law's Empire, supra note 135, at 243-44; Dworkin, Taking Rights Seriously, supra note 135, at 96-97. 

Dworkin's argument is usefully distinguished from an argument based less on legitimacy and more on institutional design. If courts by their design are not well-suited to deal with the informational intensity and polycentricity of policy decisions, this argument would go, then they should stay out of the arenas in which their particular institutional competencies put them at a comparative disadvantage. 

Even this latter argument, however, takes the design of the courts as fixed and their jurisdiction as variable. Yet in some areas--the First Amendment, for example--the jurisdiction may in some important way be fixed, in which case we might want to treat the question of design as a variable and consider how courts might better make the policy decisions they are necessarily going to make and that they may, as a practical matter, be constitutionally compelled to make as well.

  1. See Frank H. Easterbrook, Cyberspace and the Law of the Horse, 1996 U. Chi. Legal F. 207, 207-08 (arguing that there is no more a "law of cyberspace" than there is a "law of the horse"). Some of this flavor is also captured in Justice Souter's dissent in Finley, in which he characterizes a plea for an institution-specific rule for NEA patronage as an argument for "a new exception to the fundamental rules that give life to the First Amendment." Finley, 118 S. Ct. at 2193 (Souter, J., dissenting).

  2. For some, many of the ideas about institutional specificity and institutional differentiation are associated with the thinking of Niklas Luhmann. See Niklas Luhmann, The Differentiation of Society (Stephen Holmes & Charles Larmore trans., Columbia Univ. Press 1982); Niklas Luhmann, Social Systems 189-94 (John Bednarz, Jr. trans., Stanford Univ. Press 1995); Niklas Luhmann, The Autopoiesis of Social Systems, in Sociocybernetic Paradoxes: Observation, Control, and Evolution of Self-steering Systems 172 (Felix Geyer & Johannes van der Zouwen eds., 1986). For an application of the Luhmannesque idea of social differentiation to the question of arts funding from non- governmental sources, see Christoph Beat Graber and Gunther Teubner, Art and Money: Constitutional Rights in the Private Sphere?, 18 Oxford J. Legal Stud. 61, 64-65 (1998). On the connection between institutional differentiation and institutional authority, see Richard H. Pildes, Avoiding Balancing: The Role of Exclusionary Reasons in Constitutional Law, 45 Hastings L.J. 711, 722 (1994).

  3. Cf. Board of Educ. v. Pico, 457 U.S. 853, 863 (1982) (struggling with the question of the applicability of the First Amendment in the context of the removal of books from a school library). For an analysis of the problems and possibilities inherent in seeing school libraries as public forums, see Mark G. Yudof, Library Book Selection and the Public Schools: The Quest for the Archimedean Point, 59 Ind. L.J. 527 (1984).

  4. The most obvious form of this concern arises in the context of controversies regarding the delegation of the content selection process to Internet filtering technology. See Mainstream Loudon v. Board of Trustees of the Loudon County Library, No. 97-2049-A (E.D. Va. Apr. 7, 1998) (mem.) < http://www.techlawjournal.com/courts/loudon/80407mem.htm>; Joshua Micah Marshall, Will Free Speech Get Tangled in the Net?, Am. Prospect, Jan.-Feb. 1998, at 46. Yet even outside of the highly-charged and possibly unique context of Internet filtering for sexual explicitness, outsourcing of the selection process has generated controversy which is often couched in terms of professional prerogative. See George Eberhart, The Outsourcing Dilemma, Am. Libr., May 1997, at 54; Sandy Schwalb, The Ins and Outs of Outsourcing: The Changing, Evolving Scene for Information Professionals, Database, June-July 1997, at 41.

  5. This is not to deny that professional cultures have their own speech- inhibiting pathologies, a point I belabor in Frederick Schauer, Discourse and Its Discontents, 72 Notre Dame L. Rev. 1309, 1326-28 (1997). But if we raise our concern for speaker pathologies to the level of our concern for regulator pathologies, we call into question the core of the idea of free speech. Even though speakers and their professional cultures censor with regularity, the preference for this type of censorship lies at the heart of the First Amendment.

  6. 457 U.S. 853 (1982).

  7. For a sampling of litigated controversies, see Rodney A. Smolla, Freedom of Speech for Libraries and Librarians, 85 L. Libr. J. 71, 74-75 (1993).

  8. I do not claim that the so-called "institutional theory of art" is correct, or the only theory of art, but it is a widely-held theory. See George Dickie, Aesthetics 101 (1971) (defining a work of art as any "artifact ... upon which some person or persons acting on behalf of a certain social institution (the artworld) has conferred the status of candidate for appreciation"); Richard Shusterman, Pragmatist Aesthetics: Living Beauty, Rethinking Art 38-40 (1992); George Dickie, A Tale of Two Artworlds, in Danto and His Critics 73 (Mark Rollins ed., 1993).

  9. See sources cited supra note 146. See also B.R. Tilghman, But Is It Art? 56-60 (1984).

  10. Put differently, the Court might say that different social institutions, differentiated by their empirical social situations and not by any legal or philosophical principle, could because of that empirical differentiation be granted presumptive authority over certain decisions. See Pildes, supra note 140, at 721-22 (connecting institutional differentiation to questions of authority).

  11. It would be a mistake to think that institutional differentiation would necessarily have a more or less "protective" valence in First Amendment cases. First, cases like Forbes challenge the assumption that one can easily identify the outcome that better serves the First Amendment or its values, as both Forbes's claims to access and the Arkansas Educational Television Commission's claims to the journalistic prerogative to deny that access can easily be couched in First Amendment terms. In addition, institutional differentiation might at times produce greater First Amendment protection by providing better arguments for special First Amendment status for the press, for colleges and universities, for libraries, and for the institutions of the arts, and might at other times produce less protection by providing arguments for resisting the identification of all state decisions with the stereotypical censor.

  12. In generally failing to see institutional differences as making a First Amendment difference, the Court might be seen as treating institutions as commensurable. Conversely, stressing institutional differentiation could be seen as compatible with the view that values are incommensurable. On the debates about incommensurability, see Symposium, Law and Incommensurability, 146 U. Pa. L. Rev. 1169 (1998). On the connection between incommensurability and institutions, see Richard H. Pildes, Conceptions of Value in Legal Thought, 90 Mich. L. Rev. 1520, 1521 (1992) (book review).

  13. See Evans v. Newton, 382 U.S. 296, 299 (1966) (holding that a private actor performing a public function constitutes state action); Burton v. Wilmington Parking Auth., 365 U.S. 715, 725 (1961) (finding state action when private actors are entangled with state enterprises); cf. Marsh v. Alabama, 326 U.S. 501, 506 (1946) ("The more an owner, for his advantage, opens up his property for use by the public in general, the more do his rights become circumscribed by the statutory and constitutional rights of those who use it.").

  14. 484 U.S. 260 (1988).

  15. See id. at 273 n.7; see also Papish v. Board of Curators of Univ. of Mo., 410 U.S. 667, 670 (1973) (striking down disciplinary sanctions against editors of a university newspaper for publishing allegedly offensive language). Papish is an important but unacknowledged predecessor of Forbes, for in both cases a state journalistic enterprise was framed as journalism rather than as a state enterprise.

  16. 478 U.S. 675, 685 (1986) (holding that the First Amendment allows school officials to sanction students who disrupt the educational process by engaging in offensive, lewd, or indecent speech).

  17. See Grosjean v. American Press, 297 U.S. 233, 250 (1936) (invalidating a politically-inspired punitive tax on newspapers).

  18. Although I focus in this Comment on the government enterprise cases, in which the failure to take institutions seriously produces the greatest problems, most of what I say applies to First Amendment doctrine generally. Only by appreciating the cultural and institutional reasons why governmental assessments of truth and falsity are treated skeptically in some institutional contexts and sympathetically in others, for example, can we explain why the Securities Act of 1933, in First Amendment language a quite perfect scheme of prior restraint, remains entirely untouched by First Amendment doctrine. And only by appreciating the cultural and institutional differences between the workplace and other places in which words are spoken can we grapple with difficult issues of workplace speech. See Cynthia L. Estlund, The Architecture of the First Amendment and the Case of Workplace Harassment, 72 Notre Dame L. Rev. 1361, 1383 (1997) (advocating "build[ing] a separate room" within First Amendment law for workplace harassment); see also Cynthia L. Estlund, Freedom of Expression in the Workplace and the Problem of Discriminatory Harassment, 75 Tex. L. Rev. 687, 716 (1997) ("Because of their economic dependence on the job and their physical confinement in the workplace, employees cannot simply ignore or escape from discriminatory speech in the workplace as readily as the law assumes they can in the world at large.").

  19. See Vincent Blasi, The Pathological Perspective and the First Amendment, 85 Colum. L. Rev. 449, 457-58 (1985).

  20. See Karl Llewellyn, The Common Law Tradition: Deciding Appeals (1960). See generally William Twining, Karl Llewellyn and the Realist Movement 302-40 (1985) (explaining how Llewellyn's approach to the Uniform Commercial Code was based on locating and reinforcing commercially significant groupings of fact patterns).

  21. See Schauer, supra note 118, at 224-28.

  22. One commentator has observed: In ... [many] cases it has been the experience of men that discretion to discriminate between individual cases "on the merits", far from being unjust, was vital to the attainment of justice; far from being arbitrary was necessary to avoid the arbitrary operation of general rules of law. And so it may be with legislation which discriminates in advance between different concrete situations and establishes standards for administrative application to this end. 

Julius Stone, The Province and Function of Law: Law as Logic, Justice, and Social Control 262 (1973).

  1. See Frederick Schauer & Virginia J. Wise, Legal Positivism as Legal Information, 82 Cornell L. Rev. 1080, 1102-09 (1997) (tracking the Supreme Court's increased willingness to make use of policy-oriented "non-legal" sources of information). The reluctance of free speech adjudication and free speech theory to pay proper attention to empirical reality is stressed in L.A. Powe, Jr., Situating Schauer, 72 Notre Dame L. Rev. 1519, 1534-37 (1997), and L.A. Powe, Jr., The Supreme Court, Social Change, and Legal Scholarship, 44 Stan. L. Rev. 1615, 1615 (1992) (book review).