Printer Friendly

Autonomy and democracy.

Standard arguments about such fundamental matters as the nature of political obligation and the justifiability of civil disobedience, not to say revolution, revolve around conditions under which interference with autonomy is permissible. Indeed, it is a commonplace that the success of a political regime is to be measured, in significant part, by the extent to which it protects and furthers the autonomy of the individuals who constitute it.

Such arguments presuppose a relatively uncontroversial version of autonomy. But what is it to be autonomous?

A natural initial thought is that autonomy is nothing other than (negative) liberty. On this view, a person is autonomous with regard to his decisions if his decision-making procedure is unencumbered by constraints. He is autonomous with regard to his actions if, having decided to act in a certain way, he is at liberty so to act.

It is, however, unlikely that anyone seriously has espoused this view. For it is almost certainly a mistake to identify autonomy with negative liberty, if negative liberty itself is thought of as freedom from all constraints. The problem for autonomy, rather, arises when constraints are imposed on an agent's choices or actions by someone or something other than the agent. Thus, the more plausible view (to which I will refer as individualism) is that an autonomous agent is one who possesses negative liberty in the sense that he is free from any externally imposed constraints. As Robert Paul Wolff has put it, the autonomous person may be bound by a variety of constraints, as long as "he alone is the judge of those constraints."(1) That is, a person need not, at every turn, be "free" to choose or act precisely as he is inclined at that moment. A constrained choice or act can be an autonomous one, as long as, and insofar as, the source of the constraints is the person himself

Wolff thought that no citizen of a state can be fully autonomous. That is, only anarchism is fully compatible with genuine autonomy.(2) But even those who do not contest the prima facie legitimacy of political authority have employed a Wolffian conception of autonomy to constrain state authority. A prominent example is Robert Post, who has argued that the state can have no role in regulating speech under the First Amendment.(3) That discussion is the focus of this Note.

Post's target is the view, propounded by (among others) Alexander Meiklejohn(4) and Owen Fiss,(5) that the state has an affirmative obligation to regulate speech for the purpose of securing citizen autonomy. Specifically, this view ("collectivism') envisages the state stepping in when it determines that debate is not rich enough, in order to provide for its enrichment, even if doing so requires it to silence some speech in order to enable other speech. Post claims that this position threatens citizen autonomy, arguing that a fundamental concomitant of such autonomy is the citizens' power and right to decide what they will hear on public issues, and that this is precisely what collectivism seems to wrest from them. The argument for Post's position, as I will show, rests squarely on individualism.(6)

This charge against collectivism is serious. I attempt to answer it by arguing, first, that individualism generally is compatible with some external constraints on the choice set of autonomous decision makers. Second, I argue that it follows that, in the context of the autonomy of citizens in a democracy, a certain degree of state regulation of speech is permissible. In Part I, I expound, in general terms, the collectivist commitment to rich public debate. In Part II, I rehearse Post's arguments against collectivism. Parts III through V contain my defense of collectivism. In Part III, I argue against the coherence of (Wolffian) individualism. On the basis of that result, I argue that a particular kind of autonomy must be at the center of collectivism's concern. Finally, I spell out collectivism's principal commitments in terms of that conception. Part IV contains my direct response to Post, while Part V develops and answers an important objection to my response. Part VI contains some concluding remarks.


The First Amendment, according to collectivism, requires public debate to be robust and wide-ranging enough to enable "true collective self-determination,"(7) and permits the state to regulate speech to that end, where necessary. Though the concept of true self-determination may not be perfectly perspicuous, we can imagine well enough scenarios in which we are intuitively confident that public decisions are not products of true self-determination.

Consider first a scenario in which citizens have been so inundated with messages from one side of a debate - say, as a result of an extremely effective advertising campaign - that they have, as a result, not only embraced the message conveyed, but have become unable, also as a result of the inundation, to attend to or to process messages from any other perspective. (Call this the saturation scenario.) By "as a result," I mean that it has been by the sheer power of the repetition that citizens have arrived at their present convictions, rather than by way of some deliberative acts (for example, by reflection on the messages in the light of presently held views, wishes, hopes, and the like). In a word, the repetition has caused the convictions.(8)

It is clear, I think, that if a public decision is the result of votes cast by persons on the basis of convictions like these, the decision is not an instance of true collective self-determination. For the decision of the collectivity, resulting from the votes of individual citizens, is an act of determination by the collectivity only if each citizen (or, perhaps, enough citizens) has determined for him- or herself what his or her vote shall be. To determine for oneself how one will vote, however, is to deliberate with respect to the matter to be voted on.(9) If, on the other hand, some external agent were to cause citizens to vote for a certain position,(10) although the result might be unanimous, the decision really would amount to the external agent's acting through the citizens. Perhaps the vote would be an act of self-determination by the external agent - but it is no such act (and perhaps, no act at all) of the citizens.

The saturation scenario makes plain the central role of individual deliberation in true collective self-determination, and that is why I have chosen to start with it. A slightly different scenario, however, will be more centrally involved in my argument with Post. I will refer to it as the deficiency scenario. In it, citizens' convictions are provisional; they are subject to revision upon the receipt of new information. Citizens do, in fact, sift information they receive, at least insofar as they think it bears on their convictions. That is, they deliberate before they vote, and how they vote is determined by their deliberations. In the deficiency scenario, however, there exists information that would bear on the votes of (at least some) citizens in the sense that, were they to receive the information, it would figure in their deliberations. Their convictions would not necessarily change - indeed, the information might only confirm the citizens' sense that these convictions are the ones they ought to have. Even so, the citizens would critically consider their convictions in light of the new information, if they received it.

The public decisions in this scenario are the results of deliberation., In a way then, they are acts of collective self-determination, for the collectivity determines itself by virtue of each citizen determining his or her vote for himor herself. But there was information available - in the sense that it would have been disseminated had its possessor been able to disseminate it. By hypothesis, its dissemination would have made a difference to the citizens' deliberations, and hence could have altered the public decision. In this case, then, it is not clear that the votes of the citizens represented their own best judgment. It is as though no citizen had fully expressed himself in his vote, because there was a part of him that had been inactive, but that would have been active had he received the information. But then, the public decision might not have been truly collectively self-determined, for it was not a judgment by which citizens truly represented themselves.

A simple example may help to cash out this metaphorical story about "parts" of a citizen. Suppose a citizen is to vote on a presidential candidate. No issue of marital fidelity has been raised in the campaign, and the citizen has not thought to raise the question himself regarding any of the candidates. He has strong views on the matter, however. He believes that a candidate who cheats on his or her spouse shows dishonesty and a failure of commitment, and thus is not qualified for substantial public leadership positions. Suppose now that one of the candidates has indeed been unfaithful to her spouse, and that some citizens would publicly announce this information, but are unable to do so. Suppose, further, that the citizen casts his vote for that candidate. His vote is inconsistent with one of his deep convictions. A part of him - a conviction that he believes constitutes part of the kind of person he is - is not represented in his vote even though it should be, since there is public information available (in the sense that some citizens possess the information and would willingly convey it) that implicates that conviction. In this case, the citizen's vote could have represented substantially more of him than it in fact did, and therefore the vote does not well represent him. If this is true of enough other citizens, the resulting public decision is not a true reflection of the citizens, as represented by their true commitments. It is not, therefore, an act, or at least not fully an act, of true collective self-determination.

The deficiency scenario must be distinguished sharply from another scenario, in which the information (e.g., that the candidate was unfaithful) is not available-no citizen possesses it. This scenario involves, of course, a situation that citizens always face, no matter how they are otherwise situated. We have incomplete knowledge. We don't know how various policies, once enacted, will affect behavior; we don't know how the physical world will tolerate our technological incursions; we aren't sure how our national neighbors will react to what we do. It is not possible to defer public decision until all the information is in, and the result is that citizens must deliberate, and vote, knowing that they might well vote differently if they had more information than they in fact do. In this case, where citizens deliberate over all the information in fact available to them, where no available information is kept from them, but where not all the information is in, is the resulting decision an act of true collective self-determination? I think that here one must answer in the affirmative, if the concept of true collective self-determination is to have any practical force at all. It is true that the citizens are constituted such that they might have deliberated to a different vote with new information. But there is no sense in which the information in question was a public possession, even potentially, and there was thus no act practically available to the citizens, individually or collectively, by which the results of individual deliberations could have been made to differ. It is one thing for citizens to vote on the basis of all the information they could have, given the natural limitations on their knowledge, and another for them to vote on the basis of an amount of information less than the amount actually available to them. In the former case, but not in the latter, it seems right to say that the resulting public decision is not disqualified as being a case of true collective self-determination, because the individual citizen votes were based on the fullest deliberation possible for them.

Collectivism as I understand it is best articulated against the background of these scenarios. Its central tenet, as I have said, is commitment to rich public debate. But such debate is not an end in itself The collectivist is committed to ensuring that, as far as possible, important information is disseminated to the citizens, and that the state enable such dissemination if need be, to the end of enabling true collective self-determination.

Later, as the discussion dictates, I will develop this commitment more fully. For now, I have only wished to convey the flavor of collectivism; in particular, how the commitment to rich public debate is a commitment to ensuring true collective self-determination, and how true collective self-determination turns on how representative are the votes of citizens of their deliberative selves.


Post thinks that collectivism is an incoherent doctrine. For him, the true collective self-determination to which collectivism is committed presupposes citizen autonomy. But such autonomy, Post argues, is flatly incompatible with the state playing any coercive role in public debate. Thus, true collective self-determination is impossible as long as the state has any regulatory role to play in public debate.

The first part of Post's argument forges the link between autonomy and collective self-determination, and we need not pause long over it. Autonomous choice is, intuitively speaking, choice determined by the chooser, and neither forced upon him nor determined by forces outside him. Collective self-determination is the choice by citizens of what Post calls national self-identity:" It is, as I will say, the choice by citizens of a democracy of the way in which they will conduct business democratically. Crucially, collective self-determination is determination that flows from citizens and not from external causes.(12) In that simple sense, it is determination that presupposes citizen autonomy. As Post writes, "The ascription of autonomy is ... [the] precondition for the possibility of democratic self-determination."(13)

This position is clearly implied by collectivism as I have described it, and I will assume its correctness for what follows. It is Post's view on what is necessary for autonomy, and not (directly) what is necessary for self-determination, that I think is mistaken. I turn now to the details of that part of his argument.

A. The Town Meeting

Post's argument centers on Meiklejohn's conception of collectivism as it appears in the latter's town-meeting model of democracy.(14) To develop Post's argument, I need to say a little about that model.

For Meiklejohn, a central feature of the town meeting is that it does not tolerate just any speech. In the town meeting, "[w]hat is essential is not that everyone shall speak but that everything worth saying shall be said."(15) But we cannot readily discern this feature if we think of the town meeting merely as an institution where citizens congregate to discuss public issues, to vote on them, and to be bound by their votes. Nothing follows from this description about the regulation of speech.

Meiklejohn tells us what needs to be added to the description. The town meeting is "a group of free and equal men, cooperating in a common enterprise, and using for that enterprise responsible and regulated discussion."(16) Further:

[I]t is usually agreed that no one shall speak unless "recognized by the chair." Also, debaters must confine their remarks to "the question before the house." If one man "has the floor," no one else may interrupt him except as provided by the rules.... If a speaker wanders from the point at issue, if he is abusive or in other ways threatens to defeat the purpose of the meeting, he may be and should be declared "out of order." He must then stop speaking, at least in that way.... The town meeting, as it seeks for freedom of public discussion of public problems, would be wholly ineffectual unless speech were thus abridged.(17)

These rules distinguish some speech from some other speech, and they permit some speech while prohibiting other speech.(18) They thus determine, at least minimally, that some speech is not worth saying, and prohibit utterance of that speech.

Further, the very possibility of such rules presupposes an "external" means of enforcing them - what Post calls a managerial structure.(19) If they did not assume such a structure, if the interpretation of what is permissible speech within the purview of the rules was itself open to debate by the citizens at the same time as the main issue, debate over which is supposed to be constrained by those same rules, then each citizen would be entitled to speak to the interpretation. If this were the case, there would be no basis for ruling out any speech as not permitted, even according to "the rules." The result would be simply that everyone could speak.

So the debating rules presuppose a managerial structure for the town meeting whose role it is to regulate speech, at least procedurally. The structure, however, may be more expansive than that. For another part of the town-meeting conception, presumably, involves a mechanism for setting the agendas of ensuing meetings. We may assume that the citizens themselves set at one meeting the agenda for the next; in any case, the agenda having been set the citizens are not free to alter it at the next meeting. This further constraint on citizen speech imports a more substantive censorship power on the meeting manager. Now part of what is "worth saying" is determined, not merely procedurally on the basis of what participants in the meeting have already allowed to be said there, but also on the basis of the substance of the issues under discussion.(20) If the issue is nuclear power, for instance, a citizen may not speak on a referendum for a new park - at least, he may not do so if the manager judges that the park issue is irrelevant to the nuclear power issue, and thus is not worth saying in the, nuclear power debate. (That this determination must be made by the manager follows from the argument above that speech-regulating rules must be enforced by an agent external to the public debate.)

B. Town Meetings and Public Discourse

Post concedes that constraints on speech such as the ones I have described "make sense within the context of a town meeting";(21) that is, this conception may not offend true collective self-determination. In particular, the manager of the town meeting does not necessarily violate free speech norms in prohibiting certain citizens from speaking. This is precisely because, sharing a "common enterprise" and an understanding that the enterprise can go forward only if its participants have previously agreed upon rules of procedure, the citizens may be supposed actually to have agreed to the rules that the manager imposes.(22) In applying the rules then, the manager only imposes an order that the citizens had freely chosen for themselves in a previous act of self-determination.

Now, Meiklejohn "sees the exercise of democracy as analogous to an enormous town meeting."(23) For Post, however, this analogy is fatally flawed. Even though citizens may adopt the town-meeting structure for making "local" decisions (e.g., whether to build a nuclear power plant), the state "cannot control the agenda of public discourse ... [because] [tlhe enterprise of public discourse ... rests on the value of autonomy, which requires that all possible objectives, all possible versions of national identity, be rendered problematic and open to inquiry."(24) Therefore,

[an agreement like the one to adopt the managerial structure of a town

meeting] cannot be assumed without concomitantly diminishing the

arena for self-determination.... The state cannot control the agenda

of public discourse, or the presentation and characterization of issues

within public discourse, because such control would necessarily

circumscribe the potential for collective self-determination.(25)

In short, though it is not always the case that a citizen must have before her "all possible objectives" - she need not in the town-meeting setting, when she is confronting a "local" issue - she must when it is the institution of public discourse itself that is at issue. Otherwise, as the first passage teaches, she would fail to be autonomous. And thus failing, according to the second passage, her choices would not effect self-determination.

Post's claim then is that democracy is not simply an "enormous" town meeting. This is because, though town-meeting citizens are able to adopt constraints on their (future) consideration of local issues, thereby constraining the choices that they confront as they (later) deliberate over those issues, citizens of a democracy cannot adopt constraints - hence cannot freely agree to accept constraints - on public discourse itself.(26) It follows, according to Post, that the collectivist theory is inconsistent with autonomous choice, and hence is incoherent as a theory of collective self-determination. For it precisely advocates external constraints on public discourse.

But why can there be no agreement in public discourse by which citizens arrange to constrain, in the future, that very discourse? Post doesn't say, but he surely has something like the following argument in mind: Autonomous individuals develop and change over time, and their judgments over time themselves change to reflect those changes. A prior decision to constrain a later decision-making process can only be conceived as not intrusive on the decision maker's autonomy, given his changeableness, if that prior decision is ever open to revision - for the assumption that the constraints were freely chosen bears on die compatibility of the constraints with present autonomy only if the choice then relates to the constraints now. Put otherwise, the prior choice can bind now only if it is a choice that the citizens, ever changing, approve now. If that is so, however, the prior agreements must always be up for grabs. On this model, there is no room for the creation of an external managerial structure; or better, that structure can have no independent authority over citizens deriving from an antecedent decision of theirs. To suppose otherwise is to embrace the possibility that decision procedures from the past are constraining the choices of citizens now even though the procedures are currently disapproved.(27)

It is apparent, I hope, that the operative principle underlying this argument is none other than Wolffian individualism - the view that autonomy is incompatible with the existence of any externally imposed constraints on choice. Indeed, it is only with some such view in place that sense can be made of the crucial distinction between the context in which public discourse is itself, somehow, at issue, and merely "local" contexts in which it is not.

Let me explain. As I have said, public discourse, for Post, is that process by which citizens choose for themselves the way - the set of constraints under which - they will conduct their (political) business democratically.(28) Sometimes, however, citizens congregate not to choose a way of conducting their business, but rather to do some local business. When, for instance, town-meeting citizens meet to discuss nuclear power, their way of doing business - that is, the town-meeting form - is not on the table; when they set nuclear power on the agenda, they presuppose the town-meeting structure as the means by which the issue will be decided. This is, moreover, no accident. The local context is the one in which business gets done, and thus necessarily presupposes a way of getting it done.

But the mere fact that this structure, this set of constraints, is in place does not derogate from citizen autonomy. For - and here is the Wolffian point - as long as the constraints are "local," that is, as long as they do not constrain public discourse itself, citizens are free now to reject them. In short, the autonomy of the citizens is not "located in" the mere fact that the local structures constrain them, but rather in the fact that the structures doing the constraining have been freely chosen by them.

Now suppose that, as the debate on nuclear power proceeds, some citizens become dissatisfied, not with what is said, but by the constraints created by the town-meeting form itself. Suppose the focus of the meeting shifts to the issue of whether the town-meeting debate rules should constrain the debate process. Suppose, finally, that some citizens attempt to shut down that part of the debate on the ground that the town-meeting identity had been presupposed by the agenda-setting of the previous meeting, and is not now open to revision.

In this case, Post would say that a national identity - a way in which citizens conduct their business democratically - is being forced upon some citizens by others. For, as just stated, local business can only be conducted on the supposition that the decision-making structure is freely embraced by citizens subjected to it. Ignoring the citizens' objections falsifies the supposition. Indeed, if such a structure is forced upon citizens at any time, the choices made under it are not guaranteed to express the will of enough citizens, and the resulting decision cannot be said to be a case of true collective self-determination. For instance, consider that an alternative to the town-meeting structure could be much more forgiving in the amount of speech it permits. It is possible that in that alternative, views on the nuclear power issue are expressed that do not get expressed in the town meeting and that the resulting public decision would be different.

In sum, the claim that the existence of "local" decision contexts in which structures are imposed on public discourse for the purpose of getting democratic business done is compatible with citizen autonomy rests crucially, for Post, on the idea that the constraints are freely chosen. The claim rests, that is, on the assumption that only those constraints that are freely chosen are compatible with citizen autonomy: It rests on Wolffian individualism.

To recap Post's argument: Collective self-determination presupposes the autonomy of public discourse. The autonomy of public discourse is incompatible with the existence of any managerial structures relative to that discourse.(29) But the collectivist theory urges state regulation of speech in the name of collective self-determination, which regulation amounts to a managerial structure. Because collectivism would impose such regulation upon citizens, regardless of whether they have chosen to accept it (and are ever ratifying that choice), it impedes citizen autonomy and hence collective self-determination. The theory is, therefore, incoherent.

Post's argument is perhaps seductive, but it is mistaken. On one level, its error lies in its assumption that any legitimately constraining managerial structure presupposes a prior and continuing public act of acceptance, without which the constraints impede citizen autonomy. More profoundly, however, Post's error lies in the Wolffian assumption that any constraint on choice derogates from the autonomy of the chooser, and the apparent consequence that any managerial structure within the public discourse of a democracy represents an externally imposed constraint on the choices of democratic citizens, derogating from the autonomy of citizen choices and making true collective self-determination impossible. Put another way, the application of individualism to the context of true collective self-determination leads to the mistaken assertion that imposing any managerial structure on public discourse constitutes the imposition of a particular way of doing business democratically upon citizens, and hence involves keeping from them options that should instead be genuine objects of choice for them. Against this, I will argue, the constraints that a collectivist would impose on public discourse do not represent a particular kind of democracy among others; rather, they represent the minimal conditions on the true collective self-determination that is the defining mark of democracy itself.

The argument to that conclusion proceeds through three stages. In the first, against the general Wolffian conception I argue that autonomous choice of any kind presupposes the identity of the chooser, and that certain constraints on an agent's choice follow from the agent's identity. In a sense, an agent cannot choose to be, or to do, just anything. At the second stage, I argue that the specific kind of autonomy to which collectivism is committed is the autonomy of citizen as hearer rather than speaker. I then apply the general anti-Wolffian result to the context of hearer autonomy. I argue that hearer-autonomous citizens cannot choose to refuse to consider information that would make a difference to their deliberations over (and eventual vote on) public issues: Their identity entails a commitment to getting as much of such information as possible. At the third stage, I show that the state regulation of speech that the collectivist supports does not constitute a particular national identity among others for democratic citizens, because that regulation secures the possibility of hearer autonomy, and hence democracy itself, in the first place. In other words, instead of regulation representing a particular way of doing business democratically, it represents a minimal condition on democratic business being done at all. If, as Post says, a condition of true collective self-determination is citizen autonomy, then a condition of true collective self-determination is the collectivist managerial structure.


We know that autonomous choice is choice flowing from, rather than imposed upon, an agent. An agent chooses autonomously when she deliberates over the options she is presented with and chooses between them. But, without more, the idea of choice flowing from the agent really makes no sense. To see why, imagine an agent with no prior commitments, no beliefs or desires, no inclinations whatsoever, and imagine that she is offered a choice between two alternatives. The agent cannot choose, for there is nothing about her that would incline her to one option rather than another.

But when we supply something about her that makes choice possible for her, we thereby constrain the choices that she can make. In the most general case, when we supply content to an agent sufficient merely to give her an identity as a kind of person (for instance, a rational agent, as in some economic models), we constrain her to making only those choices that persons of that kind can make. (For instance, the economic agent does not, and cannot, give away more than she receives in an arin's-length bargain.) It certainly does not follow, however, that such constraints on her choices imply that she is not free with respect to the choices she is able to make. Quite the contrary, and as we have just admitted, it is only by virtue of there being some content to the agent herself that choosing is possible at all. In a word, then, true autonomy requires certain constraints on the choices "presented to" autonomous agents. Without some such constraints, choice would not be possible.

Let me illustrate this point with a simple example. Consider a baseball umpire. Behind the plate, he has a highly constrained choice set indeed: He may choose to call a pitch a ball or a strike, but not, for example, imprudent or too hard or inelegant. It would be a mistake, however, to say that the umpire is not free to pronounce the pitch imprudent. Rather, the specification of the choice set that constitutes the freedom of umpires - given the kind of thing an umpire is - just does not include pronouncements as to the prudence or imprudence of pitches. Perhaps it is true to say that the person who is the umpire is constrained not to call a pitch "imprudent" or "prudent" when he is behind the plate: He, that human being, is not free to speak in that way. But the umpire is not, for the purposes of understanding his behavior, a person, and his freedom has nothing to do with such behavior. With respect to that behavior which it is in the nature of the umpire to undertake, the structure of, the constraints on, his choice set, together with the structure of the game, enable his freedom.

It will be objected that, even if this example illuminates the general case of the Wolffian agent, it is irrelevant to the kind of autonomy that is at stake in public discourse. After all, the example presupposes the identity of the umpire: What the example illustrates is that, given an identity, certain choice constraints follow. But part of the idea of political autonomy, in particular, is that the autonomous citizen chooses the kind of (political) life she will lead, and in that sense chooses her identity (or the identity of the polity in which she lives) herself.

This point is right as far as it goes. For instance, we have agreed that democratic citizens must be able to choose, at any time, between the possible ways of doing business democratically. If, as we have supposed, the ways differ substantively (so that whichever way is chosen may have some effect on what decisions are reached), then at any time democratic citizens must be able to discard one such way and to embrace another. That is right. But what the point presupposes is that the choosers are democratic citizens, i.e., citizens who do, and must, choose their "national identity." That is, they must choose how they will conduct their business democratically. To see the force of this point, consider whether democratic citizens, as such, can choose a way of doing their business - e.g., a dictatorship - under which collective self-determination becomes impossible.

I think it is clear that they cannot. Of course, the persons who are citizens can decide to give up democracy. That is, if we think of the choice context as one in which (political) persons choose which kind of political life to lead, then democracy is one way among many of living, and is within the choice set for such persons. Indeed, were that not merely one among a range of choices for those choosers, we would not think they were autonomous. This just reiterates the general point, now contextualized to the case of political persons, that an autonomous agent chooses the way she will be from amongst the possible ways she can be. But a democratic citizen, that is, a political person qua engaged in true collective self-determination, cannot "choose" an identity whose consequence is to render such self-determination impossible. To suppose that such a choice has been made is to suppose away the identity of the chooser as a democratic citizen.

In this contextual sense, the identity of the democratic citizen constrains the choices that she confronts. Post's context is the same as the collectivist's: the context of true collective self-determination. In this context, it is impossible for citizens to elect to be ruled by a dictator. Put otherwise, dictatorship is not a possible "national identity" for democratic citizens, and to say that the option of becoming a dictatorship is a choice that may be "kept" from the citizens does not derogate in the slightest from their autonomy. Quite the contrary, it is an enabling condition on their autonomy as democratic citizens.

The application of our general anti-Wolffian result has yielded the (modest) conclusion that not every possible choice for, e.g., political persons, is a possible choice for democratic citizens. This is a fact that follows from the identity of democratic citizens. In particular, we have seen that, because a dictatorship is not a way of doing business democratically, it is not a possible object of choice for democratic citizens. The general point is that if, for whatever reason, a way of doing political business is not a particular way of doing business democratically, it is not an object of choice for the public discourse of democratic citizens. If now we can show that state regulation of speech, as the collectivist conceives it, does not represent a particular identity for democratic citizens, then we will be able to conclude that it cannot be an object of choice for democratic citizens in public discourse. If, moreover, we can show that it is an enabling condition of democracy itself, then we will be able to conclude that imposing it on democratic public discourse will not impair citizen autonomy. This is what I propose to do in the remainder of this Note.

A. The Autonomy of Democratic Citizens

I have followed Post in talking generally about the autonomy of democratic citizens. But our discussion of autonomy can no longer proceed so abstractly, for autonomy, in the context of democracy, is not a univocal notion. In particular, it is crucial for our purposes to distinguish between hearer and speaker autonomy. I draw this distinction just below. I then argue that collectivism (and hence Post's argument against it, since Post takes collectivism as he finds it) presupposes hearer autonomy. With that result in place, the case against Post is straightforward.

The notion of speaker autonomy is intuitive enough - for present purposes, it suffices to think of it as a liberty right, to be able to say what one wishes publicly.(30)

The notion of hearer autonomy, on the other hand, is rather more complex. To develop it in the requisite detail, I need to introduce some technical vocabulary.

First, public decisions are determined by votes of citizens. Ideally, each citizen deliberates over the various positions from which he must choose. He listens to arguments, sifts data, and consults his own moral and political sense. As he comes to learn new information, he deliberates anew: As new information comes in, he considers it and updates his commitments. Finally, as a result of deliberating, he chooses and publicly acts thereupon by casting his vote.

Second, let us define information as available to citizen A if it is possessed by at least one citizen other than A who would convey it to A (i.e., would convey it publicly) if he could. In the typical case, the information would be a new position, or an argument for a position already public, or data bearing on some position already public, and so forth.

Finally, let us say that information is relevant to a citizen if learning it would make a difference to his deliberations on some public issue, and hence would affect his vote. Note that the notion of "affecting a vote" does not mean "changing a vote"; some received information might only confirm a citizen more strongly in the position he has already adopted.

It is vital to see that the concept of relevance cannot be tied, in abstraction from the needs and circumstances of particular citizens, to any particular informational content. For instance, the relevance of information to a citizen is not measured by what an "ideal thinker" would wish to know or think about in order to vote on an issue. If the issue is nuclear power, we cannot say a priori that information about the long-term health effects of radioactive spills is relevant to any citizen. Of course, we may think that the information is relevant -but this is just to say that the information would make a difference in how we would vote on the issue. Relevance is determined solely by what effects the receipt of information would have on citizens' deliberations.(31) If no citizen's deliberations would be affected by learning of those health effects, then the information is not relevant to any citizen and, derivatively, is not relevant for the collectivity on that issue.(32)

Now, we have agreed that a citizen's choice (hence, his vote) is autonomous if it flows from him, if it serves the citizen in representing himself to the polity. But citizens do not express themselves, at least not fully, if they do not receive relevant information that is available to them. For to say that the information is relevant is to say that there is something about the citizen that the information brings out; the effect of the information is its ability to permit the citizen to "put more of himself' into his deliberations, and hence to permit his vote to express himself fully.

Given these points, we can say that the autonomy of a voter is a function of the information he receives. That is, the autonomy of the voter is measured, not by what he says to others, but by what he hears. Voter autonomy, the autonomy that is at stake in true collective self-determination, and hence which must be the focus of collectivism, is hearer autonomy.(33)

B. Details of Collectivism

My ultimate aim in the argument against Post is to show that the particular kinds of regulations that the collectivist urges not only do not threaten hearer autonomy, but are necessary conditions for it, and hence do not constitute a particular national identity for democratic citizens. But I cannot make that argument until we have the specifics of the collectivist proposal before us. Thus far we know only that collectivism supports some state regulation of speech. But which kinds? We can now fill out the collectivist commitments by seeing how they connect to the concept of hearer autonomy, which (as we have just seen) the collectivist must endorse.

The collectivist's commitment to rich public debate spells out the commitment to true collective self-determination by ensuring that a citizen has before her information that would make a difference in her deliberations over public issues. There are two cases to consider, in which citizens who do not possess such information might come to have it. First, suppose a citizen possesses information that would make a difference to other citizens' deliberations, but that the citizen does not have the resources to get the information out. Thus, for example, the newspapers or electronic media would be prepared to run an advertisement containing her information if she could pay for it, but she cannot. It is clear that the demands of hearer autonomy require the dissemination of this information, and therefore that the collectivist is committed to enabling the citizen to get the information out.(34)

Compare this case to the case in which all the available communication channels are taken up (e.g., all the prominent information conveyors have already sold their advertising space). Here, too, it is plausible that hearer autonomy may require that the citizen be enabled to announce the information - I will argue presently that it does require this in the case where the information is more relevant than information presently being conveyed over at least one communication channel. Assuming for the moment that this is so, collectivists must hold that the citizen must be enabled to convey his information in such cases even though this action would preempt another speaker.

One final point before we proceed: It might be argued that, even granting the collectivist claim that information that could make a difference to citizen deliberations must be released, that obligation (supposing it to be the state's, in the last analysis) is discharged by the provision of public and protected places, such as parks, street comers, public community-access TV/radio frequencies, and the like. As long as these remain protected and available to speakers, any citizen with information to convey may convey it there, and the state thereby discharges its obligation to secure rich public debate, without needing to further extend the public purse to support the dissemination of the information through privately held entities.

This objection deeply misunderstands the collectivist commitment to getting out relevant information as a means of securing autonomy, for it confuses speaker autonomy with hearer autonomy. To see this, consider that the street comers might be empty, and that the public access channels may go virtually unwatched. Even if the state spends significant amounts maintaining these institutions, even if it does so in significant measure to discharge its obligation to secure rich public debate, it does not discharge that obligation unless the information actually gets through to those citizens to whom it would make a difference. Rich public debate is informed debate, and informed debate is debate based upon information that makes a difference in citizen deliberations. Information conveyed by a lonely person on a soapbox to no one makes no difference in the deliberations of any citizen. From the point of view of the collectivist espousing hearer autonomy, such information is worthless, and the state does not discharge its obligation as the collectivist understands it by maintaining such forums.(35)

Of course, the objection may have some resonance from the point of view of speaker autonomy. Though the case is far from clear, one might think that a right to self-expression is not the same as the right to be heard, though it is the right to speak in places where the public might be. To the extent this, is so, one might argue that the state fully discharges its duty to secure and protect the self-expression rights of citizens by maintaining and protecting parks and the like, no matter how poorly attended they are. But the duty to protect and secure hearer autonomy is a more stringent duty: It is not satisfied merely when a citizen who would speak is enabled to speak in a public forum. Rather, it is satisfied only when the speaker who has something relevant to say is enabled to speak in a way that ensures he will be heard. More precisely, the duty to secure hearer autonomy is satisfied only when the state enables the dissemination of the information that the citizen possesses.(36)


Our specific inquiry has now been considerably narrowed. We must ask: Does state regulation, in enabling the announcement of available and relevant information either over an available communication channel, or over a presently occupied channel (so that enabling the announcement would entail silencing some speaker), represent merely a particular way of doing business democratically (a particular "national identity")? Or does it represent, rather, a condition on the doing of business democratically at all? If the former, then Post is right that such state regulation is incompatible with true collective self-determination, and the collectivist cannot have it. If the latter, however, Post mistakes an enabling condition of democracy with a kind of democracy, and his argument fails.

The first point to make in pursuit of the latter conclusion is this: An autonomous choice by a citizen - a vote - can reflect more or less of his identity as citizen, and will be more or less autonomous accordingly. In particular, the greater the amount of relevant information a citizen's vote is based on, the more autonomous it is.

Here is why. Autonomous choice, as we have repeatedly said, is choice determined by the agent. An agent, however, may put more or less of himself into a choice. Thus, choices over what the agent considers to be fairly trivial options are not likely to engage much of his attention or effort. If the issue is what to have for lunch on a given day, he is not likely to weigh, or weigh very seriously, his long-term health or financial interests against his present desire for some, rather than some other, particular meal. Rather, he is more likely merely to take note of his relative preferences and decide accordingly. Even if some deliberations making use of more of his interests would have yielded the same outcome, and even if we suppose that in both cases he is the only final source of his choice, it is reasonable to say that his actual choice expresses less of himself - because it is based on fewer of his interests and commitments than it could have been - than the more deliberative choice would have been.

Similarly for the case of citizen voting: Assume two voting situations for a citizen. In the first she votes on the basis of some (relevant) information, and in the second she votes on the basis of that information and some further relevant information. To say that the extra information is relevant, recall, is to say that it makes a difference to the citizen's deliberations, and hence that there is something about the citizen that the information evokes. That aspect of the citizen plays a role in her second vote that it did not in her first, yet every aspect that played a role in her first vote also played a role in the second. In that clear sense, the second vote expresses more, of the citizen.(37)

There is, then, a range of hearer autonomy, not indeed based on the degree to which external influences tend to precipitate a vote (I assume, following Post, that there are no external influences of that sort), but on the degree to which the citizen expresses himself through his vote. Ideally (hearer-) autonomous votes, as we might call them, are votes based on all relevant information; practically ideally autonomous votes are votes based on all relevant available information; and there are, of course, votes based on even less relevant information.

It might be objected that the idea that autonomy can come in degrees really makes no sense. Autonomy, according to the objection, concerns an agent's power to ensure that her choices flow from her deliberations rather than from external causes. An agent who is under the influence of a demagogue, for instance, might not be autonomous if she is unable to make choices except insofar as the demagogue tells her what to "choose," because she is unable to base her choices on her own deliberations. But either an agent has that power or she doesn't; it doesn't make sense to say that she possesses it to some degree or other.(38)

I think this point is probably right. It does not, however, bear on my claim about degrees of autonomous choice. For the concept of autonomous choice is different from, though related to, the concept of autonomy at work in the objection. The latter concept points to a condition of the agent, required in order that any of her choices flow from her. Being in that condition, however, does not guarantee that any particular choice she makes flows from her, nor does it guarantee that every choice that does flow from her possesses an equal amount of her investment. In short, my claim is that, given the autonomy of the agent, particular choices (each of which, by hypothesis, is based on her deliberations) can reflect more or less of the citizen. If autonomous choice is choice that is reflective of, flows from the agent, then even if an agent is either fully autonomous or not autonomous at all, the autonomy of particular choices can come in degrees.(39)

Now, the collectivist is immediately concerned not with autonomy in the objector's sense, but with the autonomy of particular choices. The underlying concern is with particular acts of true collective self-determination, which are themselves the result of individual choices (votes) of citizens. Though the commitment to collective self-determination presupposes citizen autonomy in the objector's sense, the proper focus of the collectivist, interested in particular acts of self-determination (and hence of Post as well), is on the autonomy of choice. The collectivist focus is on a kind of autonomy that does indeed come in degrees.

With that objection behind us, we are ready to proceed. If collective self-determination presupposes hearer autonomy, then, from the point of view of self-determination, ideally autonomous votes are preferable to practically ideal votes; more generally, the preferability of votes varies with the amount of relevant information over which the citizen has deliberated.

Imagine now two citizens, informant and voter.(40) Both are deliberating over an issue in public discourse. Informant possesses relevant information that voter lacks but not vice versa), and she would publicly announce it if she could. Finally, there are "spaces" available in communication channels that would reach voter. It follows from the present discussion that, from the point of view of true collective self-determination, the vote of informant is preferable to the vote of voter. But this is just to say that, from the point of view of collective self-determination, the state of affairs in which voter learns the information from informant is preferable to the one in which she does not. There is, therefore, no objection to enabling informant to convey (publicly) the information to voter.

This conclusion follows, we must keep in mind, from the content of the identity of democratic citizens autonomously choosing, and is thus true no matter what any such citizen believes or would agree to. Put otherwise, the following is not a choice for democratic citizens: The announcement of available relevant information over free communication channels should be enabled. For such enabling, where possible, is a condition of the collective self-determination that defines democracy, and hence does not merely specify a way in which business gets done democratically. In short, commitment to enabling such announcements is not a commitment to a particular way of doing business democratically. It is, rather, an essential feature of any polity that could count as a democracy.(41)

Does it follow that the state may enable announcements such as the one by informant? To say that the option to enable such announcements is not the object of citizen choice is to say that the permissibility of the state's enabling such announcements does not depend on the approval of any citizen. Thus, the state cannot be said to do wrong in enabling the announcements solely by virtue of its not having been authorized by citizens to do so. Moreover, to say that the option is not an object of choice for citizens is to recognize the legitimacy of a managerial structure external to citizens by which their identity is "secured." In political theory, this structure (no matter how extensive) just is the state; to recognize the legitimacy of this part of managerial structure is to affirm that the state may enable the announcements.(42) It would also seem to follow that it is permissible for the state to announce information that is not available (i.e., not possessed by any citizen) on an otherwise free communication channel, if the information is relevant. But I will not pursue this possibility further.

I turn now to the more difficult, and crucial, case, in which one citizen is silenced in order that another citizen can speak. Before I develop the case, however, I need to make a further point about the concept of relevance, viz., that relevance can come in degrees, just as autonomy can.(43) To see how this can be, consider a citizen who is highly inclined to a position for which a given piece of information (p) is an argument, but who needs to have the argument run through another time, while some new available information (q) propounds a position very remote indeed from any the citizen would be inclined to vote for. Receipt of q by the citizen would have only a marginal effect on her deliberations. But given that q would have some marginal effect, we must say that it is, to some degree, relevant to her. Receiving anew the information (p) that she already possesses, however, will have a greater effect on her deliberations than would receipt of q, notwithstanding that receipt of p will not increase the amount of information she possesses.(44) Information p, in other words, is more relevant to the citizen than is information q.

The crucial case is the one in which there are two different pieces of information (let me say: two different messages), both available and relevant to citizen A. The first is being repeated over a given communication channel, while the second is unknown to A. Both messages are relevant to her, but Oust the reverse of the illustration in the last paragraph) the repeated message is less relevant to her than is the unknown one. Finally, there is no free communication channel available for announcing the first message to which A is likely to listen.

Because this case is indeed the crucial one, it will be useful to fill it out in some detail. Let us then borrow our example of nuclear power from the town-meeting context.(45) Before citizens is the issue of whether to authorize the building of a nuclear power plant. Proponents of the plant have flooded the area's communication channels with information on the salutary effects of building a plant. The issues of potential health and environmental effects have barely been touched; the advertisements merely mention that nuclear energy is "clean and safe." Most citizens, let us suppose, are convinced, but not irrevocably. They are prepared to listen to new arguments, would consider them, and are prepared to deliberate further. Suppose that a citizen group, having done research on nuclear accidents, argues that the health and environmental effects are potentially more serious than the proponents would acknowledge. Its funds are limited. It publishes a newsletter whose circulation is modest, reaching many fewer citizens than the number for whom the information it contains is relevant (and anyway, citizens are not inclined to read newsletters that show up in their mailboxes). Finally, suppose that a good number of citizens are more likely to have their deliberations differentially affected by the group's information than by repetition of the pro-plant message.(46)

The claim that the group should not be allocated, say, television time, on the grounds that they already have a newsletter, is unavailing. For, as we have seen, what matters for hearer autonomy is that (enough) citizens hear the message, and not merely that the group be able to say it "aloud." Further, I do not mean to suggest that the group's information would have a greater effect on citizens in the sense that it would change more minds than would the rebroadcast of the pro-plant message. The picture, rather, is this: We suppose that the pro-plant message, simply a repetition of what citizens have already heard and believe, does not cause (many of) them thoroughly to think through the issue again, and in that sense makes little difference to their deliberations. The group's information, on the other hand, is likely to be sifted, considered, by many citizens. Perhaps many of them ultimately will discount it (some will judge the source not to be objective; others will think the newly discovered detriments are outweighed by the benefits). Still, more citizens' deliberations are likely to be affected by the group's information than by the pro-plant message.

May the state regulate this situation, providing for the pro-plant message to be preempted in favor of the group's message? Our discussion shows that it may. For the group's information is more relevant than is the proponents' message (now). It follows that the autonomy of the citizens' ultimate choice whether to build the power plant is enhanced by silencing the proponents in favor of the group. From the point of view of collective self-determination then, silencing the proponents is permissible.

More generally, the argument is this. The case presents the problem that enabling one speaker, A, to speak requires silencing another, B. It is permissible to silence B in favor of A if, for enough citizens, what A has to say is more relevant than what B has to say. In such a case, it follows from the identity of democratic citizens that autonomous choice is enhanced by the silencing of B in favor of A; therefore, silencing B is permissible. Of course, B is silenced. But the value of B's speaking in the first place, under the democratic model of hearer autonomy presupposed by Post, is the contribution it makes to collective self-determination. Here, silencing B in order to let citizens hear what A has to say contributes more to (hearer) autonomy than not silencing B, and is thereby permissible.

Once again, this permissibility does not rest on an agreement between citizens.(47) It is not, therefore, an object of choice for them. Thus, it is also not an option available for choice in public discourse and does not represent a way of doing business democratically. In short, rather than being a particular national identity for democratic citizens, the permissibility is a condition on doing democratic business at all. Post's argument, supposing to the contrary, is therefore mistaken.


It will be objected that this result may actually be inimical to the collectivist theory, once the latter is fully spelled out. It appears that, in practice, given that citizens saturated with a "popular" message are more likely to react, in some sense deliberatively, to a repetition of the message than to receipt of altogether different, and unfamiliar, messages, my argument appears to commit the collectivist to "drowning out" those who would announce the latter, no matter how "worthy" their message. But this, it seems, is the very position that collectivist theory would oppose, committed as it seems to be to ensuring a fair hearing for as wide a range of information as possible.

I think this cannot be right, in part because this cannot be the right take on the collectivist theory, notwithstanding the objection's initial appeal. We can admit that the theory is concerned to enrich public debate by ensuring that more information is injected into it - and to ensure that public discourse is not saturated by a small percentage of available information, continually announced by citizens with privileged access to communication channels. But the presupposition behind this position, surely, is that the information whose announcement is thus enabled will make a difference to the deliberations of citizens. The idea of rich public debate is not merely the idea of debate informed by more rather than less information. It is, rather, the idea that citizens consider as much information as would make a difference in their deliberations as possible. It is no part of the collectivist ideal to enable a citizen to announce publicly information that would fail to enter into the deliberations of any citizen. The crucial, and plausible, assumption that the collectivist theory makes is that there is some unpopular information that would have more of an impact on citizens than would the repetition of popular information. Such information, under certain conditions, permissibly may preempt the announcement of more popular, more familiar, information.

What matters to the collectivist is that market forces, or other forces, could conspire so to flood the forum of public debate that information that would make a difference to citizens-information whose receipt by citizens would enable them to express more of themselves in their voting - is drowned out by the announcing of less relevant information. Only in such cases does the collectivist countenance the state's preempting some speech in favor of some other speech.(48)


It is a mistake to identify decisional autonomy with absence of constraints on choice. With that mistake behind us, we can understand better the relation between state regulation in a democracy and democratic autonomy. We can see that not all autonomy proceeds from the citizen outward, as liberties seem to do; some autonomy depends upon inputs to the citizen from outside her, and hence depends upon a kind of assistance from others. This insight prepares the way, against much initial resistance, for the idea that the state can as much be an enabler of freedom as it can be a threat to freedom. It prepares the way for the idea that freedom may be secured, in some cases at least, by the state and the citizen conspiring together. (1.) Robert P. Wolff, In Defense of Anarchism 13 (1970). (2.) Id. The resilience of this view to criticism attests to its power. Some of its many critics include Leslie Green, The Authority of the State 29-36 (1988); Roberto M. Unger, Knowledge and Politics 277-81 (1975); and Allan C. Hutchinson & Patrick Monahan, Democracy and the Rule of Law, in The Rule of Law: Ideal or Ideology 97 (Allan C. Hutchinson & Patrick Monahan eds., 1987). (3.) This discussion appears most prominently in Robert Post, Managing Deliberation: The Quandary of Democratic Dialogue, 103 Ethics 654 (1993) [hereinafter Post, Managing Deliberation] and Robert Post, Meiklejohn's Mistake: Individual Autonomy and the Reform of Public Discourse, 64 U. COLO. L. Rev. 1109 (1993) [hereinafter Post, Meiklejohn's Mistake]. (4.) See, e.g., Alexander Meiklejohn, Political Freedom: The Constitutional Powers of the People (1960). (5.) A representative statement of Fiss' view can be found in Owen M. Fiss, Why the State?, 100 Harv. L. Rev. 781 (1987) and Owen M. Fiss, Free Speech and Social Structure, 71 Iowa L. Rev. 1405 (1986) [hereinafter Fiss, Social Structure]. (6.) See in a part III. In later work, Post makes it clear that he is not, in the end, a Wolffian individualist as to what it is to be a democratic citizen. This is, however, because he thinks that institutions designed for social order are pulled in opposing ways by principles that have incompatible "internal logics" - these include the principle that social ordering requires management structures and the principle that at least some community values must be institutionalized. These principles, he suggests, run counter to the basic requirements of both individual and collective autonomy, and yet must somehow have a place in even democratic orders.

It is beyond the scope of this Note to develop these themes. At issue here, in any case, is Post's view of the "internal logic" solely of the principles of democracy. It is here, I shall show, that Post is committed to Wolffian individualism. He writes,

The reconciliation of individual and collective self-determination entails a serious tension. On the one hand, a democratic social structure must provide an appropriate space for individual autonomy. Within that space democracy must function negatively; it must refuse to foreclose the possibility of individual choice and self-development by imposing preexisting community norms or given managerial ends. On the other hand, a democratic social structure must also function positively, to foster an identification with the processes that enable the collective experience of self-determination .... These processes presuppose forms of social cohesion that depend upon community norms and ... [they] often require strategic managerial intervention.

There is thus a paradox at the center of democracy. Democratic theorists tend to finesse this contradiction by concerning the community and managerial structures necessary for democratic social cohesion as voluntary ....

Robert Post, Constitutional Domains 7-8 (1995) (emphasis added). That is, the "managerial intervention," introducing a coercive element into the process of self-determination, can be "finessed" - can be made to seem inoffensive to democracy - if they are conceived of as voluntary. I discuss Post's (7.) Fiss, Social Structure, supra note 5, at 1411. (8.) Of course, convictions of this sort, ones that are resilient to new information (indeed, that repel new information), are not necessarily the result of external causal processes. But many, perhaps most, of them are externally caused. (9.) This is not the place to expound the notion of deliberation. Suffice it to say that I do not mean to denote by "deliberation" necessarily philosophical, or geometric argument. Deliberation involves a citizen consulting herself - considering the options in light of her preferences and her judgments about the political good. For some citizens, deliberation will be an intuitive affair, more a reflection on how they feel about the options than an empirical speculation about likely outcomes. I do not mean to suggest that such intuitive approaches are not fully deliberative for the purposes of determining self-expressive, hence autonomous, choice. (10.) The concept of cause here is difficult to spell out precisely: As I will discuss, citizens are subject to all sorts of influences, only some of which raise questions about the autonomy of their choices. (11.) Post, Managing Deliberation, supra note 3, at 660. (12.) id at 660-61. (13.) Id at 672. (14.) See Meiklejohn, supra note 4, at 24-28. (15.) Id. at 26. (16.) Id. at 25. (17.) Id. at 24-25. (18.) Debate rules are only one example of rules that bind the town meeting. As I indicate in a moment, a town meeting may adopt more substantive rules as well, ones that are more restrictive of speech. (19.) Post, Managing Deliberation, supra note 3, at 658. (20.) I do not mean to suggest that which constraints are procedural and which are substantive will be perfectly clear. Nothing will turn, however, on our being able to make any such distinction. (21.) Post, Managing Deliberation, supra note 3, at 658. (22.) I do not mean to suggest here that Post would inquire an explicit act of agreement, as though the entire population had voted on the procedures at issue. Rather, it is not incompatible with his position that there could be behavior over an extended period of time that could be construed as agreement by citizens, even though no explicit vote had ever taken place. In short, I do not mean to saddle Post with any theory of the requirements of citizen agreement. As will become apparent, my argument applies to whatever theory Post would use to fill out his view. (23.) Post, Managing Deliberation, supra note 3, at 659. (24.) Id. at 662 (emphasis added). (25.) Id. (26.) Post, of course, does not hold merely that no such actual agreement has ever been made - as though the occurrence of a prior agreement would validate the existence of managerial structures governing public discourse now. One can imagine that a suitably framed referendum could provide grounding for the collectivist theory. Yet, Post surely is not urging merely that the collectivist theory can legitimately be implemented, but only via some referendum process. For, not only does he insist that at every point in public discourse "all possible objectives ... be open to inquiry," and that autonomy, "properly understood, signifies that within the sphere of public discourse and with regard to the suppression of speech the state must always regard collective identity as necessarily open-ended," id at 666 (emphasis added), but also that "[w]e can decide, within public discourse, to form and set in motion specific organizations of order and instrumental rationality, like town meetings. But it would be a grave mistake to confuse these discrete institutions with the sea of tumult and discord that is public discourse itself" Id. at 663. It would, that is to say, be a grave mistake to suppose that orders representing instrumental rationality could ever be adopted by public discourse to constrain itself. (27.) This view does not require that change has taken place. It suffices that, at any point, it could have. (28.) It is a "communicative process[] ... [that] instill[s] in citizens a sense of participation, legitimacy and identification." Post, Meiklejohn's Mistake, supra note 3, at 1115. (29.) It is worth noting that Post himself is unclear on just how radical a result he intends here, although the abstractness of his argument masks its ambiguity. On the one hand, he writes:

While acknowledging that "the voting of wise decisions" must be kept from


interference, [Meiklejohn's model] ... authorizes the censorship of public

discourse on the

basis of assumptions about function and procedure. Meiklejohn cannot appeal

to a neutral

distinction between substance and procedure to justify this contraction of

the scope of

democratic self-government, for the procedural assumptions he wishes to

enforce, no less than

substantive ones, are ultimately grounded upon a distinctive and

controversial conception of

collective identity. His paradigm of the town meeting specifically

presupposes that the function

of American democracy is to achieve an orderly, efficient, and rational

dispatch of common

business, and it consequently implies that aspects of public discourse

incompatible with that

function are constitutionally expendable. To the extent public discourse

is thus truncated, a

particular concept of national identity is placed beyond the reach of the


processes of self-determination. Id. at 1117. On this view, even the "debate rules" that in part constitute the town meeting, representing a way in which (but not the only possible way) business gets done democratically, constitute a national identity and thus must ever be up for grabs in public discourse. Though the adoption of such rules may be supposed to remain in place when citizens meet, acknowledging their identity as town-meeting citizens, they cannot remain in place (i.e., be imposed by an external manager) where the issue of national self-identity (whether or not to be town-meeting citizens at all) is itself up for grabs - as it always is in public discourse.

On the other hand, Post's target seems at times to be narrower:

What follows ... is not that public discourse can never be regulated, but

that it ought not to

be managed in ways that contradict its democratic purpose. This purpose need

not preclude

"time, place and manner regulations" that function as "rules of the road" to

co-ordinate and

facilitate expression within public discourse. Nor need it rule out

government action designed

to supplement or augment communications within public discourse, as for

example by

establishing state supported forums to enhance public debate. Id. at 1121. Here, the idea seems to be that if there are certain "merely procedural" rules that would regulate public discourse without imposing thereby an identity on the citizens constrained by the rules, citizens may be subjected to them without losing their autonomy. According to this view, perhaps a town-meeting structure that is sufficiently thin, consisting only of "time, place and manner" restrictions, might not need to be subject to continual citizen approval as a condition of their being enforceable by a manager. Only thicker rules, for example, rules stipulating which types of decision are subject to majority rule and which are not, that reflect a "substantive" character of public discourse, must ever be open to citizen (re)approval.

At some points, Post appears to reject the procedure/substance distinction altogether, and this suggests the more radical posture - indeed, I suspect this is his preferred position. At any rate, we know that the sort of regulation that collectivism supports does represent, for Post, a particular national identity. To put the point starkly: Post thinks that collectivism represents one possible national identity for a democratic polity but that there are other possible identities. I argue in the text, however, that this is one source of his mistake. There are no other possible identities that collective self-determiners can have; hence, collectivism does not represent one national identity among many for democracies. (30.) Of course, there are constraints on such rights - such as that one speaker cannot, all things being equal, exercise his right to say what he wishes publicly in such a way as to deprive another of being able to exercise her right to do the same. (31.) Thus, in any particular situation, the relevance of given information will be determined in part by its content. This is because citizens react differentially to information precisely based on its content. But what content a citizen reacts to, i.e., whether or not a piece of information will make a difference to a citizen's deliberations, is dependent upon the states (beliefs, desires, etc.) of the citizen. Though relevance does have to do with informational content (that is, a given piece of information will or will not make a difference to a citizen's deliberations depending in part on its content), there is no "privileged" content that can be said to be, in the abstract, relevant.

To stress what is perhaps obvious, this technical notion of relevance has at best a contingent connection to the ordinary notion of relevance, which is intimately connected to considerations of (relatedness of) meaning. The technical notion is a strictly causal notion: Information is relevant if it has a certain causal effect on its recipient. Of course, in the context of this discussion, where we have assumed away ideological oppression and the like, but see in a note 48, information that is likely to have an effect on a citizen's deliberations is also information that is most likely to be relevant, in the ordinary sense, to information that she presently possesses and which figures in her deliberations. But perhaps not - the crucial point is that, as an analytical matter, this need not be so. (32.) The saturation scenario rears its head at this point, though it will not be pertinent to the discussion of autonomy that follows in the text. For in the case of "informationally oppressed" citizens, it may well be that the cause of the fact that some information is not relevant to them is simply that they have been manipulated in a certain way. We might think that, in some sense, the information is nevertheless relevant to them: But for their being manipulated, the information would have made a difference in their deliberations. This claim, however, raises deep and troubling problems about how to think about the citizen who does the deliberating. For instance, is the "real" citizen, i.e., the real deliberator, the one that we imagine, shorn of undue influences? Or is he the one who actually is before us? I think a full treatment of collectivism must take this issue on and settle it. But because Post operates in the rarefied air of autonomy, where there are no external causes of citizen deliberation (except, possibly, external causes adopted by the citizens themselves-this will be the point of contention between Post and me), we need not pursue this point further. (33.) Thus, Fiss, somewhat misleadingly, writes: Speech is protected when (and only when) it [enriches public debate], and precisely because it does, not because it is an exercise of autonomy. In fact, autonomy adds nothing and if need be, might have to be sacrificed, to make certain that public debate is sufficiently rich to permit true collective self-determination. What the phrase "the freedom of speech" in the first amendment refers to is a social state of affairs, not the action of an individual or institution. Fiss, Social Structure, supra note 5, at 1411.

Hearer autonomy is itself, of course, a social notion - it implicates both an information recipient and an information provider. Fiss, it is plausible to say, is saying here only that speaker autonomy, that autonomy most naturally associated with the value of self-expression, might have to be sacrificed in the name of rich public debate. There is good reason to drink he would reserve the same role for hearer (that is, decisional) autonomy as I do. But to write generally of autonomy as Fiss does here at least, I think, gives the appearance of a more radical claim than collectivism needs to make. In particular, putting the point this way threatens to obscure the vital link between the decisions of democracy and the actual decisions of individual citizens. (34.) How much of a difference the information must make in the citizen's deliberations, i.e., how relevant it is, and to how many citizens, will make a difference to the collectivist commitment on this point. Though these are important details, they may be skipped over for our purposes. (35.) If the parks and streets were teeming with wilting listeners, however, the case would be very different. (36.) There may be good reason for permitting the speaker himself to convey the information. For there is no doubt that the impact of information is a function of, among other things, the way it is presented. If the full import of some information is less likely to be conveyed to citizens by, e.g., a state-produced announcement than it is by the speaker - if, that is, the state-produced announcement, on the same matter, is likely to make less of a difference to citizens' deliberations than is an announcement by the speaker - then hearer autonomy may well require not only that the information be disseminated, but that the speaker himself be enabled to disseminate it. (Perhaps some intuitions about self-expression coalesce with intuitions about rich public debate around this point. This would be a topic worth pursuing.) (37.) Admittedly, the illustrative case of choosing what to eat for lunch is not on its face strictly analogous to the case just described. In the lunch case, the agent is in "control" of how much of himself he is involved in his choosing, while this is not so in the second case, where the citizen simply does not possess (yet) one relevant proposition. But that difference is really beside the point, at least on our assumption (following Post) of the autonomy of the citizen. For the assumption requires the stipulation that whether or not the extra proposition is received is not determined by "external" controls that implicate the citizen's freedom.

We have glimpsed this point already, in Part I, and I will not develop it further here. The distinction between "natural" constraints on information and "coercive" constraints is important and would need further development in a full treatment of collectivism. Suffice it to repeat here that, unless the notion of citizen autonomy is to have no practical import at all, the absence of some information from a citizen's deliberations as a result of natural constraints no more implies that her vote lacks autonomy than does the absence of information about her long-term health interests imply that her decision about what to eat for lunch lacks autonomy. (38.) Post has suggested this criticism to me in correspondence. (39.) I do not mean to suggest a metaphysical view of the self as recipient of information, as though the self remains the same throughout deliberations, and as though the only variability in its expression depends on variations in its empirical circumstances, as it comes to know of them. Sometimes, of course, this is roughly the correct picture - as the example of the citizen voting for the unfaithful candidate shows, see supra part I; for there the information that the candidate was unfaithful brings into play commitments that the citizen already had. At other times, however, deliberations on new information will yield, for the citizen, new commitments. In that case, his vote will be, in the sense required in the text, expressive of the citizen only insofar as it is based on a commitment that the information itself helped to bring about. In short, relevant information can serve to shape the self that expresses itself through its vote as well as it can serve to engage, and express, a self already "in place." (40.) Informant is also, of course, a voter. (41.) I do not mean to suggest that such enabling should occur no matter the cost. A fuller treatment of these matters would have to consider how costs would, in some way or other, have to be weighed against the interest democracies have that the votes by their citizens be informed by information relevant to them. (42.) Perhaps there are no such "pure" cases of state enabling, for in any real case, the state will have to expend public funds to enable the announcement, and thus will expropriate the speech of some citizens for whom the information is not relevant. This may be right; but it simply makes the point that every case of enabling of the announcement of available information is in fact a case of silencing some speech to enable some other speech. I treat this case immediately below. (43.) This point was implicitly relied upon above. See supra note 34 and accompanying text. (44.) This example shows that, given our implicit assumption that the amount of deliberation by a citizen is reflective of the degree of autonomy of her resulting choice, the degree of autonomy of that choice does not necessarily vary with the amount of information that the citizen possesses. This is because, in some cases, deliberation is enhanced by the receipt of already-possessed information, rather than by the receipt of new information. (45.) I develop this example roughly along the lines of the facts in Syracuse Peace Council v. FCC, 867 F.2d 654 (D.C. Cir. 1989). (46.) What counts as a "good number, of citizens? More generally, how is a state to take the crucial measure of relevance that is the basis for its speech regulation?

These are questions of detail that the collectivist must work out. It is at least important for us to see, however, that it will not do for the state to assume some sort of "representative" citizen, and to conduct its relevance inquiry with respect to him or her ("how would a representative citizen respond to this information?"). For the crucial part of the present story is that relevance is pegged to actual citizens, whose deliberations vary in ways that reflect the differences between them. The state as regulator must take these differences seriously, and thus must make its relevance assessments across actual citizens.

Finally, it is worth noting that if the anti-plant message is more relevant to most citizens, the issue of preempting its promulgation would be settled: Such preemption would be impermissible. (47.) See text preceding note 41. (48.) What if, in the real world, the continual receipt of given information by citizens desensitizes them to new information, and makes them especially sensitive to repeated information? That is, what if in the real world, the more familiar information is, the more relevant it is, and the less familiar, the less relevant? On this assumption, the collectivist theory, under the present argument, would very rarely permit the state to silence the purveyor of often-repeated information in order to enable the announcement of new information. This was the problem presented in the saturation scenario of Part 1: We can call it the problem of ideological oppression. Its assumption is that the sheer repetition of messages itself desensitizes citizens - by impairing their deliberative abilities. The result is that the repeated messages are more relevant to the citizens than are alternative messages that, intuitively speaking, we think they should hear.

Though a full treatment of ideological oppression is beyond the purview of this Note, it bears pointing out that my argument does not lead to the consequence that, for the collectivist, the state may not preempt the repeated messages for alternative ones, even though, by hypothesis, the former are more relevant than the latter. For the assumption that drives the argument is the existence of basic autonomy (i.e., the power to base decisions on deliberation), which precisely is absent in the situation of ideological oppression. Where there is no autonomy, a supporter of collective self-determination is plausibly committed to producing it. That commitment may well have very different consequences for the regulation of speech than the commitment to hearer autonomy, in a context in which basic autonomy is presupposed.

It is worth bringing up the matter of orthodoxy in this context. A strain sometimes heard-in the Supreme Court, and repeated in collectivist literature, is that the robustness of public debate includes, as a fundamental element, ensuring that ideas challenging orthodoxy are in circulation. The concept of orthodoxy, however, is troubling. Consider that it is no part of the hearer-autonomy model that all ideas, no matter how marginal, must have their place in public discourse. Perhaps truly marginal ideas have a role to play in self-expression, but not so for collectivism. This is because, by their nature, truly marginal ideas are not relevant - they make no difference in citizen deliberations.

That is clear enough. Thus, if collectivism holds that ideas challenging orthodoxy must be injected into public debate, it must be because it supposes that, for hearer-autonomous citizens, ideas that challenge are always relevant. Perhaps, indeed, the driving assumption is that orthodoxy itself causes laxity of thought and a degeneration of deliberation, and hearing heterodox views will serve to combat that tendency. But then, the worth of heterodox ideas to the collectivist will lie, not in the mere fact that they are heterodox and that they therefore "enrich" the mix of ideas in public debate, but rather in the fact that they do figure in citizen deliberations.

What if, however, orthodoxy, in a given context, represents citizens' best judgments about what is worth deliberating about? In such a context, the fact that an idea is orthodox is a sign that it is relevant; and if an idea is not orthodox, then it is probably not relevant. Admittedly, this is a remote possibility. But it does serve to show that the connection between hearer autonomy and the presence of heterodox ideas in public discourse is not a necessary one.

Unless, of course, the idea is that orthodoxy causes deliberations to cease. Then the injection of "marginal" ideas into public debate is not premised on deliberation, but on the hope that deliberation will be enabled. To the extent this is so, the collectivist must stand both in the world of hearer autonomy, where state regulation of speech protects autonomy, and in the world of ideological oppression, where regulation produces autonomy. These are very different bases for regulation, which understand the citizen very differently. A thoroughgoing collectivism must develop both bases; I have, in this Note, attended only to the first.

Finally, a related problem concerns the state's reliability in making relevance determinations. In a sense, this is the very great problem of tyranny, to which collectivism must also have an answer. It is enough to note here, however, that the problem does not serve as an objection to the present argument. It is a problem, rather, for democracy itself.
COPYRIGHT 1995 Yale University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1995 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Lipson, Morris
Publication:Yale Law Journal
Date:Jun 1, 1995
Previous Article:The Rushdie incident as law-and-literature parable.
Next Article:Compound-complex criminal statutes and the Constitution: demanding unanimity as to predicate acts.

Related Articles
Homes Not Jails.
China, Hong Kong, Taiwan (China).
The truth about freedom.
China official fails to rule out more rulings on H.K. constitution.
Women's control over their sexual lives: most women in Namibia are not free to make their own decisions about their sexual and reproductive lives.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters