公 法 评 论

et revelabitur quasi aqua iudicium et iustitia quasi torrens fortis


volume 13, number 1| fall 2000


Liberalism and Democracy in Habermas, Rawls, and Constant
Robert S. Taylor

A. Introduction

In the past quarter century we have witnessed a remarkable transformation of the world's political landscape: authoritarian regimes in Latin America, Africa, East Asia, and the former Soviet Bloc have given way to democratic ones. This global wave of democratization continues apace and promises even further reductions in interstate military conflict, owing to the historical tendency of democratic regimes not to go to war with one another. [1] Democratization has also contributed to an increased respect for human rights, though the relationship has not been a consistent one. Political regimes in Russia, Peru, and other parts of the world that are now characterized by universal suffrage and competitive elections have often shown disregard for such fundamental human rights as freedom of speech and security of property. In short, liberalism and democracy have advanced together, though not in lockstep.[2]

Despite such evidence that liberalism and democracy can sometimes diverge, two of the greatest liberal political theorists of our time-John Rawls and Jürgen Habermas-have argued that these two concepts are in fact equivalent. In his book Between Facts and Norms, Habermas develops what he calls the "Co-Originality Thesis" (hereafter referred to as "CT"), which holds that liberalism and democracy are internally related and mutually supporting. Rawls strongly endorses the CT in his "Reply to Habermas" in Political Liberalism. Their shared belief in the conceptual equivalence of liberalism and democracy contrasts sharply with the position of, among others, Benjamin Constant. Constant, a Franco-Swiss novelist and political writer of the early nineteenth century, argued that "we can no longer enjoy the liberty of the ancients, which consisted in an active and constant participation in collective power. Our freedom must consist of peaceful enjoyment and private independence," which he called the "liberty of the moderns." [3]He insisted upon the distinction between liberalism and democracy, between human rights and popular sovereignty, and would therefore have denied their equivalence.

In this paper I critically examine both Habermas's and Rawls's defenses of the CT and contrast their positions with that of Constant (with whose work Rawls is quite familiar). Part B focuses on Habermas and is composed of two subsections. In the first subsection I argue that Habermas succeeds in making democracy perfectly consistent with liberalism only by building liberalism into its very definition. I show that Habermas's use of these two terms is at variance with contemporary English usage and that such usage allows us to make certain useful political distinctions, distinctions made impossible by Habermas's (re)definitions. Then in the second subsection I suggest that both Habermas and Rawls confuse idealized conceptions of popular sovereignty with popular sovereignty as it is actually practiced. This confusion leads them to ascribe liberal qualities to popular sovereignty that it only necessarily possesses in its idealized form.

Part C centers on Rawls and consists of two subsections and an excursus. In the first subsection I argue that Rawls's defense of democracy, which regards democracy as primarily an instrumental good (i.e., as a means to promote liberal political values rather than as something valuable for its own sake), sits uneasily with his endorsement of the CT. This tension arises because an instrumentalist defense of democracy makes the relationship between liberalism and democracy a contingent one, whereas the CT denies any such contingency, as we shall soon see. In the second subsection I argue that Rawls's support for political representation, which follows directly from his instrumentalist defense of democracy, succeeds in making human rights and popular sovereignty compatible only at the cost of draining the latter of much of its substance. In an excursus to this subsection, I examine Jean-Jacques Rousseau's and Immanuel Kant's sharply contrasting views on the issue of representation and show that Kant's attitude toward popular sovereignty bears a striking resemblance to that of Rawls, who has always been a devoted student of Kant's moral and political philosophy.

Part D concludes by reiterating the theme of the paper-namely, that real and important differences between liberalism and democracy exist, differences that may at times make them incompatible with one another-and by considering the relevance of this theme to the structure of both domestic and foreign political institutions.

B. Habermas and the Co-Originality Thesis

1. The Thesis Itself: Building Liberalism into the Definition of Democracy

Habermas offers a succinct statement of the Co-Originality Thesis in the following passage from Between Facts and Norms:

The sought-for internal relation between popular sovereignty and human rights consists in the fact that the system of rights states precisely the conditions under which the forms of communication necessary for the genesis of legitimate law can be legally institutionalized.[4]

In other words, a regime is not genuinely democratic if it fails to protect liberal rights, because such rights are the very preconditions of democracy. The common term "liberal democracy" is, for Habermas, a redundancy.

Two points need to be made regarding this conflation of liberalism and democracy: first, Habermas's use of these terms is at variance with contemporary English usage; second, such usage is (at least in this case) quite sensible. As to the first point, consider the following definitions of liberalism and democracy from Merriam-Webster's Collegiate Dictionary:

Liberalism-2 b: a theory in economics emphasizing individual freedom from restraint and usually based on free competition, the self-regulating market, and the gold standard; c: a political philosophy based on belief in progress, the essential goodness of the human race, and the autonomy of the individual and standing for the protection of political and civil liberties.

Democracy-1 a: government by the people; especially: rule of the majority; b: a government in which the supreme power is vested in the people and exercised by them directly or indirectly through a system of representation usually involving periodically held free elections.[5]

As commonly used, liberalism and democracy are possible answers to two very different questions: respectively, "what is government to do (subject to what limits)?" and "who is to control government?" Habermas wants to argue that the first question is really part and parcel of the second, a move that creates more heat than light.

Continuing with the second point, the common definitions of liberalism and democracy allow us to make some helpful distinctions that are lost in Habermas's account. Consider, for example, the following two "hybrid" forms of government: illiberal democracy and liberal nondemocracy. Habermas would consider both of these categories to be empty ones, but such a claim would be merely confused:

Illiberal democracy: We can easily imagine regimes that, while democratic, fail to protect some human rights (e.g., religious liberty) because the citizens are themselves intolerant and their illiberal policy preferences are efficiently realized through the democratic process. Consider, for example, the ancient Athenian polis, which had a political regime described by George Klosko as "probably the most democratic Western civilization has seen." [6]All male citizens of a certain age were expected to participate in the Assembly and serve on juries. This direct democracy, characterized by robust and open debate, was also responsible for the execution of Socrates on grounds of impiety: while political speech was generally protected, speech that called into question the tenets of the civic religion was often not. Such time-intensive political participation was made possible, of course, by the unremunerated labor of slaves. Thus, the ultimate expression of popular sovereignty (direct, participatory democracy) was made possible by the ultimate degradation of human rights (slavery)-a terrible irony not lost on Rousseau. [7] Less extreme, more contemporary examples of illiberal democracy might include such countries as Russia and Peru.
Liberal nondemocracy: We can also easily conceive of regimes that, while basically liberal, are nondemocratic. The "enlightened despotisms" of the Philosophe imagination come to mind, though real-world examples are also readily identified. Consider, for instance, Hong Kong under British rule, which long enjoyed a very liberal economic policy in addition to the rule of law and the protection of fundamental civil liberties (e.g., free speech and press). Only in the last years of colonial rule, under Governor Christopher Patten, were the residents of Hong Kong granted any degree of political autonomy.
None of this is meant to deny, of course, that liberalism and democracy are intimately related to one another. As Benjamin Constant noted (perhaps a bit too strongly), "Political liberty is [individual liberty's] guarantee, consequently political liberty is indispensable." [8] Constant is offering an instrumentalist defense of democracy: the political empowerment of citizens is not an end in itself but is rather a means to promote liberal values. Giving citizens some method short of revolution to check the behavior of government officials is often an efficacious way to curtail the abuse of power and protect human rights, and competitive electoral systems provide one such check. No reasonable person would deny this point. What is instead denied is Habermas's radical notion that liberalism can and should be subsumed within the allegedly more general category of democracy. The two are conceptually distinct, and their conflation by Habermas is ill-advised.

2. Popular Sovereignty: Confusing the Actual with the Ideal

Both Rawls and Habermas have a tendency to confound two other things that should be kept distinct: their idealized notions of popular sovereignty (the Original Position and the Ideal Speech Situation, respectively) [9] and actual popular sovereignty (popular sovereignty as realized in existing representative democracies). Consider the following passage by Habermas:

According to the Discourse Principle, just those norms deserve to be valid that could meet with the approval of those potentially affected, insofar as the latter participate in rational discourses. Hence the desired political rights must guarantee participation in all deliberative and decisional processes relevant to legislation and must do so in a way that provides each person with equal chances to exercise communicative freedom to take a position on criticizable validity claims.[10]

The first sentence in this passage describes how norms can be validated (i.e., how moral/legal principles can be substantiated and ratified) through an Ideal Speech Situation consisting of all affected parties; the second sentence recommends that extensive rights of political participation be granted to citizens in the real world. The key word in the quotation is, of course, "hence": Habermas seems to think that the second sentence follows from the first, whereas in fact the second sentence follows only if political participation in the real world could reproduce (or very closely approximate) conditions under idealized popular sovereignty. The likelihood of such a correspondence is an empirical issue, and one can readily identify situations (e.g., contemporary Russia) where this likelihood is quite low. The danger here to Habermas's CT is clear: if actual popular sovereignty can deviate substantially from idealized popular sovereignty, and idealized popular sovereignty is the one and only avenue to norm validation (as asserted by the Discourse Principle), then actual popular sovereignty may generate norms that are invalid (e.g., hostile to human rights). The unreflective analogizing contained in the above passage obscures this essential point.

Analogizing of a similar kind can be found in the following passage from Rawls, which defines and justifies his "principle of participation":

[The principle of participation] requires that all citizens are to have an equal right to take part in, and to determine the outcome of, the constitutional process that establishes the laws with which they are to comply. Justice as fairness begins with the idea that where common principles are necessary and to everyone's advantage, they are to be worked out from the viewpoint of a suitably defined initial situation of equality in which each person is fairly represented. The principle of participation transfers this notion from the Original Position to the constitution as the highest-order system of social rules for making rules. If the state is to exercise a final and coercive authority over a certain territory, and if it is in this way to affect permanently men's prospects in life, then the constitutional process should preserve the equal representation of the Original Position to the degree that this is practicable. [11]

This passage closely parallels the one above-popular sovereignty as actualized by the principle of participation is justified by analogy to an idealized conception of popular sovereignty (the Original Position)-and may therefore by subjected to the same line of criticism. In this passage, Rawls is more circumspect than Habermas (notice the last seven words), but the tendency is the same: confounding the ideal and the actual.

see figure 1

In order to make the nature of this confusion and its implications for the CT clearer, Figure 1 illustrates the interconnections (or lack thereof) among Human Rights, Idealized Popular Sovereignty, and Actual Popular Sovereignty. At a suitable level of abstraction, Human Rights and Popular Sovereignty are indeed mutually supporting: this is the lesson of Rawls's Original Position, Habermas's Ideal Speech Situation, and Kant's Third Formulation of the Categorical Imperative. [12] Actual Popular Sovereignty, on the other hand, has a more tenuous connection with Human Rights and can fail to support them (as the last subsection indicated). Actual and Idealized Popular Sovereignty are obviously distinct, and the distance between them may be rather large in practice. Both Habermas and Rawls appear to pursue a strategy of analogizing Actual to Idealized Popular Sovereignty, which allows them to ascribe the liberal implications of the latter to the former. As the diagram indicates, however, this strategy is highly questionable: Actual and Idealized Popular Sovereignty are not only distinct, but they also are related to Human Rights in fundamentally different ways.

Before moving on to a more careful analysis of Rawls's position on the CT, we should ponder a question that has been left unanswered up until now: what is the content of Actual Popular Sovereignty for Habermas and Rawls? For Habermas, the answer is unclear. He speaks of "basic rights to equal opportunities to participate in processes of opinion- and will-formation" [13] and, more provocatively, of a desire to "reignite the radical democratic embers of the Original Position in...civic life," [14] but he is not forthcoming with institutional details. Rawls speculates that Habermas may have in mind forms of participatory democracy similar to those recommended by Jefferson in his 1816 letter to Samuel Kercheval, including a "ward scheme" and constitutional conventions every generation.[15] Still, the specifics of Habermas's "Deliberative Politics" remain something of a mystery. Rawls, on the other hand, is anything but mysterious. As we shall see in the following section, this admirable openness makes him more vulnerable than Habermas is to certain criticisms, especially regarding the contested relationship between political representation and popular sovereignty.

C. Rawls and the Co-Originality Thesis

To say that Rawls buys Habermas's CT lock, stock, and barrel would be only a slight exaggeration. In his "Reply to Habermas," Rawls denies Habermas's claim that "the two-stage character of [Rawls's] political conception leads ... to the liberal rights of the moderns having a priori features that demote the democratic process to an inferior status"; he argues instead that "in liberalism properly interpreted, public and private autonomy are co-original and of equal weight." [16] This endorsement of the CT presents Rawls with two challenges-both of which, I will argue below, he fails to meet. First, he must distance himself from thinkers (like Benjamin Constant and Isaiah Berlin) who offer strongly instrumentalist defenses of democracy, because such defenses of democracy make the relationship between liberalism and democracy a contingent one, whereas the CT denies any such contingency. Second, he must explain how the highly qualified form of popular rule that he advocates-one characterized both by constitutional constraints on the exercise of power and by the mediating influence of representation-is really consistent with Habermas's radical notion of public autonomy. That is, in his own words, he must answer "the question of how exactly the political institutions associated with constitutional democracy can be understood to be consistent with the idea of popular sovereignty." [17] We will examine Rawls's response to each of these challenges in turn.
1. Rawls and the Instrumentalist Defense of Democracy

In Footnote 39 of his "Reply to Habermas," Rawls seems to take umbrage at Habermas's suggestion that his defense of democracy is largely instrumental in character. [18] Why is Rawls so anxious to defend himself from this charge? The reason for his concern is clear. If extensive citizen participation in politics were wholly (or even largely) instrumental, then its status would be empirically contingent: under certain circumstances, democracy might have to be curtailed in order to safeguard more effectively the ends it serves, i.e., liberalism and its associated individual liberties, both negative and positive. Rawls cannot afford to build his justification for democracy on such unstable grounds, however, if he is to be concurrently committed to Habermas's CT, because the CT asserts the existence of an intrinsic, internal relationship between human rights and popular sovereignty, a relationship not subject to the vagaries of the external world.

A careful reading of Rawls suggests that Habermas has a point: Rawls's defense of democracy, though not purely instrumental in character, is predominantly so. Consider, for example, the following passage from §37 ("Limitations on the Principle of Participation") of A Theory of Justice: "we should narrow or widen [the principle of participation's] extent up to the point where the danger to liberty from the marginal loss in control over those holding political power just balances the security of liberty gained by the greater use of constitutional devices." [19] This passage is a classic instrumentalist defense of democracy. The function of the principle of participation here is the defense of liberty, and the extent of such participation is determined by an explicit cost/benefit calculus. Imagine the following scenario: a country with a civic-minded, educated liberal elite but highly illiberal (though obedient) masses. Using Rawls's own logic, we might be able to justify political meritocracy for such a country, with participation rights perhaps limited to those with a college education.

Is such a reading fair to Rawls? A little later in §37 Rawls provides what can only be described as a sympathetic reading of Mill's proposal that "persons with greater intelligence and education should have extra votes in order that their opinions may have a greater influence." [20] The conclusion of this section of §37-whose nautical imagery is reminiscent of Plato's in The Republic (488b-489c)-is instructive on this point:

The passengers of a ship are willing to let the captain steer the course, since they believe that he is more knowledgeable and wishes to arrive safely as much as they do. There is both an identity of interests and a noticeably greater skill and judgment in realizing it. Now the ship of state is in some ways analogous to a ship at sea; and to the extent that this is so, the political liberties are indeed subordinate to the other freedoms that, so to say, define the intrinsic good of the passengers. Admitting these assumptions, plural voting may be perfectly just. [21]

Whatever Rawls's views on this particular proposal, he plainly indicates in this passage and preceding ones that the extent of participation can be substantially curtailed for instrumentalist reasons. Whether such limited-suffrage regimes are still characterized by "popular sovereignty" in any meaningful sense will be addressed later.

These instrumentalist arguments for democracy can be found not only in Theory but also in Political Liberalism, though in the latter work they usually appear in less explicit form. One particularly relevant passage, cited by Habermas, discusses the views of Constant and Berlin:

One strand of the liberal tradition regards the political liberties as of less intrinsic value than freedom of thought and liberty of conscience, and the civil liberties generally. What Constant called "the liberties of the moderns" are prized above "the liberties of the ancients." In a large modern society, whatever may have been true in the city-state of classical times, the political liberties are thought to have a lesser place in most persons' conceptions of the good. The role of the political liberties is perhaps largely instrumental in preserving the other liberties. [22]

Habermas interprets the last line as an endorsement of the Constant/Berlin position, but the language is somewhat ambiguous, and Rawls himself wants to deny any such endorsement. [23] There are other parts of Political Liberalism that point less ambiguously to an instrumentalist understanding of democracy, however. Consider, for example, Rawls' argument in favor of maintaining the "fair value" of the political liberties: "unless the fair value of these liberties is approximately preserved, just background institutions are unlikely to be either established or maintained." [24] Notice that he focuses here on the role political liberties play in the maintenance of liberal institutions; such liberties are viewed here primarily as a means to an end.

Despite Rawls' frequent flirtations with, and even endorsements of, instrumentalist defenses of democracy, he adamantly denies the attraction, [25] as just noted. In this footnote, he points to two locations in his writings-one in Political Liberalism, the other in Theory-where he recognizes the intrinsic value of the political liberties. The former reference is to a section of Political Liberalism dealing with the "The Good of Political Society." [26] This section is only incidentally about political participation; its main focus is on the good of citizenship. One source of unity in a well-ordered society, says Rawls, is that citizens "share one very basic political end, and one that has high priority: namely, the end of supporting just institutions and of giving one another justice accordingly, not to mention many other ends they must also share and realize through their political arrangements." [27] But surely this sort of unity can potentially be achieved in a nondemocratic but liberal regime. Rawls provides no reasons here to make us believe otherwise. He does discuss classical republicanism and civic humanism at the end of the section but endorses neither: he firmly rejects the latter and says of the former merely that it is consistent with, not required by, justice as fairness. The closest Rawls comes here to an argument for the intrinsic value of political participation is the following quotation: "it remains to say only that justice as fairness does not of course deny that some will find their most important good in political life, and therefore that political life is central to their comprehensive good." [28] This statement is hardly a ringing endorsement of widespread political participation and might even be consistent with the strict limitations on suffrage discussed earlier.

The closest Rawls comes to a full-fledged defense of the intrinsic value of participation is in the last two paragraphs of §37 of Theory. Given the strongly instrumentalist bent of the rest of the section, they appear to be something of an afterthought. Rawls summarizes his arguments in the following passage: "equal political liberty is not solely a means. These freedoms strengthen men's sense of their own worth, enlarge their intellectual and moral sensibilities, and lay the basis for a sense of duty and obligation upon which the stability of just institutions depends." [29] A question immediately arises: why is democracy required to realize these goods? A liberal society protects freedom of association, free speech, etc., so that citizens may engage in political protest and organize into special-interest groups in an attempt to influence those in power. So long as a regime is liberal in this sense and responsive to the desires of its citizens (perhaps through something like the "decent consultation hierarchy" described by Rawls in §9 of The Law of Peoples), these goods can at least potentially be realized in a nondemocratic society. Few would deny that widespread and direct political participation can also contribute to these goods, but if such participation has high costs (in terms of threatening important liberal values), then the foregoing suggests that suffrage might be curtailed in a manner that is still consistent with their realization.

One final point: even if we were to ignore Rawls's largely instrumentalist defense of democracy, which sits uneasily with his endorsement of Habermas's CT, he still explicitly leaves the door open for such a defense, which is itself inconsistent with the CT. Consider the following two quotations: "[the priority of liberty] allows although it does not require that some liberties, say those covered by the principle of participation, are less essential in that their main role is to protect the remaining freedoms"[30]; "it is not supposed that the basic liberties are equally important or prized for the same reasons."[31] Compare these to the following quotation from Rawls's "Reply to Habermas," in which Rawls implicitly endorses Habermas's CT: "the ancient and modern liberties are co-original and of equal weight with neither given pride of place over the other."[32] If Rawls is serious about his commitment to the CT, he cannot simultaneously hold that political liberties are just as important as other basic liberties and that (classical) liberal doctrines that relegate these same liberties to a lower tier are admissible.

2. Rawls and Representation

We now move on to the second challenge facing Rawls-namely, reconciling his highly qualified form of popular rule with Habermas's radical notion of public autonomy. The exact relationship between constitutional democracy and popular sovereignty is of concern to Rawls, but he mentions this concern only at the end of his "Reply to Habermas," evidently at the urging of his colleagues Ronald Dworkin and Thomas Nagel:

There is one related question that I have not discussed in detail, and that is the question of how exactly the political institutions associated with constitutional democracy can be understood to be consistent with the idea of popular sovereignty. If we associate popular sovereignty with something like majority rule following free, open, and wide discussion, then there is at least an apparent difficulty. This difficulty may be an aspect of what Habermas is referring to when he says that "[t]he form of political autonomy...does not fully unfold in the heart of the justly constituted society." The consistency of constitutional democracy with popular sovereignty...requires an account and explanation of the institutions for the exercise of the constituent power of a democratic people in making constitutional decisions, as opposed to the institutions of ordinary democratic politics within the framework set up by those decisions. [33]

Rawls's point here is fairly straightforward: constitutional constraints on democratic governance are consistent with popular sovereignty so long as these constraints are themselves chosen by the people. Setting aside for a moment the question of why such constraints would be necessary, we will focus instead on the meaning of the phrase "constituent power of a democratic people." [34]

Rawls uses this phrase and variations on it throughout his "Reply to Habermas," but its meaning is less than clear. Consider the use of it that occurs during Rawls's discussion of "the three most innovative periods in American constitutional history...the founding of 1787-91, Reconstruction, and in a different way the New Deal," where he says that "the liberties of the moderns are subject to the constituent will of the people." A national referendum would seem to be the most obvious way of discerning the people's will, yet referenda were not used during any of these innovative periods. The US Constitution was drafted by delegates appointed by state legislatures (representatives of representatives, so to speak) and ratified by special state conventions. The Reconstruction amendments were drafted by Congress and were then ratified by state legislatures (including those of newly readmitted Southern states). Finally, the New Deal was saved from a hostile Supreme Court by threats from Roosevelt and Congress to "pack the Court" with more sympathetic justices. All of these events took place at several removes from "the people."

This passage and similar ones throughout the "Reply to Habermas" indicate that, for Rawls, the constitutional decisions of representative institutions (e.g., state legislatures, ratifying conventions, Congress, etc.) are mere reflections of the will of the people, i.e., representation effects popular sovereignty. Rawls' unreflective endorsement of representation is most easily seen in his treatment of the Original Position:

Are the parties [of the Original Position] as citizens' trustees not selecting principles of justice to specify the scheme of (basic) liberties which best protect and further citizens' fundamental interests and which they then concede to one another? Here, also, the ancient and the modern liberties are co-original and of equal weight with neither given pride of place over the other. The liberties of both public and private autonomy are given side by side and unranked in the first principle of justice. [35]

The notion of the parties being "trustees" may seem completely innocuous in this context, but Rawls extends this idea of representation to real-world institutions: "the constitutional process should preserve the equal representation of the original position...." [36]

An additional, related point should be made regarding the above quotation. In this quotation and throughout his works, [37]Rawls associates the ancient and modern liberties with public and private autonomy, respectively-but this usage does not follow that of Constant. For Constant, the ancient and modern liberties both included political liberties: the modern liberties include "everyone's right to exercise some influence on the administration of government, either by electing all or particular officials," whereas the ancient liberties include "exercising collectively, but directly, several parts of the complete sovereignty." [38] In other words, one of the key differences between the ancients and the moderns is not the presence or absence of public autonomy but rather the presence or absence of political representation.

Rawls's belief that representation merely effects popular sovereignty would have seemed peculiar to the American Founders. Consider the following quotation from The Federalist #10, where James Madison, after criticizing the idea of direct democracy, goes on to explain the merits of representation as a "filtering" device:

The two great points of difference between a [direct] democracy and a republic are: first, the delegation of the government, in the latter, to a small number of citizens elected by the rest.... The effect of the first difference is, on the one hand, to refine and enlarge the public views, by passing them through the medium of a chosen body of citizens, whose wisdom may best discern the true interest of their country, and whose patriotism and love of justice will be least likely to sacrifice it to temporary or partial considerations. Under such a regulation, it may well happen that the public voice, pronounced by the representatives of the people, will be more consonant to the public good than if pronounced by the people themselves, convened for the purpose.

The Electoral College, the indirect election of US Senators, etc., were institutional devices designed not to effect the popular will but rather to refine it and, where necessary, to block it. The tendency of such devices to distance the federal government from the popular will was not lost on the Anti-Federalists. Their demand for a Bill of Rights reflected not a fear of majoritarian tyranny (a concept that would have been as alien to them as to other classical republicans) but rather a fear of a remote federal government not under adequate popular control.

These concerns over representation are also reflected in the canon of modern political theory. For example, Jean-Jacques Rousseau and Immanuel Kant, discussed by Habermas, had radically different views on representation, which is surprising given Rousseau's strong and readily noticeable influence[39] on Kant's political philosophy (see Taylor 2000). In the following excursus we will examine their respective views for two reasons: first, to derive a baseline conception of popular sovereignty with which Rawls's conception can be compared; second, to show that on this issue Rawls was most likely influenced by Kant, especially by his unreflective endorsement of representation and by his willingness to sacrifice popular sovereignty to protect human rights.

3. Excursus: Rousseau and Kant on Representation

At the beginning of his discussion of Rousseau and Kant, Habermas says that "Kant suggests more of a liberal reading of political autonomy, Rousseau a republican reading." [40] This view is essentially correct, and nowhere is it more obvious than in their respective positions on representation. Rousseau was adamantly opposed to any representation of the popular will, whereas Kant was not only comfortable with the idea but considered it to be a vital feature of republican government.

First consider Rousseau's position, which is stated succinctly in The Social Contract: "sovereignty cannot be represented, for the same reason that it cannot be alienated; its essence is the general will, and will cannot be represented-either it is the general will or it is something else; there is no intermediate possibility.... Any law which the people has not ratified in person is void; it is not law at all. [41] His reasons for opposing representation are several: he believes that direct citizen participation in lawmaking reinforces patriotism, reduces corruption, etc. [42] His main concern, however, is with the realization of the popular will, a process that he believes is stymied by representative systems.

Does Rousseau believe in any constraints at all on the sovereign will of the people? As far as formal constraints are concerned, few are mentioned. Rousseau does insist that the people be "properly informed" and that they be prevented from (privately) communicating among themselves before voting-the latter presumably to prevent the formation of "sectional associations." [43] One major informal constraint is discussed at length: the Lawgiver. [44] The Lawgiver of a society provides it with a distinctive set of ethical beliefs and cultural practices, which he inscribes upon "the hearts of the citizens." [45] These shared beliefs and practices catalyze the formation of the general will, but only at the expense of ethical and cultural pluralism. [46]Thus, Rousseau's vision of democratic governance is both highly participatory and fundamentally illiberal.

To understand Kant's position on representation, we must begin with the Groundwork and, more specifically, with the Third Formulation of the Categorical Imperative. As noted earlier, the Third Formulation demonstrates that Human Rights and Popular Sovereignty are mutually supporting-at least, that is, at a high enough level of abstraction:

And what is it, then, that justifies a morally good disposition, or virtue, in making such high claims [to dignity]? It is nothing less than the share it affords a rational being in the giving of universal laws, by which it makes him fit to be a member of a possible kingdom of ends, which he was already destined to be by his own nature as an end in itself and, for that very reason, as lawgiving in the kingdom of ends-as free with respect to all laws of nature, obeying only those which he himself gives and in accordance with which his maxims can belong to a giving of universal law (to which at the same time he subjects himself). [47]

This "share...in the giving of universal laws" is given concrete form through the institutions of republican government. In The Doctrine of Right (Part I of The Metaphysics of Morals), Kant defends popular sovereignty with language reminiscent of Rousseau:

The legislative authority can belong only to the united will of the people. For since all right is to proceed from it, it cannot do anyone wrong by its law. Now when someone makes arrangements about another, it is always possible for him to do the other wrong; but he can never do wrong in what he decides upon with regard to himself. Therefore only the concurring and united will of all, insofar as each decides the same thing for all and all for each, and so only the general united will of the people, can be legislative.... From [the citizen's] independence follows his civil personality, his attribute of not needing to be represented by another where his rights are concerned.[48]

The clear implication of this passage is that legislative authority must be exercised by the united citizenry-Kant appears to follow Rousseau in rejecting representation. A little later he confirms this by speaking of "whatever sort of positive laws the citizens might vote for," implying that the citizens themselves are the legislators. [49]

Then, in a genuinely perplexing move, he begins to refer to the "people's sovereign (legislator)," [50] and he is soon writing sentences like the following: "therefore a people cannot offer any resistance to the legislative head of a state which would be consistent with right, since a rightful condition is possible only by submission to its general legislative will." [51] The sovereign people has been mysteriously replaced by a sovereign body of legislators; representation has entered through the back door. [52] This casual transmutation of sovereignty presumably occurs because Kant, like Rawls, thinks representation to be an unexceptionable principle, one not in need of justification. On this point, Kant and Rousseau most decidedly part company.

Kant's slippery introduction of representation is accompanied by a very explicit defense of property qualifications for voting. Kant urges exclusion from the electorate of apprentices, domestic servants, minors, women, and anyone else "whose preservation in existence (his being fed and protected) depends not on his management of his own business but on arrangements made by another (except the state)." [53] He does not spend much time defending this exclusion but simply asserts that those in a position of dependence should not be "active" (i.e., voting) citizens. This restriction of suffrage might be justified in two different ways: first, one might reasonably suppose that dependents will be unduly influenced by those who provide their maintenance and that allowing them the franchise will therefore dangerously increase the political power of the rich, who have many retainers; second, one might suspect that dependents (being poor) will vote for redistributive policies and that allowing them to vote will therefore destabilize property relations. The former explanation has classical republican overtones, while the latter has classical liberal ones; both justifications were likely on Kant's mind. With suffrage as with representation, Kant's commitment to popular sovereignty is more evident in theory than in practice; as Habermas suggests, his primary concerns seem to be with protecting the integrity of the political process and the security of liberal rights (e.g., private property).

Now that we have examined the respective views of Rawls, Rousseau, and Kant, how should we go about defining Rawls's "constituent power of a democratic people," i.e., popular sovereignty? Rawls provides part of the answer when he says that "the most extensive political liberty is established by a constitution that uses the procedure of so-called bare majority rule (the procedure in which a minority can neither override nor check a majority) for all significant political decisions unimpeded by any constitutional constraints." [54] The other part of the answer is provided by Rousseau: the people should rule directly, not through the "refining" intermediation of representatives. The resulting system-unconstrained direct democracy-will serve as a useful baseline conception of popular sovereignty. In the United States, this baseline conception might be approximated through a system of national referenda (both obligatory and optional) and national initiatives (for both constitutional changes and regular legislation).

Rousseau's model of political participation comes very close to this baseline conception, but it achieves this by sacrificing ethical and cultural pluralism. It belies the rosy reconciliation of human rights and popular sovereignty promised by Habermas's CT. Kant's model, with its use of representation and property qualifications for voting, deviates substantially from the baseline conception in order to protect the stability and integrity of the political system as well as liberal rights. In the final analysis, Rawls follows the path laid down by Kant. His implicit endorsement of representation, as well as his instrumentalist sympathies for restrictions on the extent of political participation, [55] indicate a willingness to sacrifice popular sovereignty for liberal ends. Whereas Habermas makes popular sovereignty and human rights mutually supporting by simply redefining popular sovereignty to include human rights (a semantic sleight-of-hand), Kant and Rawls do so by draining popular sovereignty of much of its meaning. They offer instead a venomless popular sovereignty, one incapable of threatening liberal values, which are of course their primary concern.

D. Conclusion: The Wisdom of Benjamin Constant

Rawls's instincts are impeccable: despite his assurances that he is not offering a largely instrumentalist defense of democracy, he is drawn again and again to Constant's distinction between the liberty of the ancients and that of the moderns. [56] According to Constant, "we can no longer enjoy the liberty of the ancients, which consisted in an active and constant participation in collective power. Our freedom must consist of peaceful enjoyment and private independence." [57] As noted earlier, Constant considers political liberty to be a vital bulwark of individual liberty, and he supports its realization, not through direct democracy, but rather through "the representative system." [58]

Why does Constant offer such a strongly instrumentalist defense of democracy? [59] The reason is simple: contrary to Habermas's CT, human rights and popular sovereignty are not always compatible, and an instrumentalist defense suggests the need for various safeguards to protect liberal rights from an often illiberal popular will. A concrete example may be helpful. Consider the following Gallup survey results, based on randomly selected national samples of American adults: in June of 1989, 71% of those polled supported "a constitutional amendment to make flag burning illegal," and in June of 1999, well over two-thirds of those surveyed believed that public schools should be allowed "to display the Ten Commandments," to organize "daily prayer...spoken in the classroom," and to teach "creationism along with evolution." [60] As these results indicate, large and continuing majorities of the American public would like to modify First Amendment protections (more specifically, the speech and establishment clauses), and under a system that even approximated the baseline conception of popular sovereignty discussed above, they would have their way. We obviously do not live under such a system: innumerable features of our constitutional system, including judicial review and a difficult amendment process, thwart the will of such majorities with great and ongoing success. They are key to preserving our highly principled (some would say extreme) brand of civil libertarianism.

American liberals like Rawls are uneasy about this fact and reassure themselves that the entire system is really consistent with popular sovereignty-in some highly indirect and tortuous way. A better approach might be to recognize that all good things are not compatible and that tradeoffs are sometimes necessary, as is the case here. The American constitutional system deals with the frequent incompatibilities between popular sovereignty and human rights generally by sacrificing the former on the altar of the latter. This approach is not the only one available, and some commentators appear to question whether it is the best one available, [61] but no amount of wishful thinking will make the tradeoff itself disappear.

Finally, the existence of such a tradeoff has important implications for the design of political institutions in countries going through a democratic transition. Mere implementation of a competitive electoral system is no guarantee that human rights will be protected, as we saw in the earlier discussion of illiberal democracies. Institutional mechanisms such as a written bill of rights, division of powers, an independent judiciary, federalism, etc., can all help safeguard individual liberty against the depredations of both "evil-minded rulers" (Louis Brandeis) and illiberal majorities.

[1] For historical evidence regarding the so-called "democratic peace," see Doyle (1997), Russett (1993), and Oneal and Russett (1997).

[2] Throughout this paper, the terms "liberalism," "human rights," and "private autonomy" will (unless otherwise indicated) be used interchangeably, as will the terms "democracy," "popular sovereignty," and "public autonomy."

[3] LAM: 316. See Bibliography for notation

[4] (BFN: 104; see also 118, 122.) He later makes this point even more emphatically: human rights, according to Habermas, are "necessary enabling conditions; as such, they cannot restrict the legislator's sovereignty, even though they are not at her disposal. Enabling conditions do not impose any limitations on what they constitute" (BFN: 128; emphasis added).

[5] Merriam-Webster's Collegiate Dictionary. For further reading on the contested definition of democracy, see Held (1997) and Collier and Levitsky (1997).

[6] Klosko 1993: 3.

[7] SC: III-15; 143.

[8] LAM: 323.

[9] Here are rough definitions of these terms of art: 1) Original Position (Rawls): a hypothetical initial bargaining position in which citizens debate and ultimately ratify the political principles that will govern their society. This initial bargaining position is made fair through the device of the "veil of ignorance," which hides from citizens any knowledge of their respective socioeconomic statuses, natural talents and/or handicaps, etc., in order to equalize bargaining power. 2) Ideal Speech Situation (Habermas): an idealized moral dialogue in which the interlocutors participate both under fair conditions (e.g., equal opportunity for participation, the right to attack or defend any moral or factual claim, and the absence of constraining status differences) and in good faith (i.e., with the purpose of reaching consensus on moral truths).

[10] BFN: 127; see also 454-5, 458, emphasis added.

[11] TJ: 221-2, emphasis added.

[12] "The principle of every human will as a will giving universal law through all its maxims" (GMM: 4/432; 40).

[13] BFN: 123.

[14] Quoted in PL: 399.

[15] PL: 408. Interestingly, though, Habermas appears to endorse judicial review (BFN: 167-8, 240-2), an institution that was anathema to Jefferson and his Republican compatriots (see Ellis 1997: 224-5, 277).

[16] PL: 396, 412.

[17] PL: 433.

[18] PL: 404.

[19] TJ: 230.

[20] TJ: 232.

[21] TJ: 233. Such a system existed in the United Kingdom until 1949: "business owners and university graduates received two votes apiece: one for their residence and one for their business or university affiliation" (Palmer 1997: 90).

[22] PL: 299 (Berlin reference in footnote), emphasis added.

[23] PL: 404, Footnote 39.

[24] PL: 327-8.

[25] PL: 404, Footnote 39.

[26] PL: 201-6. The reference in Footnote 39 is to V:6 ("Is Justice as Fairness Fair to Conceptions of the Good?"), but this section seems irrelevant to Rawls1 argument. The intended reference is likely to V:7 ("The Good of Political Society").

[27] PL: 202.

[28] PL: 206.

[29] TJ: 234.

[30] TJ: 230.

[31] PL: 298-9.

[32] PL: 413, emphasis added.

[33] PL: 433-4, emphasis added.

[34] PL: 406.

[35] PL: 413, emphasis added.

[36] TJ: 222, emphasis added.

[37] E.g., TJ: 222.

[38] LAM: 311.

[39] BFN: 100-3.

[40] BFN: 100.

[41] SC: III-15; 141. Rousseau softens his position ever so slightly in The Government of Poland, where he suggests that deputies might be used if they are changed frequently and if their hands are tied by "letters of instruction" to which they will be held accountable by their constituents (GP: 36). (Compare this notion of representation to that of, say, Edmund Burke.) Still, Rousseau views this as a second-best alternative; he prefers direct democracy where it is feasible.

[42] SC: III-15; 140-3, GP: 35.

[43] SC: II-3; 73.

[44] See especially SC: II-7; 84-8.

[45] SC: II-12; 99.

[46] As Habermas notes?BFN: 102.

[47] GMM: 4/436; 43.

[48] MM: 6/314; 91. Cf. Rousseau: "no longer ask if the law can be unjust, because no one is unjust to himself" (SC: II-6; 82)

[49] MM: 6/315; 92.

[50] MM: 6/317; 94.

[51] MM: 6/320; 96.

[52] One possible explanation is that while only the "united will of the people" is competent to effect constitutional change, representative institutions are adequate for the task of promulgating statutes. Despite some intervening text discussing the "original contract," this interpretation fails to tie together the many loose threads of Kant1s argument.

[53] MM: 6/315; 92.

[54] TJ: 224.

[55] TJ: §37.

[56] TJ: 201, 222; PL: 5, 206, 299, and throughout his "Reply to Habermas."

[57] LAM: 316.

[58] LAM: 325.

[59] An interesting parallel: Constant mentions the intrinsic value of political liberty at the very end of his speech before the Athéné Royal (LAM: 326-8), just as Rawls does at the very end of §37 in Theory.

[60] Gallup Organization, 1999.

[61] Consider, for example, Orlando Patterson: "while the notes of negative and positive freedom ring gloriously in America, the third note of its cultural chord - democracy - has acquired a tinny, hollow timbre" (1999: 63). He does not stop to consider whether the latter phenomenon may be causally related to the former.



The parenthetical abbreviations are used in the paper as referents.

Collier, D., and S. Levitsky. 1997. Democracy with adjectives: conceptual innovation in comparative research. World Politics. 49: 430-51.

Constant, B. 1988 [1819]. The liberty of the ancients compared with that of the moderns (LAM). The Political Writings of Benjamin Constant. Cambridge: Cambridge University Press. 307-28.

Doyle, M. 1997. Ways of War and Peace. New York: Norton.

Ellis, J. 1997. American Sphinx: The Character of Thomas Jefferson. New York: Alfred Knopf.

Gallup Organization. 1999. <http://www.gallup.com/>.

Habermas, J. 1996. Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy (BFN). Cambridge, MA: MIT Press.

Held, D. 1997. Models of Democracy. Stanford: Stanford University Press.

Kant, I. 1997 [1785]. Groundwork of the Metaphysics of Morals (GMM). Trans. M. Gregor. Cambridge: Cambridge University Press.

-----. 1996 [1797]. The Metaphysics of Morals (MM). Trans. M. Gregor. Cambridge: Cambridge University Press.

Klosko, G. 1993. History of Political Theory: An Introduction (Volume 1: Ancient and Medieval Political Theory). Fort Worth, TX: Harcourt Brace.

Madison, J. 1787. The Federalist #10. <http://www.law.emory.edu/FEDERAL/federalist>.

Oneal, J., and B. Russett. 1997. The classical liberals were right: democracy, independence, and conflict, 1950-85. International Studies Quarterly. 41(2): 267-94.

Palmer, M. 1997. Comparative Politics: Political Economy, Political Culture, and Political Independence. Itasca, IL: F. E. Peacock Publishers.

Patterson, O. 1999. The liberal millennium. The New Republic. November 8. 221(19): 54-63.

Plato. 1991. The Republic. Trans. A. Bloom. New York: Basic Books.

Rawls, J. 1999. The Law of Peoples. Cambridge, MA: Harvard University Press.

-----. 1993. Political Liberalism (PL). New York: Columbia University Press.

-----. 1971. A Theory of Justice (TJ). Cambridge, MA: Harvard University Press.

Rousseau, J-J. 1985 [1772]. The Government of Poland (GP). Trans. W. Kendall. Indianapolis, IN: Hackett Publishing Co.

-----. 1968 [1762]. The Social Contract (SC). Trans. M. Cranston. New York: Penguin Books.

Russett, B. 1993. Grasping the Democratic Peace. Princeton, NJ: Princeton University Press.

Taylor, R. 2000. Rousseau and Kant: the general will, cosmopolitanism, and the psychological foundations of politics. Critical Sense. 8(1): 9-50.