THE PROBLEM OF AGREEMENT IN REPUBLICANISM, PROCEDURALISM, AND THE MATURE DEWEY: A "TWO MOMENTS OF DISCOURSE ETHICS" ANALYSIS

This essay refines Honneth's analysis by focusing on the agreement problem: discourse ethics' demand that norms be agreed to by all. This problem is not overcome by either republicanism or democratic proceduralism. Turning to "the sphere of concrete production" [etc.] only translates this problem to the new level. However, Honneth's / Dewey's approach can overcome the agreement problem by using Chilton's (forthcoming) "two-moments" approach to discourse ethics. Honneth's linkage of the spheres of politics and concrete production neatly parallels the vertical decalage between how large groups and dyadic pairs approach agreement, and so the "two-moment" solution to the agreement problem can be applied equally to both of Honneth's spheres.
 

Honneth (1998) advances a perspective that overcomes the opposition between two critical alternatives to liberalism: "republicanism" and (democratic) "proceduralism". As argued later in more detail, his perspective depends on various normative assumptions: that we know what is just, that we can distinguish free cooperation from coerced submission, and that we can find common goals. His argument does not redeem these claims, however. The present essay supplies the necessary redemptions by use of discourse ethics and in particular the "two-moment" understanding of discourse ethics (Chilton forthcoming). This analysis supports, makes concrete, and slightly emends Honneth's work.
 

We start (§§ I & II) not with Honneth's work but instead with laying out the necessary prior understanding of discourse ethics. The final section (III) applies this machinery to Honneth's argument by making explicit its normative demands and then reconstructing its logic in discourse ethics terms.
 

I. The Agreement Problem in Discourse Ethics

. . . every valid norm has to fulfill the following condition:  . . . All affected can accept the consequences and the side effects its general observance can be anticipated to have for the satisfaction of everyone's interests (and these consequences are preferred to those of known alternative possibilities for regulation) (Habermas 1990a:65).

Only those norms can claim to be valid that meet (or could meet) with the approval of all affected in their capacity as participants in a practical discourse (Habermas 1990a:66).⁽¹⁾

On the other hand, discourse ethics is frustrating, because it demands more agreement than we seem capable of. Even if it captures our central understanding of the nature of morality, as Habermas claims, its justification in terms of the presuppositions of argumentation does not provide us with concrete, practical answers of how to reach agreement. In fact, there seems good reason to believe that it cannot be reached. And if it cannot, if no norm can be admitted as binding, then discourse ethics is only a glorious failure, not a practical guide to political action. This hurdle faced by discourse ethics is the "agreement problem".
 

Republicanism and democratic proceduralism can each be interpreted in light of this problem as alternative attempts to reach agreement. Thus republicanism focuses on the impulse to achieve agreement, finding the attempt worthy even in the face of inevitable failure. Metaphorically, it finds the journey important even if we never reach our destination. And there is certainly much to approve in this: the equality implicit in everyone's participation and the liberatory force this equal participation unleashes; the social solidarity and mutual recognition implicit in the free exchange of views; the intelligence implicit in seeking and considering as many perspectives as possible; and the self-understanding and even self-realization arising from mutual, respectful - "loving" is not too strong a word - examination of one other's motives.
 

On the other hand, democratic proceduralism does not deny discourse ethics' demand for agreement but rather focuses on the need for regulating our intercourse through actual, even if imperfect, norms. Metaphorically, it recognizes that our journey is to serve a practical need; an infinite journey, however enjoyable, still does not get us to our destination. And of course there is much to approve in this too: it ensures a decision; it prevents participants from employing the strategy of infinite delay; and it avoids having our energies consumed by infinitely-long discussions of an infinite variety of issues.
 

Honneth sees the two approaches in slightly different terms:

Today, these key terms ordinarily designate two normative models of democracy whose common goal it is to give democratic will formation a greater role than is usual in political liberalism. Instead of limiting the participatory activity of citizens to the function of periodically legitimating the state's exercise of power, this activity is to be a permanent matter embodied in the democratic sphere and should be able to be understood as the source of all political decision-making processes. The differences . . . between the two models . . . follow first of all from the different ways in which the principle of the democratic public sphere is normatively justified in each case: . . .in republicanism, the democratic public sphere is . . . regarded as the medium of a self-governing political community, in [proceduralism] as the procedure with whose help society attempts to solve political problems rationally in a legitimate manner (1-2).

In sum, both republicanism and democratic proceduralism testify to discourse ethics as a normative foundation, their differences arising simply from their different approaches to solving the intractable agreement problem. The hope of reconciling and synthesizing the two approaches therefore requires a direct attack on that problem. The next section of the paper outlines a perspective on discourse ethics - the so-called "two-moment" perspective - that restructures and solves the agreement problem.
 
 
 

II. Solving the Agreement Problem through the "Two-Moment" Perspective⁽³⁾

The ideal speech situation does not solve the agreement problem, however. I will not dwell on the problem raised by its counterfactual nature, i.e., the problem that we cannot know in advance of the actual creation of an ideal speech situation whether universally-agreed-upon norms will actually always appear.⁽⁴⁾ Instead, I will merely point out that almost all normative decisions must be made within a limited time frame, so that the leisurely discourse contemplated by the ideal speech situation is not possible. Without denying that the ideal speech situation is worth striving for, we can still recognize that it does not solve the agreement problem.
 

We can solve the agreement problem by regarding discourse ethics as divided into two moments, termed M1 and M2. The first moment, M1, is in essence the ideal speech situation, allowed to continue for as long as our time and resources allow. In M1, (U) is merely an injunction to pursue free agreement, that is, agreement in which no one feels h/her sense of the Good is being violated.
 

This agreement could arise in any of several ways. Here are three:

(1) Open discourse reveals to me (us/you) problems with my (our/your) sense of authenticity, and the resulting transformation in my (our/your) sense of the Good eliminates the conflict between my and your sense of the Good. For example, I may recognize that my homophobia originated in my father's irrational fears, not in my own experience, and - reconstructing my sense of self in light of what I have experienced and understand - I may come to accept gay marriage, ending the conflict with what was already your sense of the Good.

(2) Open discourse may reveal new options that satisfy both our senses of the Good. For example, I may not want to spend a lot of money on a cruise ship vacation through the Norwegian fjords, while you want to see them. Enquiring further, we may discover that we can cruise very cheaply on board a Norwegian mail packet.

(3) Open discourse may clarify the nature of our Goods, revealing the absence of true conflict or opening mutually agreeable possibilities. Fisher et al. (1991:40) give the example of how the conflict was resolved between two library patrons: X, who wanted the window open, and Y, who wanted it closed. The librarian, discovering that X wanted fresh air and Y wanted to avoid drafts, simply opened a window in another room.⁽⁵⁾
 

Unfortunately, as argued above, M1 will most likely not produce free agreement; sweet reason may well not convince each of us that we really do want to adopt the candidate norm. This means that if we are to agree, someone's sense of the Good will necessarily be violated. At this point (U) ceases to be a request to pursue agreement and, "recoiling" on us, becomes instead a demand for agreement. But what could possibly induce people to agree to a norm that violates their sense of the Good? If (U) demands agreement, and the norm is not consonant with my sense of the Good, why should I agree to it?
 

The answer depends on an understanding of the nature of the relationship among people on which Habermas's "presuppositions of argumentation" are predicated. In essence, Habermas (1990a) argues that the act of communication presupposes a relationship among people, a relationship that presses them toward a discursive solution to (U) but exists prior to that solution. (U) expresses our recognition of that relationship; it does not exist independently of it. This relationship is here denoted with a capital R: a "Relationship", as opposed to a "relationship", is a connection between people based solely on communicative action, on a mutual orientation to achieve understanding.
 

Given the fact of disagreement, (U) still requires that the operative norm be agreed-upon by all. Discourse then turns from M1 to M2, in which - with special constraints noted below - we decide on our norm out of a process of power and accident. Because agreement is required, we are forced into agreement with whatever norm emerges from whatever decision-making system is in place. M2 is that procedure for producing a decision.
 

I can hear - and in fact agree with - the outraged objections to this, from any one of several perspectives. One perspective holds that this is simply a disguised perpetuation of whatever oppression already rules society.⁽⁶⁾ Another perspective is, more simply, that this is not even a disguised oppression but just the way things already occur: we argue, we find we can't agree, we observe that some norm must be adopted, and we (or the dissenters, at least) are simply coerced into submitting to it. Still another perspective holds that this procedure creates not good reasons for agreement but instead only an enormous social pressure to conform. If my argument were to stop here, these objections would indeed be devastating. But one more piece needs to be put in place; bracket these objections until the framework is complete and then judge if they apply.
 

So far, I have treated M1 and M2 as independent, isolated moments: finding M1 uninhabitable, we vanish from it and magically reappear in M2, with no structured connection between the two. The final piece of my argument has to do with the passage from one to the other. First, norms in M2 are limited to those that do not violate the Relationship that projects us into M2. Let me clarify, because there are two senses of M2 that might get confused. One sense of M2 is that of our current political strife, where power and strategic action rule supreme. In this view, nothing "projects" us into M2 - it simply exists; it is simply the default condition when discourse fails. It is not so much a "moment" as the absence of one. However, there is another view of M2, which is the one taken here. In this view, M2 is the consequence of (U) under the condition that no agreement is reached in M1. M2 thus carries within it the presuppositions that led to our projection into it,⁽⁷⁾ specifically, the assumption that we are still in the Relationship referred to above. In M2, therefore, we can only behave in ways that preserve the Relationship bringing us to M2 in the first place; in other words, the process by which decisions are made in M2 is subject to certain constraints.⁽⁸⁾ These constraints are termed the "transition conditions" of M2. They are discussed at greater length in Chilton (forthcoming), but examples of them are listed briefly below, falling into the categories of: (1) attitudes to be adopted in M2; (2) how we engage each other in M2; (3) what norms we may consider; and (4) how we behave after the process in M2 results in a norm being established.
 

Ex. 1: Each person accepts that someone - a someone who might be h/herself - will be hurt by the norm emerging from M2, in the sense that the norm will violate h/her sense of the Good. For example, if we adopt a norm forbidding same-sex marriage, then many gays will feel their sense of the Good violated, but if we permit such marriages, then the norm will violate the sense of the Good of many people who believe being gay is a sin. We can speculate that further discourse could convince everyone on one side or the other to shift their sense of the Good, letting us all happily agree, but in the absence of the possibility of such discourse, we must accept the existence of the conflict.
 

Ex. 2: In M2 people deal with one another in the recognition that no one has a privileged position to say what is Right. Each feels h/herself to be Right, but recognizes that s/he can establish this Rightness only in persuasion of the other to h/her own position, a possibility that has been attempted in M1 and finally abandoned. Acceptance of this gives M2 a tolerant character despite the pressure from forcing a decision. While all can press their views as vigorously as possible, this is done without the burdensome and disruptive attitude of moral supremacy.
 

Ex. 3: Not all norms may be considered in M2. An important and obvious constraint is that candidate norms are nonstarters if they prevent people from expressing their disagreement or from continuing the discursive process in M1. People agree to follow the norms that emerge from M2, but the very basis of their agreement is the Relationship that accepts all parties as being free to argue for their ethical understandings.

Another important constraint is that no one (or no identifiable group collectively) can consistently be subject to greater or more frequent violation than others.⁽⁹⁾ In M2 there is no longer any way to favor one party over another in the Relationship bringing M2 into existence. One's agreement to the norms emerging from M2 is predicated on one's being treated in terms of the Relationship. If the norms actually adopted consistently violate one person's sense of the Good and not another's, then there is some power differential that needs to be corrected.
 

Ex. 4: The transition conditions contemplate how people are to act even after a norm has emerged from M2. The process of coming to a decision in M2 puts a strain on people's sense of the Relationship, particularly those people who end up with the short end of the stick - those whose sense of the Good is violated by the norm. As a practical matter then as well as an ethic one, we need to attend to this strain. For example, the winners need to resist their sense (or at least mute their huzzahs) that since their Right has been victorious, they need not acknowledge those who were Wrong, i.e., the losers. Instead, the winners need to explicitly, publicly recognize that in being willing to endure the sense of violation, the losers are in fact contributing to the underlying Relationship. And a beneficial consequence of this would be to support the winners' acceptance of losing in subsequent discussions.⁽¹⁰⁾
 

Even from this limited discussion the reader should see that M1 and M2 are coordinate moments of a single process, each with its own task and organization but connected to the other by their mutual connection - an umbilical connection, so to speak - to the Relationship whose existence Habermas (1990a) establishes.
 

Understanding that decisions in M2 are constrained by the transition conditions, we now consider how, as promised earlier, we are to actually find and agree upon a norm. In what follows, I make two points: that such norms are achieved in ways that may be arbitrary or accidental but that are nonetheless morally neutral; and that achieving agreement on social norms is closely analogous to achieving it on norms of dyadic relationships. I will do this by showing the method by which norms are to be made in dyadic relationships and then how this method carries over to the domain of social norms.⁽¹¹⁾
 

What can be said about a process in which we seem to have neither justice, reason, nor agreement to guide us but only a sense of Relationship? I start with a discussion of this in terms of decision-making within a marriage.⁽¹²⁾

For concreteness I will call the two spouses Sergio and Susanna. We assume that the two have to make a decision - say, whether to go to a movie this afternoon or this evening - and that a period of open discourse (i.e., M1) has failed to produce an agreement, and if they do not make a decision, it will be made for them by default as the afternoon passes away. At this point they move into M2, in which the decision gets made by the moral and strategic contingencies of their specific situation, e.g.:

•  Sergio got to have his way the last time they disagreed on when to go.
•  On this occasion, at least, Susanna has a stronger time-preference than Sergio.
•  Sergio dislikes going to this theatre in the afternoon, when children tend to disrupt the movie.
•  If they go to the afternoon movie, they won't be able to get to the bank before it closes, and Sergio doesn't feel comfortable about "borrowing" food money.
•  Sergio has a heat rash and would prefer to go out in the cool of the evening.
•  Susanna is opposed to just going their separate ways this afternoon.
• Sergio has the only set of keys to their car.

So at least for dyadic relationships it is possible to achieve agreement. If free discourse fails, then decision-making under the constraints of the transition conditions will ensure a just outcome insofar as it is possible under the circumstances to know what justice is. Moreover - and this is the point of the long discussion of marriage - the two-moment process of decision-making will equally produce agreement in the larger domain of society as a whole.⁽¹³⁾ I am not suggesting that society and marriage are equivalent; one central difference is the presence of systemic forces in the dynamics of society as a whole.⁽¹⁴⁾ My basic point is that the agreement problem is the central issue in both situations: a norm must be agree to by all, regardless of whether "all" means the partners in a marriage or the citizens of some large society. This requirement of agreement in our finite circumstances produces M1, M2, and the transition conditions, which apply regardless of the size of the society.
 

It is surely true that it is easier to leave a marriage than a country, and this will affect whether people choose to submit or exit in the face of norms that violate their sense of the Good. But as Habermas points out in a similar context for similar reasons, "any universalistic morality is dependent upon a form of life that meets it halfway. There has to be a modicum of congruence between morality and the practices of socialization and education" (1990:207). It is here that discourse ethics needs proposals like Honneth's for practical implementation, just as Honneth needs the normative grounding provided by discourse ethics. Habermas's (1990a) argument is that even a complete skeptic can be driven into an acceptance of discourse ethics. But this "driving" arises not from some magical force (a la Hegel's "Spirit") but from concrete arguments about concrete problems of coordination. We must work to create these practical-political conditions congruent with the two moments of discourse ethics. So changes in the two-moment perspective to account for social dynamics do not alter the basic issue. In a larger society it may be harder to see how a given norm affects people or whether it satisfies the transition conditions, but agreement and the transition conditions remain the criteria for norms.
 

With this understanding of the agreement problem and how it can be overcome, and with this allusion to the mutual dependence of discourse ethics and Honneth's concerns, we are now in a position to examine Honneth's treatment of republicanism, proceduralism, and the theories of the mature Dewey.
 
 
 

III. Honneth's Analysis in Light of the Two-Moment Perspective

From a broad perspective, there are two basic critical responses to liberalism: republicanism and proceduralism. (1)

These two responses have different normative justifications: republicanism through citizens' virtues; proceduralism through morally justified procedures. (2)

The two responses see different relations between citizens and the State. In republicanism, State politics derives from public discourse; it has no separate existence. In democratic proceduralism, State politics is in the final analysis autonomous, in order to provide "universally binding decisions" in the face of continuing discursive conflict. (2)

The two responses also differ in their conceptions of law. For republicans, law is society's self-expression.⁽¹⁷⁾ For proceduralists, law is to preserve fair decision rules with a view to commanding the consent of the governed.⁽¹⁸⁾(2-3)

Republicanism and proceduralism are not the only alternatives to liberalism. John Dewey's work presents us with a third alternative, even though Dewey is claimed by both republicans and proceduralists. (3)

Dewey's alternative is not part of these critical traditions, however, because it is based on social cooperation, not "communicative consultation". (4)

There is an internal connection⁽¹⁹⁾ between cooperation and democracy, a connection arising from the necessity of social problem-solving, for which Dewey uses the model of scientific problem-solving.⁽²⁰⁾ Just as scientific problems are more easily solved when a variety of perspectives are brought to bear, so also are social problems better resolved through the participation of a variety of people. (3-4, 17, esp. 19 & 21-22)

In complex modern society it is difficult in practice for people to see the necessity of democracy to a cooperating community. (19)

To learn the connection between democracy and cooperation, people require a "prepolitical association" (19) that shows how democratic procedures are "the means for a political solution to their problems of social coordination" (20).

Note that this parallelism is also present in the two-moment perspective. Whether in the small community or the large society, agreement is still required on norms, and the search for this agreement in either society must still pass through the open discourse of M1 and the transition-condition-constrained decision processes of M2. And implicit in this comment is the observation that the problem of agreement is still unresolved. Whatever prepolitical association Dewey and Honneth may have in mind, it must be one in which (U) is satisfied.
 

This prepolitical association can be found in "a fair and just form of a division of labor", which "give[s] each individual member of society a consciousness of cooperatively contributing with all others to the realization of common goals" (20). (20-21)

As indicated in previous comments, this solution is good as far as it goes, but it does not yet go far enough to solve the agreement problem. Without such a solution, Dewey's / Honneth's perspective remains hypothetical, contingent on the yet-to-be-demonstrated existence of a "just division of labor" (as opposed to the exploitative relations of production we find around us), of "common goals" (as opposed to the diverse senses of the Good comprehended in modern society), and of peoples' "cooperative contribution" to these goals (as opposed to their coerced or unreflective submission to their roles in this division of labor).⁽²⁷⁾ Progressive political organizing is aided by Dewey / Honneth pointing to the potential significance of small-community relations, but there is one further step necessary: to be able to recognize (or if necessary create) just relations of production. It is of no use to propose local relations of production as our training ground if those relations simply recapitulate in miniature the injustices of the society.
 

The two-moment perspective is useful here, fleshing out Dewey's / Honneth's theoretical skeleton with concrete considerations that allow us to evaluate even such prepolitical associations as productive processes. Let me just summarize some ways in which it does so:

 Such associations must permit initial discourse (i.e., M1), with the presuppositions of such being identified and understood not a priori / transcendentally but instead through iterations of a reconstructive process (Habermas 1990b).

 Members of such associations must recognize and accept that such discourse will in practice be limited by constraints of time, resources, and psychology.

 The decision procedures subsequently employed (including the tactical advantages arising from the more or less accidental circumstances of their world) must meet the transition conditions, e.g., by the participants: accepting with equanimity (if not pleasure) when, at the final decision, their sense of the Good is violated; maintaining a tolerant attitude toward each others' senses of the Good, even if those senses are in conflict; remembering and honoring the Relationship by which they are bound to one another, and considering only norms that do not violate it; paying attention to the claims by identifiable groups that the decision methods disproportionately or egregiously violate their particular sense of the Good; and acknowledging their mutual dependence through rituals and concrete actions that show respect for those whose sense of the Good is violated.

 Increasing understanding and refinement of the transition conditions through a reconstructive process, just as (and for the same reason as) there is a continuing reconstructive process of greater understanding of the Relationship defined by the presuppositions of argumentation.
 

With these, I believe, we can systematically go about constructing Dewey's / Honneth's conception of a fair and just division of labor, where each individual member of society is conscious of cooperatively contributing with all others to the creation of the fundamental common goal of maintaining their Relationship with one another. The two-moment perspective is thus not in conflict with Dewey's / Honneth's perspective; its real role is to keep our attention on the agreement problem, to indicate practical ways to solve it, and to evaluate existing associations in those terms. Honneth's proposal will be justified and effective to the extent that we can find, create, or encourage production processes that are indeed "fair and just divisions of labor".
 

• Seyla Benhabib, ed. (1996). Democracy and Difference: Contesting the Boundaries of the Political. Princeton, NJ: Princeton University Press.
• Stephen Chilton (1988). Defining Political Development. Boulder, CO: Lynne Rienner.
• Stephen Chilton (forthcoming). A Second Moment of Discourse Ethics. [A relatively current version is available here.]
• Ann Colby and Lawrence Kohlberg (1987). The Measurement of Moral Judgment (2 volumes). New York: Cambridge University Press.
• Roger Fisher, William Ury, and Bruce Patton (1991). Getting to Yes: Negotiating Agreement Without Giving In (2^nd edition). New York: Penguin Books.
• Jürgen Habermas (1990). Moral Consciousness and Communicative Action. Cambridge, MA: MIT.
• Jürgen Habermas (1990a). Discourse Ethics: Notes on a Program of Philosophical Justification. In Habermas (1990:43-115).
• Jürgen Habermas (1990b). Reconstruction and Interpretation in the Social Sciences. In Habermas (1990:21-42).
• Axel Honneth (1995). The Struggle for Recognition: The Moral Grammar of Social Conflicts. Translated by Joel Anderson. Cambridge, MA: MIT.
• Axel Honneth (1998). Democracy as Reflexive Cooperation: John Dewey and the Theory of Democracy Today. Typescript.
• Kohlberg, Lawrence (1981). Essays on Moral Development: Vol. I. The Philosophy of Moral Development. San Francisco: Harper and Row.
• Kohlberg, Lawrence (1984). Essays on Moral Development: Vol. II. The Psychology of Moral Development. San Francisco: Harper and Row.
• Iris Marion Young (1996). Communication and the Other: Beyond Deliberative Democracy. In Benhabib (1996:120-135).
• Christopher F. Zurn (1998). Recognition, redistribution, and identity. Typescript.

FOOTNOTES

Approximate number of words: 8,000.

Prepared for the workshop, "Democracy and Social Cohesion", Amsterdam, Netherlands, September 5-6, 1998. I am indebted to Tom Bacig, Mark Gonzalez, Kate Pearson, Paul Sharp, Rene von Schomberg, and Janelle Wilson for encouragement and intelligent commentary. Grants from the University of Minnesota-Duluth Chancellor's Office and College of Liberal Arts helped support this work. An ancestor of this work, "Retreat from Interpersonal Harmony to Merely Fair Procedures? or, Can't We All Just Get Along?", was delivered at the First Reading XVII Conference on April 18, 1998.

This work is dedicated to V. O. & B. R., who survived it 

1. This work will not be compulsively precise about what "all affected" means, since the precise demarcation of who is affected is not germane to this analysis. Honneth (18) seems to hold that this imprecision is a central problem for discourse ethics, but I don't believe so, at least in the context of this paper. Within broad limits of how we define "all", we still face the problem of how to reach agreement.

Unless otherwise indicated, citations (like the above) are to Honneth (1998). Since Honneth's manuscript is likely to be revised, the pagination and possibly even the quotations may be inaccurate. 

2. The demand for agreement also forestalls the opposition of various postmodern positions by in effect recognizing in advance the legitimacy of their concerns and pledging to effectuate only norms that address them. 

3. The material in this section is adapted from "A Second Moment of Discourse Ethics", Chapter 1 of Stephen Chilton (with Maria Wyant Stalzer Cuzzo) The Two Moments of Discourse Ethics (in preparation). The remainder of the present essay is intended to become a later chapter in that work. 

4. Certainly our history of long-standing philosophical disagreements should not make us optimistic. One can argue that such persistent disagreements merely reflect the historical action of oppression and false consciousness, but this response merely highlights the counterfactual nature of the ideal speech situation referred to above. Even though we can point in concrete ways to the presence of repression and false consciousness throughout history, this does not prove the positive assertion that their absence would produce agreement. 

5. Chilton (forthcoming) deals with M1's virtues at greater length. I appreciate Michael Neblo for pressing me on this issue. 

6. See, for example, Young (1996); other authors in the same volume make similar points. 

7. Another way of saying this is that certain norms would represent a performative contradiction, in that their content would violate the very conditions by which we are brought to agreement with them. 

8. Let me repeat: if we were to ignore the continued Relationship with each other, then M2 would indeed be an anarchy, a mere default situation. But if Habermas (1990a) is correct, the fact that we are in this Relationship can be brought home to us by the method implied by his argument. That argument is really a "cornering" of a skeptic; while our interlocutors might not realize the existence of the Relationship, Habermas (1990a) shows how they can be brought to recognize it. 

9. Ideally, we would take turns winning and losing. One couple I know makes decisions by flipping a coin when they cannot agree. 

10. I am indebted to Eileen Theimer for this observation. 

11. In Chilton, Defining Political Development, 49., I have discussed this similarity with respect to the structural parallelism ("vertical decalage") between Kohlberg's Stages 1-3 and 4-6. But I need to clarify that this parallelism only obtains for the points made below. The actual process of M2 will be significantly different between marriage and the larger society, and I note explicitly that M2 in the larger society needs to recognize the unique, systemic forces at play. 

12. The arguments here can be applied to any mutually maintained, dyadic relationship: friendship, same-sex partnerships, etc. I discuss marriage not to privilege it over such other relationships, thus perpetuating the oppression of nonstandard marriages, but rather because it provides a clear, elaborated, and widely understood example - and one with which I have some experience. At the dyadic level, at least, the example is unproblematic. I am indebted to Ken MacKendrick for pressing me on this issue. 

13. Some readers might bristle at this use of marriage as a model for society, arguing that, for example, marriages can be dissolved. This general objection might be valid even though this specific objection is ill-founded, since one can renounce one's citizenship. 

14. Systemic forces are those forces that are not observable from the perspective of any single individual or group of individuals. A classic example is the oppression of capitalism, which is not apparent from the perspective of any single relationship, even that of wage negotiations between an owner and a worker. The worker's complaint about low wages ("If you don't pay me better, I can't afford health insurance") is met by the employer's reference to market pressures ("If I pay you better, I'll just go broke, and then you won't have a job at all"). To both of the participants, the market appears as an impersonal, "natural" force, its root in the overall system of relationships being invisible. Such an oppression can only be uncovered by systemic analysis, a la Marx. A major task of critical theory is to uncover, understand, and evaluate such systemic forces. 

15. Witness the emphasis on communication in Honneth's description of republicanism:

Because the tradition of republicanism assumes there is a solidary citizenry that is in a position to organize society itself through processes of communicative consultation and negotiation, state politics itself can be grasped here only as the implementation of publicly negotiated programs; the government and the parliament are . . . the institutional spearhead of the progressively rejuvenating communication process which has its real center in the citizens' democratic public sphere (2). 

17. ". . . political republicanism by nature has a certain tendency to understand legal norms as the social instrument through which the political community attempts to preserve its own identity" (2). 

18. ". . . basic rights according to the proceduralist conviction represent a kind of security for the continued existence of the interplay of the democratic public sphere and political administration. . . . [Law] represents state-sanctioned but morally legitimated precautionary measures to protect the democratic procedure in its entire complexity" (2-3).

I hope I am not over-interpreting (or even misinterpreting) Honneth here. To me, the proceduralist view of law still goes back to the need for agreement, and so I am implicitly adding what I believe is implicit in Honneth's statement above: "Protection is needed for this democratic procedure 'in its entire complexity' (i.e., including 'the interplay of . . . the public sphere and political administration' (2), because people will agree with the dictates of political administration only when these procedures protect their ability to have their concerns considered."

19. "Internal connection" means a logically necessary relationship, i.e., a relationship inherent in the logical structure of the concept itself, not based on the empirical observation of a (necessarily contingent) association.

20. The connection is associated with Honneth's use of such terms as "rationality" (15, 16), "intelligence" (15, 28), and "balanced, comprehensive, intelligent solutions" (15).

21. Scientific inquiry is not an appropriate model in any case. As Habermas (1979a) has argued, the validity claims of scientific truth, moral rightness, and personal sincerity (including knowledge of one's aesthetic reactions) are redeemed by distinct methods: the methods developed by the philosophy of science, discursive processes, and psychoanalytic methods, respectively. The scientific method might provide Dewey with a metaphor for social problem-solving, and certainly how we establish scientific truth will be part of the context of social issues, but it cannot provide the direct justification Dewey desires.

22. Cooperation refers to arrangements that are more local and more production-oriented than those of democracy (or so Dewey speaks of them).

23. Thus we avoid a consensus theory of truth; we do not consider that our consensus or dissensus over the nature of nature renders our theories true or false.

24. This is not to say that small communities can have no oppression or that organizing against it is easy.

25. When this reasoning is applied to the relationship between family structure and political structure, political scientists and sociologists know it as the "correspondence hypothesis".

26. In objecting that productive processes do not provide normative grounding, Zurn (1998:9) misinterprets Honneth's purpose. Honneth is trying to find a practically useful link between people's actual experiences and normatively justified social arrangements.

27. In fact, Honneth ends his essay by referring to the normative idea of democracy as "first and foremost a social ideal" (26, emphasis supplied).