This article was downloaded by: [Northeastern University]On: 04 November 2014, At: 18:24Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Inquiry: An Interdisciplinary Journal of PhilosophyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/sinq20

Grasping the Force of the Better Argument: McMahonversus Discourse EthicsWilliam Rehg aa Saint Louis UniversityPublished online: 06 Nov 2010.

To cite this article: William Rehg (2003) Grasping the Force of the Better Argument: McMahon versus Discourse Ethics,Inquiry: An Interdisciplinary Journal of Philosophy, 46:1, 113-133, DOI: 10.1080/00201740304524

To link to this article: http://dx.doi.org/10.1080/00201740304524

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe Content. Any opinions and views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Review Discussion

Grasping the Force of the BetterArgument: McMahon versus DiscourseEthics

William RehgSaint Louis University

A book like Christopher McMahon’s Collective Rationality and CollectiveReasoning* offers such a rich and thought-provoking store of arguments thatthe reviewer must either settle for a superficial overview or go for depth at theexpense of coverage. In this review essay I opt for the latter, focusing inparticular on his analysis of collective reasoning and critique of discourseethics. Indeed, McMahon seems to have reached his views partly through hiscritical scrutiny of Jurgen Habermas’s discourse ethics, which he finds eitherimplausible or not very distinctive.1 At the heart of McMahon’s argument,however, lies an individualistic account of the normative force of reasons,which ultimately undermines his attack on discourse ethics. After some briefremarks on his broader project (Section I), I clarify the competing models ofcollective reasoning in McMahon and discourse ethics (Sections II and III);we can then test their implications for moral objectivity (Section IV).

I. The Principle of Collective Rationality

McMahon situates his account of collective reasoning in a general treatmentof cooperative rationality, which he treats in the first four chapters. Theanalysis turns on questions regarding the reasons that people might have forcooperating, that is, ‘whether an individual has sufficient reason to contributevoluntarily to a cooperative enterprise’ (p. 1). Game theory answers thisquestion on the assumption that each individual follows the ‘Principle ofIndividual Rationality’ (PIR): each ‘player’ chooses the option (cooperate ordefect) whose consequences maximize the satisfaction of its own interests. Inmaking the choice, that is, each player anticipates the consequences (the

* Christopher McMahon, Collective Rationality and Collective Reasoning (Cambridge:Cambridge University Press, 2001), ix + 251 pp., $54.95, pb. $19.95. Intralinear citations referto pages in this book.

Inquiry, 46, 113–133

� 2003 Taylor & Francis

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

possible outcomes or payoffs) in the expectation that others also want tomaximize individual utility, and then ranks the outcomes according to its ownvalues and interests. Notice that nothing in the formal choice situation (or‘game’) excludes the possibility that an individual’s interests includealtruistic concerns or other moral values. The problem is that no one whohas moral interests can assume a similar stance on the part of others. As aresult, in certain types of games, PIR leads individuals to choose suboptimaloutcomes, defecting from mutually beneficial cooperative schemes (i.e.,schemes that yield more benefits for each individual, if each cooperates, thandoes mutual defection). Such problems run deep: in the end, game theory failsto explain social order, or even shared linguistic meanings.2

Conceding the intractability of these problems from the standpoint of PIR,McMahon reformulates the leading question: we should simply assume thereis a requirement to cooperate and ask, not whether cooperation is rational, butrather how, or under what conditions, ‘cooperatively disposed’ individualshave more reason to cooperate than not. This shift in perspective rests on adifferent principle of rationality, the ‘Principle of Collective Rationality’(PCR). According to PCR:

One has sufficient reason to contribute as provided to a cooperative scheme thatproduces something that one regards as good if the value to one of the outcome of thescheme, when one’s contribution is added to the others that will actually be made,exceeds the value to one of the noncooperative outcome (pp. 21–22).

If, conversely, the noncooperative outcome has greater value to one, then onehas sufficient reason to defect (p. 22).

Notice that both PIR and PCR have the agent assess outcomes in light of hisor her interests and values – the ‘value to one’. How then does PCR differfrom PIR? In terms of game theory, PCR redefines payoff schedules,lessening the rewards of defection and free-riding to the point wherenoncooperators cannot benefit at the expense of cooperators (pp. 13–16, 21–23). This redefinition in effect builds the value of cooperation into the formalchoice situation: as ‘cooperatively disposed’, one has no interest inadvantageously exploiting the cooperative tendencies of others by defectingat their expense. If one’s assessment of the actual choice situation leads one tobelieve that enough people are similarly disposed to cooperate, then onechooses according to PCR rather than PIR.

PCR remains a formal principle in the sense that the in-built cooperativeinterest leaves open the specific values by which individuals evaluate anygiven cooperative scheme and the contributions it requires of them. Oneperson might evaluate a scheme according to the fairness of its consequences,another from a religious standpoint, and so on. Because he makes noassumptions about what further values cooperatively disposed individualshave, McMahon can get his account of social cooperation going without

114 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

having to defend the priority of a substantive value such as fairness. Thosewho take the latter approach (e.g., John Rawls) must suppress or sweep asidethe possibility that for reasonable persons other moral values mightsometimes compete with fairness. McMahon’s model does not introducesuch a priority claim, and so his analysis of multicultural conflict has a morerealistic flavor. In particular, his model gives us a sharper sense of thedilemma that can arise in situations and institutions that take fairness as theprimary standpoint from which to adjudicate social conflicts. In suchsituations, which are not untypical, fairness requires participants in effect toreduce other moral values to interests that can then be fairly adjudicated,normally through the use of a decision-making procedure based on the idea ofpublic reason. McMahon’s model allows one to see how cooperativelydisposed individuals might nonetheless rationally opt to defect in suchsituations – or why they might resort to some sort of power or bargainingadvantage, thereby subverting the procedures (pp. 40-41, 90-95).

By framing the challenge of pluralism in terms of the above dilemma,McMahon can approach multicultural conflict within a more inclusiveframework – one more open to people whose comprehensive conceptionssometimes put them at odds with democratic procedures – than theorists whofollow Rawls. His interesting analyses of such topics as authority,governmental legitimacy, and democratic deliberation thus have a realisticedge, honed by the problem of the ‘minority democrat’ who disagrees withthe majority opinion.3 Similarly, his treatment of moral discourse underconditions of value-pluralism is particularly interesting as an attempt torespond to the foregoing dilemma (chap. 7).

Behind the analysis of conflict resolution, however, lies an individualisticconception of moral discourse and the cogency of moral reasons: the finalarbiter of cogency is, in the end, the individual reasoner.4 This individualismhas fateful consequences for McMahon’s approach to collective reasoningand his critique of discourse ethics.

II. Two Models of Collective Reasoning and Judgment

In Chapter 5 McMahon turns to the cooperative scheme he calls ‘collectivereasoning’, which he characterizes in problem-solving terms:

Collective reasoning, as I understand it, . . . is a cooperative effort, involving linguisticexchange, to answer a question or solve a problem confronting a group. It presupposesthat each member of the group wants the question answered or the problem solved (p.105).

Like reasoning in general, collective reasoning ‘seeks to identify, anddetermine the force of, the reasons relevant to some question’. Although

Grasping the Force of the Better Argument 115

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

agreement may issue as a by-product of collective reasoning, the ‘benefit’ atwhich such reasoning specifically aims is each participant’s ‘justifiedjudgment’ (ibid.). More precisely, ‘each expects to acquire a judgment thatis better justified than the best justified judgment she could have made byreasoning independently’ (p. 106). Here I am concerned above all withcollective reasoning as it bears on moral questions, and in what follows I usethe phrase ‘moral discourse’ as equivalent to moral collective reasoning.5 Iconstrue the term ‘judgment’ broadly here: depending on the question underdiscussion, the judgments we hope to justify can be about either the correctmoral norm (what McMahon calls a ‘moral principle’) or about how aparticular person should act in a concrete situation.

How exactly does discourse lead to better justified moral judgments?Because he understands morality as oriented toward resolving conflicts ofinterest, McMahon says moral discourse aims ‘to identify the appropriatecourse to follow in situations where interests differ’ (p. 136). To resolve suchconflicts requires participants to agree on the values by which one identifiesthe morally relevant facts (p. 133).6 Collective reasoning need not involvedisagreement, but only an uncertainty or question about what the moral normsor values require. Disagreement arises when participants find themselvesreaching different answers to such questions. McMahon distinguishes twomain sources of such differences (pp. 106–08). Disagreement may arisebecause of ‘cognitive malfunction’ within the group, say when somemembers lack or mistake the facts, lack or misunderstand certain concepts orvalues, or fail to appreciate the implications of the available reasons. In othercases, disagreement can arise without malfunction, say because of valuepluralism or the inherent ambiguity of moral concepts. To get at McMahon’sdispute with discourse ethics, I focus on his analysis of discourse to overcomemalfunction. The question, then, is how discourse helps a group to overcomemalfunction so as to arrive at better justified moral judgments. The answerdepends on how one interprets the outcome of collective reasoning as a‘collective product’ from which individuals gain the benefit of better justifiedjudgments (p. 109).

McMahon examines two interpretations of this product. Both assume thatdiscourse produces ‘a common pool of criticized arguments germane to theissue’ under discussion (p. 109). ‘Criticized’ arguments are critically testedarguments, what survives after the participants have subjected their variousinputs regarding putative relevant facts and their implications to objections,further questions, and possible modifications. As a shared pool of argumentsor reasons relevant to answering some question, this pool represents a publicgood upon which individuals can draw when they deliberate before making(and justifying) a judgment about the correct or reasonable answer to thequestion.7 Notice that the pool also contains the various participants’judgments about which answer the relevant reasons justify: the ‘fact that an

116 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

apparently competent reasoner has drawn a certain conclusion from a pool ofreasons . . . is itself a presumptive reason to believe that conclusion correct’(p. 110). McMahon dubs these ‘interpersonal reasons’ in contrast tosubstantive ones. The latter include primary reasons that directly support aview as well as debunking reasons for rejecting some position (p. 115).

The two interpretations part company over the nature of the cooperationrequired for discourse to improve judgments. On one view, discoursefunctions primarily as a source of information and corrective againstindividual biases, but each individual makes up his or her mind independently– judgment itself remains an individual responsibility, so that if a consensusemerges, it does so in a ‘piecemeal’ fashion. The alternative view holds that agreater cognitive benefit, ‘some extra increment of justification’, results ifjudgment is a collective act. Hence ‘each withholds final judgment until aconclusion all can regard as correct has been identified’ (pp. 111–12). Anyjudgments added to the pool prior to agreement must be regarded as merelytentative; when consensus emerges, it emerges as an integral act in which allthe participants together make a final judgment. The group acts as a whole insome sense, as a kind of collective agent, when it issues its judgment.8 As heexplains the second view, ‘it is only through collective judgment – that is,integral consensus – that the true force of all the relevant reasons can begrasped’. Consequently, ‘no one can attain a fully justified judgment in theface of disagreement’ (pp. 112–13, emphasis added).

McMahon seems to have gotten this idea – call it the ‘integral consensus’(IC) model for short – from discourse ethics, in particular from a certainreading of Habermas’s principle of universalization (U).9 According to (U), amoral norm or principle is valid ‘just in case the foreseeable consequencesand side-effects of its general observance for the interests and value-orientations of each individual could be jointly accepted without coercion byall concerned’.10 In any case, McMahon’s analysis of the IC model leads himto reject it as a plausible interpretation of Habermas’s discourse ethics. Healso rejects the reading of discourse ethics as a ‘constitutive’ model of moralimpartiality, according to which discourse first creates moral reasons out ofnonmoral inputs, thus constituting what it means for a norm to count as agenuine moral principle (pp. 141, 143–46). McMahon opts instead for a‘standard epistemic’ interpretation: discourse helps us to figure out the trueprinciples of morality, but does not itself define or constitute those principlesas moral.11 Rather, what makes something a true moral principle is its relationto the relevant reasons (pp. 139–40). As McMahon interprets discourse ethics,then, its distinctive significance lies ‘solely’ in the emphasis it places oncollective reasoning as allowing ‘members of a group to acquire a betterunderstanding of the force of the relevant reasons’ (p. 149). But that emphasisdoes not require collective judgment; in the end, only the piecemeal model ofconsensus holds up.12

Grasping the Force of the Better Argument 117

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

The thought that discourse ethics requires each person to suspend finaljudgment until everyone can participate in a single collective act of judgmentwould seem to make that theory implausible on its face. However, everythingdepends here on what we mean by ‘withholding final judgment’ and how thisfunctions as a requirement. To clarify matters, I distinguish three theses inMcMahon’s characterization of the IC model:

(T1) The Better Justification Thesis: An integral consensus bestows ‘someextra increment of justification’ on our judgments – either because it allows a‘better grip’ on the ‘true force’ or cogency of arguments or because thedisagreement of others weakens one’s assessment of cogency.(T2) The Full Justification Thesis: A fully justified judgment must be acollective judgment.(T3) The Suspension Thesis: if we want to maximize the benefits of discourse,then each individual must suspend final judgment until the group achievesintegral consensus.

McMahon’s various formulations of (T1) admit of different interpretations.Inasmuch as the cognitive benefits of discourse depend on an integral act ofjudging together as a group, then (T1) requires more than the merecoincidence of agreeing judgments; that is, those making the judgments mustknow they agree. But (T1) would also seem to require more than the mereknowledge that we agree: my merely knowing that you agree with myjudgment does not of itself improve my understanding of the justifyingreasons, though it may increase my subjective confidence in my judgment. Toanticipate the next section, (T1) says that others’ agreement improves ourhold on the objective cogency of arguments as justifications for judgments.The ‘our’ is important here. Our mutual agreement, as a collective result,provides the better – more objective or more reliable – measure of the trueforce of those reasons than do our individual assessments taken singly.

The Suspension Thesis (T3) also requires clarification. In his earlier essay,‘Discourse and Morality’, McMahon interpreted (T3) as the flip side ofcollective judgment, which he took as a kind of requirement: Do not judgeuntil you can do so together, consensually, with everyone else. As McMahonhimself recognized, the idea of suspending judgment should not be taken tooliterally, as though it required us indefinitely to postpone decisions ondisputed questions. Rather, integral consensus is a metaethical idealization,‘part of a theory of what we mean when we claim that a principle is a genuinerequirement of morality’.13 So if Habermas’s (U) posits collective judgmentas a condition for full justification, then this functions as an idealization, adirection for improving our discursive endeavors rather than a fully realizablestate. We can acknowledge such an idealization and still make individualjudgments, provided we recognize the associated cognitive provisos and

118 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

obligations – that our judgments are not fully justified as long as reasonablepeople disagree with them, and that we should to the extent feasible seekothers’ agreement.

Nonetheless, even as an idealization, integral consensus might be so strongas to render many of our everyday moral judgments somehow defective orunreasonable. In fact, this is just what McMahon argued in ‘Discourse andMorality’, where he reached the conclusion that according to (U), judgmentsmade in the face of disagreement ‘are necessarily rationally defective’.14 Thisinterpretation of discourse ethics – call it the ‘strong IC model’ – depended ona particular way of construing cases in which people disagree over the moralprinciple(s) by which they ought to resolve some conflict of interest, and toresolve the disagreement they make claims about people’s needs. The ‘needs-based scenario’, as I shall call it, assumes participants strive to identify needsas a way to determine which interests have a kind of priority, and thus cansupport a moral claim on others. More precisely, when one calls an interest a‘need’, one links it with one or more basic values (goods or ends) of sufficientimportance to merit some kind of principled moral recognition from thecommunity, which therefore ought to protect or even positively support thepursuit of that interest. If health is such a value, for example, then one mightargue that the interest in health care insurance counts as a need whosesatisfaction ought to be guaranteed by a universal right to some minimumlevel of insurance.

Although it does not capture all aspects of moral reasoning in general, theneeds-based scenario introduces some further issues in McMahon’s critiqueof discourse ethics. The crucial point concerns the status of need-claims.According to McMahon, the IC model assumes that each individual has aprivileged access to what counts as the satisfaction of his or her needs. As heexplains:

the individuals whose interests are at stake have a special role to play in the process ofneed determination. They have a certain authority regarding need claims that pertainto them. To be sure, an individual with an interest at stake must convince the othersthat it has legitimacy if it is to be promoted to the status of a socially recognized need.But such an individual is in a privileged position with respect to the question ofwhether her needs are being met. Thus others cannot rationally reject any claim shemight make about this unless they can convince her that she is mistaken. Without herconcurrence, they must suspend judgment.15

McMahon’s conclusion in this passage depends on an undifferentiated notionof rational conviction that I once proposed: ‘I can be rationally convinced ofthe worthiness of a norm only if I suppose that others are rationallyconvinced, which in turn depends on their supposing that I am rationallyconvinced. If this is not to be a vicious circle, then rational conviction must besomething that we arrive at together’.16

Grasping the Force of the Better Argument 119

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

McMahon seems to understand ‘rational conviction’ as an all-or-nothingaffair: if a rational conviction requires a single collective act of judgment,then any independent judgments taken in the face of dissent must count as‘necessarily rationally defective’. Hence his strong-IC reading of theSuspension Thesis as an overly demanding idealization that renders thejudgments we often must make, in the face of practical necessities, rationallydefective. This interpretation puts discourse ethics on a collision course withthe requirements of everyday personal morality. For many issues, we expect(and the law permits) individuals to act on the basis of independent judgmentswith which others might reasonably disagree. If the individual’s choicefollows upon due reflection and consultation, then we do not consider others’disagreement to make that choice necessarily defective or irresponsible.

But we need not accept this all-or-nothing conception of rationalconviction and the strong IC model it entails. If rational convictions arejustified judgments, then the two Theses on justification (T1 and T2 above)suggest a weaker IC model that distinguishes between full and partialjustification. In the face of disagreement, then, individuals could makepartially justified judgments that are, not ‘necessarily rationally defective’ (ina pejorative sense) but reasonable.17 At the same time, the model holds on toan idealized notion of integral consensus as a conceptual limit in whichindividuals who make partially justified independent judgments would(counterfactually) achieve the full measure of the ‘extra increment[s] ofjustification’ that come with others’ agreements. This idealized limit sets adirection for improving moral discourse as a cognitive endeavor; (U), in otherwords, articulates the demand internal to discourse that participants strive formutually acknowledged agreement. But this demand does not imply aSuspension Thesis in any literal or strong sense.

This weaker interpretation of discourse ethics – call it the ‘mutualagreement’ (MA) model – incorporates elements of McMahon’s piecemealview. That is, the MA model allows us to envision a consensus that firstemerges piecemeal and then gradually becomes integral as various parties andscattered pockets of consensus recognize their mutual rational agreement. Inthe next section I further clarify the MA model in contrast to McMahon’sposition, which I’ll label the ‘PA model’ for ‘piecemeal agreement’.

III. The MA Model as an Interpretation of Discourse Ethics

In his book McMahon characterizes discourse ethics in terms of the moredifferentiated model of justification expressed by the two Theses onjustification. So when he revisits the needs-based scenario he casts hiscriticism of discourse ethics along different lines from those in ‘Discourse and

120 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

Morality’. Before I examine that argument (Section IV), however, I want toclarify the competing ideas of justification in the MA and PA models.

The key difference lies in McMahon’s rejection of the Theses onjustification, which are central to the MA model. To see this, consider hisanalysis of radical dissent – cases in which a single individual disagrees witha group whose members appear to be competent reasoners (pp. 116–18). If, asMcMahon holds, others’ judgments are further (interpersonal) reasons in thecommon pool, then it seems that the dissenter should defer to the majorityopinion. Not so, McMahon objects. If the group’s members are simplyfollowing an opinion leader, then the majority judgment rests on a singleindividual’s assessment of the substantive reasons for the majority view. Thusthe majority view has rational force for a dissenter only to the extent thatthose in the majority have each made independent judgments based on theirown personal assessments of the substantive reasons. But if a group’sconsensus has force only insofar as it comprises a set of independentjudgments, then piecemeal consensus should normally suffice for gaining thecognitive benefits of discourse. McMahon concludes that a dissenterconfronted by a majority of competent reasoners should at most reconsider– but not necessarily change – her position in view of the substantivearguments. If the majority arguments do not convince her after furtherreflection, she ‘should stick to her guns’ (p. 116).

This situation poses a puzzle for McMahon, for it suggests that‘interpersonal reasons are stronger the less attention the members of a grouppay to them in deciding what to believe’ (p. 116) – so our consensus, it seems,has cognitive significance only if we disregard one another’s final judgments.But the puzzle dissolves if we notice that, according to the PA model,discourse primarily serves two cognitive ends: information-pooling (theaddition of substantive reasons to a shared pool) and criticism (counter-arguments and debunking reasons that alert participants to possible mistakes,biases, etc.) (pp. 117–18). Once I have these benefits, the mere fact thatothers’ final judgments agree with mine adds nothing to the reasons thatactually justify my judgment, nor does it add anything to my grasp of thosereasons. Hence the conclusion McMahon considers decisive against discourseethics: if ‘collective judgment can genuinely grasp the force of a certain bodyof reasons, so can a judgment made by an independent individual’ (p. 231,note 19). Disagreement does not of itself diminish a dissenter’s ability tograsp the force of the available reasons, though it may weaken her initialconfidence in her position. Radical dissenters can not only make substantivelycorrect judgments, they can grasp the cogency of arguments as well as agroup can when it achieves rational consensus.

The MA model accepts the first half of this conclusion (i.e., that a dissentermight be correct, more on which momentarily), while rejecting the secondhalf. Even after one has engaged in discourse and reconsidered one’s views in

Grasping the Force of the Better Argument 121

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

light of others’ contributions, one cannot, as an individual, assess theobjective cogency of reasons as well as a group can when it has arrived at arational consensus. According to the Better Justification Thesis (T1), areasonable consensus of reasonable individuals on some moral claim providesceteris paribus a better reflection or measure, a more accurate assessment, ofthe objective cogency of the relevant arguments. One can, I suspect, cast theidea as a matter of degree: a judgment is better justified to the extent thatreasonable individuals in discourse agree on it.

To motivate (T1), imagine that the participants in some discourse attemptto follow McMahon’s advice and disregard one another’s final judgments.Thus, after their discourse has worked up all the relevant information andarguments, the participants disperse and make their final judgments in private.Now suppose that one of these dispersed participants subsequently talliedeveryone’s final judgments. According to the MA model, our pollster wouldhave a better sense of the objective force of the arguments, precisely becauseshe would see how persuasive or cogent those arguments were, in the end,across the different perspectives represented by the various group members.She would not necessarily have a better grip on, or understanding of, thereasons themselves, but the collective picture that emerged from her tallywould give her a more accurate or objective estimate of their force (assuming,of course, that the group members are reasonable). She could thus claim thather judgment – now reinforced by this improved estimate – is better justifiedthan those of the other members. As long as the latter remain ignorant ofothers’ judgements, they lack the reinforcement of this objective measure;and to that extent their estimates of cogency remain subjective, merely theopinion of one individual. Now extend the pollster’s position to everyoneelse. If what holds for her holds for everyone, then a group whose membersrecognized their mutual agreement would possess a superior justification thana group of privately judging individuals.

One might object that since both the pollster and the isolated individualsjustify their judgments with the same substantive reasons, both are equallyjustified. Moreover, because the pollster does not have a better understandingof the reasons themselves, both she and the others have the same ‘grip’ on thereasons. In the next section I argue that one can, in fact, conceive mutualagreement as affecting the very content of arguments. But the objection alsomisses the real thrust of the MA model of justification and its link withagreement. In the remainder of this section I elaborate on that connection.

Interpreted in terms of the MA model, discursive idealizations setconditions on full justification rather than on judgment per se. This holdsboth for (U) and for the broader (and less controversial) Discourse Principle(D), according to which a practical norm (moral, legal, etc.) is valid – that is,fully justified – only if all the concerned parties could agree to it in a rationaldiscourse.18 As conceptions of justification, such principles summarize what

122 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

it means for the grasp or estimate of cogency to be as reasonable and accurateas it could ideally get, relative to the given information, reasons, andarguments. These principles, therefore, define the ‘objective cogency’ ofarguments and what it would mean fully to grasp that cogency, that is, toassess it with complete accuracy. A group can never fully achieve such anassessment, and even if it could, no one could confirm or guarantee that it had.Nonetheless, precisely our experience in actual discourse can lead us to thedefeasible presumption that, were we able to discuss and test the availablearguments to such a point where our assessment could not get any better, thesubstantive outcome we’ve actually agreed upon would probably not beoverturned.

We might refer to such full justifications as ‘conclusive’, in the relative orindexical sense that they allow participants to reach a definite conclusionbased on the recognition that one set of the currently available arguments isdecisively stronger or winning – even if fortunes could later turn in the otherdirection as new information becomes available. In this context ‘conclusive’does not mean ‘logically compelling’ but rather has a dialectical sense: Iregard a particular view or argument as conclusive when I hold that one candissent only at the cost of adopting positions that competent reasonersconsider ‘beyond the pale’, even if those positions do not contain logicalinconsistencies. To avoid connotations of unrevisability, one might prefer theterm ‘decisive’. In any case, (U) attempts to define the conditions on such aconclusive or decisive set of arguments.

Such situations, in which all reasonable participants can regard thearguments for a particular position as decisively superior – as the argumentswe would agree upon were we ever to achieve full justification – differ fromthose in which interlocutors who disagree with one another can stillacknowledge alternative positions as reasonable, or at least as notunreasonable. In the situation projected by the ideal of full justification,dissenters would appear radical indeed, even unreasonable in light of the bestcurrently available practices of justification. But this is quite consistent withour everyday intuitions about morality. We generally recognize that lessradical forms of dissent, particularly on difficult moral questions, can bereasonable. But we would most likely doubt the reasonableness or moralcompetence of anyone who today seriously attempted to justify slavery, forexample.19

However, the projection, on the basis of actual discourse, that a majorityconsensus would meet the demands of full justification does not guaranteethat the majority view is substantively correct or true, or that it will not beoverturned later on the basis of new information. Thus even the lone dissentercould make a substantively correct moral judgment, though that judgmentwill be at best only partially justified and may even be rationally defective atthe time the dissenter makes it, depending on how radical the dissent. The

Grasping the Force of the Better Argument 123

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

difference between a correct judgment based on a full, consensualjustification and one based on a partial justification is analogous to thedifference between executing a sports maneuver with masterful ease andsimply getting the job done, awkwardly and with as much luck as skill. Thedifference gains further support if we distinguish truth from the concept offull justification.20

One might raise a troubling objection at this point. The MA model impliesthat a group should at least have a better grasp of the true force of reasons thana lone dissenter. Even if no group can ever fully satisfy the idealizedconditions of rational discourse, the fact that these conditions set a directionfor improving discursive justifications suggests we can make such acomparison. If we assume a situation in which apparently reasonableinterlocutors disagree because of a cognitive malfunction, then the majorityview (and possibly both sides) involves some mistake regarding the pool ofrelevant reasons and/or the import of those reasons. How then can a group dobetter at grasping the force of reasons than a dissenter who judges correctly?In making a judgment, one does not simply guess at the answer but ratherassesses the various considerations and arguments for their relative cogency.So a dissenter who gets it right has, presumably, made a better estimate of thatcogency than the group, even if neither side can claim to have a fully justifiedposition.

This objection deserves a closer analysis than I can provide in the space ofthis review. But notice that its force depends on the type of case one has inmind. In simple cases that involve rather straightforward reasoning, the MAmodel can allow that a dissenter has a better grasp – in the sense of a morecomplete understanding – of an issue than does a group, simply because sheknows some crucial fact of which the group is ignorant but which casts awhole new light on the shared pool of reasons. The fact represents a new pieceof information that nullifies the previous consensus and calls for a reopeningof the discourse. If the issue is so straightforward that we cannot see how anyreasonable person who knows this fact could fail to draw the dissenter’sconclusion, then the model allows us to presume counterfactually that thegroup would agree with the dissenter if it knew this particular fact, and if itwere not to agree, we would have to doubt its competence. But many cases ofmoral conflict are more complex. Typically both sides lack information orimportant conceptual resources for fully understanding the information theyhave. In such situations, it is possible that the group better appreciates the trueforce or import of the available information and reasons than the dissenter,yet the latter nonetheless judges correctly.

To illustrate the basic idea, imagine a trial situation in which an innocentman is the victim of some unlikely coincidences such that the availableevidence indicates guilt, and so the jury issues a verdict of ‘guilty’. Nowsuppose that the parents of the accused man insist on his innocence, invoking

124 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

as grounds their intimate knowledge of his upright character. If the parents areover-estimating their son’s strength of character – if they are subject to somebias or even self-deception regarding their son – then the jury has reached abetter justified conclusion (relative to the available information) than that ofthe parents, who nonetheless are judging correctly (relative to some notion oftruth).

IV. The Objective Force of Reasons

If the MA interpretation of discourse ethics holds up, then that theory emergesas more distinctive than McMahon would have it. Discourse ethics wouldcount as ‘strongly dialogical’ insofar as the full justification of moraljudgments is defined in relation to an ideal limit in which everyone mutuallyagrees. Pace McMahon, however, ‘strong dialogicality’ does not require aproblematic suspension of independent judgment – a suspension that, takenliterally, would paralyze action and, taken as an ideal, would render choicesrationally defective and thus irresponsible.21 The MA model, rather, reducesthe force that independent judgments have in the face of reasonabledisagreement. So we might say that the model requires disagreeingindividuals to suspend the ‘conclusive force’ of their judgments as the onlyreasonable conclusion.

Yet McMahon has a further reason for rejecting discourse ethics, one basedon his conception of the objectivity of moral reasons. In this last section, weexamine that conception and the argument based on it.

(1) Cases in which discourse aims to overcome cognitive malfunctionpresuppose ‘that there is a single correct way of understanding the rationalsignificance of a given body of facts’ (p. 132). In moral discourse, however,values determine which ‘factual features of the world’ are relevant to beginwith (p. 136). Thus objectivity implies that ‘there must be one set of valuesfrom the standpoint of which the questions at issue are appropriatelyaddressed’ (p. 133). As noted earlier, McMahon’s epistemic approach todiscourse holds that the substantive reasons, and not discourse as such, arewhat make a moral principle the correct one for resolving some problem.Discourse helps participants identify the pool of relevant reasons and theirimplications; for a moral question, this pool would include the correct valuesthat allow one to determine the relevance and weight of the other reasons thatare put forward in arguments. But the stronger arguments or reasons – andthus the values by which these are identified – supply the objective standardthat constitutes principles as genuinely moral (pp. 138, 141).

But how, exactly, should we understand such objectivity? McMahon’sanswer emerges from a discussion of two models of the normativity of

Grasping the Force of the Better Argument 125

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

reasons. What he dubs ‘normative realism’ treats reasons as corresponding toan ‘independent realm’ of ‘abstract normative entities’ (p. 113, also pp. 122–23). To avoid the metaphysical and epistemological baggage that encumbersthis view, one might opt for a more Wittgensteinian conception, according towhich ‘reason-constituting rules are grounded in practices’ whose normativeforce ‘is simply a matter of the existence of a particular history of judging’. Inthe eyes of some interpreters, however, this conception merely explains the‘illusion of normativity’ (p. 114). The result is a kind of rule-skepticalhistoricism.

Before elaborating his own approach, McMahon notes that neither of thepure alternatives implies the superiority of collective judgment in graspingthe force of reasons (pp. 113, 114f). To say this he must presuppose that eachindividual can have something like full access to the objective force ofreasons, that is, an access that renders reasons transparent, thereby giving theindividual complete mastery of their import. But why accept that assumption?McMahon’s arguments on this point overlook the perspectival character ofthe individual’s ‘contact’ with the pool of reasons. Whether one conceives therealm of reasons in normative-realist or practice-based terms, it seemsplausible to hold that full access to that realm – the contact necessary to gaugethe true force of reasons – must emerge collectively from the combinedcontributions of the full range of actor perspectives, which no singleindividual can occupy. In that case, no single individual can grasp the force ofthe arguments in the sense of having command or mastery of all the relevantconsiderations and their implications – which, again, does not preclude apartial grasp that issues in a correct judgment.

In any case, McMahon proposes his own ‘hybrid’ of the two puremodels. In doing so, he seeks a conception of objectivity that wouldexplain how a lone dissenter ‘who disagrees with everyone else about howa particular concept is to be employed may nevertheless be correct’ (pp.125, 122–27). This hybrid should explain the objective force of reasons, inparticular of values, without falling into normative realism, that is withoutappealing to a representationalist epistemology more suited to the grasp ofempirical facts (pp. 133–34). McMahon singles out two elements ascrucial. On the one hand, judgmental practices are social practices, whichmeans that disagreement should lead dissenting individuals to re-examinetheir position. On the other hand, socially competent individuals haveacquired the capacity to participate appropriately in these practices. Helocates the normative standard for distinguishing between appropriate andinappropriate participation in the idea of ‘proper functioning’ under‘favorable conditions’ for grasping ‘conceptual transitions’, that is forproperly employing the relevant concepts (p. 128). More specifically,through socialization the individual acquires ‘judgmental inclinations’ or‘dispositions’ regarding such transitions, which cash out the normative

126 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

import of reasons. The socialized individual’s inclinations possess ‘apresumptive claim to correctness grounded in his presumptive status as aproperly functioning participant in the social practices grounding theconcepts employed’ (pp. 126–27). Consequently, once discourse hasestablished the pool of relevant criticized arguments, the morallysocialized, favorably positioned individual possesses in principle all theresources for correctly assessing its objective implications even if everyoneelse disagrees.

However, this standard built into properly functioning inclinationspresupposes an objectivity inherent in the judgmental practice itself, namely‘a form of normativity that is prior to and independent of’ any particulardiscourse or outcome (p. 128). To get at this deeper level of objectivity, whichmust involve more than mere history, notice first how McMahon’s analysisdraws on Philip Pettit’s account of rule-following, albeit in a way that steersclear of the latter’s representationalist framework. To account for theobjectivity of judgmental practices – that is, to explain how rules for makingjudgments can be sufficiently determinate to allow for the possibility of error– Pettit too appeals to the properly functioning ‘extrapolative inclinations’ ofpractitioners.22 The favorable conditions under which our inclinationsgenerate presumptively correct judgments are those in which we expect thejudgments of socialized practitioners to converge. In some cases, such assimple perceptual judgments, these conditions boil down to the ‘normalcircumstances’ in which no ‘perturbing factors’ interfere; in other cases, onemust define the conditions ideally, for example as circumstances in whichagents have all the relevant information.23 Pettit, however, anchors theobjectivity of judgments partly in the world itself. Correctly predicating someproperty of an object, for example, depends not simply on linguisticconventions but on ‘the constraint represented by the property itself’; desiresare subject to a similar objective standard, namely ‘that the actual world isone at which [sic] the property or option is desirable’.24

Applied to the objectivity of value judgments, Pettit’s line might lead to anormative realism. To avoid this, McMahon grounds the justifying force ofvalues and other moral concepts in those rule-following practices by whichwe employ such concepts to assign relevance to facts and make moraljudgments. On a Wittgensteinian reading, such judgmental practices aresocial realities. To avoid falling back into a rule-skeptical historicism, weshould say that the objectivity of reasons stems from such social realities asnormative ‘all the way down’.25 But McMahon’s individualistic conceptionof discourse – the idea that an individual can enjoy full mastery of the realmof reasons – blocks just this move, for it ultimately requires him to reject theirreducibly social character of judgmental practices. This individualismdistinguishes his approach from Pettit’s ‘holist’ model, which takes socialrelations as a constitutive condition for the individual’s capacity to think.26

Grasping the Force of the Better Argument 127

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

More importantly for our purposes, McMahon’s individualism underlies acrucial revision in Pettit’s idea of commonability – a revision that undercutsMcMahon’s use of this idea as an argument against the MA model.

(2) Pettit’s concept of commonability builds on the analysis of rule-following and the notion of favorable circumstances as those in whichindividuals should converge in their judgments. He argues that the rules bywhich individuals form judgments must be intersubjectively knowable –convergence presupposes that I can, in principle, know which rule you areapplying when you make some judgment. Thus our concepts (as rules forjudgment) must be ‘commonable’, a ‘common possession’ accessible toevery socialized individual. For Pettit, this means that each has ‘equal access’,so that each ‘share[s] in the determination’ of what counts as the correctextension of a concept to a new case.27 When he takes up the needs-basedscenario in his book, McMahon invokes commonability as a reason forrejecting the assumption that the individual has a privileged access to ‘whatcounts as a genuine need of hers’. As he explains, ‘each of the conceptsemployed by anyone is in principle accessible to everyone else, so thatanyone can make correct judgments employing any concept. If this is right, Acould be correct about what B’s needs are, even if B disagrees’ (p. 148). Anumber of problems undermine this argument.

To begin with, the MA model allows precisely for the possibility thatMcMahon describes, namely that an interlocutor can be correct about whatone really needs even when one disagrees. Because ‘privileged access’ doesnot mean self-transparency, the claims one makes about one’s needs and theirsatisfaction are both fallible and open in principle to rational criticism. In theMA model, ‘privileged access’ simply means that claims about thesatisfaction of A’s needs cannot be conclusively justified so long as A doesnot agree (assuming A is reasonable and morally mature). This is because theassessment of needs partly depends on particularistic kinds of evidence(feelings, life-history, background aspects of one’s self-understanding, etc.).Thus the relevant facts or circumstances for applying some need-relatedconcept are to some extent unique for each individual, such that each has a(somewhat) unique perspective on those facts, a perspective no one else canfully share. This uniqueness cuts both ways, of course. Consequently,accurate assessments of the force of needs-based arguments must factor in theestimates of everyone involved. So understood, privileged access does notconflict with Pettit’s idea of commonability as equal access, so long as weunderstand this as equal but incomplete access.

The foregoing suggests a bolder reply to the objection to my pollsterscenario as motivation for the Better Justification Thesis (Section III), that is,the objection that merely knowing others agree with you gives you neither abetter grip on the justifying reasons nor a better justification based on newsubstantive reasons. The reply links perspectival uniqueness with something

128 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

like an ‘indexical reason’ that has a substantive effect on justifications.Consider first the significance of disagreement. If I take you as a mature,reasonable individual, then your disagreement with my assessment of yourneeds, similar to substantive counterarguments or debunking reasons, castsdoubt on my arguments. That your expressed reasons for disagreeing do notconvince me does not fully erase this doubt, for the fact remains that myarguments have failed to hold up within your unique, indexically qualifiedperspective. Inasmuch as publicly expressible reasons are intended to havesuasive force for their audience, they should hold up or ‘transfer’ acrossvariations in individual life circumstances. Your disagreement thus signals acognitive failure – and so long as I consider you reasonable I cannot, evenafter I reconsider my position, dismiss the possibility that the failure is mine.And if it is, then reasons I take as objectively cogent support for my judgmentin fact are not.

Conversely, mutual agreement adds an additional increment of justifica-tion, an addition that has the net effect of a further substantive reason, evenafter a discourse has put all the substantive reasons and arguments on thetable. Each person’s unique perspective on the relevant circumstances forapplying a concept illuminates those circumstances in a somewhat differentway, which generates the discursive equivalent of a new reason or fact. Ifthese differences are indexically linked to the given perspective, and if thatperspective is unique, then the individual’s attempt to publicly articulate theseindividually indexed considerations must always leave out something – anelement whose relevance and force can only be grasped existentially fromwithin that individual’s circumstances. The individual’s assent signals toothers that this particularistic element has been adequately considered.

Far from undermining this analysis, Pettit’s concept of commonabilitysuggests a way of generalizing the analysis beyond the needs-based scenario.According to Pettit, each person shares in the determination of how we shouldextrapolate the concepts that allow us to make correct judgments. As he putsit, we each defer to each other’s equal authority, so that ‘we collectivelyintroduce a determinacy of content that is unavailable at the purely individuallevel’.28 The individual’s authority, however, involves only a partialcommand of the force of reasons. This is because each individual, inprivately assessing the reasons in relation to a dispute over the correctextension of some value, need, or other moral concept, can look only to his orher own extrapolative inclinations – which presumably involve a measure ofparticularistic uniqueness. These individual inclinations supply at most ‘anintimation of the communal voice that firms up in the convergence ofdifferent extrapolative inclinations . . . an intimation that is validated only inthe achievement, perhaps after negotiation, of a concerted response’.29 Thusthe particular perspective that the socialized individual brings to a moralquestion is a source of authority, albeit not a decisive authority based on

Grasping the Force of the Better Argument 129

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

complete access – hence the importance of convergence through discourse or,as Pettit puts it, ‘negotiation’. Each person’s agreement is a necessary but notsufficient condition for a fully justified moral judgment.

To invoke commonability against the MA model, then, McMahon mustrevise Pettit’s analysis and assume that each individual has, in principle, notsimply an equal access but a complete access to, or full mastery of, therelevant concepts that justify judgments about others’ genuine needs. Fullmastery means that the individual, suitably reflective and informed, can inprinciple fully (but fallibly) grasp the objective force of moral arguments.This individualistic revision not only undermines his critique of discourseethics, it threatens to suck his model into the very normative realism he hopesto avoid. If the realm of reasons is in principle completely accessible to theindividual, then it seems that the objectivity of reasons does not inherentlydepend on judgmental practice as a socially constituted reality. Rather,objectivity must reside in the substantive reasons simply as reasons – as justthose ‘abstract entities’ to which normative realism appeals.

(3) The link between discourse and more experiential, particularistic formsof evidence suggests a further elaboration of judgmental practices as socialrealities. In closing, I briefly sketch the core idea. At one level, the idea ofgrounding the objectivity of moral reasoning in judgmental practices as socialrealities combines aspects of the constitutive and epistemic approaches. Inline with the epistemic approach, substantive moral values are consideredessential to justifying correct judgments – moral discourse does not constitutetrue moral principles from premoral scratch.30 In line with the constitutiveapproach, the capacity of moral reasons to justify inheres not simply in thereasons as such but in their capacity to convince participants collectively –even if this first emerges in a piecemeal fashion. The objective force ofsubstantive reasons, in other words, is measured by the capacity of thosereasons to convince a group of morally socialized participants who engage indiscourse under favorable discursive conditions. Thus principles such as (U)represent attempts to articulate the inherent normativity of our justificatorypractices, as grounding the objectivity of reasons.

At this first level, then, the ‘social reality’ of a judgmental practice refers tofeatures of moral discourse as a dialogical consideration of relevant reasons.However, moral objectivity also has a source that lies outside discourse perse. Christina Lafont has proposed ‘unrenounceable general interests’ as acandidate for this role.31 In my estimation, a less controversial candidateemerges if we notice that the social reality that grounds moral objectivity, inparticular the objectivity of moral concepts as reasons, lies in the socialsituations and institutions that make up the substance of everyday life.Judgmental practices, after all, are about such concrete social practices andinstitutions. The latter thus represent something like potential falsifiers for theoutcomes of moral discourse as attempts to solve moral problems.

130 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

In the present context, the key idea lies in the observation that the valuesthat function as reasons inside discourse refer to features of the concretesocial practices involved in everyday life. When a group engages in moraldiscourse, its members want to answer a moral question about how they oughtto treat one another in some area of social life. Imagine a group confronted bythis question in regard to hiring practices. Suppose they draw upon theconcept of equality (or equal treatment) as the appropriate standpoint fromwhich to assess the relevant considerations (e.g., to identify some situations as‘discriminatory’), and they then go on to agree on some particular norm orpolicy. Even if everyone has agreed in discourse, under discursive conditionsthat appear sufficiently favorable, the possibility remains that their attempt tocarry out the norm in practice could backfire. This might happen in a numberof ways. The agreement might involve empirical mistakes, as occurs when thegroup fails to anticipate problematic side-effects of the norm’s observance.But the group could also find, through actual practice, that their consensuspresupposed an inadequate conception of equality. Indeed, one couldinterpret the evolving discourse over women’s equality in the last centuryand a half in these terms, as a case in which our understanding of the conceptof equality itself has improved.32

Because moral concepts such as equal treatment refer to social realitiesoutside discourse per se, such realities can generate new information thatfalsifies – in the broad sense of provoking us to doubt and revise – even thosejudgments we once projected as fully justified relative to the informationavailable at the time. To be sure, for concrete practices to function asfalsifiers, participants must work up the relevant experiences into questions,objections, and criticisms that generate a moral problem and thereby provokecollective reasoning. My point, however, is that the source of such problems,where they first arise, lies outside discourse. This source of moral objectivitydeserves closer attention from theorists interested in the analysis of moraldiscourse.33

NOTES

1 See McMahon’s ‘Discourse and Morality’, Ethics 110 (2000), pp. 514–36 (hereafter citedas DM).

2 For more extensive argument on these points, and a sense of the various attempts of gametheorists to address the problem of cooperation, see Joseph Heath, Communicative Actionand Rational Choice (Cambridge, MA: MIT Press, 2001), chaps. 2–4.

3 His discussions of these issues further develop views he put forth in his Authority andDemocracy (Princeton: Princeton University Press, 1994).

4 Note that McMahon does not ‘deny the existence of social factors in shaping many featuresof individuals, including the values they hold’ (p. 24).

5 For present purposes I also take the terms ‘reasons’ and ‘arguments’ as more or lessequivalent, inasmuch as both function as justifications for judgments.

Grasping the Force of the Better Argument 131

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

6 We can read the term ‘interests’ in McMahon’s definition of morality as having thenarrower meaning of premoral desires and aims; moral norms and values then determinethe conditions on the moral pursuit of such desires and aims.

7 A pool of reasons is a ‘public good’ in the technical sense: the good is nonexclusive (noindividual can acquire the good without its being available to all) and allows non-rivalconsumption (one person’s use of the pool for improving judgment does not hamper others’ability to draw on the pool to improve their judgments) (pp. 110, 224, note 10).

8 McMahon (pp. 39–40) defines ‘collective agent’ broadly to include any group that comestogether to perform a collective task.

9 See DM.10 Jurgen Habermas, ‘A Genealogical Analysis of the Cognitive Content of Morality’, in C.

Cronin and P. DeGreiff (eds), The Inclusion of the Other (Cambridge, MA: MIT Press,1998), p. 42; translation amended. Although McMahon relies on earlier formulations, thequoted passage is particularly apt in view of its explicit reference to ‘joint’ acceptance.

11 McMahon’s use of ‘epistemic’ differs, therefore, from that of Habermas, ‘Richtigkeitversus Wahrheit: Zum Sinn der Sollgeltung moralischer Urteile und Normen’, Wahrheitund Rechtfertigung (Frankfurt am Main: Suhrkamp, 1999), pp. 271–318, where an‘epistemic’ conception of rightness is one that defines the meaning of rightness in terms ofideal warranted assertibility.

12 Notice that McMahon’s analysis resonates with an ongoing debate among discoursetheorists provoked by the criticism that substantive reasons, and not consensus, justifynorms; see Bernard Peters, Rationalitat, Recht und Gesellschaft (Frankfurt am Main:Suhrkamp, 1991), pp. 235–36, 259–61; for replies, see Habermas, ‘Remarks on DiscourseEthics’, in Justification and Application, trans. C. P. Cronin (Cambridge, MA: MIT Press,1993), pp. 58–59; also Kenneth Baynes, ‘Practical Reason, the “Space of Reasons,” andPublic Reason’, in W. Rehg and J. Bohman (eds), Pluralism and the Pragmatic Turn(Cambridge, MA: MIT Press, 2001), pp. 65–66.

13 DM, p. 519.14 DM, p. 534.15 DM, p. 525.16 William Rehg, ‘Discourse and the Moral Point of View: Deriving a Dialogical Principle of

Universalization’, Inquiry 34 (1991), pp. 27–48, here pp. 44–45; quoted in DM, p. 521.17 I believe this move also undercuts a Wittgensteinian objection that McMahon considered

decisive against discourse ethics (DM, pp. 527–30), but I cannot argue that here. Indeed, inthe book he relegates the argument to a note, and he appears a bit unsure of it himself (p.231, note 19).

18 See Habermas, ‘Genealogical Analysis’, pp. 41–42; also his Between Facts and Norms,trans. W. Rehg (Cambridge, MA: MIT Press, 1996), pp. 107–09. For a critical survey of theinternal debates over (U), see Niels Gottschalk-Mazouz, Diskursethik (Berlin: Akademie,2000), pt. 1; influential critiques of (U) include Albrecht Wellmer, ‘Ethics and Dialogue’, inWellmer, The Persistence of Modernity, trans. D. Midgley (Cambridge, MA: MIT Press,1991), chap. 4, and Seyla Benhabib, Situating the Self (New York: Routledge, 1992), chap.2. On the relation between the two principles, see Habermas, Between Facts and Norms, pp.107–09, 459–60.

19 DM, p. 530, considers the example of a radical dissenter in a society of slaveholders, but inthis case the majority’s reasoning is corrupted rather than reasonable. The MA model alsohas implications for conscience-formation, as I have argued in ‘Discourse Ethics andIndividual Conscience’, in N. Gottschalk-Mazouz (ed.), Perspektiven der Diskursethik(Wurzburg: Konigshausen and Neumann, 2003).

20 Habermas rejects this last move when he identifies truth with ideal warranted assertability;see especially ‘Richtigkeit versus Wahrheit,’ pp. 284–85, 296–97; this contrasts with theposition of Lutz Wingert, Gemeinsinn und Moral (Frankfurt am Main: Suhrkamp, 1993),pp. 264–81, and has been criticized by Christina Lafont, The Linquistic Turn inHermeneutic Philosophy, trans. J. Medina (Cambridge, MA: MIT Press, 1999), chap. 7.

21 McMahon, DM, p. 519ff, uses the term ‘strong dialogicality’ to characterize the strong ICmodel.

22 Philip Pettit, The Common Mind (New York: Oxford, 1993); see especially chaps. 2 and 4.

132 William Rehg

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014

23 Pettit, Common Mind, pp. 93–94, 177; note that convergence is also intrapersonal: oneexpects a consistency in one’s own judgments over time.

24 See Pettit, Common Mind, pp. 81, 67, respectively.25 Such an approach is taken by Robert Brandom, Making It Explicit (Cambridge, MA:

Harvard University Press, 1994).26 McMahon notes the difference (p. 228, note 41); for Pettit’s holism, see Common Mind,

chap. 4.27 Pettit, Common Mind, pp. 115, 180.28 Pettit, Common Mind, pp. 217–18; here I extend Pettit’s analysis of perceptual judgments to

morality.29 Pettit, Common Mind, p. 190; see also pp. 187–93.30 Cf. Joseph Heath, ‘The Problem of Foundationalism in Habermas’s Discourse Ethics’,

Philosophy and Social Criticism 21 (1995), pp. 77–100.31 Lafont, Linquistic Turn, chap. 7.32 Cf. Habermas, Between Facts and Norms, pp. 409–27.33 I thank Garth Hallett, James Bohman, and James Swindal for feedback on earlier versions

of this essay.

Received 28 September 2002

William Rehg, Department of PhilosophySaint Louis University, 3800 Lindell Blvd., St. LouisMO 63108-3414, USA. E-mail: rehgsp@slu.edu.

Grasping the Force of the Better Argument 133

Dow

nloa

ded

by [

Nor

thea

ster

n U

nive

rsity

] at

18:

24 0

4 N

ovem

ber

2014