The impact is clash – authentic disagreement produced by the labor of the negative is the only unique feature of debate – that requires limits and ground. Their model can’t explain why we switch sides and why there has to be a winner and a loser.

Hegel’s dialectical thinking powerfully exploits the idea of negation. It is a central feature of spirit and consciousness that they have the power to negate. The spirit “is this power only by looking the negative in the face and tarrying with it. This […] is the magical power that converts it into being.”102 The tarrying with the negative is part of what Hegel calls the “labour of the negative”103. In a loose reference to this Hegelian notion Gerald Postema points to yet another feature of disagreements as a necessary ingredient of the process of practical reasoning. Only if our reasoning is exposed to contrary arguments can we test its merits. We must go through the “labor of the negative” to have trust in our deliberative processes.104

This also holds where we seem to be in agreement. Agreement without exposure to disagreement can be deceptive in various ways. The first phenomenon Postema draws attention to is the group polarization effect. When a group of like‐minded people deliberates an issue, informational and reputational cascades produce more extreme views in the process of their deliberations.105 The polarization and biases that are well documented for such groups106 can be countered at least in some settings by the inclusion of dissenting voices. In these scenarios, disagreement can be a cure for dysfunctional deliberative polarization and biases.107 A second deliberative dysfunction mitigated by disagreement is superficial agreement, which can even be manipulatively used in the sense of a “presumptuous ‘We’”108. Disagreement can help to police such distortions of deliberative processes by challenging superficial agreements. Disagreements may thus signal that a deliberative process is not contaminated with dysfunctional agreements stemming from polarization or superficiality. Protecting our discourse against such contaminations is valuable even if we do not come to terms. Each of the opposing positions will profit from the catharsis it received “by looking the negative in the face and tarrying with it”.

These advantages of disagreement in collective deliberations are mirrored on the individual level. Even if the probability of reaching a consensus with our opponents is very low from the beginning, as might be the case in deeply entrenched conflicts, entering into an exchange of arguments can still serve to test and improve our position. We have to do the “labor of the negative” for ourselves. Even if we cannot come up with a line of argument that coheres well with everybody else’s beliefs, attitudes and dispositions, we can still come up with a line of argument that achieves this goal for our own personal beliefs, attitudes and dispositions. To provide ourselves with the most coherent system of our own beliefs, attitudes and dispositions is – at least in important issues – an aspect of personal integrity – to borrow one of Dworkin’s favorite expressions for a less aspirational idea.

In hard cases we must – in some way – lay out the argument for ourselves to figure out what we believe to be the right answer. We might not know what we believe ourselves in questions of abortion, the death penalty, torture, and stem cell research, until we have developed a line of argument against the background of our subjective beliefs, attitudes and dispositions. In these cases it might be rational to discuss the issue with someone unlikely to share some of our more fundamental convictions or who opposes the view towards which we lean. This might even be the most helpful way of corroborating a view, because we know that our adversary is much more motivated to find a potential flaw in our argument than someone with whom we know we are in agreement. It might be more helpful to discuss a liberal position with Scalia than with Breyer if we want to make sure that we have not overlooked some counter‐argument to our case.

It would be too narrow an understanding of our practice of legal disagreement and argumentation if we restricted its purpose to persuading an adversary in the case at hand and inferred from this narrow understanding the irrationality of argumentation in hard cases, in which we know beforehand that we will not be able to persuade. Rational argumentation is a much more complex practice in a more complex social framework. Argumentation with an adversary can have purposes beyond persuading him: to test one’s own convictions, to engage our opponent in inferential commitments and to persuade third parties are only some of these; to rally our troops or express our convictions might be others. To make our peace with Kant we could say that “there must be a hope of coming to terms” with someone though not necessarily with our opponent, but maybe only a third party or even just ourselves and not necessarily only on the issue at hand, but maybe through inferential commitments in a different arena.

f) The Advantage Over Non‐Argumentative Alternatives

It goes without saying that in real world legal disagreements, all of the reasons listed above usually play in concert and will typically hold true to different degrees relative to different participants in the debate: There will be some participants for whom our hope of coming to terms might still be justified and others for whom only some of the other reasons hold and some for whom it is a mixture of all of the reasons in shifting degrees as our disagreements evolve. It is also apparent that, with the exception of the first reason, the rationality of our disagreements is of a secondary nature. The rational does not lie in the discovery of a single right answer to the topic of debate, since in hard cases there are no single right answers. Instead, our disagreements are instrumental to rationales which lie beyond the topic at hand, like the exploration of our communalities or of our inferential commitments. Since these reasons are of this secondary nature, they must stand up to alternative ways of settling irreconcilable disagreements that have other secondary reasons in their favor – like swiftness of decision making or using fewer resources. Why does our legal practice require lengthy arguments and discursive efforts even in appellate or supreme court cases of irreconcilable legal disagreements? The closure has to come by some non‐argumentative mean and courts have always relied on them. For the medieval courts of the Germanic tradition it is bequeathed that judges had to fight it out literally if they disagreed on a question of law – though the king allowed them to pick surrogate fighters.109 It is understandable that the process of civilization has led us to non‐violent non‐ argumentative means to determine the law. But what was wrong with District Judge Currin of Umatilla County in Oregon, who – in his late days – decided inconclusive traffic violations by publicly flipping a coin?110 If we are counting heads at the end of our lengthy argumentative proceedings anyway, why not decide hard cases by gut voting at the outset and spare everybody the cost of developing elaborate arguments on questions, where there is not fact of the matter to be discovered?

One reason lies in the mixed nature of our reasons in actual legal disagreements. The different second order reasons can be held apart analytically, but not in real life cases. The hope of coming to terms will often play a role at least for some time relative to some participants in the debate. A second reason is that the objectives listed above could not be achieved by a non‐argumentative procedure. Flipping a coin, throwing dice or taking a gut vote would not help us to explore our communalities or our inferential commitments nor help to scrutinize the positions in play. A third reason is the overall rational aspiration of the law that Dworkin relates to in his integrity account111. In a justificatory sense112 the law aspires to give a coherent account of itself – even if it is not the only right one – required by equal respect under conditions of normative disagreement.113 Combining legal argumentation with the non‐argumentative decision‐ making procedure of counting reasoned opinions serves the coherence aspiration of the law in at least two ways: First, the labor of the negative reduces the chances that constructions of the law that have major flaws or inconsistencies built into the arguments supporting them will prevail. Second, since every position must be a reasoned one within the given framework of the law, it must be one that somehow fits into the overall structure of the law along coherent lines. It thus protects against incoherent “checkerboard” treatments114 of hard cases. It is the combination of reasoned disagreement and the non‐rational decision‐making mechanism of counting reasoned opinions that provides for both in hard cases: a decision and one – of multiple possible – coherent constructions of the law. Pure non‐rational procedures – like flipping a coin – would only provide for the decision part. Pure argumentative procedures – which are not geared towards a decision procedure – would undercut the incentive structure of our agonistic disagreements.115 In the face of unresolvable disagreements endless debates would seem an idle enterprise. That the debates are about winning or losing helps to keep the participants engaged. That the decision depends on counting reasoned opinions guarantees that the engagement focuses on rational argumentation. No plain non‐argumentative procedure would achieve this result. If the judges were to flip a coin at the end of the trial in hard cases, there would be little incentive to engage in an exchange of arguments. It is specifically the count of reasoned opinions which provides for rational scrutiny in our legal disagreements and thus contributes to the rationales discussed above.

2. THE SEMANTICS OF AGONISTIC DISAGREEMENTS

The agonistic account does not presuppose a fact of the matter, it is not accompanied by an ontological commitment, and the question of how the fact of the matter could be known to us is not even raised. Thus the agonistic account of legal disagreement is not confronted with the metaphysical or epistemological questions that plague one‐right‐answer theories in particular. However, it must still come up with a semantics that explains in what sense we disagree about the same issue and are not just talking at cross purposes.

In a series of articles David Plunkett and Tim Sundell have reconstructed legal disagreements in semantic terms as metalinguistic negotiations on the usage of a term that at the center of a hard case like “cruel and unusual punishment” in a death‐penalty case.116 Even though the different sides in the debate define the term differently, they are not talking past each other, since they are engaged in a metalinguistic negotiation on the use of the same term. The metalinguistic negotiation on the use of the term serves as a semantic anchor for a disagreement on the substantive issues connected with the term because of its functional role in the law. The “cruel and unusual punishment”‐clause thus serves to argue about the permissibility of the death penalty. This account, however only provides a very superficial semantic commonality. But the commonality between the participants of a legal disagreement go deeper than a discussion whether the term “bank” should in future only to be used for financial institutions, which fulfills every criteria for semantic negotiations that Plunkett and Sundell propose. Unlike in mere semantic negotiations, like the on the disambiguation of the term “bank”, there is also some kind of identity of the substantive issues at stake in legal disagreements.

A promising route to capture this aspect of legal disagreements might be offered by recent semantic approaches that try to accommodate the externalist challenges of realist semantics,117 which inspire one‐right‐answer theorists like Moore or David Brink. Neo‐ descriptivist and two‐valued semantics provide for the theoretical or interpretive element of realist semantics without having to commit to the ontological positions of traditional externalism. In a sense they offer externalist semantics with no ontological strings attached.

The less controversial aspect of the externalist picture of meaning developed in neo‐ descriptivist and two‐valued semantics can be found in the deferential structure that our meaning‐providing intentions often encompass.118 In the case of natural kinds, speakers defer to the expertise of chemists when they employ natural kind terms like gold or water. If a speaker orders someone to buy $ 10,000 worth of gold as a safe investment, he might not know the exact atomic structure of the chemical element 79. In cases of doubt, though, he would insist that he meant to buy only stuff that chemical experts – or the markets for that matter – qualify as gold. The deferential element in the speaker’s intentions provides for the specific externalist element of the semantics.

In the case of the law, the meaning‐providing intentions connected to the provisions of the law can be understood to defer in a similar manner to the best overall theory or interpretation of the legal materials. Against the background of such a semantic framework the conceptual unity of a linguistic practice is not ratified by the existence of a single best answer, but by the unity of the interpretive effort that extends to legal materials and legal practices that have sufficient overlap119 – be it only in a historical perspective120. The fulcrum of disagreement that Dworkin sees in the existence of a single right answer121 does not lie in its existence, but in the communality of the effort – if only on the basis of an overlapping common ground of legal materials, accepted practices, experiences and dispositions. As two athletes are engaged in the same contest when they follow the same rules, share the same concept of winning and losing and act in the same context, but follow very different styles of e.g. wrestling, boxing, swimming etc. They are in the same contest, even if there is no single best style in which to wrestle, box or swim. Each, however, is engaged in developing the best style to win against their opponent, just as two lawyers try to develop the best argument to convince a bench of judges.122 Within such a semantic framework even people with radically opposing views about the application of an expression can still share a concept, in that they are engaged in the same process of theorizing over roughly the same legal materials and practices. Semantic frameworks along these lines allow for adamant disagreements without abandoning the idea that people are talking about the same concept. An agonistic account of legal disagreement can build on such a semantic framework, which can explain in what sense lawyers, judges and scholars engaged in agonistic disagreements are not talking past each other. They are engaged in developing the best interpretation of roughly the same legal materials, albeit against the background of diverging beliefs, attitudes and dispositions that lead them to divergent conclusions in hard cases. Despite the divergent conclusions, semantic unity is provided by the largely overlapping legal materials that form the basis for their disagreement. Such a semantic collapses only when we lack a sufficient overlap in the materials. To use an example of Michael Moore’s: If we wanted to debate whether a certain work of art was “just”, we share neither paradigms nor a tradition of applying the concept of justice to art such as to engage in an intelligible controversy.