1 A Biased Lay of the Land

The adversariality of argument debate has two main axes.Footnote 1 The first is that of identifying the potential that adversariality in argument has for producing inappropriate argumentative performance. Call this the practical domain of the adversariality question. In the practical domain, there are those who hold that all adversarial argument is inappropriate, those who hold that some is inappropriate, then those who hold that some is inappropriate and some is in fact salutary. No one in the discussion, to my knowledge, holds that all adversarial argument is salutary.Footnote 2 Anyone who has complained that the combative comportment of some discussant stifles exchange or who has wondered whether the desire to defend one’s view at all costs perverts the ends of critical discussion can feel the pull of this issue. As can anyone who has seen that the exchange of critical reasons on our deeply held views deepens our understanding and may be a requirement of sharing a polity with those with whom we disagree. Further, anyone who has seen that dissent and expressions of outrage are important argumentative gestures can see the stakes for this discussion.

The second axis of critical discussion is about whether argument must be adversarial. Is adversariality essential to argument, or could argument be otherwise? With the question of essential adversariality, there are the intrinsic adversarialists, who hold that argument is adversarial at its core. The rejection of intrinsicism comes in two strengths, modest and strong. The modest rejection of intrinsicism is that argument is not essentially adversarial. The strong rejection of intrinsicism is that argument is essentially not adversarial.Footnote 3 Considerations in favor of intrinsicism are that argument’s primary felicity condition is disagreement and that others’ arguments are impingements on one’s beliefs. Considerations in favor of weak non-intrinsicism are that, for as often as there are argumentative adversaries, there are argumentative collaborators. Strong non-intrinsicism is motivated by the insight that argument, as a joint enterprise, must be cooperative.

There are further questions bearing on these debates, such as whether how we talk or conceive of argument changes our performances in argument and whether non-adversarial conceptions of argument recapitulate the problems of adversarial conceptions.Footnote 4 But these are beyond the scope of this paper, and they clearly depend on answers to the two core questions for their salience.

My plan for this essay is to make the case for a new form of intrinsicism and then argue that given the intrinsic adversariality of argument, we can explain a number of features of argumentative error and excellence that would be difficult to explain otherwise. The big idea to start is that reasons are contrastive, that instead of merely being for something, they are for something instead of something else. They have a sorting function, and these contrasts that do the sorting makes argument structurally adversarial, not between people necessarily, but between viewpoints. Clarity on this point then makes the connection between dialectical and epistemic objectives of argument more fecund. This connection between dialectical and epistemic elements of argument then provide tools to defend intrinsicism against the most trenchant objection against it, what I call the winners-and-losers problem.

2 Reasons as Contrastive

Contrastivism about reasons is the view that all reasons do the work they do indexed to a contrast class, so reasons function not just as reasons for something, but rather as reasons for something as opposed to something else. A few examples may help make this point clear.

  • Practical reasons: Should we go to the Burger Hut for dinner? If I say, “Yes, because it’s close,” that may be a good reason if the alternative is the far-away Taco Palace, but it does not work if the alternative is the equally close Curry Café (from Aikin and Talisse 2020).

  • Moral reasons: Was Adrianna running into the burning building and saving Tiny Tim praiseworthy? If we say, “Yes, because she saved a life at risk of harm to herself,” that reason holds so long as it was as opposed to her staying out of the fire altogether. But if Adrianna could have also saved Tiny Tom in the process without any more trouble, then it seems the praise is mitigated (from Snedegar 2015).

  • Explanatory reasons: Why is the sky blue? If we say, “because the sun is up,” that explains why the sky is blue instead of dark (as in, at night). If we say, “because blue light’s shorter wavelength makes it more diffused in air than yellow, red, or green,” then we explain why it’s that color as opposed to those (from Sinnott-Armstrong 2008).

  • Epistemic reasons: How do I know you are having iced tea with lunch? If my reason is that what I see in your glass is a light-brown and non-carbonated beverage with lots of ice, then that’s a justifying reason, assuming the other choices to drink with lunch are water, sodas, coffee, and so on. But not quite if we’re in a bar and Long Island Iced Tea is on the menu (from Aikin and Talisse 2020).

The big idea here is that a central feature to evaluating the quality of a reason is in evaluating what it’s a reason for from a set of alternatives. If we change the alternatives, the reason’s quality also changes. What reasons do, then, is sort the best of the bunch out, and that (given the options) is what we should accept. Reasons-evaluation, then, is triadic, between (i) the reason, (ii) what the reason favors, and (iii) the alternatives, or what’s commonly called the contrast class.

There are a number of benefits from the contrastivist take on reasons. The first (as noted by Schaffer 2004) is that we have the tools with this program to explain what seemed mysterious about contexts—why does reason-quality change when we change interests or with whom we are talking? Contrastivism explains that change in quality with the change in contrast classes—in some discussions the contrast class is different from others, so the reasons themselves, to do the work of reasons, must be different, too.

A further benefit of contrastivism is that, in epistemology, it allows us to explain the appeal of various skeptical scenarios but also retain the idea that even against skeptical challenges, we can identify better and worse cognitive performances. For example, a person’s kinaesthetic and visual experiences of their hands may give them good reasons to believe that they have hands rather than wings or flippers, but it does not give them reason to believe that they have hands rather than being deceived by an evil demon (as noted by Sinnott-Armstrong 2004; Dretske 2013; Baumann 2015).

An important lesson of contrastivism, consequently, is that issues for critical evaluation are clarified by not only what is in question, but what the range of alternatives is. And so, if the question is whether I should, say, ride my bicycle to campus today, what reasons we can bring to bear on the issue depend on what the alternatives are. So if the issue is between bike or bus, that I need exercise is a reason that performs contrasting work. But if the choice is between bike or walk, then that reason doesn’t (from Snedegar 2013). The key is that before we start deliberating, it’s best to identify our options so we can then aggregate and evaluate our reasons appropriately. And it’s a familiar enough experience for many that, as deliberations proceed, new options may arise, and so returning to previously settled matters is necessary. Contrastivism explains why that’s good policy.

3 Contrastivism as Dialectical Adversariality

Minimal Dialectical Adversarialism is the view that adversarial relations are intrinsic to argument, as reason-giving is done with the objective of posing or answering critical challenges among competing theses (Aikin 2017). This controversy-focused feature of dialectical adversariality is anticipated by Govier’s (1999) approach, but importantly, Govier’s program requires there be speakers or representatives of the views in contrast to be addressed. Minimal dialectical adversariality does not require particular individuals to be representatives to represent contrastive positions—there only need to be intelligible alternatives for the issue to be live and for reasons to be needed. As Govier rightly notes, disagreements are the prime locus for identifying the range of alternatives and thereby activating the sorting function of reasons, but on the minimally dialectical approach, one need not have actual flesh and blood adversaries, but rather intelligible alternatives in need of critical attention. Devil’s advocacy, when there are not actual representatives of the range of options can sharpen our capacities, makes sure the reasons given do not fail to take the relevant contrasting views into consideration (as noted by Stevens and Cohen 2020; Aikin and Clanton 2010). Further, ensuring that options, and those that may hold them as relevant, receive critical attention is a requirement of epistemic justice (Alsip Vollbrecht 2020).

Consider the following feature of the adversariality debate as a data point for the contrastivist thought behind minimal dialectical adversariality. Non-adversarialists about argument go out of their way to argue against the adversarialist position. This seems, at least on its face, curious. Ironic perhaps. Maybe even a performative self-refutation. Whatever it is on this sliding scale of tension between theory and practice, it’s worth noting that these folks are arguing, and their arguments pick out target theses, and contrasting ones. And then they argue against those contrasting views with reasons that show that they are wrong and their view is right. Again, the curiosity of this phenomenon shouldn’t be understated. Take Foss and Griffin, criticizing the culture of critique:

Even discursive strategies can constitute a kind of trespassing on the personal integrity of others when they convey the rhetor’s belief that the audience members have inadequacies that in some way can be corrected if they adhere to the viewpoint of the rhetor (1995).

Not only does this non-adversarial view prohibit protest and dissent as argumentative contributions (a high price to pay, for sure), but it seems clearly to prohibit its own critical program (as noted by Fulkerson 1996). A more recent contribution to the non-adversarialist case is Bailin and Battersby’s dialecticalist model of controversy that emphasizes that adversariality yields my-side bias, aggression, a focus on winning that “may well eclipse the goal of coming to a reasoned agreement, undermining cooperation, open-mindedness, and willingness to concede the strongest reasons” (2020a, p. 45). The key here is that the reasons Bailin and Battersby offer work only against the background of a contrast—they are reasons-for only because they are also reasons-against. And in the process of articulating the range of options, they cite a range of adversarialists to occupy those positons. Aren’t their arguments given with the purpose of winning the debate with the adversarialists?

Perhaps this is all too quick, and a little unfair. Bailin and Battersby, for their part, outline a program of dialectical inquiry wherein “the exploration of conflicting views is at the centre of the inquiry process, but the process is a collaborative rather than adversarial endeavor” (2020a, p. 47). The result, as they put it, is that “confrontation is really between views and not between people” (2020a, p. 48). What seemed a tension from the adversarialist perspective is now seen as perfectly consistent from the perspective of the properly articulated dialectically-savvy non-adversarialist. So what just happened? Here’s my view on the matter. Bailin and Battersby hold that inquiry is dialectical, that evaluation is meaningful only against a backdrop of comparisons, and so once we see this properly, what was initially adversarial in these comparisons becomes “tamed through adherence to appropriate dialectical norms” (2020a, p. 45). The problem I see is that tamed adversariality is still adversariality. On analogy, if boxers otherwise would just brawl unless their fighting spirit were not tamed by the rules of boxing, I nevertheless do not see why their exchange is not adversarial. They box each other, within the rules, bound at the end by the judge’s decision—that’s structured, rule-bound adversariality. Bailin and Battersby hold that this arrangement is cooperative, but still allows for “critical probing.” But, as I see it, that’s like interpreting the other boxer’s jabs at my chin as helpful reminders to keep my gloves up.

My initial thought was that non-adversarialist views, when argued for, have a kind of performative self-refutation. That’s looking less plausible, given the current state of dialectical play. The problem with self-refutation arguments is that they are wont to produce more heat than light, and that is probably the situation here with my case against non-adversarialism. That said, let’s not lose sight of the fact that there surely is something odd about non-adversarialists coming out and arguing against a view, naming names, and not working too hard to repair those views when their criticisms come a little too easily (surely, anyone familiar with this literature has their favorite straw man of adversarialism). Again, that’s just weird, right? The key is that adversarialism about argument can explain all that perfectly consistently, but non-adversarialism has a harder time. That’s not the same as the self-refutation argument before, but there’s a little more light (and maybe a little less heat) than before. The key to it all is how to interpret what role dialecticality plays in this debate. So far, both sides lay claim to its function. So some contrasts are necessary.

4 Dialecticality, Controversy, and Contrasts

The challenge for any dialectical notion of argument is explaining why dialecticality matters. If the objective of argument is primarily epistemic, that of getting the truth, then why must one address others at all? Johnson (2000) famously made the pragmatic case for the dialectical tier, and it is baked into the pragma-dialectical program that dialecticality is a norm of finding resolution (van Eemeren and Grootendorst 2004). But these objectives are more about garnering adherence than approaching truths, and, at least for epistemic reasons, that’s insufficient. More purely epistemic reasons have ranged from the Wohlrapp (2014) thought that testing from the perspective of others is a requirement for validity and knowledge to the view that grounds for others’ disagreement and critical questioning, if unaddressed, are defeaters for one’s justification (Aikin 2017, 2020). It’s here that we can see the pragmatic and epistemic objectives coalesce. Dialecticality is a desideratum of argument precisely because argument is appropriate in contexts of controversy. Dialecticality, then, addresses the alternatives and the reasons available that count for or against any view favored in the end. And given that reasons are reasons against contrast classes, dialecticality provides the friction for the wheels of reason.

The critical thing about all this procedural social-epistemic talk is that what results from argument exchange is a commitment that survives the exchange (and maybe, further downstream, a belief). One thing we have to acknowledge is that we not only leave our argumentative exchanges with commitments as to what the reasons support, but we arrive with those commitments too. Many times those commitments are the results of prior argumentative exchanges, with the same folks or with others. Sometimes, strangely enough, we just find ourselves assenting to things because they just feel right. For whatever reasons we have for those views when we arrive at argument, those views and reasons will be scrutinized anew in the exchange. But notice something important about participating in the dialogue. First, what reasons we give depends on the breadth of the options under consideration. So, for example, my reasons for supporting a particular candidate for office will differ depending on whom I am addressing, because I will be offering reasons for my choice as opposed to, for example, the far-right gun-nut my uncle prefers, the centrist caretaker my father likes, the crypto-Maoist my daughter likes, or just opting out of the process altogether, which my anarchist wife prefers. Second, the options have their salience because there are others for whom they are plausible, and they deserve being addressed in ways that they see as reasons and they see as addressing their reasons (as noted by Alsip Vollbrecht 2020). Third, and finally, if I don’t work hard to think up good responses to their criticisms, or I don’t hold their replies to a level of scrutiny, I’ve failed in a crucial way. We, in the end, may agree that my reasons for supporting Bucky McBuckerton’s candidacy for city council were not very good, but if I didn’t try hard to think up defenses for Bucky’s tax policies to libertarian critique, I failed him and the discussion. Let’s call it a fiduciary duty of critical discussions—views, often in the opening stages of discussions, are entrusted to arguers for their elaboration, defense, and promotion. Again, this distribution and claiming of roles and responsibilities often tracks what commitments and beliefs with which we arrive at the critical discussion. But sometimes, it emerges as one thinks of a defense for a view in the midst of the exchange or a new argument against another view. It regularly happens organically in the course of a critical conversation or just in watching a debate—suddenly, you find you have a view and thereby a new kind of stake in the critical discussion. That’s just what it is to be a thinking person—you live with your perspective, and you live it from the inside. Others are doing the same, and that’s what creates issues and controversies to begin with. And that explains why norms of dialecticality in argument are so important—we not only need to make our reasons clear, but we need to make them clear to those who disagree, so they too can see them as reasons. If we weren’t bound by that rule of dialecticality, then it’s unclear how argument would be much different from browbeating. That’s true even if we meet the purely epistemic goal of getting the truth—we won’t be able to possess it properly unless we have cleared up what critical challenges undercut our command of the issue.

What’s important here is that this fiduciary and stake-holding account of argument’s arrangements explains a good number of the ways that argumentative and broader intellectual virtues are what they are. Consider the notion of intellectual courage (as noted by Casey and Cohen 2020). On the one hand, there may be those who stick up for under-represented views, suffer scorn of peers to defend something unpopular but nevertheless plausible by their lights. On the other hand, there may be those willing to submit their views to a wide variety of criticism, for the sake of revising and improving the view (or even rejecting it altogether). The courage one instantiates in those kinds of intellectual endeavors is hard to capture without the thought that argument and all the critical back-and-forth is dangerous in some way. It is costly, as Casey (2020, p. 101) notes, to change one’s mind, if one’s views do not survive scrutiny. There may be intellectual backfire, as a view that’s actually true may have been undercut by some wily rhetoric, an uninspired defense, or misleading evidence (Cohen 2005). One may suffer loss of status for one’s unpopular view being made public or for the simple fact that one has been criticized (Paglieri and Castelfranchi 2010; Kidd 2020b). Or one may lose the argument about who should walk the dog on a rainy day, and then have to do it oneself (Stevens 2019). Arguments are risky, because controversies have stakes. Any theory of argumentation that makes it inappropriate for someone to take her own side in an argument has lost sight of the point of it all.

5 The Winners-and-Losers Problem

So far, I’ve focused on what might be called the intellectual self-defense element of argument and its dialectical backdrop, but it’s important to note that the self-defense works only if argument has an inquirational edge to it—the norms of argument trace and make manifest where we have shared reason for assent and action. Legitimate defenses in argument trace what we ought to agree upon. And the problem for the adversarial model of argument is that sometimes the results of argument run contrary to what at least one participant holds. On the adversarial notion of argument, at least with the objectives that animate it, there are winners and losers, and this relation distorts our relationship with the norms, performances, and results of argument. Let’s call it the winners-and-losers problem for the adversarial view of argument. It has both a practical and theoretical side. The practical side is that this attitude yields badly motivated and so badly performed argumentative exchanges, and the theoretical side is that the adversarial orientation misconstrues a central argumentative good, that of a shared reasoned view in the end. I will not address the practical version of the challenge here, as I believe the matter of a practical difference hangs on the question of whether argument could be different in the first place. In this regard, the theoretical question is prior.

The theoretical winners-and-losers problem for the adversarial view of argument is an old one. Epicurus’s Vatican Saying #74 runs: “the one who loses a philosophical dispute gains more the more he learns.” What’s important in Epicurus’s observation is that a second perspective is required to make sense of what happens when one loses an argument – that there is a kind of epistemic improvement that’s a result from losing the argument that itself couldn’t be appreciated from the perspective prior to the losing argumentative exchange. That’s a discovery, and it’s for the sake of that discovery and its ilk that the norms of argument are what they are – namely, those in pursuit of epistemic improvement.

By my lights, the winners-and-losers problem is the best objection to the adversarial view of argument. Again, it has both pragmatic and theoretical versions, but they both come to the same thought – adversariality requires that arguers are posed as potential winners and losers, which occludes other goods that arise from the exchange, in particular the goods of cognitive improvement. The argument, on this objection is, as noted by Bailin and Battersby (2020a, b), Rooney (2010), Hundleby (2013), Stevens and Cohen (2019), and Kidd (2020a, b), that argument is a zero-sum game on the adversarial model. Since we have good reason to think that argument is not a zero-sum game, we have reason to reject the adversarial model.

My defense of the adversarial view will be in two stages, a dialectical and then a unifying stage. The dialectical stage starts with the following point: with argument, we are managing a variety of objectives, two of which are getting the truth and increasing adherence. And it’s important to note that we can achieve one without the other. They are non-identical objectives. But notice that we’ve stated this in a way that requires two takes on the matter. Saying that an argument gets truth without also getting our adherence, or our adherence without also getting truth isn’t consistent from the first person perspective. Precisely what it is to assess an argument as being worthy of your adherence is to accept it as tracking truths, and vice versa. Taking it as tracking truths just is seeing it as worthy of your adherence. Notice that it’s this internal relation between (a) reasoning things through from the perspective of accepting the results and (b) seeing the reasoning as good that makes it so that argument is effective at all upon us. And it’s the reason why the rule of dialecticality, that of addressing arguments to audiences with reasons they can see as reasons, is binding on arguments. But we, especially when we are surrounded by others who disagree or just have questions, must realize that the first-personal reasoning we do can be seen from the outside in second- or third-personal forms. And the internal connection between assent and taking-as-true from these perspectives is broken. So, statements in first-personal form, like P is true, but I don’t believe it, have a tension, but statements like P is true, but you don’t believe it and P is true, but they don’t believe it do not have that tension.Footnote 5

This first-person perspective on reasoning explains why and how my-side bias, rationalization, and polarization work. Contrary to the anti-adversarialists, these cognitive distortions come as a result not of looking outward at disagreeing others, but by looking at the people who believe as you believe. They are not problems with disagreement, but rather, problems with agreement. Further, it explains why, when disagreement appears, there is the temptation to see all disputes as resolvable in terms of simple truths and common sense—those in the dispute are resistant to reason, given that they don’t agree.Footnote 6 What yields these inclinations, again, is not adversariality, but simply the fact that you have a view and endorse your reasons for it. All of that is prior to adversariality with any particular other person.

The key to this insight is that understanding that the issue is not about giving arguments, but about how they are received. Seeing oneself as defeated in argument requires a kind of double-vision, one where we take that third-person perspective on ourselves and our reasons. Notice that successfully seeing oneself as defeated in argument means that one must see those reasons and commitments of the defeated side as one’s own, but the easy internal transparency between believing in the first person and taking those commitments as true is broken. One sees oneself, really, from a perspective that isn’t one’s own, but it must in the end be one’s own.

This double-vision is a part of what it is to be a social rational creature. And it yields plenty of internal conflict for us beyond argument. It’s a global feature for us as rational beings. Here’s a short list of conflicts that range from the deeply philosophical to the quotidian:

  • Grit: I want to develop perseverance, but I do not want to fail.

  • Fallibilism: Everything I believe I hold is true (otherwise, I wouldn’t believe it), but I don’t think I’m right about everything.

  • Free will: It’s true that my life and my choices are products of causal processes that can’t be otherwise, but I nevertheless proceed as though I am free.

  • Value: Once you realize that it all will end with the sun swallowing the Earth, it all seems pretty pointless, but I still endeavor to be on time to meetings.

  • Goodness: I think I’m a pretty good person, but there are others who disagree, and they have a point.

Double vision here is seeing oneself from the first- and third-person perspectives, switching back and forth. Seeing that one has benefitted from losing an argument requires a shift along these lines, seeing oneself from outside. And the key is that once we’ve seen things from this perspective, we wish to find a way to integrate it with the original one. But it doesn’t work that way, does it? What we do, instead, is toggle back and forth between the perspectives when needed.Footnote 7 So, when we lose an argument, we’ve learned a lesson, but that underdetermines what we do next. We can revise our view, find new reasons for the old view, find undercutting reasons against the argument against us, or just suspend judgment. In a way, that all depends on what connection we have with our original views – we may be tied to them in ways that make it so that we may wish to put up with bad weather, argumentatively, to save them.Footnote 8 So we might shore them up, find better defenses, build better battlements. Or we may not care too much about them, and so go over to the other side. But either way, we still can look at the failure in a way that might make a future victory sweeter, or at least that the challenge improves our character.

The key with argument is that both perspectives bear on the practice – it has to be acceptable from the first- and third-person perspectives. The epistemic theory of argument is, for the most part, devoted to addressing the third-personal perspective, because we are out to secure the truth in a way that approximates knowledge. But because arguments are addressed to audiences, bearing on issues, they must also be appropriate from the first person perspective, too. Again, argument’s dialectical norms highlight this feature – what the objections, worries, and standing reasons are in the circumstance bear on whether the reasons given actually do what they purport to do. That’s the contrastive work of reason. The truth of an argument’s conclusion may not depend on how it hooked up with the alternatives in a deliberative circumstance, but whether the argument was dialectically adequate does.

The adversariality of argument arises not just because there are regularly conflicting perspectives, so reasons as contrasts, must sort them out (though that surely is the primary site of adversariality). In addition to that, adversariality lies in the tension between the perspectives we must toggle between when we think things through at all. In fact, that’s how argument can be adversarial when one’s arguing with oneself. We break our own spells, we call our own bullshit, and we whisper in our own ears that we are not gods. Others are better at those jobs, of course. It’s because they don’t have to work too hard to see the backs of our necks. The argumentative agreement is that we’ll do the same for them. But every time we do so, we are exposed to the risk that this time we touch on the reasons that mean too much, that are too comfortable. And that’s why all the skills of non-adversarial argumentative exchange are so important – they are de-escalating techniques for circumstances gravid with tension. It’s not because argument isn’t adversarial that we need de-escalating skills there, but because it’s adversarial at its core.

The unifying phase of the defense of adversariality against the winners-and-losers problem, now, should be clear. We wouldn’t be able to make sense of the notion of losing an argument in the first place without the notion of adversarial exchange. And it’s because we take both the first- and third- person perspectives on the matter that we now side with the winners in the argumentative exchange when we concede the loss. That’s precisely what it is to see yourself as having an argument that’s failed critical scrutiny—you’re no longer on that team, but the name for that team is still your name. That’s strange, but that’s just what it is for creatures like us—we live our lives from the inside, but we have these moments when we can see ourselves from the outside and make corrections. Now this may come across as merely heavy-breathing profunditas, but I think it’s a pretty deep point about our relationship with ourselves and the norms that we see bear on us. Plato’s Republic, with the question Why be Just?, would not be what it is if we didn’t both acknowledge the oughts of a moral life but also just want to do what we wanna do. And Plato is right that a perfect soul and city have a harmony between rationality’s demands and what they desire. That’s the ideal, though, and then there’s us. Sure, we come around with reason, but it takes some time, some cajoling, some argumentative losses, and the odd win. But the reason it takes such work is because we thought we were right in the first place. That’s just what it is to have a view.

Here’s another way to look at this toggling phenomenon in argument. Everyone who rightly emphasizes the cooperative elements of argument has in mind a critically conversant participant. Cooperation in argument isn’t just rolling over when some reasons get given. Cooperative arguers aren’t yes-men. So what are they? They object when appropriate, request clarification when needed, add to the case when useful, and they may even help repair some arguments. And the winners-and-losers problem for adversarialism runs that it’s hard for adversarialists to explain all that, especially the stuff about improvement and repair. And the non-adversarialist view can handle all of those things quite well. That’s why the problem is such a serious one for adversarialism. Here’s my reply: first, contrastivism about reasons and minimal dialectical adversariality explains a good deal of it, as clarifying views makes it also clear what reasons will do the sorting work and objections prompt further arguments. But what about argument extension and repair? Here is where the toggling is useful—what we are when we argue together is a cooperating adversary against our respective worse selves. When I’m a good arguer with you, I’m an adversary of the part of you that’s happy to rationalize, who thinks it’s all too easy. And you’re the same for me. And notice that as we play these roles for each other, we toggle back and forth between our first- and third-person perspectives on the commitments and reasoning.Footnote 9 So even if we agree on something, we work to step outside and see ourselves from another perspective, one that would say that they agreed on this, but they forgot about this consideration or that. And notice that failing that kind of adversariality toward ourselves is what yields complacency and dogmatism. Argument extension and repair under those conditions is more architecture for rationalized ramparts. An adversarial stance is necessary to play even these cooperative roles properly.

6 Conclusion

I’ve argued here for two things. First, that given the constrastivity of reasons, it follows that argument, as reason-exchange, must be dialectically adversarial. This accounts for some intellectual virtues that should be honored, and it explains why the debate over argument and adversariality goes the way it does. Second, I’ve argued that the best objection to the adversarial theory of argument, what I’ve called the winners-and-losers problem, can be answered by reminding ourselves that the point of argument’s dialecticality norms are to test our views from the perspective of the range of alternatives. This is a guard against not only the rationalizations others are tempted by, but our own.Footnote 10