The attraction of the ideal has no traction on the real: on adversariality and roles in argument¹

Abstract

If circumstances were always simple and all arguers were always exclusively concerned with cognitive improvement, arguments would probably always be cooperative. However, we have other goals and there are other arguers, so in practice the default seems to be adversarial argumentation. We naturally inhabit the heuristically helpful but cooperation-inhibiting roles of proponents and opponents. We can, however, opt for more cooperative roles. The resources of virtue argumentation theory are used to explain when proactive cooperation is permissible, advisable, and even mandatory – and also when it is not.

Keywords:

• Adversarial
• cooperative
• argument roles
• virtue argumentation

Acknowledgments

We would like to express our sincere gratitude to the editor and to an anonymous referee for their extraordinarily conscientious, constructive and insightful criticism.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Katharina Stevens is an assistant professor at the University of Lethbridge in Alberta, Canada. Her research focuses on rhetoric, virtue argumentation theory, legal reasoning and ethics in argument. Daniel Cohen is a professor of philosophy at Colby College in Waterville, Maine. His recent research focuses on virtues, metaphors, and arguments.

Notes

1 An earlier, shorter draft of this paper has appeared in the proceedings of the 2nd ECA (European Conference on Argumentation), 2017 in Fribourg (Oswalt and Maillat, 2018).

2 Gascon (2016a, 448 ff). We see virtue-based approaches to argumentation as providing valuable resources for theorizing in these areas, although we do not want to commit to agreeing with those who suggest that this is where virtue theories should confine their attentions, for example Paglieri (2015), Godden (2016), and Gascon (2016a, 2016b).

3 Compare, for example, the way in which Walton and Krabbe illustrate the way downward from deliberation over negotiation to quarrel (Walton and Krabbe 1995, 107) or Gilbert, who warns the readers of his textbook Arguing with People that they should be aware that more adversariality on their side will always be matched from the other side (Gilbert 2014, 66).

4 Cohen (2013a, 2014) critiques the language of proponents, opponents, antagonists, and combatants used to refer to the participants in arguments (although in the end, he acquiesces to common usages). He goes so far as to describe the judges, juries, kibitzers, and other participants as “the extras” and “the supporting cast.” Stevens (2016) champions the value of argumentative roles as tools to understand virtue argumentation.

5 That we grant that epistemic betterment is the telos of argumentation means, in this context, that for the purposes of this paper, we stipulate that epistemic betterment is the capital G Good that argumentation accomplishes. Given that the virtuous person is motivated by the Good, the virtuous arguer is then motivated by the epistemic telos of argumentation. We are aware that this will not be enough to satisfy critics of the notion that argumentation can have a Good or even a function [e.g., Goodwin (2007), see also footnote 12] nor does it give an answer to the question whether virtue is a value in itself or derives its value from the role it plays in realizing other values [see, e.g., Godden (2016)]. However, the attempt to deal with either of these important issues adequately would lead us too far afield.

6 In fact, Bailin and Battersby make the conscious choice to favor the epistemological aspect over the social aspect of argumentation (Bailin and Battersby 2017, 7).

7 Following Annas (2011), we think virtues have a motivational component, though we understand that this will not satisfy someone like, e.g., Godden (2016) who wants a more foundational account. Engaging with Godden’s ideas directly would take us too far afield here. We do not think, however, that adopting Godden’s idea according to which virtue must be anchored by reference to other kinds of values would change much about the arguments we will make in this paper. At most, we would have to adapt them slightly. See also footnote 4.

8 See, e.g., Bailin and Battersby (2017) and Cohen (1995).

9 For an example of this model, see one of the ground-laying works of the school of pragma-dialectics, van Eemeren and Grootendorst (1984).

10 Arguers might do this for various reasons. Here are a few examples: Arguers do so because they hope to gain an advantage from winning the argument, because they deeply believe that they are right, or because they want to be right so deeply that they cannot deal with the possibility that they might be wrong. We believe and will argue that arguers might also do so because they believe that taking an adversarial stance will ultimately promote the epistemic telos of argumentation better than taking a cooperative stance (arguers might take an adversarial stance out of a globally cooperative mindset).

11 Strictly speaking, the opposite of taking a cooperative stance is to take a non-cooperative stance. Taking an adversarial stance is just one possible way of taking a non-cooperative stance. Another one (there are probably more than a few) would be, for example, to refuse engaging with any arguments that are not to one’s liking at all (depending on the situation, this could also be done out of the recognition that the epistemic goal is better served through non-engagement, a globally cooperative mindset). We are here interested in the cooperative-adversarial contrast and will disregard the other ways in which one might take a non-cooperative stance.

12 Or they are confederates, if they happen to find themselves on the same side. We would like to distinguish adversarial argumentation from what Aikin calls “belligerent argumentation.” Belligerent argumentation is a very extreme form of adversarial argumentation with no holds barred-belligerent argumentation is argumentation in which the arguer is willing to do anything necessary to win, including the use of all available fallacies, intimidation, and emotional violence (Aikin 2011, 250). We would also like to distinguish adversarial argumentation from what Gilbert calls “eristic argumentation” in his Coalescent Argumentation (1997), as well as from Walton and Krabbe’s eristic dialogue (1995). Eristic argumentation is connected to the “desire to achieve one’s strategic ends at all, or at least some, moral costs if necessary” (Gilbert 1997, 43). Adversarial argumentation can be eristic, but it does not have to be. One could argue adversarially by making all moves that are (a) necessary to achieve one’s objective but (b) do not cross any moral boundaries. Finally, we want to point out that under our definition, merely making an objection does not constitute an adversarial move. Objections can be presented by those taking a cooperative stance as well, if they think that the objection is relevant and therefore helpful.

13 Defining the adversarial and cooperative stance with reference to the arguer’s attitude towards the epistemic telos would mean that cooperativeness is analytically connected to argumentative virtue and adversariality to argumentative vice. We realize that it would be possible to define the concepts like this. In the literature, we have perceived a tendency to waver between this analytic connection on the one hand, and the habit of ascribing certain behaviors to being adversarial/cooperative on the other: On the one hand, cooperativeness seems to be connected to taking the epistemic telos seriously; on the other, it is simply the stance people take in an inquiry [this is, for example, the impression we get from Bailin and Battersby (2017)]. We think that these blended understandings of adversariality and cooperativeness are problematic and confusing. In this paper, we solve the problem by adopting an understanding of adversarial/cooperative stance that is independent of the arguer’s attitude towards the epistemic telos. We hope that we will have the opportunity to give a detailed examination of the definition of adversarial and cooperative argumentation in a later project. We would like to thank our reviewer, who also noticed this difficulty and pushed us to be very clear about the way we talk about cooperativeness and adversariality.

14 Bailin and Battersby (2017, 8). Cf. Aberdein (2010, 173). We understand talk about the telos of argumentation as basically normative. Argumentation is supposed to further epistemic or cognitive betterment of those involved in the argument (importantly, since cognitive betterment comes in degrees, the goal can still be achieved even when truth is not reached). The telos is different than the goals arguers might individually or together adopt, which is a descriptive question. We understand that the idea of a telos of argumentation is at odds with the arguments presented in Goodwin (2007), but we will accept its existence for engaging with Bailin and Battersby here. Goodwin, who argues that there is no single function to argument, is correct from a descriptive standpoint. However, the idea of an argumentative telos helps with thinking about argumentation normatively.

15 Granted, arguments are occasioned by differences. Pragma-dialectical analyses have always made differences of opinion an integral part of the initial confrontation stage of a critical discussion (van Eemeren and Grootendorst 1984, 85–88). Aikin (2011), along with most other argumentation theorists, concludes that adversariality is a necessary part of argumentation, although he recognizes that this means adversariality must be managed, not necessarily removed. We disagree: difference does not imply adversariality. Hundleby (2013, 254) is right when she points out that we may argue in completely cooperative ways. However, when people choose to argue (instead of doing other things), it is often in situations that are already set up as adversarial, such as law courts, thus, ensuring that the argument is also adversarial. Cohen and Stevens (2016) shine a light on how the evolution of argumentative contexts inform argumentative practices.

16 For the argument that we gain (and not lose) if we turn out to be wrong or have to revise our position, see, e.g., Rooney (2010, 221), Govier (1999, 254). Both have been cited by Bailin and Battersby (2017, 4).

17 Walton and Krabbe (1995, 66) distinguish the purpose of a dialogue type, and the goals of a dialogue’s participants.

18 This argument is better described as a persuasion dialogue or a negotiation than an inquiry into how to divvy up house-hold chores fairly. Of course, arguers do not always or only argue to further personal interests. When we argue to figure out the best solution to a problem or to gain clarity on an issue, our goals in the argument align with the telos of argumentation. Those cases are closer to the paradigm that Bailin and Battersby have in mind—pure inquiry or its practical sister pure deliberation (Bailin and Battersby 2017). Empirically, however, inquiry and deliberation are more the exception than the rule for argumentative encounters. Gilbert (1997, 74/116) thinks that pure inquiry is so rare that building an argumentation theory around it would be idle.

19 See footnote 9 on difference and disputes, and on confrontation and argumentation.

20 Though, or so we will argue, this does not mean that taking the adversarial stance must always mean that we are not interested in the epistemic telos.

21 This point is made in Aikin (2011, 266). Something similar is asserted in Aikin and Anderson (2006, 20). John Searle makes this point on a more general level; according to Searle, many adversarial activities require some degree of cooperation. A prize-fight, for example, could not take place if the parties did not agree to abide by the rules of boxing (Searle 1995, 24).

22 See also Gilbert on heuristic and eristic argumentation (1997, ch. 3) and Williams (2004, ch. 1).

23 Admittedly, the neutral phrase “in accordance with” finesses several important debates. For helpful and relevant general accounts of virtue, see, e.g., Zagzebski (1996), Battaly (2008), Annas (2011); for the responsibilism–reliabilism debate, see, e.g., Code (1984), Greco (1999), Axtel and Carter (2008).

24 Cohen, commenting on an earlier version of Stevens (2016), does indeed still refer to proponents and opponents, despite his own sustained critiques of the DAM account of argumentation beginning in (1995) and continuing in an unbroken chain that includes, notably, (2013a) and (2014a), both of which problematize those roles. Cohen (2003) and Cohen (2017) offer justifications for the continued use of the conceptual apparatus of the adversarial metaphor despite its objectionable elements.

25 Of course, in the abstract, there is no such thing as losing an argument. Every argument would be an inquiry or a pure deliberation, and all the arguers would be interested in nothing but the truth or the best solution. However, not every argument is an inquiry or pure deliberation. For an exploration of the different ways of losing an argument in the real world, see Aberdeen (2016).

26 This is a modification of the scenario, the “Noble Chess Player,” from Cohen (2015) to which Bailin and Battersby (2017) respond.

27 Compare, e.g., Cassady (2006).

28 The real-life case that inspired this example was Rowe v. Gibson. In this case, Judge Posner did his own internet research about the effects of a drug to rectify what he perceived as an unacceptable power-imbalance: Only one side could afford to pay an expert-witness. Posner’s decision to break with judicial impartiality and enter his own argument about the facts of a case into the trial resulted in a lively discussion about the boundaries of judicial interference. (See, e.g., http://www.abajournal.com/news/article/posners_internet_research_in_inmates_gerd_suit_leads_to_sparring_in_7th_cir; https://law.stanford.edu/2015/09/14/posner-vs-hamilton-on-google-happy-judges/). See also Wein (2017). Note the differences between the roles of prosecutors in criminal case in US courts and their counterparts, “the Crown,” in Canadian courts on exactly this point. It is worth noting that the judicious judge only oversteps the boundaries of his roles in legal systems that have adopted the adversarial method. In inquisitorial systems, the judicious judge is doing exactly what he should be doing.

29 The “Freedom Rule” ensures that all arguers get their say, but it is in combination that the rules further the epistemic goals of argumentation. See Eemeren, Grootendorst, and Snoek-Henkemans (2002, 182–183).

30 See e.g. Moulton (1989), Gilbert (1994), Rooney (2003, 2010).

31 We do not mean to say that our discussion is exhaustive. Additional factors play a role, e.g., the face-goals to which Michael Gilbert has drawn attention. Gilbert points out that arguers have goals related to their relationships to the other arguers and the way they will be perceived by them. These goals exist in addition to the strategic goals that motivate them to enter an argument and their (possible) commitment to the epistemic telos of argumentation. They can interfere with both the strategic argumentative goals of the participants and with the epistemic telos of argumentation, even if the arguers are committed to it. That we need to preserve face while we argue with others is true in general, especially if we need to interact with them outside the argument as well. [See Gilbert on familiars (1997, 2016).] But preserving face is also necessary simply to stay in the argument as a full participant that others take seriously. If the participants of an argument find it hard to take each other seriously, arguing will be futile and no-one will be epistemically better off, no matter what the goals of the participants are. Therefore, the virtuous arguer should pay attention to the ways in which the context of an argument will shape the face-goals of the participants as well as the conditions under which these face-goals can be fulfilled. This might sometimes necessitate adversarial behavior.

32 Some scholars seem to think so. When Mercier and Sperber’s (2011) paper appeared, Zarefsky quickly pointed out that these findings suggest that argumentation could be most useful when set in an overall adversarial context (Zarefsky 2012). Being aware that in an adversarial context people tend to cheat, Zarefsky suggested that this context should be guided by rules that forbid unfair means. That a battle in which the strongest arguments will ultimately survive serves the epistemic telos of argumentation best is quite possibly the main idea behind the adversarial model of argumentation. Stevens (2016) contrasts this model with the cooperative model.

33 However, when they agree—or when subgroups agree, and animosity exists—Sunstein’s polarization effect sets in and people simply find more reasons for the view they already favor (Sunstein 2000).

34 In addition, the basic adversarial structure of the trial is meant to make sure that the best arguments for each side are produced, leading to a balanced presentation of the two sides for the judges (see Fuller 1978).

35 Hence, the “starting point” in Eemeren, Grootendorst, and Snoek-Henkemans (2002). See also, e.g., Govier (1993) and Walton (1999).

36 This is the argumentative analog to the simple, but surprisingly successful, “tit-for-tat” strategy in Prisoner’s Dilemma tournaments.

37 Cohen draws on this shared knowledge when he points out (2015) that none of the argumentative roles we typically recognize seem to include pointing out missed opportunities as one of the key-associated tasks.

38 See, e.g., Walton (1998) or Walton and Krabbe (1995). Bailin and Battersby’s (2017) preferred structure is the inquiry, in which each arguer plays the same role that includes all argumentative tasks. Another well-known structure is that of the critical discussion as the pragma-dialectical approach describes it, with well-differentiated roles for proponents and opponents. See, e.g., Eemeren and Grootendorst (2004, 51).

39 The other one is associated with a division of tasks because only those with special skills and knowledge that take a lot of time and commitment to learn can perform the tasks. Turner cites the roles of physician and attorney as examples; these roles are differentiated because the amount of special skills needed to carry them out effectively is too great for a single person to acquire them all. We do not think that this principle applies to normal instances of argumentation. At least the basic skills associated with argumentation are so vital that everyone should acquire them.

40 Several argumentation theorists, for example Aikin (2008) and Zarefsky (2012), have made a related point that commitment motivates and that arguers should therefore take the roles of proponents and opponents. Marcin Lewinski (2017) recently has defended the idea that (adversarial) models of dialogues in which participants take on the roles of proponents and opponents can lead to rational decisions in practical reasoning—as long as proponents and opponents stay within the boundaries of certain rules. He also cites Mercier and Sperber to emphasise the advantages of the division of work (Lewinsky 2017).

41 Don’t get us started on the harms of the role of gender roles in argumentative roles—that’s for another paper!

42 See, e.g., Rooney (2010) and Hundleby (2013). Aikin points out that “Sessions at the American Philosophical Association are regularly described as ‘blood sport’, and many paper panelists have referred to the commentator as the paper's ‘assigned assassin’” (Aikin 2011 p. 255/56). Here, the role of “opponent at an APA-session” is clearly associated with behaviors detrimental to the telos of argumentation. In so far, as such associations are learned for argumentative roles in general, adopting these roles is dangerous.

43 The fluidity of roles should not be overstated as Bailin and Battersby (2017) do. They are not fluid to the point of nonexistence. For example, the role of opponent, in its most primitive form, simply contains all argumentative behavior directed against the acceptance of a certain claim or thesis. The arguer does not leave the role simply because an objection must be supported by an argument. Claiming that in this moment, the opponent becomes a proponent is exploiting an ambiguity: In some theoretical contributions to argumentation theory, the opponent role is differentiated more strictly to include only the raising of objections (see, e.g., Wohlrapp 2014, 86). It might or might not be a problem that no unified concepts of proponent, opponent, et cetera exist in argumentation theory. But the ambiguity of these terms should not be used to make it appear as if roles do not really exist and are not played within real argumentative encounters.

44 This is already what happens in the system we have built to deal with those disagreements that have the greatest potential for adversariality because the stakes are typically high and the mutual trust very low: the legal, adjudicative system. Legal scholars discuss the argumentative roles for defense attorneys, prosecutors, civil lawyers, and judges with great vigor. The results of their research are then taught in law schools. See, e.g., the discussion in Fuller (1978) of the roles of adjudicators, which carefully examines and evaluates the norms associated with judges.