33.1 2. Bowell and Kingsbury © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. Virtue and Argument: 
 Taking Character into Account
 TRACY BOWELL Philosophy Programme University of Waikato Private Bag 3105 Hamilton New Zealand taboo@waikato.ac.nz JUSTINE KINGSBURY Philosophy Programme University of Waikato Private Bag 3105 Hamilton New Zealand justinek@waikato.ac.nz Abstract: In this paper we consider the prospects for an account of good argument that takes the character of the arguer into consideration. We conclude that although there is much to be gained by identifying the vir- tues of the good arguer and by con- sidering the ways in which these virtues can be developed in our- selves and in others, virtue argumen- tation theory does not offer a plausi- ble alternative definition of good argument.
 
 
 Résumé: Dans cet article, nous ex- aminons les avantages d'une notion de bon argument qui prend en con- sidération le caractère d’un argu- mentateur. Nous concluons que bien qu'il y ait beaucoup à gagner à iden- tifier les vertus d’un bon argumenta- teur et à tenir compte des façons dont ces vertus peuvent être développées en nous-mêmes et chez les autres, la théorie des vertus dans l’argumentation n'offre pas une al- ternative plausible aux critères de bon argument. Keywords: argument, argumentation, character, virtue argumentation theory, virtue epistemology, virtue.
 
 
 1. Introduction
 
 In this paper we consider the prospects for an account of good argument that takes the character of the arguer into considera- tion. Andrew Aberdein (2010) has defended a virtue-theoretic Taking Character Into Account © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 23 approach to argumentation that builds on work in both virtue ethics and virtue epistemology, and a trend in this direction can also be seen in Battaly (2010) and Cohen (2009). Virtue episte- mology defines knowledge partly in terms of the exercise of epistemic virtues by the knower; virtue argumentation theory hopes to define good argument partly in terms of the exercise of argumentational virtues by the arguer. We conclude that al- though there is much to be gained by identifying the virtues of the good arguer and those of the good evaluator of arguments, and by considering the ways in which these virtues can be de- veloped in ourselves and in others, virtue argumentation theory does not offer a plausible alternative to a more standard agent- neutral account of good argument. 
 Here is the view of argument, and in particular of good argument, which we take as our starting point. When we put forward an argument, we seek to rationally persuade others to accept our conclusion.1 Given this, it seems natural for an ac- count of good argument to center on the ability of an argument to provide its intended audience with good reasons to accept its conclusion. A good argument is an argument that provides, via its premises, sufficient justification for believing its conclusion to be true or highly probable, or for accepting that the course of action it advises is one that certainly or highly probably should be taken. This account of good argument has both logical and epistemic elements. 
 Of course, this is not an account of good argument that a virtue argumentation theorist would accept. The virtue theorist thinks that what makes an argument good is that the person pre- senting it has argued well, whereas we think that what makes it the case that an arguer has argued well is that they have pre- sented an argument that is good in the sense described in the previous paragraph. It would be question-begging to assume that ours is the correct account. However, we think that the intuitions 1 This view of argument is widespread. One well-known source is Govier 1989. We realize, however, that it is not a conception of argument that con- nects with all projects in argumentation theory. In particular, it does not neatly align with the view of arguments as best defined in terms of the dia- logical and dialectical contexts in which they are presented (see, for example, Walton 1990). Note however that there are some affinities between our view and the dialogical view in that we do define argument in terms of the inten- tions of an arguer with respect to their interlocutor, which opens up the pos- sibility that matters of character and trust are relevant (which they could not be if we considered an argument to be, for example, just an abstract structure of propositions). A discussion of the relationship between our approach and the dialogical view of argument is beyond the scope of this paper. Tracy Bowell and Justine Kingsbury © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 24 about the purpose of arguments from which it arises are deep- seated, and that any agent-centered account that cannot accom- modate them will be unable to offer a complete account of good argument. 2. Virtue epistemology
 A character-based account of knowledge explains what it is for an agent to know that p in terms of that agent’s exercise of the relevant epistemic virtues in coming to truly believe that p. Vir- tue-based epistemic theories divide into two main strands, reli- abilist and responsibilist, differentiated by what they count as the relevant epistemic virtues. The former, the roots of which lie in more standard reliabilist approaches to knowledge, was first developed by Ernest Sosa. According to a virtue-reliabilist ac- count of knowledge, the relevant epistemic virtues include per- ceptual abilities, introspection, memory and the ability to reason deductively and inductively (Sosa 1991). The responsibilist al- ternative was first developed by Lorraine Code (who coined the label) and by James Montmarquet (1987). For the principal ele- ments of her account of the overarching virtue of epistemic re- sponsibility, Code returns to the Aristotelian intellectual virtues of wisdom, intelligence and prudence (1984, 40) fleshing her central idea out thus:
 The intellectually virtuous person… is one who finds value in knowing and understanding how things really are. S/he resists the temptation to live with partial explanations where fuller ones are attainable, the tempta- tion to live in fantasy or in a world of dream or illusion, considering it better to know, despite the tempting com- fort and complacency that a life of fantasy or illusion (or well-tinged with fantasy or illusion) can offer. (Code 1984, 44) 
 The responsibilist approach is delineated from the reli- abilist one by its emphasis on intellectual virtues the exercise of which involves the agent’s choice (consequently, the agent is responsible for exercising or not exercising them). But the de- lineation is not clear cut, because in the case of some reliabilist virtues, one might choose whether or not to exercise them, and making the right choice in such a situation (or failing to make it) might itself constitute a virtuous (or vicious) act. (For example, one might be a good perceiver and have a good memory, but on Taking Character Into Account © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 25 some particular occasion fail to make the effort to mentally re- trieve the details of some event which one has observed.) While the responsibilist and reliabilist strands have developed as com- peting accounts of intellectual virtue, some contributions to the literature explore the possibility of an account which draws upon both reliabilist and responsibilist resources (e.g. Battaly 2000, Lepock 2010). Linda Zagzebski has developed perhaps the most complete account of intellectual virtues, and although she identifies her theory as a responsibilist one, she includes a number of characteristics that are better classified as reliabilist (Zagzebski 1996, 114). 3. Virtue argumentation theory 
 Recently, as agent and character-based strands of epistemology have become established, some philosophers interested in argu- mentation have begun to consider whether a similar agent- oriented turn has the potential to improve upon our understand- ing of what constitutes good argument. Among these, Andrew Aberdein has developed an initial framework for a virtue-based account of argumentation that has some similarities with virtue- theoretic approaches to knowledge in a (roughly) responsibilist vein (2010). On Aberdein’s account, the proper exercise of the argumentative virtues would become constitutive of the norms of argument. To anyone comfortably embedded in a more or less agent- neutral approach to argument appraisal, a move towards includ- ing consideration of the characteristics of arguers is likely to seem counter-intuitive. Indeed, many arguments often regarded as fallacious are held to be so precisely because they involve appeals to claims about arguers, and such claims are deemed to be irrelevant to the truth of their conclusions. On the face of it, an argument that is structurally strong and has premises that the appraiser of the argument has good reason to accept is not weakened by claims about the character or competence of the arguer. Any shift in the direction of an agent-based approach may itself appear to commit some kind of illegitimate ad hominem move. Although the established approach to argument evaluation leans strongly towards agent-neutrality, it does allow that there are circumstances in which facts about a person’s character are relevant to whether or not we should believe what that person says: there are legitimate ad hominem arguments. If we have reason to think that X is habitually dishonest, we should not ac- Tracy Bowell and Justine Kingsbury © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 26 cept X’s claim solely on the grounds that X has made it. If we have reason to think that Y is unreliable about the kind of thing about which she is testifying—for example, if Y is making a claim about distances on the basis of her perceptions and we know that Y’s depth perception is defective—and we have no independent reason to believe it, then we should not accept Y’s claim. Sometimes, then, pointing out a characteristic of a speaker (her lack of some relevant epistemic virtue, either re- sponsibilist or reliabilist) can undermine her claim. Likewise, when we accept someone’s claim on the grounds that they are an authority on the subject in question and we are not competent to establish the truth of the claim more directly, the character of the person appealed to is in play. A legitimate appeal to author- ity requires not just that the person is genuinely an authority on the subject in question and that there is a high degree of agree- ment amongst authorities in that area, but also that there is no reason to think that the person in question is biased, insincere or untrustworthy. These latter conditions open the way to an evaluation of her epistemic character. Should that evaluation turn out a certain way, it will provide good reason to reject the argument from authority.2
 Legitimate ad hominem arguments provide reasons to doubt the truth of a claim on the basis of facts about the person making it. It is commonly supposed that it is never reasonable to reject an argument on the basis of such facts, however. If the CEO of a brewing company provides an argument to the conclu- sion that the legal drinking age should not be raised, we should not reject his argument solely on the grounds that the arguer has a vested interest in the conclusion’s being accepted: he has pro- vided an argument, and we should evaluate that argument on its merits. On the face of it, if it is a good argument—either valid or inductively forceful, with premises that we have good reason to believe—then no facts about the arguer will make it cease to be so. However, when someone presents an argument they are in general asserting the premises and suggesting, implicitly or ex- plicitly, that the premises provide good reason to accept the 2 Notice that a principle of charity applies here—if we have no good reason to do otherwise, we take honesty, sincerity and an intention to believe and communicate truths as the default position to be attributed to an arguer. It is only when we have good reason to think that one of these characteristics is absent that we explicitly treat it as relevant. Taking Character Into Account © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 27 conclusion. To the extent that we accept the premises purely be- cause the arguer has asserted them, we should care about the veracity of the arguer. In other words, part of what someone does when she gives an argument is make substantive claims, and so all of the considerations above, about how the character of a speaker legitimately influences whether we should believe her claims, apply. This is not to say that the conditional that ap- pears at the end of the previous paragraph is false: merely that considerations involving the character of the arguer may already be packed into the “premises that we have good reason to be- lieve” part of its antecedent. What about the “valid or inductively forceful” part? Is it ever the case that facts about the arguer legitimately influence our evaluation of the structure of an argument? On the face of it, no. If the conclusion logically follows from the premises, or if given the premises, the conclusion is almost certain to be true, no fact about the arguer can change that. Likewise, an argument which is structurally weak is not redeemed by any facts about the epistemic virtues of the arguer. However, consider the fol- lowing cases.
 Case 1. It is surely true that facts about the arguer cannot undermine the validity of a deductive argument, or make an in- valid argument valid. However, perhaps they can be relevant to the evaluation of an inductive argument, and not just by casting doubt on the truth of the premises. Suppose someone tries to convince me that Tom is not fluent in German, on the grounds that Tom is a New Zealander and only 2% of New Zealanders are fluent in German. This looks like a good enough inductive argument. However, there could be information that I lack which would undermine the argument without falsifying the premises; for example, the information that Tom is the New Zealand ambassador to Germany. Given this, facts about the ar- guer might matter. The arguer has given me no information about Tom other than that he is a New Zealander. Is the arguer the sort of person who would tell me if he knew that Tom was the New Zealand ambassador to Germany, or is he the sort of person that would delight in tricking me into thinking that the New Zealand ambassador to Germany doesn’t speak German? Does he have some particular interest in getting me to believe the conclusion which might have lead him to leave out facts which, if presented, would make it seem less likely that the con- clusion is true? What sort of person I take him to be seems to matter to whether I should accept his conclusion on the basis of his premises. Note that this would not be the case if the arguer provided a deductively valid argument (substitute “No New Tracy Bowell and Justine Kingsbury © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 28 Zealanders speak German” for “Only 2% of New Zealanders speak German”). There, provided the premises are true, the con- clusion must be true too, and so the issue of character can only come in if we are deciding whether or not to believe the prem- ises purely on the arguer’s say-so. Case 2: Just as there are areas in which I defer to experts about matters of fact, there might be areas in which I defer to experts about matters of logic. Some kinds of reasoning might be just too complicated for the untrained to follow. Then it makes epistemic sense to defer to someone who is an expert in this kind of reasoning. Consider the Monty Hall puzzle. A game-show host indi- cates three doors and tells a contestant that there is a prize be- hind one of the doors. Each of the other doors, he says, has a booby prize behind it. The contestant is then asked to choose a door. When the contestant has made a choice, the host opens one of the doors that has not been chosen and reveals that there is a booby prize behind it. The host then asks the contestant whether she wishes to revise her choice or not. The contestant makes a decision and then the host opens the door that the con- testant settles on. If the prize is behind that door, the contestant wins; otherwise not. The puzzling question is this: should the contestant change doors when given the opportunity to do so? When the contestant first picks a door the probability that it has the prize behind it is one third. Before the host opens one of the unchosen doors, the contestant already knows that one of the unchosen doors has no prize behind it; at least one of the unchosen doors must be a loser. So when the unchosen door is opened, she learns nothing new that is relevant to the probability that she has already chosen the winning door: that remains at one third. If she swaps she will not choose the door the host has just revealed to be a loser; so the opportunity to swap is equiva- lent to the opportunity of opening both the other doors instead of the one she has picked, which doubles her chances of winning. So she should swap.3 This is highly counterintuitive, and the reasoning is diffi- cult to get a grip on. Imagine someone who, when the reasons to swap are explained to her, kind of understands; just after the ex- planation has been given, she thinks she can see why swapping is the rational thing to do, but she can’t keep the reasons in her head—ten minutes later, although she still thinks the rational 3 There is an extensive literature on the Monty Hall problem. See, for exam- ple, Franco-Watkins et. al. (2003) on why the contestant should swap and why people find the problem difficult. Taking Character Into Account © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 29 thing to do is to swap, she can no longer see exactly why that’s so. However, she is informed by reliable experts that the argu- ment that the contestant should swap is absolutely watertight. It would be better if she could firmly establish this for herself, but if she is incapable of it, it is reasonable for her to defer to rele- vant experts. Case 2 does not seem to be a case in which facts about the arguer are legitimately influencing someone’s evaluation of the structure of the argument. Rather, she is accepting the conclu- sion of the argument even though she cannot evaluate the struc- ture of the argument: because of facts about the arguer, she trusts that the arguer would not put forward an argument that was structurally bad. How about Case 1? With inductive argu- ments, by definition, the fact that the argument is structurally strong and the premises are true does not guarantee the truth of the conclusion—it is always possible that there is further infor- mation that would undermine the support that the premises give to the conclusion. There seem to be a couple of different ways in which one might think about this case. On the one hand, one might think that the structure of the argument as stated (2% of As are Bs, X is an A, therefore X is a not a B) makes it a strong inductive argument, in the sense that if all that you know about the situation is what is contained in the premises, you have good reason to believe the conclusion. If we think of the argument in this way, then facts about the arguer are not relevant to the evaluation of the structure of the argument. On the other hand, one might think that the argument has an unstated premise, something along the lines of “There is nothing unusual about Tom that bears on the likelihood of his speaking German,” which makes it a much stronger argument. If we take the argu- ment this second way, then facts about the arguer legitimately influence our evaluation, not of the structure of the argument, but of the plausibility of the unstated premise. The upshot is that although characteristics of the arguer can be relevant to the question of whether or not we should ac- cept the premises of her argument purely because she has as- serted them, they do not appear to be relevant to our evaluation of the structure of her argument. What does this mean for the virtue argumentation theorist's attempt to give an account of good argument that appeals to the character of the arguer? A character-based account of knowledge explains what it is for an agent to know that p in terms of that agent’s exercise of the relevant epistemic virtues in coming to truly believe that p. A full-blown character-based account of good argument would explain what it is for an argument to be good in terms of the ar- Tracy Bowell and Justine Kingsbury © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 30 gumentational virtues displayed by the arguer in putting it for- ward. Thus, on such an account, the virtues exercised by the good arguer would become constitutive of the goodness of their argument. There are important disanalogies between good ar- gument and knowledge that suggest that we ought to resist this move. While it seems plausible that someone who displays all the relevant characteristics, both reliabilist and responsibilist, will be able to construct good arguments and be successful at appraising the arguments of others, surely a good argument could be put forward by someone who lacked those characteris- tics. Consider someone who puts forward a valid argument with true premises but doesn’t see that it is a good argument— someone who, for instance, has learned to recite a valid syllo- gism, or someone who doesn’t understand the premises of her own argument, or someone who mistakenly thinks that the premises of her own argument are false. We would not deny that the argument is a good argument; rather, we would say that the arguer has accidentally put forward a good argument.4 This con- trasts with what we would say in the parallel case regarding knowledge: we would deny that the person who accidentally ar- rives at a true belief that p knows that p. There is a way to restore the parallel with virtue episte- mology: deny that accidentally sound arguments are good argu- ments, just as accidentally true beliefs are not knowledge. How- ever, this does not fit with the sense of argument being used here. It would force us to say that if two different arguers pre- sented the same sentences in order to reach the same conclusion in the same way, there are actually two different arguments pre- sent. There are rhetorical and dialectical grounds for embracing that conclusion, but not for the purposes at hand. When it comes to argument evaluation, the standard view is standard for a rea- son. It allows us to say that if one person presents a good argu- ment, another person can use that same argument, confident that it will remain so. Virtue argumentation theorists see the differ- ence but are blind to the sameness. The fact that a good argument can be put forward by an argumentationally unvirtuous arguer suggests that in those cases 4 While the argument remains a good argument, if the arguer doesn’t have good reason to accept her own premises or does not see that the syllogism is valid, the arguer herself ought not to be rationally persuaded by her own ar- gument. However, someone who hears her argument, understands and be- lieves the premises with good reason should be rationally persuaded by it. Taking Character Into Account © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 31 in which a good argument is put forward by a virtuous arguer, the goodness of the argument is not constituted by the virtues displayed by the arguer. An agent-based account cannot replace a conventional account of argument. However, we do not take this to mean that there is no future for virtue argumentation the- ory. It is useful to think about the virtues of the good arguer; whether someone has them or not makes a difference to whether or not we should accept her premises. More broadly, the virtues of the good arguer and those of the good argument evaluator are ones we should want to develop in ourselves and encourage in others, and so we should think about what they are and about why they are virtues.
 Which characteristics of an arguer might usefully be taken into account in the evaluation of her premises? We have already mentioned some: reliabilist virtues such as perceptual acuity, responsibilist virtues such as honesty. Linda Zagzebski provides a list of epistemic virtues which include further candidates: the ability to recognize the salient facts; sensitivity to detail; open- mindedness; fairness; epistemic humility; perseverance; dili- gence, care and thoroughness; the ability to recognise reliable authority; intellectual candour; intellectual courage, autonomy, boldness, creativity and inventiveness (1996, 114). These are epistemic virtues (the reliabilist and the responsibilist ones alike) because possessing them tends towards the believing and assert- ing of truths. A person who is open-minded and epistemically humble will be prepared to revise her beliefs, even cherished beliefs, in the light of new evidence. A person who has intellec- tual courage will be prepared to consider the possibility that un- popular and unpalatable claims are true, and if the evidence for them is compelling, to put them forward even when it is obvious they will not be well-received. A person who is epistemically diligent (and skilled at argument evaluation) will carefully con- sider the evidence for a claim before accepting it or passing it on to others. Knowing about these kinds of characteristics of an ar- guer may in some cases legitimately influence our evaluation of their argument, because someone who lacks these characteristics is less likely to be someone who believes and asserts truths. However, in spite of its virtues as a theory informing the peda- gogy of argumentation and a framework for thinking about spe- cific instances of arguing, virtue argumentation theory cannot be the whole story when it comes to argument evaluation.5
 5Thanks to Jonathan McKeown-Green, Tim Dare, Stephanie Gibbons and the audience at a Waikato University Philosophy Seminar for their feedback on this paper. Thanks also to two anonymous referees for this journal, whose Tracy Bowell and Justine Kingsbury © Tracy Bowell and Justine Kingsbury. Informal Logic, Vol. 33, No. 1 (2013), pp. 22-32. 32 References
 Aberdein, A. (2010). Virtue in Argument. Argumentation 24, 165-179. Battaly, H. (2000). What is virtue epistemology? Proceedings of the Twentieth World Congress of Philosophy. 
 Battaly, H. (2010). attacking character: Ad Hominem argument and virtue epistemology. Informal Logic 30, 361-390.
 Code, L. (1984). Toward a ‘responsibilist’ epistemol- ogy. Philosophy and Phenomenological Research, 45, 29– 50. Cohen, Daniel H. (2009). Keeping an open mind and having a sense of proportion as virtues in argumentation. Cogency 1, 49-64. 
 Franco-Watkins, A., Derks, P. and Dougherty, M. (2003). Rea- soning in the Monty Hall problem: examining choice behav- ior and probability judgements. Thinking and Reasoning 9, 67-90.
 Govier, T. (1989). Critical thinking as argument analysis. Argu- mentation III, 115-126. Lepock, C. (2011). Unifying the intellectual virtues. Philosophy and Phenomenological Research 83, 106–128. Montmarquet, J. (1987). Epistemic virtue. Mind 96, 482–497. Paul, Richard W. (1992). Teaching critical reasoning in the strong sense: Getting behind worldviews. In Richard A. Ta- laska (Ed.), Critical Reasoning in Contemporary Culture (135-156), Albany: State University of New York Press. 
 Sosa, E. (1991). Knowledge in Perspective. Cambridge: Cam- bridge University Press. Walton, Douglas N. (1990). What is reasoning? What is an ar- gument? Journal of Philosophy 87: 399-419. Zagzebski, L. (1996). Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge. Cambridge: Cambridge University Press. careful reading of the paper and insightful suggestions about how to improve it are much appreciated.