1 Introduction

An earlier version of this article was presented as a keynote lecture at the Argage conference 2021, 10-12 November, University of Neuchâtel.

Argumentative topoi from classical rhetoric provide us with an interesting angle for studying contemporary argumentative discourse. This is particularly true for the topos of eikos, which has been translated variously as an “eikotic argument” and “plausible reasoning” (Walton et al. 2014; Walton 2019), “(early Greek) probability argument” (Gagarin 1994; Kraus 2007), or “argument from likelihood” (Tindale 2010). The term eikos refers to what is likely from the viewpoint of social expectation (Kraus 2007, p. 5; Hoffman 2008, p. 7; Tindale 2010, p. 82).Footnote 2 Generally, eikos arguments appeal to customary behavior, to how people would act in particular circumstances, i.e., to “what a man might logically be supposed to do” (Kennedy 1963, p. 100). According to Walton (2019, p. 72), these arguments are omnipresent in contemporary discourse. Be that as it may, their modern use has been studied only scarcely.Footnote 3

The best known ancient examples of eikos are the arguments that, according to Aristotle, the Sophists would have recommended to a weak man and a strong man who had ended up in a fight. The weak man was advised to say that he is not to blame, since it is unlikely that a weak person would attack someone who is obviously much stronger:

(…) if a man is not likely to be guilty of what he is accused of, for instance if, being weak, he is accused of assault and battery, his defense will be that the crime is not probable[.] (Rhetoric 1402a17–22; see Freese 1975, p. 335; cf. Kennedy 2007, pp. 188–189)

In turn, the stronger man should say that he knew his physical superiority would make him a suspect, therefore it is implausible that he acted as the initiator:

[B]ut if he is likely to be guilty, for instance, if he is strong, it may be argued that the crime is not probable, for the very reason that it was bound to appear so. (ibidem)

While the weak man’s argument has been described as an “ordinary” (Gagarin 2001, p. 284) or “straight” (Kraus 2007, p. 3) eikos argument, the strong man’s argument is referred to as a “reverse” eikos argument. A reverse eikos argument reverses the justificatory power of an opponent’s argument—in the weak and strong man case, their different builds.

Analyses of both arguments have been presented by Walton (2019) and Walton et al. (2014). The main purpose of these authors is to show that these arguments are instances of plausible reasoning that can be analyzed and evaluated with the computational argumentation tool the Carneades Argumentation System (CAS). While Walton, Tindale and Gordon also analyze other types of plausible reasoning in their article, Walton concentrates on the weak and strong man case and has the additional aim of comparing the relative strength of the weak man’s and strong man’s arguments, in order to find an explanation for Aristotle’s suggestion that the strong man’s argument is the weaker one (2019, p. 47).Footnote 4 Proposing two reconstructions of the strong man's argument and being unable to decide which of them was the best, he concludes his article with a call to further study which analysis “is best justified by the text and context of the example” (ibidem, p. 68).

In this article I take up that call and present an alternative analysis and evaluation of this type of argument. I argue that both the weak man’s and strong man’s arguments belong to a subclass of the general class of eikos arguments, which I call “I’m not stupid” arguments. My reconstruction of the argumentative pattern of this type of argument is based on a collection of ten contemporary argument examples that have a line of reasoning similar to the weak man’s and strong man’s arguments.Footnote 5 These ten cases may seem a small collection, but the wide variation in the argumentative elements that are expressed explicitly and those that have been left implicit offers enough clues for a full-fledged representation of their argumentation structure. My analysis identifies five new key elements compared to the previous ones. This result shows that we need empirical data to strengthen our analyses of argument schemes and argumentation structures, and ultimately the typologies that are based on them.

In the following, I will first explain the concept of an eikos argument on the basis of its characterization in classical rhetorical theory (Sect. 2). Then I will turn to my analysis of an “I’m not stupid” argument (Sects. 3 and 4), after which I will address the critical questions that can be drawn from the analysis and discuss how they work out in an assessment of these arguments (Sect. 5). These last three sections also contain a comparison of my own analyses with those of Walton et al. (2014) and Walton (2019). All of the reconstructions can be found in the appendices.Footnote 6 As Walton et al. (2014) and Walton’s (2019) analyses were originally visualized in the diagram style of CAS, I converted theirs into a pragma-dialectical argumentation structure to allow for a proper comparison with my own, which are based on the pragma-dialectical way of schematizing argumentation (van Eemeren and Grootendorst 1992; van Eemeren and Snoeck Henkemans 2016).

2 Eikos Arguments and the Context of Criminal Law

Eikos is the present participle of the Greek verb eoika (“look like”). It is asserted that two basic meanings can be derived from the use of the verb eoika in old Greek texts (Kraus 2007, p. 5; Hoffman 2008, p. 13; see also Schmitz 2000, p. 69): (1) likeness/being similar, and (2) appropriateness/fittingness. In an eikos argument these two meanings come together: because an event is in accordance with what is socially expected (second meaning), it looks like what is considered to be true (first meaning) (Hoffman ibidem, p. 21).

Eikos arguments appeal to how people are supposed to act in particular circumstances. They derive their persuasiveness from generally accepted assumptions about the universal characteristics of human nature (Kennedy 1994, p. 24).Footnote 7 These are assumptions about human motivation in terms of earlier life, custom, habit, or circumstances such as motive, means, or aim, and the values of human society, i.e., morality and justice (Kraus 2007, p. 8). In the Rhetoric to Alexander human motivation is described in terms of emotion, custom, and advantage (1428a35–1428b10; cf. Kraus ibidem). For instance, the defendant was in need of money, and we therefore infer that it was he who killed his rich father. Inferences like this one lead Braet (2004, p. 136) to describe an eikos argument as “a plausible causal generalization concerning human behavior, on the basis of which, given certain causes, the probable occurrence of criminal behavior can, as it were, be predicted with hindsight.”

It is assumed that theorizing about eikos started with the development of rhetorical theory in the fifth century BC, probably by Corax and/or Tisias in Sicily (Gagarin 1990, p. 30). In this period, the development of the Sicilian—but also, for that matter, the Athenian—democracy yielded a growing interest in proof and argumentation.Footnote 8 In the rhetorical handbook Rhetoric to Alexander—our oldest such source, reflecting the sophistic and pre-Aristotelian doctrine—eikos is mentioned as one of a group of seven types of arguments, called pisteis (proofs) (Braet 2004, pp. 129–130). Although the Rhetoric to Alexander presents pisteis as arguments that can sustain or attack a standpoint in either a juridical or political speech, it only elaborates on their legal usage. Ancient argumentative practice shows that judicial oratory is indeed the primary domain where eikos arguments were used (Kraus 2007, p. 6). More in particular, eikos is a subargument supporting the first of the Hermagorean lines of defense developed for criminal law, i.e., the conjectural status: the issue of whether or not the defendant did indeed commit the crime they have been accused of. Eikos arguments were typically used as additional proof or in cases where hard proof, such as written documents or testimonies, was absent or indecisive (Gagarin 1994, pp. 53 ff.; Kraus 2007, p. 1, 3; Tindale 2010, pp. 78–79).Footnote 9

3 Reconstruction of the Main Argumentation

The reconstruction of the “I’m not stupid” argument that I present in this and the next section takes the shape of an “argumentative pattern.” An argumentative pattern consists in an argumentation structure that displays a constellation of argumentative moves in support of a certain type of standpoint dealing with a certain difference of opinion that typifies the argumentative context at hand, i.e., the argumentative activity type (van Eemeren 2017, pp. 18–19). The previous section made clear that the ancient use of eikos arguments predominantly occurred in the legal domain. When we look at contemporary instances of eikos, we find that they too are typical of the genre of adjudication. However, my ten examples of “I’m not stupid” arguments show that this need not be the formal context of a legal trial. Any situation in which someone has been accused of illegal or improper conduct and in which the accused feels inclined to offer a defense against this accusation can be regarded as what I call an “informal legal context”.Footnote 10 This context is legal in the sense that it concerns a communication activity resembling a legal procedure of accusation and defense; it is informal since the accusation and defense take place between citizens, outside a court of law.

The difference of opinion at stake in the informal activity type of accusation and defense concerns the question of whether the defendant is blameworthy. A judgement about this question depends on the four requirements for punishment according to ancient rhetorical status theory: (1) that the defendant did what they are accused of, (2) that what they did is illegal, (3) that they cannot rely on justification or grounds for being excused, (4) that the legal procedure has been properly conducted. These requirements are more or less the same as those that need to be fulfilled for a conviction in a formal court case in criminal law, albeit that in modern legal systems the fourth issue is a preliminary requirement, which can lead to dismissal of the case (at least in the Netherlands). The defendant’s standpoint in an “I’m not stupid” argument is that they did not commit the act they are accused of. Like the general class of eikos arguments, “I’m not stupid” arguments therefore also typically function as support in relation to the first status.

The first level of defense of an “I’m not stupid” argument consists in an appeal to the harmful consequences that would result from the act the arguer has been accused of. The weak man appeals to the consequence of being beaten up, the strong man to the risk of becoming the likely suspect. For this reason, Walton et al. (2014, pp. 93–95) base their analysis on the general structure of the “argumentation scheme for argument from negative consequences” presented in Walton, Reed and Macagno (2008, p. 332):

Premise:If A is brought about, then bad consequences will occur.

Conclusion:Therefore A should not be brought about.

Below I show, by making use of the pragma-dialectical notation method, how Walton, Tindale and Gordon apply this scheme to the argument of the weak man (whom they call the “smaller” man)Footnote 11:

1.The smaller man did not attack the bigger man, becauseFootnote 12

1.1a If the smaller man were to attack the bigger man, he would be beaten and humiliated, and

1.1b Being beaten and humiliated are negative consequences.

I do not endorse this analysis, because it seems to lack an essential element.Footnote 13 What is missing, is the defendant’s attitude to the negative consequences: the alleged bad consequences are only unlikely to happen if the arguer made an attempt to avoid them. Walton must have noticed this too, because in his follow-up analysis (2019, pp. 61 ff.) he substituted premise 1.1b with another one:

  1. 1

    The weak man did not assault the strong man, because

  2. 1.1a

    If the weak man did such a thing there would likely be negative consequences, and

  3. 1.1b

    Foreseeing negative consequences is a reason for not doing something.

In my view, this new premise is definitely better than the one stating that the predicted consequences are bad. However, its problem is that, because of the general drafting, it cannot adequately link the appeal to consequences with the standpoint that one did not do the act one has been accused of. After all, if you have a reason for avoiding negative consequences, this does not necessarily mean that you actually want to avoid those consequences, i.e., that you intend to act on this reason.

Some real discourse data from my collection of examples of “I’m not stupid” arguments indicate the shape that a more adequate formulation of the second coordinatively linked premise might take. Both extracts below are the arguments of two professional cyclists, defending themselves against the accusation of using performance enhancing drugs. In (1) the arguer says “I would never risk my life again”—spoken by seven times Tour de France winner Lance Armstrong, referring to his surviving cancer. The second cyclist says, in (2), that he is not willing to risk all the things he’s got in his life:

  1. (1)

    (…) why would I (…) dope myself up and risk my life again? (…) I would never do that

  2. (2)

    If I doped I would potentially stand to lose everything. It’s a long list. (…) I am not willing to risk all those things I’ve got in my life.

These arguments boil down to the line of reasoning: (1) I did not dope, because (1.1a) if I doped, I would risk negative consequences, and (1.1b) I would not risk those consequences (see appendix 2). The claim that the arguer would not risk the alleged harmful consequences leaves no doubt about their intentions. This premise therefore adequately fills in the missing link between the appeal to consequences and the standpoint in a logical way. If you expect that certain consequences would result from an act, and if you do not want these consequences to happen, then this gives at least some credibility to the standpoint that you would not commit that act.

My collection of real discourse data shows that premises 1.1a and 1.1b can be drafted in all kinds of ways. For one thing, they do not necessarily contain the word “risk”. Moreover, they are often drafted in a way that has been considered particularly suitable for eikos arguments, i.e., with a rhetorical question (cf. Bonner 1927, p. 227; Gagarin 2002, p. 29; Fairchild 1979; Kraus 2007, p. 7).Footnote 14 We see one such typical drafting in the following arguments, where the rhetorical question implies premise 1.1b:

  1. (3)

    I know what the consequences could be then; they could take away my practice. Would I let that happen?

  2. (4)

    In my whole life I have earned more money than I can spend. Why would I take a risk, then?

In (3) a lawyer denies the accusation of helping clients while he was suspended. The rhetorical question “Would I let that happen?” can be taken as the contention that the arguer would not let that happen (premise 1.1b)—“that” being the consequences mentioned in the preceding sentence, which instantiates premise 1.1a. Example (4) concerns a denial of money laundering. Here too, the rhetorical question can be taken as denying that the arguer would take the risk. In this latter example, premise 1.1a—the appeal to the negative consequences—has remained implicit. Since the arguer was already in jail, he may have judged it superfluous to explicitly state this harmful consequence in his argument.

There are also examples in which the rhetorical question expresses the coordinatively linked premises 1.1a and 1.1b simultaneously:

  1. (5)

    Suppose I did it, would I then set a dachshund onto a bull?Footnote 15

    1.1a: If I did X, I would have set a dachshund onto a bull

    1.1b: I would not risk that

  1. (6)

    Why cheat [use drugs] and put in a tortuous ride [in the Tour de France] knowing there was a good chance of being disqualified?

    1.1a: [If he cheated], [he] knew that he had a good chance of being disqualified

    1.1b: He would not risk being disqualified

In example (5), setting a dachshund onto a bull means making oneself an easy target for prosecution. Example (6) appeals to the negative consequence of being disqualified in the Tour de France. In both examples, the rhetorical question not only appeals to negative consequences (premise 1.1a), but at the same time implies the arguer’s denial of wanting to risk these consequences happening (premise 1.1b). Note that the latter argument is not voiced by the accused cyclist himself, but by a fan who stood up for him on an internet forum.

I conclude this section with two remarks. The first is about the standpoint of an “I’m not stupid” argument, i.e., the denial of having committed the act one has been accused of. In my collection of examples, the standpoint often remains implicit. And this is only natural, as it is clearly implied, if not by one or both of the main arguments, then by the activity type of accusation and defense. The second remark is that Walton et al. (2014) and Walton (2019) reconstructed a main argumentation that is slightly more complex. They added an intermediate step between the standpoint and the coordinatively linked premises (see appendices 3, 4, 5, 6). I do not include this intermediate step in my analysis. In my reconstruction such a premise is superfluous because the implausibility of the accusation is already conveyed by the subargumentation of premise 1.1b; this will be discussed in the next section.Footnote 16

4 Reconstruction of the Subargumentation

According to van Eemeren (2017, p. 20), an argumentative pattern results from an arguer’s attempt to make the strongest possible defense of their case by anticipating all relevant critical questions. In my understanding, this means that a pattern represents a certain mode of argumentation in its most full-fledged version; that is to say, when all the potential critical responses of an antagonist are included as premises in the model. Such a full-fledged version would in this case require that both the appeal to the negative consequences (premise 1.1a) and the contention that the arguer would not risk them happening (premise 1.1b) should be supported with reasons that argue, on the one hand, why the alleged consequences are likely to occur and, on the other hand, why we should believe that the arguer would not risk these consequences happening. Section 4.1 will discuss the support for premise 1.1a; Sect. 4.2 the support for premise 1.1b.

4.1 Support for the Claim That Harmful Consequences Will Occur

Walton et al.'s (2014; see appendix 3) reconstruction has one subpremise, i.e., a premise supporting the first coordinatively linked premise appealing to the harmful consequences. We have already seen that in the weak and strong man case, this subpremise consists in the men’s respective builds. The predicted negative consequence of being beaten up is justified by the weak man’s small build, whereas the strong man’s big build would have led to the negative consequence of becoming the likely suspect. I have reconstructed a similar subpremise in the argumentative pattern of an “I’m not stupid” argument (1.1a.1; see appendix 2); it also occurs in my collection of examples:

  1. (7)

    What bothers me is reading on the internet that I hire actors for my shows. A ridiculous idea. Because it would not only cost me lots of money, but I would be risking a lot as well. If just one actor went to the media, my reputation would be in tatters.

  2. (8)

    Imagine that five years from now a rider came out and said that I encouraged him to take drugs. Or that I bought a blood machine for him. What do you think would happen then? I’d be finished.

Extract (7) is from an interview with a TV host who claims that he can speak with dead people; in his shows he acts as a mediator between his guests and their deceased loved ones. The TV host had to defend himself against the accusation that the guests in his shows are hired actors. He denies this allegation by pointing out the negative consequences for his career and reputation, and supports this with the subpremise that actors could reveal what is going on in the TV shows. The same fear of people talking can be found in (8)—an argument of cycling coach Rasmussen, answering an interviewer who asked him whether he allows his riders to use performance enhancing drugs.

Both of Walton’s (2019) analyses of the strong man’s argument have a left branch that includes more premises. These are intermediate premises between the appeal to consequences and the lowest level premise(s). In my view, the intermediate premises of both reconstructions are of a repetitive and redundant nature. For instance, both reconstructions have a premise literally saying that negative consequences will occur, while a specification of the exact consequences is only made in a subpremise. The other premises all elaborate on the likelihood of the consequences occurring. For instance, reconstruction 2 (appendix 6) has the following line of reasoning: there would likely be negative consequences, because the strong man could be convicted, because people would think that the strong man was likely to do it, because the stronger man would be more likely to do it and the strong man was visibly stronger.

4.2 Support for the Claim That the Arguer Would Not Want to Risk the Predicted Consequences

The right branch of the appeal to consequences, i.e., the second coordinatively linked argument at the main level, remains unsupported in both Walton et al.'s (2014) and Walton’s (2019) analyses. This is completely understandable, considering the respective contents of the second coordinatively linked main premise in these analyses. Walton, Tindale and Gordon may have thought that the typification of the predicted consequences as being bad is obvious and needs no further support. Walton’s premise that foreseeing negative consequences is a reason for not doing the action that causes them takes an impersonal perspective and suggests that—generally—people consider foreseeing bad consequences as a reason for not doing the act that causes them. Since this does indeed seem, in general, to be a true observation, Walton may have regarded this premise as an expression of common ground, which would explain why he leaves it unsupported. However, the weak man’s and strong man’s arguments do not argue about what people generally do: they argue about what these men would do. Would they really not have taken the risk?

To find out how the right branch of an “I’m not stupid” argument can be supported, let us take another look at some of the examples. In (1) ex-cyclist Lance Armstrong argues that he would not risk the consequence of potentially losing his life, because he had already almost lost his life, namely when he suffered from brain cancer in the period preceding his seven victories in the Tour de France. This has been expressed in the part where he says that he came back from a death sentence.

  1. (1)

    (…) And if you consider my situation: A guy who comes back from arguably, you know, a death sentence, why would I then enter into a sport and dope myself up and risk my life again? That’s crazy. I would never do that. No. No way.

In (4) a Dutch businessman, who has been imprisoned in Thailand, claims that he has been falsely convicted for having laundered drugs money in that country. He argues that he would not have wanted to take the risk of such a crime because he does not need the money:

  1. (4)

    In my whole life I have earned more money than I can spend. Why would I take a risk, then? I might sometimes look a bit stupid, but I’m not an idiot.

In (1) and (4), the arguers refer to their respective personal circumstances to make it look plausible that they would not risk the negative consequences of the act they have been accused of.

Example (6) contains another type of reason. In this extract a fan stands up for cyclist Floyd Landis, who was accused of drug abuse in the 2006 Tour de France. Landis had lost 8 min in stage 16, but then announced that he would win the next day, which he did after performing a long solo ride. However, a shadow was cast on this victory after a urine test showed an unusually high level of testosterone. The fan argues that Landis knew he would be tested if he won the stage, and would thus have known he would be disqualified if he had indeed taken drugs:

  1. (6)

    (…) So he thought he could win Stage 17… If he won the stage, he knew he would be tested… Why cheat and put in a tortuous ride knowing there was a good chance of being disqualified? How would that prove anything to anyone, least of all himself? It does not make sense.Footnote 17

In this example, the parts in bold seem to boil down to a gain–loss calculation. The gain would be to prove something, i.e., that he (Floyd Landis) is capable of winning. The loss would be to have tortured himself with a solo ride. Since the gain could never be achieved, because taking drugs would always be revealed, and since the loss would still be there, this gain–loss calculation leads to the conclusion that it is unlikely he would have taken the risk of being disqualified (the second part of the first rhetorical question, which expresses both premises 1.1a and 1.1b at the same time).

On closer inspection, the circumstances referred to by Armstrong and the Dutch businessman can also be interpreted as being part of a gain–loss calculation. Although these two arguers do not explicitly weigh potential losses and gains, a calculation thereof can be regarded as being presupposed, in that they address one of these elements. Cyclist Lance Armstrong addresses the greatest imaginable loss: one’s life—and exacerbates it by referring to his medical past, implying that no gain could ever offset this potential risk. The businessman addresses (the absence of) a potential gain: because he is rich enough, money laundering would not bring him any significant benefit—thus implying that potential gains could not have outweighed the risk of ending up in jail. It is only logical that a calculation resulting in high losses overruling potential gains is used as a reason why one would not take a risk. After all, if the arguer is expected to gain great benefit from taking a risk, this would surely diminish the credibility of their claim that one would not want to take this risk.Footnote 18 In fact, Landis himself explicitly refers to a gain–loss calculation:

  1. (9)

    People say I had nothing to lose. I don’t understand that. I lost a lot. I had a great deal to lose. There was very little to gain by doing that [using drugs].

Emphasizing the costs of taking the risk and downplaying its benefits are two important premises that make up the subargumentation in the right branch of the argumentative pattern. However, my collection of real discourse data shows that these premises do not directly support the premise that the arguer would not want to risk the predicted harmful consequences. There is an another, intermediate set of premises that forms the bridge between the arguer not wanting to take the risk and the gain–loss premises. The first premise of this intermediate set says that it would be stupid to risk the predicted consequences; the second contends that the arguer is not stupid (again, see appendix 2 for the whole argumentative pattern). I take these premises to express the implausibility of the arguer having committed the act of which they have been accused: one cannot be so stupid, meaning that it is implausible that one did it. 

Let us have a look at some explicit versions of one or both of these two linked premises, expressing the stupidity of taking the risk and contending that one is not stupid. Example (1) conveys the first of this set of premises; (4) expresses both in an explicit way; and the rhetorical question in (10) expresses both in an indirect wayFootnote 19:

  1. (1)

    (…) And if you consider my situation: A guy who comes back from arguably, you know, a death sentence, why would I then enter into a sport and dope myself up and risk my life again? That’s crazy. I would never do that. No. No way.

  2. (4)

    In my whole life I have earned more money than I can spend. Why would I take a risk, then? I might sometimes look a bit stupid, but I’m not an idiot.

  3. (10)

    Chris [Froome] knew he would be tested. How stupid can you be to do this deliberately?

This section and the preceding one have shown that almost any element of the argumentative pattern (appendix 2) can be traced back to the examples in my collection of real discourse data. The only element that was not found is the premise supporting the “I’m not stupid” premise, i.e., premise 1.1b.1a’. And this seems only logical, because, from a rhetorical point of view, it would be rather stupid to give reasons why you are not stupid. An eikos argument’s rhetorical force lies in the arguer’s attempt to present oneself as someone who behaves according to common expectation. In the case of an “I’m not stupid” argument, the created expectation is that the arguer acts in accordance with the generally accepted idea that people usually try to avoid doing stupid things. Arguing that you are not stupid would highlight the fact that an individual could behave stupidly after all; it does not seem sensible to draw attention to this fact. This does not mean, however, that premise 1.1b.1a’ is useless: one can imagine that support for the “I’m not stupid” premise could be provided if the “I’m not stupid” premise is called into question by an opponent and the arguer is challenged to prove that they are not stupid. In such a case the arguer could refer to how they behaved in previous situations or to positive attestations of family and friends.

5 The Evaluation of an “I’m Not Stupid” Argument

5.1 Critical Questions Derived from the Argumentative Pattern

The idea of an argumentative pattern as a full-fledged argumentation structure of a particular type of argumentation implies that it can serve as an ideal model, which has both a heuristic and a critical function. In their heuristic function, argumentative patterns are a template that helps an analyst interpret and analyze an actual piece of argumentation containing implicit elements. Their critical function is that they allow for a systematic evaluation of the type of argumentation they represent. Since a pattern makes explicit all the elements that are presupposed in a certain type of argument, it generates the critical questions that are relevant for its assessment. Walton too justifies the aim of his analyses by saying that they lay bare implicit premises, so that the argumentation can be properly evaluated (2019, p. 50).

Let us now start with the critical questions concerning the first coordinatively linked main argument, i.e., the appeal to the predicted negative consequences. Are these consequences really likely to occur? Is it likely that, in the case of the TV host, an actor would go to the media? Would this actor then not be ruining their own career? One could certainly doubt this claim. While it is true that ex-employees might be inclined to talk, they could be kept quiet if the TV host has enough money in his pocket to buy them off. And this is also true for all those doping sinners: they had enough money to keep people silent and to hire the best doctors who could help to keep them safe and healthy. The evaluation of the likelihood of the predicted consequences occurring should therefore concentrate on the arguer’s resources to minimize the predicted risk.

The assessment of the second coordinatively linked main argument centers around the question of whether there could be reasons that make it likely the arguer still took the risk. This depends on several critical questions, which can be derived from the subargumentation of the right branch of the argumentative pattern. The first question to ask here is: are the assumed consequences really so bad, or in other words, is the loss really so high? The second question should concern the gains that could be made from doing the act: might benefits outweigh the risk of the negative consequences occurring? Related to these questions, there is an important third question, which can be derived from the “I’m not stupid” premise. This question requires that the preceding questions should be considered from the arguer’s perspective. How would they have weighed the losses versus the gains? Consideration of the arguer’s perspective in estimating costs and benefits, and their respective chances of occurring, reveals that an “I’m not stupid” argument is less convincing when the arguer is a highly ambitious person and when the gains of the contested act are high and satisfy those ambitions. It is also less convincing when this person has the intelligence and the means to minimize the chances of getting caught.

These critical questions also allow for an assessment of a more general nature: is an “I’m not stupid” argument a reasonable type of argument? It seems clear enough that this argument can by no means provide conclusive proof. An “I’m not stupid” argument merely expresses a plausible course of contemplation and acting by the arguer. But the fact that the arguer was not likely to do the act of which they have been accused, does not mean that they did not do it. Actually, an “I’m not stupid” argument requires that an interlocutor take the arguer’s word for it. Most often we just do not know the arguer well enough to know how this person would have weighed losses, gains, and their respective chances of happening. And even if we think we know this person, we can never be sure that they acted in accordance with these known character traits. It can therefore be concluded that an “I’m not stupid” argument does give some plausibility, and may be more or less plausible in a given case, but even the more plausible versions are still weak arguments.

5.2 Walton et al.’s (2014) and Walton’s (2019) Assessments of the Weak Man’s and Strong Man’s Arguments

How does my evaluation differ from Walton et al.’s (2014) and Walton’s (2019) respective assessments of the weak man’s and strong man’s arguments? Walton, Tindale and Gordon do not elaborate greatly on this issue. They state that “[t]he plausible inference (…) only carries some weight” and consider that it could be defeated if new evidence comes in, for instance if the small man turned out to be a pugilist (2014, p. 95). For the rest, these authors make some remarks about the acceptability of the general class of eikos arguments. They say, for instance, that the acceptability of these arguments depends on the internal consistency of the narrative on which they are based, in relation to the evidence (ibidem, p. 96).

Without giving further explanation, Walton judges the weak man’s argument to be “an instance of a reasonable use of plausible reasoning” (2019, pp. 64–65, 72). His second reconstruction of the strong man’s argument receives a similar assessment, and he judges it to be equally plausible as the weak man’s argument (ibidem, pp. 67–68).Footnote 20 This leads Walton to conclude that comparing the weak man’s argument with the second version of the strong man’s argument results in a “deadlock” (ibidem, p. 68).

As for his first reconstruction of the strong man’s argument, Walton points out two problems affecting its evaluation. The first concerns the likelihood of the predicted consequences occurring. According to Walton, the negative consequence to which the strong man appeals is less obvious than the one predicted in the weak man’s argument. That is, “the likelihood of conviction is more indirect” than the weak man’s chance of being beaten up: “it is far from automatic that the stronger man would be convicted” (ibidem, p. 65). For this reason, Walton judges the strong man’s argument to be less plausible. Confusingly, Walton does not explain why the observed problem does not affect his second reconstruction, where the strong man appeals to similar negative consequences.

The second problem that Walton identifies in his first reconstruction concerns a premise that says that nobody would commit a crime without a reasonable hope of getting away with it (premise 1.1.1d; appendix 5). Walton (ibidem, p. 62) calls it as a key premise of the strong man’s argument because of a remark made in an article on eikos by Kraus (2007). Kraus notes that the strong man’s argument is based on the belief “that nobody would commit a crime without any reasonable hope of getting away with it unsuspected and undetected” (ibidem, p. 3); Walton’s premise is a verbatim presentation of these words.Footnote 21 The problem with this premise, according to Walton, is that it is contradicted by reality. We all know that crimes are often committed, even in cases where culprits had little chance of escaping suspicion and detection (2019, p. 67). Moreover, psychological insights show that criminals usually only consider short-term consequences, rather than long-term ones (p. 70). Walton therefore concludes that the contested premise should be regarded as weak, which makes reconstruction 1 a weaker version of the strong man’s argument than reconstruction 2 (ibidem, p. 68).

In my view, a premise referring to what “nobody” would do should not be part of an “I’m not stupid” argument. It is true that such a premise seems to be a natural element of the argumentation, because an eikos argument draws its power from appealing to customary behavior, i.e., to the behavior we generally expect from people in particular circumstances. However, rhetorical considerations should not guide a dialectical analysis. Although an assessment of an “I’m not stupid” argument’s persuasiveness relies on how people generally behave in a certain situation, its rationality should be assessed in terms of how the defendant would behave in that situation.

6 Conclusion and Discussion

In this article, I proposed a new analysis of the ancient weak man’s and strong man’s arguments by regarding them as instantiations of an “I’m not stupid” argument. Walton asked for an interpretation that “is best justified by the text and context of the example” (2019, p. 68). I could not agree more with the supposition that underlies this wish, which is that the text and context of an argument are indispensable ingredients for reconstructing argumentation. Walton himself had the problem that he could only rely on Aristotle’s text, which was also Walton et al.’s (2014) problem. Aristotle’s text presents, as we can now ascertain, an extremely concise drafting of two “I’m not stupid” arguments. I made use of a small but sufficient collection of ten examples, which all share the line of reasoning of the weak man’s and strong man’s arguments. These examples vary in terms of elements that are explicit or have been left implicit, and they gave me enough clues to ascertain the elements that make up a fully developed argumentative pattern of this type of argument. To me, this has shown the importance of studying empirical data when working on argument types and examining the structures they are based on.

Walton was unsure about reconstructing the strong man’s argument, not the weak man’s: he found that the latter was unproblematic and gave rise to only one interpretation. My analysis shows that both the weak man’s and strong man’s arguments are based on the same line of argument and fit the proposed argumentative pattern. They both appeal to negative consequences, containing (1) a premise pointing out what these consequences would be, and (2) a coordinatively connected premise reading that the arguer would not risk these harmful consequences. No previous analysis brought to light this second coordinatively linked premise, and hence the whole line of subargumentation supporting this argument was also overlooked. The right branch of the argumentative pattern, consisting of these newly identified elements, expresses the stupidity of risking the harmful consequences because of high losses and low gains, and the arguer’s denial of being stupid. The reconstruction of this branch shows that, in contrast to the previous analyses, there is no need for a main argument expressing the implausibility of the defendant’s guilt, since the implausibility is already expressed by this subargumentation.

The way an argument is analyzed has consequences for its evaluation. The previous analyses generate only some general criteria, like the requirement that the whole body of evidence in the given case should be taken into account, and that one should ask whether people can indeed oversee the long-term consequences of a criminal act. In pragma-dialectics, the function of an argumentative pattern is to allow for a systematic evaluation, because its premises generate the critical questions that should be used to evaluate an actual instance of that type of argument. My pattern revealed that the crucial point of an “I’m not stupid” argument is the arguer’s supposedly rational character in making a gain–loss calculation. This generates the following critical questions: Are the assumed consequences really as harmful as the arguer posits, and how likely are they to occur? Could the gains of the criminal act outweigh its losses and the risk of their occurrence? And, most importantly, how do we think the arguer—considering their character—would have answered these questions? Depending on its actual use, an “I’m not stupid” argument will be found more convincing in one case than another. However, the critical questions also allow for a general assessment of an “I’m not stupid” argument as a type of argument, and show that they are inherently weak arguments, because we can never be sure whether the arguer acted in accordance with what we expect of their character.

Argumentative patterns should be brought to light by means of qualitative empirical research (van Eemeren 2017, p. 22, footnote 27). In this article, I only concentrated on part of this research, i.e., the study of empirical data of the argumentation concerned. I used the concept of a pattern as an ideal model displaying a full-fledged version of the argument, while neglecting another interesting and important feature of patterns, namely that they “can be explained by the institutional preconditions prevailing in the communicative activity type in which they occur” (ibidem, p. 20). Patterns result from modes of strategic maneuvering that are supposed to convey the strongest potential defense in the relevant macro-context. I did not elaborate on this feature because the purpose of this article is to show how real discourse data can be helpful in reconstructing types of argument. However, I would like to conclude with a brief remark about the strategic function of an “I’m not stupid” argument.

The institutional point of the activity type of accusation and defense is to obtain a judgement about a defendant’s blameworthiness, and the four ancient rhetorical statuses mentioned at the beginning of Sect. 3 are preconditions of this activity type. The first, factual status (Did the defendant do it?) is presumed to be the strongest line of defense, provided that it succeeds, of course. And this is precisely where the problem lies in an informal legal context: the accuser and the accused are usually not equipped with sophisticated forensic detection methods, hence they cannot rely on conclusive evidence. This means that defendants have to come up with something else to refute the accusation, such as an appeal to alibi, lack of motive, physical inability, lack of means, or (other) eikos arguments.Footnote 22 If these do not apply, what remains is an appeal to ethos, i.e., to the arguer’s claim that they would never commit a crime because their character would not allow it.

In my view, an “I’m not stupid” argument is a disguised appeal to ethos. Defendants use it only when there is no better alternative. Nevertheless, it is a sophisticated appeal to ethos. It does not consist in an outright boast of the arguer’s good character, which would never be capable of committing a crime; rather, it appeals to the arguer’s rational character. This is strategic, because attacking an “I’m not stupid” argument boils down to saying that the arguer is lying about their gain–loss calculation. It is precisely because people generally prefer to stay away from this type of face-threatening communication that the arguer stands a chance of getting away with their argument.