5 Two kinds of ad hominem attack

In the following section I make a suggestion in order to disarm the problem of the ad hominem argument with regard to the positive reason component. Does an ad hominem attack always include fallacious reasoning? Walton (2008, p. 170) claims that “the argumentum ad hominem is not always fallacious, for in some instances questions of personal conduct, character, motives, etc., are legitimate and relevant to the issue.” Even though cases of ad hominem arguments might sometimes be informally fallacious, there are some highly relevant cases in which a particular ad hominem attack could be committed in order to remain epistemic vigilant. Since you, as a hearer, have a set of positive reasons—for instance being aware of the usual verbal espousals of politicians during an election campaign—you are forced to commit a personal attack. Did your reliability assessment rely on fallacious reasoning? In the literature there is a common distinction between three types of the ad hominem argument: the abusive, the circumstantial, and the tu quouqe argument (Groarke 2011; Walton 1998, 2008).[11] These kinds of personal attacks describe various pragma-dialectical reasoning in interpersonal communicative relationships. Below (cf. section 5.1, 5.2) I want to draw a further distinction between two kinds of ad hominem attack that are closely connected to communicative situations explicitly involving knowledge transmission. Hence I want to provide a framework that fits well with social epistemological dualism.

Before this, I want to address the Bayesian argumentation model, which is presupposed by the following account of ad hominem arguments. Few things have been said about the inductive reasoning model which is the underlying mechanism of an ad hominem attack. Roughly, one has to consider past experiences with regard to reliability, constituted by contexts and speaker properties, to adjust this experience for future communication. Harris et al. (2012) provide an account that fits well with an inductive model of reasoning, because, potentially evidence is not provided by deductive reasoning. They claim that the evaluation of a proposition or a given testimony is based on an individual’s probabilities, which could be described formally using Bayes’ Theorem. A big advantage of this account is that it describes our subjective evaluations in daily experiences very well. Often, when we are asked “do you believe S?” we are inclined to say something like “I am not sure. I guess not”. This could be well explained with the Bayesian model, where an individual’s belief does not have a truth-value of 0 or 1. The relevant belief is instead estimated in one’s subjective degree of that belief, as a probability between 0 and 1. Let us embed this in our current considerations. The utterance type of the politician is already embedded in a kind of bias or in posterior beliefs about politicians in general: this is called the hypothesis h, and has a particular probability P(h) in isolation from evidence e. Evidence e in this particular case is constituted by personal characteristics, properties, and circumstances of the utterance. The receiving of the new evidence e should update P(h), the probability of the hypothesis. Individuals ought, according to the normative stipulation, if they receive any evidence, to let it influence the probability of the proposition: “[this] normative procedure by which individuals should update their degree of belief in a hypothesis h upon receipt of an item of evidence e is given by Bayes’ Theorem:

”(Harris et al. 2012, p. 316)

P(h|e) describes the conditional probability of a hypothesis being true after one has received evidence e. P(e|h) terms the conditional probability of receiving evidence e, given hypothesis h. P(e) just describes the evidence in isolation from the truth-values of the hypothesis h (cf. Harris et al. 2012).