I agree with Jacob that some aspects of Grice’s theory of meaning are in a broad sense incompatible with Millikan’s account of communicative agency. The focus of this commentary so far has been communicative agency in epistemic contexts and its implications, and in particular the personal attack. I will now evaluate whether Millikan’s account of direct perception or Grice’s account of mindreading could account for ad hominem arguments in epistemic contexts. My answer is that the Gricean mindreading thesis is more plausible. I then compare the separability thesis with the cooperative devices. The separability thesis fits best with the practices of ad hominem fallacies. The presupposition of cooperative devices is less plausible.
Recall from section 1 that the mindreading thesis relies on the twofold account of informative and communicative intention. First, the speaker has to recognize or understand the speaker’s informative intention, which is the speaker’s communicative intention. For the fulfillment of the informative intention, the trustworthiness of the speaker has to be accepted. In other words, in order to fulfill the informative intention, the hearer commits neither a local nor a global ad hominem argument (or she would not accept it). But in order to be an epistemically vigilant agent, the hearer has to make some further inferences, which are inductive (as well as the ad hominem fallacies). This inductive inference model involves some kind of mindreading that could affect the reliability judgment.[15] Millikan claims that the acceptance of a given testimony as a source of knowledge is a form of direct perception without any kind of inference (Millikan 1984; Sperber et al. 2010). She talks about translation instead of inference. The hearer translates the utterance via direct perception into a new belief (Millikan 2004):
Forming a belief about where Johnny I on the basis of being told where he is I just as direct a process (and just as indirect) as forming a belief about where Johnny is on the basis of seeing him there. (Millikan 2004, p. 120) There is no reason to suppose that any of these ways of gaining the information that Johnny has come in requires that one perform inferences. (Millikan 2004, p. 125)
It is doubtable that these circumstances explain our everyday communicative agency, especially with regard to epistemic conversations. According to Millikan, the acceptance of a new belief does not involve any representation of the speaker’s intention. But in order to assess the reader as benevolent and competent (or reliable), one has to rely—as argued in section 5—on inductive inferences which are of course derived representations manifested in beliefs about the speaker’s intention.
In epistemic contexts of communication only the mind-to-world direction is involved, qua descriptive utterances. One criticism offered by Jacob is that Millikan’s account of perception could only account for descriptive utterances, hence only for the mind-to world direction. Another issue is closely related to this kind of criticism. It concerns testimony that has very little to do with perceptual capacities. With regard to very complex utterances like (1), I agree with Jacob (this collection, p. 9) that “it does not make much sense to assume that either the speaker or her addressee could perceive what the speaker’s utterance is about.” Consider the nature of testimonial reports. Even some direct perception of a testimony about some state of affairs is perceptually impoverished compared to directly perceiving the state of affairs in question. Imagine some testimonial reports that have been heard through the radio. In such a case, you are not in a perceptually close relationship to the reported state of affairs. If you evaluate the credibility of the speaker, it is very likely that you would run through different processes of inductive inference in order to commit an ad hominem argument or avoid one. The more abstract the testimony, the more implausible it becomes that it has anything to do with direct perception. It becomes even more complicated with complex indexical utterances or a group of different but equally eligible interpretations of a particular testimony. Consider again example (1). Here it is very likely that a hearer represents some intentions of the speaker that are linked to her psychological states. If one representation is that the speaker could deceive the hearer in particular circumstances, the hearer will probably commit a local ad hominem attack. To sum up: Ad hominem arguments are ascriptions that result from inductive inferences that also depend on belief-desire psychology, because the hearer gains a representation of the second-order representation of the sentence expressed by the speaker. The representation of the hearer is a third-order representation of the second-order linguistic representation of the speaker (cf. Jacob 1987).[16]
The problems addressed so far are closely related to the separability thesis. The separability thesis is the claim that the hearer and the speaker could have different interests, which are causes of the informative intention remaining unfulfilled, because there are two cases that suggest that the interests of both parties fall apart. In the first, the hearer gains a new belief that is not true, because the speaker has the informative intention to deceive the hearer. So her informative intention has the aim that the speaker gains a false belief and not one about some states of affairs, as described in section 2. In the second, the sentence, uttered by the speaker, is true, but nonetheless denied by the hearer on the basis of an ad hominem argument (Sperber & Wilson 1986).[17] These two cases do not support Millikan’s (2005) claim that the interests of both speaker and hearer are balanced. If a hearer commits a global ad hominem argument, it is even harder to ascribe balanced interests to speaker and hearer. Sperber (2001) defends a plausible weak version of coincidence of interests. It is only necessary that they overlap in the long term in order to establish successful practices of social knowledge transmission.
Cases of global ad hominem arguments could only occur if a speaker understands the informative intention which she combines with some particular personal properties of the speaker in order to reject the testimony in question. Hence the speaker succeeds in establishing the communicative intention, but fails to fulfill the informative intention. In Millikanean terms, the direct proper function of the speaker is that the hearer gains a new belief. If the hearer commits an ad hominem attack, the direct proper function remains unfulfilled. But the communicative intention is still fulfilled, and that is all that is required for successful communication according to the separability thesis of communicative intentions (Sperber & Wilson 1986). The hearer recognizes that the speaker wants to inform her of her informative intention, which means that the communicative intention has been fulfilled. But the informative intention—which is that the hearer gains some new information or a true belief—fails, because an ad hominem attack has been committed. This circumstance could be well explained with the separability thesis and the weak account of communication that we addressed in section 6.1. To sum up, and in agreement with Jacob, if an ad hominem attack has been committed, even a weaker version, communicative agency is violated in the Millikan (2004, 2005) framework because the cooperative conventional transmission is violated in the first place.[18]
The picture I draw with regard to the ad hominem arguments rests on the assumption that trustworthiness has to be assessed by the hearer in order to be counted as epistemic vigilant, which would be the reductionist component. On the other side, the speaker has to utter a true sentence to transmit knowledge to a speaker, which would be the anti-reductionist-component. This view is supported by Lackey’s dualism. As can be seen, the establishment of a dualistic account and all its implications for the inductive reasoning model can be better explained with the separability thesis.