2. THE ARGUMENTATIVE VIEW

Our ability to reason is manifested in both individual reflection and discursive argumentation. It is typically assumed that the former underpins the latter – that argumentation is merely the external expression of an individual cognizer's reasoning capacities. But this does not address the question of the ontogeny of our reasoning skills, a question that is not obviously to be answered in terms of any supposed gain in reliability that reasoning might afford. Reversing the traditional order of dependence, Mercier and Sperber argue that “[r]easoning has evolved and persisted mainly because it makes human communication more effective and advantageous” (Reference Mercier and Sperber2011: 60b). More specifically, according to the argumentative theory, “reasoning is a tool for epistemic vigilance, and for communication with vigilant addressees. Its main function is to enable communicators to produce arguments designed to convince others, and addressees to evaluate arguments so as to be convinced only when appropriate” (Sperber et al. Reference Sperber, Clement, Heintz, Mascaro, Mercier, Origgi and Wilson2010: 378). Here again is the thinking:

Just as an ability to detect certain of its features enables a creature to cope with its natural environment, linguistic communication enables a creature, not simply to pass on this modest sort of knowledge (though it might do that), but to cope with other humans – to pass on information with a view to shaping their minds and, thence, their behavior.

The potential benefits of receiving information from others, however, must be weighed against the costs of acquiring misinformation. Hence, there arises the need for hearers to evaluate the reliability of both individual speakers and the messages they convey. And among the most obvious ways of doing so is to attend to both the internal coherence of the message and to its external coherence with things already believed.

Meanwhile, from the point of view of the hearer, the next move in “the evaluation-persuasion arms race” (Sperber Reference Sperber2001: 410; Sperber and Mercier Reference Sperber, Mercier, Landemore and Elster2012) would be to anticipate such ‘coherence-checking’. In the case of things not likely to be taken simply on trust by the addressee,Footnote ⁴ she can provide an explicit presentation not merely of a given message, but of its internal and external coherence: she will “try to convince her addressee by offering premises the addressee already believes or is willing to accept on trust, and showing that, once these premises are accepted, it would be less coherent to reject the conclusion than to accept it” (Mercier and Sperber Reference Mercier and Sperber2011: 60b). A speaker's thus offering reasons for p, her advertising its coherence with believed or credible things, functions as a kind of ‘honest display’, and is liable to make the hearer more likely to believe what's conveyed, thus furthering the communicator's goal of persuading, and so manipulating, his audience. (So, in the beginning anyway, the principal benefit of having reasons is that we're able to give them.)

Through successive iterations, there arises out of this process skills – at times explicitly articulated, in the form of rules of logic and rhetoric, etc. – centering around the effective presentation and evaluation of arguments, skills which both speakers and hearers are then able to exploit in individual cognition. In this way, while it arises simply as a means of manipulating the beliefs and behavior of others (and avoiding being unduly manipulated oneself), the ability to reason can come to serve as an instrument for both effective interpersonal information-transmission and examining, regimenting, and extending one's own stock of knowledge.

3. NAGEL'S PROPOSAL

The argumentative theory offers a novel and intriguing way of thinking about the ontogeny of our ability to reason. But it also promises a new way of thinking about some familiar features of the traditional epistemological landscape. Among the things the argumentative theory might shed light on, according to Nagel, is “why we commonly think of perceptually and testimonially supported judgments as justified despite feeling worried, on reflection, that only what is internally available can justify”.

On the argumentative theory, recall, the impulse to offer reasons for one's beliefs is liable to kick in when a given claim is expected not to be taken simply on trust. And, Nagel suggests, it is the differential susceptibility of beliefs based on perception, testimony, inner sense, and inference, to meet with such resistance that explains why judgments grounded in perception and testimony can seem, on reflection, to be on shakier ground than those grounded in inner sense and inference. Importantly, this is not because of any difference, perceived or real, in the reliability of these types of judgments; rather, it is because some of them, by their nature, are going to be more susceptible to public scrutiny and social challenge.

Thus, Nagel writes:

This, again, is not to be confused with the claim that reasoning is particularly reliable: “on reflection, we can appreciate that explicit reasoning remains vulnerable to distorting factors such as confirmation bias, and to the standing risk that we have taken faulty premises on trust”.Footnote ⁵ The point rather, as the Mercier–Sperber theory predicts, is that “we are instinctively driven to accept the premise-conclusion patterns that others are instinctively driven to produce”.

Beliefs based on inner sense also have “a sheltered social status” – not, again, because they are hyper-reliable (arguably, they're not), but because their content largely shields them from corrective scrutiny, and because neither their content nor their accuracy is, in any case, something that generally holds much interest for others.

By contrast, judgments of perception do often matter to others, and do concern a public world readily observable by others. (Many beliefs based on testimony are like this as well.) So mistakes here are much more likely to be noticed, and to matter; because of this, those judgments are much more likely to meet with resistance – to not simply be taken on trust. Thus, even though we tend to take our perceptual beliefs to be unproblematically justified, “[o]ur natural attitude to justification shifts when we expect resistance from others”: knowing that others may well not simply accept some judgment based on perception or testimony, we may well feel the urge to find some argumentative basis for them. (“[M]uch reasoning is done in anticipation of situations where an opinion might have to be defended” (Mercier & Sperber Reference Mercier and Sperber2011: 68b).) And it may even seem to us, upon reflection, that absent supplementation by reasons such as might suffice to persuade others to accept them, those judgments are epistemically wanting.

4. SECOND THOUGHTS

As Nagel notes, an obvious objection to the foregoing line of thought is that “we can feel [the relevant] instability in our instincts about justification even without a live audience: reflecting in solitude, it can seem to me that I don't have immediate perceptual knowledge of the printed page in front of me, but only have an interpreted awareness of inner seemings that are compatible with the absence of the objects I take myself to see”. However, even reflection in solitude may not be an appropriate test. As Sperber says,

In any case, such concerns as I have about Nagel's discussion don't have to do either with the extent to which social factors can shape private reflection or with the argumentative theory itself. Rather, as indicated, they have to do with the account she provides of how exactly that theory might help explain the instability in our thinking about the justification of certain beliefs, and a natural asymmetry in our attitudes towards the basic sources of judgment described. The issue again is “why we commonly think of perceptually and testimonially supported judgments as justified despite feeling worried, on reflection, that only what is internally available can justify”. Nagel's suggestion is that it is the greater susceptibility to public challenge of perceptual and testimonial judgments that explains this. They lack the kind of “sheltered” or “special social status” that judgments of inner sense, and judgments supported by explicit argument, enjoy.

One wonders whether this is so, however. The point here is not the trivial one that “[a]lmost any judgment one makes may encounter resistance from an audience”. Rather, it is that it's just not clear that, in terms of susceptibility to challenge, there is in general a marked difference between judgments based on perception, for example, and those accompanied by argument.Footnote ⁶ It may well be true, as M&S say, that “much reasoning is done in anticipation of situations where an opinion might have to be defended” (Reference Mercier and Sperber2011: 68b); and it may be the (or, a primary) function of explicit, reasoned argument to make the acceptance of a given claim more likely. But both of these points – like the argumentative theory itself – are neutral as to the asymmetry in question. While perceptual judgments may sometimes meet, or be expected to meet, resistance from the addressee, many times they do not – often, they are taken on trust.Footnote ⁷ This is what one would expect, given that we naturally think of such judgments as justified. And it is because they are often taken on trust that such judgments can be deployed, as they are, as mutually accepted premises in arguments designed to persuade.

As to argument-supported judgments, as Nagel says,

However, it is a separate issue whether “we are instinctively driven to accept the premise-conclusion patterns that others are instinctively driven to produce” – or, at least, whether the instinctive drive to acceptance is any stronger here than in the case of expressions of judgments acquired via perception or testimony. In fact, the arguments others produce often don't conform to the ideal of obviously legitimate inference from mutually obvious premises. And even when they do, it's not guaranteed that the conclusion will be accepted – reasoning, after all, is not monotonic: “Realizing that our previous beliefs entail some implausible consequence we had not thought of before may give us reason to revise our beliefs rather than accept this consequence” (Sperber and Mercier Reference Sperber, Mercier, Landemore and Elster2012: 380; cf. Sperber et al. Reference Sperber, Clement, Heintz, Mascaro, Mercier, Origgi and Wilson2010: 374). So again, that the function of explicit argument is to increase the chance of message-acceptance does not ensure that it is often met. (It is the function of assertoric utterances generally, including assertions of judgments based on perception, to produce belief [Bach and Harnish 1978]; but that doesn't ensure that they often succeed.) Often the presentation of a given argument is just one step in the dialogical exchange, the back-and-forth of advertising and scrutinizing coherence, which may or may not end in the hearer's being persuaded of what the speaker says.Footnote ⁸

5. AN ALTERNATIVE PROPOSAL

For these reasons, I am not sure that it is the prospect of challenge from others – the relative ease with which we can expect to get various claims accepted on trust – that leads us to sort our basic sources into the familiar ‘easy’ and ‘hard’ cases. However, even if judgments backed by explicit argument do not have a particularly privileged social status, if the Mercier–Sperber theory is correct, argumentation certainly does. And this, I think, holds greater promise of shedding light on the asymmetry, and instability, in our epistemic judgments that's a familiar fixture in epistemological theory. Here, briefly sketched, are two lines of thought exploring this idea.

First, if the argumentative view is correct, then even though we naturally take beliefs based on perception and testimony to be justified, and even though reasoning holds no special promise of enhancing the accuracy of our judgments, argumentation becomes the public face of legitimate belief. The argumentative view – or rather, the practices that it describes and predicts – encourages thinking of legitimate belief in terms of supportability by reasons that any reasonable addressee would accept. (Instances of reports being accepted simply on trust are likely not to grab our attention: since the claims made are not disputed, and since they don't involve reflective engagement, they're likely to slide by unnoticed.) So, if the argumentative theory is correct, we should expect that people will find an argumentational view of justification (the property, not the activity), or well-founded belief, to be a natural one to take.Footnote ⁹ So too, it will be natural to move into thinking that “what justifies a belief [must be] somehow available to the subject” (quoted from Audi Reference Audi2001: 22). For, of course, factors beyond one's grasp or ken aren't fit to serve as reasons, and can't form part of a justifying argument (Bonjour Reference Bonjour1978).

In short, where argumentation is prized – not to mention, where it's the focus of our effort and attention, as reflective epistemic agents – it will be very natural to think about legitimate belief along internalist lines. Backing by accessible reasons becomes the coin of the realm.Footnote ¹⁰ And once we are thinking in such terms, judgments based on perception or testimony, unadorned by reasoned argument, can come to seem epistemically lacking. The point, then, is not so much that judgments backed by argument are epistemically easy, as it is that they're obviously in the game,Footnote ¹¹ whereas judgments based on perception or testimony (see note 6) aren't in the game at all.

A second, closely related line of thought is this. If, in our reflective moments, judgments based on perception or testimony can begin to look precarious, presenting such can also put us in a precarious position, and not just because our claims might not be taken on trust. Though it doesn't figure prominently in her discussion, one notion that Nagel mentions is accountability:

Accountability here is not a purely epistemic phenomenon – it's tied to punishment and rewards, reputation, responsibility, desert (see Audi, quoted in Nagel Reference Nagel2015), and so on, where these are not solely or purely epistemic.Footnote ¹² Because it places reasoning squarely in the social sphere, the argumentative theory brings epistemic matters into direct contact with social values and concerns such as these. And it is plausible that the general importance of such values and concerns encourages reflective worries that only what is internally available can justify, even when we have no specific reason to think that a given report will be challenged. For where we are concerned about such things as accountability, reputation, and responsibility, we're more apt to care about control. In the epistemic sphere, engaging in reflective processes and dialogical argumentation – having accessible reasons for the beliefs we hold, and for which we might be held to account – can appear to give us that. The materials of reflection and reasoning are propositions and states available to the subject “‘through consciousness or reflection’” alone (Audi, quoted in Nagel Reference Nagel2015). So engaging in such processes can seem to put one's epistemic fate (and so one's extra-epistemic reputation, etc.) in one's own hands: one need only carefully consider the relevant states and propositions, and one can ‘just see’ whether and why something is true or likely to be true, and so on.Footnote ¹³

Of course, it's another question whether such thoughts survive close scrutiny – whether, that is, the processes in question are any more reliable, and any more free of an element of trust, than their perceptual or testimonial counterparts. I'm among thoseFootnote ¹⁴ who think that they aren't. But the goal here is explanatory, not evaluative. And it should not be controversial that, given a concern with being in control of one's intellectual goings-on, a bare relianceFootnote ¹⁵ on perception or on another's testimony is apt to feel like an uncomfortable form of heteronomy. Insofar as it helps explain a concern for cognitive control, the argumentative theory once again sheds some light on the latter belief.

6. CONCLUSION

It is ironic, of course, that social considerations might help to explain some of the allure of having beliefs based on reasons, reasons to which one has unproblematic access. The latter is the internalist's stock-in-trade; and internalism is characterized by a movement away from having the epistemic status of one's beliefs depend on factors ‘external’ to oneself or one's first-personal point of view, including social factors. While I have departed from the details of Nagel's account of how the argumentative theory might encourage internalistic thinking, she is surely right that its doing so is one of the theory's interesting features.Footnote ¹⁶