By Tillman Vierkant, University of Edinburgh
(See all the posts in this series here.)
I am very grateful to Gloria Andrada, Paulius Rimkevicius and Ting Huang for their very insightful comments. I hope that my replies will help to clarify some of the issues raised and stimulate further discussion.
Let me begin by addressing Andrada’s very important worry that my distinction between direct and indirect components in belief acquisition processes is not as clear as it could be and that it is also not clear why it is as central as I make it out to be. Perhaps, as she writes, all that this distinction marks is a phenomenological difference. It feels to us as if belief acquisition events happen to us automatically, but does this mean there really is a genuine difference in process?
Of course, I agree that it would be bad news for my account if the difference were really only phenomenological, but I think there are a number of good reasons to suspect that this is not the case.
The standard reason in the literature for why the acquisition of a specific belief cannot be under the intentional control of the agent is that the belief is acquired when a specific evidence threshold is crossed, and it is not up to the agent whether a specific bit of evidence will achieve that. As far as I can see, there is a reasonably broad consensus about this fact, so I do not take myself to argue for a radical claim here. In addition, as far as I can see most of the cases that seem to undermine the idea that we never intentionally acquire a specific belief are not a problem for my account. This is because I am particularly interested in rational belief acquisition and all of the cases that I know of where one could argue that belief acquisition does not follow the evidence are cases where the belief acquisition would not happen for first order rational reasons. Suppose for example that Pascal manages to condition himself to believe in God. In that case, he acquires the belief by intentional means, and he might well have metacognitive reasons to acquire it. However, on the first order level the process is clearly not driven by rational considerations.
The last step in belief formation in rational deliberation seems to be uncontroversially non voluntary. However – and this brings us to Andrada’s second point – the more important question is how significant this event is. Andrada suspects that despite my protestations about being sympathetic to the extended mind, my direct/indirect distinction brings in a form of internalist core cognition. There is a kernel of truth here. I admit that I can see how someone with internalist sympathies might use the distinction this way, but I don’t see that it is necessary or even more plausible than externalist alternatives. This is because even avowed doxastic voluntarists like Robert Lockie (2018) are happy to accept the distinction. They simply argue that the last involuntary direct step is given undue significance by people like Strawson. Accepting that this last step exists, and that it is direct, does not mean that one has to accept that it is the most important step. At the same time, even probably the main proponent of extended cognition Andy Clark accepts that the brain might well occupy a privileged role in selecting cognitive strategy with his hypothesis of organism-centred cognition (HOC) (2007). I don’t think my account is more internalist than HOC.
This takes us to Andrada’s last worry that my account might revive the idea of a stoic passive thinker, especially in the context of willpower, the discussion of which takes up a substantial part of the second half of the book.
But the worry clearly does not hold, at least not if by willpower we mean something like voluntary epistemic agency. In fact, the very point of the chapter (chapter 6) on tying to the mast strategies is that it is impossible to think of such strategies as less direct than any skull internal intentional exercises of willpower. If one were worried about the indirectness of the former, then one has to accept the implication that there simply is no intentional willpower at all. In that case, all willpower ever could be would be a form of evaluative control. To me this seems implausible, because intentional epistemic agency clearly plays a crucial role in achieving self-control aims.
Talking about willpower also takes us straight to the comments by Rimkevicius and Huang.
Rimkevicius has the opposite worry to Andrada on willpower. He argues that my account might obscure that there still is a form of internal “real willpower”. This is he argues because internal “toughing it out” is in some ways really superior. In contrast to external strategies, it is not dependant on specific environmental conditions and therefore always available. Internal strategies also allow one to be genuinely open to considerations that might make what looks like temptation into the genuinely better choice.
I am very sympathetic to the point that internal strategies can in some respects be different to most external strategies, but the important point to me is that this is only insofar as the ties employed in such strategies (like rehearsals or distractions) are still strategies. And this brings me to the point where there might be some genuine disagreement. I admit that these internal strategies also might leave the agent more flexible to take new reasons into account, but this is only because they are more flimsy ties. A distraction does not have the same binding force as Odysseus’ ropes. What however does not change is that on my account all intentional strategies are not a genuine rational weighing of the evidence, because that would go against the core argument of the account. In this respect, all tying to the mast strategies from the flimsiest to the most durable are made of the same stuff: intentional not evaluative epistemic agency.
Apart from the point on willpower, Rimkevicius has comments on my accounts of choice and freedom. Dealing with the choice part is very easy. Rimkevicius explains that I describe choices as either intentional or evaluative epistemic agency but that there is a third alternative, according to which both elements play a role. In short, I would agree with that. My point was not to describe my options as mutually exclusive, and in everyday use I fully suspect that we use the term in ways which describes both.
This brings us finally to freedom, where both Rimkevicius and Huang suspect that on my tinkering account we have less freedom than we might have previously thought. Huang rightly insists on an important clarification of my account. It is easy to misunderstand my position as one that aims to give humans back full voluntary control over our thinking by means of managerial control. That is emphatically not my intention. As Ting Huang rightly points out, the intentions that drive managerial actions are formed involuntarily. The only way that agency would not peter out in a passive intention formation process is if we assume something like evaluative agency. Nevertheless, that does not mean that the ability to intentionally manipulate the mind might not give us freedom in some sense. It is a very powerful trick in the cognitive toolbox and allows morally relevant forms of self-control that would otherwise be impossible. I discuss this in detail in chapter eight of the Tinkering Mind.
Sure, there is no (libertarian I would add) freedom in the loop as Huang states, but tinkering does explain at least in part why humans have a far greater control over their mentality than any other creature we know, and it does not seem farfetched to me that this might well explain at least in part why practices of responsibility ascription make much more sense in humans than in animals. In the same vein, when Rimkevicius claims that we have less freedom because we thought “mere strategy” is less free and responsible, I want to answer that this is only true in the same sense that Dan Dennett once quipped about Wegner’s claim that conscious will is an illusion (Dennett, 2003). Dennett compares conscious will to real love and Wegner to a scientist who pronounces that real love is not real, because we know now that there is no Cupid shooting arrows to set our hearts alight. In the same vein if the freedom we thought we had never made any sense in the first place, then it seems much more sensible to say that we have not discovered that the tinkerer is not really free but rather that we finally understand what freedom really could mean.
References:
Clark, A. (2007). Curing cognitive hiccups: A defense of the extended mind. Journal of Philosophy, 106, 163–192.
Dennett, D. C. (2004). Freedom evolves. Penguin UK.
Lockie, R. (2018). Free will and epistemology: A defence of the transcendental argument for freedom. Bloomsbury Publishing.