Confabulation and Rationality of Self-knowledge

Sophie Keeling is currently a philosophy PhD student at the University of Southampton. She primarily works on self-knowledge which has allowed her to research a range of topics in epistemology, philosophy of mind, and philosophy of psychology. Sophie’s thesis argues that we have a distinctive way of knowing why we have our attitudes and perform actions that observers lack. She gives a brief overview here.

This post summarises my paper ‘Confabulation and Rational Requirements for Self-Knowledge’ (forthcoming in Philosophical Psychology). The paper argues for a novel explanation of confabulation:

Confabulation is motivated by the desire to have fulfilled a rational obligation to knowledgeably explain our attitudes by reference to motivating reasons.

(Following others in the epistemological literature, I term the reason for which we hold an attitude our ‘motivating reason’ for it).

I shan’t seek to define confabulation here (a task in its own right) but instead note the subtype I’ll explain. I’m interested in cases whereby subjects falsely explain their attitudes (e.g. beliefs, desires, preferences) in response to prompting. We see a paradigm example of this in Nisbett and Wilson’s (1977) experiment in which they arranged four pairs of identical stockings on a table and asked individuals which they preferred and why. Subjects picked a pair generally towards the right of the table. Instead of noting the real cause of their preference – the position of the tights – or admitting ignorance, subjects gave incorrect explanations. That is, they confabulated an answer, such as the pair’s supposedly superior ‘knit, sheerness, and weave’. Indeed, this is a commonplace phenomenon. We’ve all at one point adopted a stance which we’ve rationalised after the fact. (E.g. I kid myself that I prefer the expensive branded yogurt over the supermarket offering because it’s tastier, and nothing to do with the clever marketing).

The paper then introduces three explananda for our explanation of this phenomenon, and argues that the two main options in the literature fail to account for all these. For example, confabulation is first-personal – we make these sorts of mistakes more readily with ourselves than others. (Here I draw on work such as Pronin et al. 2002 concerning the ‘bias blind spot’). Yet some accounts (e.g. Nisbett and Wilson 1977, Carruthers 2013, and Cassam 2014) struggle to address this important asymmetry in our mistaken self-ascriptions.

I propose an explanation which does account for all three explananda. It appeals to what I call the knowledgeable reasons explanation (KRE) obligation:

The obligation to knowledgeably self-ascribe motivating reasons when explaining one’s own attitude.

We shouldn’t confuse this rational obligation with moral ones. It just captures the thought that I ought to, for example, explain my belief that it will rain by citing a motivating reason, such as the weather forecast. That we bear the KRE obligation is independently plausible: I seem to be doing something irrational and criticisable if I instead answer the question ‘why?’ with ‘no reason’, ‘I don’t know’ or ‘I’m generally pessimistic’.

I use KRE in the following explanation:

We confabulate, and indeed confabulate with the content we do, because we desire to have fulfilled the KRE obligation (i.e. the obligation to knowledgeably explain our attitudes by reference to motivating reasons)

We can now explain the stockings experiment in the following way. The desire to have fulfilled the KRE obligation leads the subjects to confabulate an answer in the absence of a true one they can provide – they did not form their preference on the basis of reasons. And further, they specifically self-ascribe the reason that the stockings were sheerer, say, because it is a plausible motivating reason. This proposal accounts for the explananda in a non-ad-hoc way. For example, confabulation is first-personal because we desire to have fulfilled the obligation to knowledgeably explain our own attitudes by reference to motivating reasons, not other people’s.

The final section raises an upshot for understanding self-knowledge. Contrary to popular assumption, confabulation cases give us reason to think we have distinctive access to why we have our attitudes. What exactly our special access amounts to, though, must be left for further papers!