In the last post I offered examples of confabulatory explanations as attempts to give reasons for attitudes whose source might escape introspection or be otherwise difficult or impossible to access. The interesting philosophical question for me is whether confabulation carries any epistemic benefit.
My next project concerns precisely that question and I have very little to say about how to answer it at this stage. But I want to mention one line of argument which I hope it will be worth pursuing as part of the project.
In a paper where I review the recent debate on the limitations of self-knowledge in philosophy and cognitive science, I describe an apparent conflict. Philosophers (e.g., Richard Moran in Authority and Estrangement) argue that providing what we take to be our best reasons for our attitudes contributes to a form of self-knowledge (authorship) that is distinctively first-personal.
However, psychologists (e.g., Tim Wilson in Strangers to Ourselves) maintain that in many circumstances providing reasons for attitudes impairs self-knowledge, due to our vulnerability to evidence manipulation, and, in general, to our failure to identify the “real reasons” for our attitudes (introspection effects). As a result, our attitudes may be fickle and change after an explanation is given, and other people – especially if trained in psychology – may have a better insight into our attitudes than ourselves (see an argument to this effect by philosopher Krista Lawlor).
My sense of the debate is that the two positions complement each another. Providing reasons for attitudes is not always an exercise in rationality and is hostage to standard reasoning mistakes and self-serving biases – the psychological data help us acknowledge that. At the same time, providing confabulatory reasons for our attitudes does not necessarily compromise our knowledge of those attitudes, and seems to have epistemically relevant benefits too – the notion of authorship highlights those.
Consider the Nisbett and Wilson study again where people choose between identical pairs of socks and they say they chose those which were brightest and softest, blind to the influence of positioning effects on their choice. One could say that, by giving a confabulated reason for a choice whose source is opaque to them, they become the authors of that choice in the sense that they know the choice is theirs, not based on behavioural evidence, but based on what they take to be their best reason for it (even if that reason does not map onto what originally caused their choice and happens not to track the truth).
Now, being able to tell why we choose a pair of socks is not important in the big scheme of things, but psychological evidence tells us that we also confabulate when we are asked why we support a certain political candidate or why we think that our current romantic relationship will last. Providing a reason for self-defining attitudes enables us to explore the connections of those attitudes with other attitudes we have, to integrate them in an overall narrative that shapes our self-conception, and to go back to those attitudes and change them on the basis of new evidence and further reasons.
Isn’t the self-conception we construct in this way just a convenient illusion? As Valerie Tiberius argues in The Reflective Life, some creative construction cannot be escaped when we think about ourselves in a reflective and systematic way, and it does not always lead to deception. My view is that we can exercise some control over our attitudes by giving reasons, even if we are not fully aware of how they came about.
Where standard limitations in introspection are combined with further deficits in perception, reasoning and memory, then mechanisms for integrating attitudes may become even more important in order for the person to hang on to some coherent sense of self.
