We all know that we do things against our own better judgment: snap at friends, overeat or overdiet, or ignore climate change. Do we also believe things against our own better judgment? Do we believe that we are fat or unliked, that black cats bring bad luck, or that other minds exist, while believing we shouldn’t believe it?
Some philosophers think such ‘akratic’ beliefs are impossible; as Susan Hurley puts it, “the unavailability of the akratic structure is… constitutive of belief.” Many debate whether such beliefs can be rational, while assuming that they are possible. I think akratic belief is both possible and widespread, but I think it takes argument to establish this possibility. Akratic belief is puzzling to many people in a way akratic action is not.
In an earlier paper, “Moore’s Paradox and Akratic Belief” (Philosophy and Phenomenological Research, May 2016), I considered parallels between akratic belief and Moorean beliefs such as ‘I believe it’s raining, but it isn’t’. I argued that appeals to Moore’s Paradox in denying the possibility of akratic belief offer only an illustration of an underlying puzzlement about akratic belief, rather than an independently compelling argument.
In “How Can Belief be Akratic?” (Synthese, 2021, available here), I explain and address that puzzlement. Rather than appealing to a theory of belief, as some discussions do, I argue that akratic belief should, like akratic action, be treated as a pre-theoretical datum: a recognizable phenomenon that theories of belief should be able to accommodate.
I distinguish four ways of arguing for this view: intuitive arguments from plausible examples; defensive arguments that respond to arguments against the possibility; systematic arguments that appeal to more general considerations about belief; and diagnostic arguments that explain why akratic belief might seem puzzling and even impossible. I think these are stronger together, and I offer an argument that combines all four.
Its crux is an Argument from Belief Attribution, which looks to typical marks of belief such as sensitivity to evidence, recall in relevant circumstances, felt conviction, reporting or assertion, and use in further reasoning. An anorexic in treatment might insist that he is fat, while acknowledging that he should not believe it, given the overwhelming evidence of his malnutrition. In some cases, I argue, both component beliefs in an akratic state manifest these marks to an extent we can recognize as belief, while nevertheless conflicting with and partly undermining each other.
Akratic believers are still puzzling in several ways. They do not have a single, unified point of view, or at least not one that makes sense. They are difficult to understand and frustrating to interact with. What is the point of trying to convince them they should not believe they are fat, if they already agree? But like climate change denial, this puzzling phenomenon is nevertheless real.
Recognizing these cases leads to a more lifelike picture of ordinary cognition, more compassionate and resolute interaction with akratic believers, and an increased ability to recognize one’s own akrasia. And it prevents us from drawing a misleading disanalogy between theoretical and practical reasoning. In both, the conclusions we believe we should reach can differ starkly from the ones we actually come to.