| Summary: | <p>Bayesian epistemology provides a promising framework for a theory of epistemic rationality. But the way in which this framework has been built upon thus far yields an unfortunately mechanical picture of rationality, on which rational agents are mere data crunchers who receive evidential input and spit out numeric credal output. This picture is rightly criticized, most prominently by Bas van Fraassen, for being too narrow and restrictive and thus failing to account for certain features which rationality plausibly has, such as a degree of permissiveness, and for certain unconventional rational phenomena, such as conversions. Unfortunately, van Fraassen’s apt criticism of mechanistic rationality overshoots its mark in seeking to topple the entire Bayesian framework. Bayesian epistemology suffers a guilt by association with the robotic picture. This dissertation aims to restore Bayesianism from the mechanistic but often implicit assumptions which corrode it, and to rebuild, from the Bayesian foundation, an alternative picture of rationality as a property of sentient agents who are capable of understanding and mentally engaging with the objects of their credences. Along the way I account for some basic Bayesian objects such as credence and evidential input. I also accord a central role to the ability of representational experiences, largely sidelined in many Bayesian discussions, to give rise to surprising evidence. On these building blocks I develop theory of rationality, Expansive Bayesianism, which evades the criticisms launched at the robotic picture and shows that Bayesianism itself is a fruitful and powerful framework for a theory of rationality.</p>
|
|---|