Keynesian uncertainty and the instability of beliefs

In an influential critique, Coddington (1982) argues that Keynes uses the theme of uncertainty in an opportunistic way in the General theory emphasizing it when it suits his purposes and ignoring it when it does not. This reading is examined in the light of Keynes's earlier Treatise on probability. The first part of the paper argues that Keynes's epistemology differs in important respects from the one Coddington is attacking, and in particular, that Keynes's account contains conceptions of uncertainty not considered by Coddington. The two premises of Coddington's argument are then assessed. These are: 1) that Keynes provides no account of the relative impact of uncertainty in different decision-making contexts; and 2) that uncertainty is not a sufficient condition for unstable beliefs about future outcomes of current (investment) decisions. It is argued that Coddington is mistaken in asserting 1) and, although 2) is correct in principle, that there is nevertheless a strong presumption that uncertainty does lead to unstable beliefs in the particular situations Keynes describes. The conclusion is that the connection between uncertainty and unstable beliefs cannot be regarded as either analytically destructive or somehow beside the point, even on Coddington's terms.