It may be that Keynes did not use the term "uncertainty" in the
Treatise on Probability in this context, but he certainly had the
concept and used the term later to relate to it. Thus he identified
four cases. One is the straightforward one that Knight would call
risk, and with Keynes not denying the frequentist element in it, even
if that is not how he saw probabilities, which he defined at that
time as being due to logical relations. Flipping a coin is the
canonical example. Then he had a case that might now be labeled as
one of a "mixed distribution," a combination of several probability
distributions, possibly with weights. This is linked to the matter
of the weight of belief as well. He then recognized where there may
in fact at the logical level exists some probability distribution,
but which observers will not be able to discern because of, well,
epistemological reasons, inability to gather or estimate data or
whatever. Finally then one reaches the case where there is no
probability distribution at all, which he later identified with the
term uncertainty. This analysis was actually carried out in his
undergraduate thesis many years earlier than 1921. Of course, he
used uncertainty in connection with investment, but when that is not
merely financial but real investment, it is closely linked to entrepreneurship.
It should be kept in mind that while Keynes is often labeled a
subjectivist and lumped in with Bayesians, he did clearly state the
objective existence of probability distributions, even if this were a
Platonic existence, in the TOP, again with the coin flipping case an
example. People often simply ignore or forget this. He is not so
easily folded into the Bayesian camp, although he clearly viewed pure
frequentism as having serious limits.
Ironically, given various people here asserting that Keynes was all
about epistemological problems, Paul Davidson has long argued just
the opposite, that for Keynes uncertainty was an ontological issue,
and that for people like Knight and many others it was merely
epistemological, that they were stuck in that third case of Keynes,
although that may be unfair to Knight.
Barkley Rosser
|