Re: Attempt at a Schema

["Goncalo Fonseca" <fonseca@xxxxxxxxxxxxxxxxxx>]

Ted Winslow writes:

> The objectivity of probability to which the Treatise points derives from
the
> foundational assumption made there that there are directly perceivable
> objective "logical probability relations".  The Treatise, therefore, does
> allow for objective probabilities though not for "objective randomness".
> Keynes, however, subsequently explicitly abandoned the foundations of his
> Treatise argument.
>

With this I am 100% in agreement.   It is for the danger of ascribing
objectivity to randomness that I would prefer not to use the term "objective
probabilities" but rather "probabilities based on (objective) formal logic
applied to (personal) information".  If this is also what Barkley was
getting at (but I'll let him speak for himself), then perhaps we are seeing
eye-to-eye after all.

> In his 1931 review of Ramsey's *Foundations of Mathematics* he explicitly
> abandons the idea of logical probability relations.  He substitutes for it
> the idea that rational degrees of belief are "useful mental habits" having
> their basis in something analogous to perception and memory rather than in
> formal logic and adopts Ramsey's term - "human logic" - to designate the
> field of study concerned with the analysis of these habits.  He rejects,
> however, Ramsey's attempt to elaborate the idea in terms of Charles
Peirce's
> pragmatism.

But does he accept Ramsey's theory under any random circumstances
or only in circumstances where his own theory doesn't apply (i.e.
"uncertainty")?

> The description of "what men actually believe" differed from this; it was
> "descriptive psychology" rather than human logic. Ramsey proposed a system
> based on betting for doing this.  The interpretive literature on Keynes
(and
> Ramsey) frequently ignores the fact that this system is a method of
> "descriptive psychology" rather than of human logic.  The mistaken
> implication is then drawn that human logic designates a subjective rather
> than an objective approach to probability.
>

Again, 100% agreement.  For Ramsey, if a die rolls alone in the middle of a
forest and nobody bets on it (even if many people see it), then there is no
probability.

Thank you for expanding the passage Barkley quoted.  It seems, in my
reading, to reinforce my argument that Keynes never really let go of his
1921 theory and he still thought it was the (only?) "scientific" approach to
probability assessment, all else being largely erroneous shortcuts.  It
strikes me as a bit strange, however, that he was so intent on placing forth
a theory of probability which he admits, at the same time, is practically
useless -- or rather, is not the way people "really" go about forming
probability assessments.  It would be interesting to see how he would react
to modern models of "conventions".

If Barkley does not raise further objections, then I think we are mostly in
agreement.

On a final interesting note, there is a Levy working paper by Hyman Minsky
(from 1996) online which examines the parallels between Keynes's and
Sargents's uncertainty theories:

Hyman Minsky (April, 1996) "Uncertainty and the Institutional Structure of
Capitalist Economies", http://www.levy.org/docs/wrkpap/pdf/155.pdf

It might of particular interest to some people on this list.


Goncalo