RE: uncertainty

[Clifford Poirot <cpoirot@xxxxxxxxxxx>]

I have a slightly different take:

Ontological uncertainty does indeed mean "non-ergodicity" in the sense that
Paul Davidson has used it. This means that the future is fundamentally
unknowable, no matter how much information you gather and how much you
refine the model. Future stock prices for example are unpredictable because
I cannot know how other actors will respond to information no one know as of
yet.

Epistemological uncertainty means that outcomes are "knowable" at some
level, but that the models and data we have are not sophisticated enough to
fully capture all the variables. For example, th National Weather Service
predicts 6-12 inches of snow in Columbus, and freezing rain for my town a
scant 90 miles south of Columbus today. A small change in wind direction,
the path of the storm, etc. (unknown to the National Weather Service) could
alter this prediction. Thus the model may predict wrongly and i may be
uncertain as to whether it will rain or snow, but with a better model, I
could be certain.

-----Original Message-----
From: Lawrence Boland [mailto:boland@xxxxxx]
Sent: Wednesday, December 13, 2000 1:33 PM
To: Forstater, Mathew; pkt@xxxxxxxxxxxxxxxx
Subject: Re: uncertainty


Without knowing the context where the distinction between epistemological
and
ontological uncertainty was used, I think one must be careful when defining
these
terms. The usual distinction raised is Knight's uncertainty vs risk. When
Keynes
(QJE 37) said we cannot know the price of copper in the future, he was
talking about
uncertainty (is this epistemological uncertainty?). To talk about risk
rather than
uncertainty one would focus on a probability distribution where we know the
parameters of the distribution but do not know what event will occur next
(e.g., a
winning lottery ticket). Risk thus could be what is meant by ontological
uncertainty
since the probability distribution exists and is known (hence not uncertain)
but
there is uncertainty about any single event in the distribution.

Helpful?

--
Lawrence A. Boland, Department of Economics
Simon Fraser University, Burnaby BC Canada V5A-1S6
ph: 604-291-4487, web: www.sfu.ca/~boland