About Belief LO20088

Fri, 4 Dec 1998 11:17:53 -0800

Replying to LO19955 --

Richard Karash writes:
>> > 3. Some beliefs are more valid than others. That is, some beliefs will be
>> > more widely effective for more people than others. Therefore, I want to be
>> > conscious of making my own judgments of what beliefs are more valid.

{Still Richard}
> I was using the term as it is used in management sciences... A model or
> theory is valid when it is successful in application in the world. That
> is, when it shows power in explaining observations and when it shows
> predictive power. Note that validity is not black and white, models can be
> more valid and less valis; no model or theory is universally "true".

See my signature block below for a different way of putting it. When
talking about models, "truth" isn't really, well, ummm, useful.

(By the way, I've attributed the quote to a different source than the
similar quote used by Scott Simmerman a few messages ago; he named Deming
as the source. In a private exchange, we both admitted that we couldn't
personally verify the attributions, that we'd gotten them second hand. So
here's another puzzle: can anyone identify the source with sureness?)

> When I said, "effective for more people.." I think that confused the
> issue. I didn't mean "believed by more people"; I meant, "works more
> widely."

I.e., more useful.

> In the management sciences, we distinguish a concept called "face
> validity" which means "makes good sense" or something pretty close to
> "well grounded in it's internal logic". Face validity helps us believe in
> a model or theory but can be misleading. Face validity usually occurs when
> the theory proposes a generative mechanism that sounds reasonable.

I think this is what mathematicians would call "internal consistency" of a
model. Of course, a model can be consistent but not useful, as your
examples indicate.

It might be useful to offer some alternate definitions of "belief",
appropriately tagged. For example, belief(Karash) seems to be something
like "holding a model to be valid". Belief(deLange) is characterized, if
not defined, as a second order emergent of creativity. Belief(Eskow), as
used in his dialogue with At, seems to be similar to the sense that Eric
Hoffer gave it in his book "The True Believer". (Richard, At, Steve, if
I've done violence to what you intended, please correct me.)

In the spirit of Scott, I'll offer one more quote in closing: "a man has
to believe in something. I believe I'll go have a beer."


Don Dwiggins "All models are false, SEI Information Technology but some are useful" d.l.dwiggins@computer.org -- George Box, "Statistics for Experiments"

Learning-org -- Hosted by Rick Karash <rkarash@karash.com> Public Dialog on Learning Organizations -- <http://www.learning-org.com>