(sorry if this is a repeat)
A robust theory would then be one that is accessible by many
explanations, unifying them by showing how they could make equivalent
paths through an heuristic. It would serve to maintain open questions by
allowing them to be more local. A theory with only one explanation would
be a crappy theory; mistakes would propagate more globally instead of
getting metabolized more locally.
I liked Phil's second question, which I take to lead more towards using
models to make sense of the present, rather than to "predict" ;
The second question is more about individual complex systems in a
particular circumstance requiring one to start from the limited
information that raises the question and to go to the trouble of
expanding your understanding while exploring possible patterns in the
environment until you find one that seems to fit.
Another one would be "who's environment?", which I think leads one back
to ontology formation/niche construction. Is it not so much that
prediction is "bad" but rather that it is quaint for the types of
questions we want/need to ask?
Carl
Phil Henshaw wrote:
Why prediction fails does not seem to be just believing your own
script.. as it were. I’m suggesting that “a theory of some sort” is
generally the same thing as “a statement of what generally has
happened”. The real question may be sort of the opposite of “but who
would believe such a thing?!!!!!” since believing in a theory with
little or no way of checking how what is actually happening is
different from it, seems to be nearly everyone’s preference. To do the
latter you need to maintain the open questions of your induction, and
not cast them off as soon as you have made something useful with it.
So I don’t think it’s a “fallacy of induction” per se. I think it’s
more just that any handy tool can be greatly misused if you don’t keep
asking “how does it apply here”. There is also an all too common
preference for absolutist rules that contributes to our dodging any
information about how they might not quite apply too… but I guess
that’s not just a matter of clumsiness.
So, is that saying “it is so” or “it isn’t so”, I’m confused… ;-)
Phil Henshaw
*From:* Nicholas Thompson [mailto:[EMAIL PROTECTED]
*Sent:* Tuesday, December 02, 2008 12:32 PM
*To:* Phil Henshaw; The Friday Morning Applied Complexity Coffee Group
*Cc:* [EMAIL PROTECTED]; [EMAIL PROTECTED]
*Subject:* RE: "what generally happens here"
Ah, Phil. If you are correct that the answer to "what generally
happens here?" is regarded by some as an "explanation", then the
source of the confusion underlying this conversation becomes
immediately evident.
But who would believe such a silly thing?!!!!! "What generally happens
here" is just a summary statement of past experience One cannot
rationally pass from such a summary of past events to any statement
about the future without a theory of some sort that models the world
as the sort of place where "what generally has happened"" is what
happens in the future.
Could it be the case that all this talk about the evils of prediction
has occured because Epstein and a few others woke up yesterday to the
Fallacy of Induction?!!!! Oh, my. Say it isnt so!
N
Nicholas S. Thompson
Emeritus Professor of Psychology and Ethology,
Clark University ([EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>)
----- Original Message -----
*From:* Phil Henshaw <mailto:[EMAIL PROTECTED]>
*To: [EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>;The Friday Morning Applied
Complexity Coffee Group <mailto:[email protected]>
*Sent:* 12/2/2008 8:28:03 AM
*Subject:* RE: [FRIAM] Wedtech to Friam: earthquakes
There’s always the difference between the kind of question you ask
and the type of prediction and explanation for it. For example,
you might ask either “what generally happens here” or “what is
happening here”. The first asks for a simple explanation and a
rule of thumb type prediction. It might be helpful for responding
to the second question, or not. The second question is more about
individual complex systems in a particular circumstance requiring
one to start from the limited information that raises the question
and to go to the trouble of expanding your understanding while
exploring possible patterns in the environment until you find one
that seems to fit.
I think there are lots of differences between any kind of
explanatory causation and the instrumental causes. Maybe
explanations become useless if they try to include all the
complexity of the instrumental processes, but also often loose
their value by ignoring the underlying complexity too.
Re: earth quakes, I went to a lecture at Columbia recently that
was just great on the physics of ‘slow slips’ in a shearing crust,
large horizontal zones of gradual internal tearing within the
crust, having leading vibration events and propagation fronts, etc.
Phil Henshaw
*From:* [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] *On Behalf Of *Nicholas Thompson
*Sent:* Monday, December 01, 2008 3:04 PM
*To:* [email protected]
*Subject:* [FRIAM] Wedtech to Friam: earthquakes
Dear All,
We have been having a discussion on a SF Site called Wedtech about
the relationship between explanation, simulation, and prediction.
If you want to get a sense of the starting point of that
discussion, have a look at Josh Epstein's forum entry in the
current JASSS, which seems to be just about as wrong headed as a
piece of writing can be. In it, he makes a radical separation
between prediction and explanation, implying that the quality,
accuracy, scope, and precision of predictions that arise from an
explanation is no measure of that explanation's value.
In the course of trying to discover where such a silly idea might
have come from, I was led to literatures in economics and
geophysics where, indeed, the word "prediction" has taken on a
negative tone. These seem to be both fields in which the need for
knowledge about the future has overwhelmed people's need to
understand the phenomenon, so that predictive activities have way
outrun theory.
However, acknowledging the problem in these literatures is not the
same thing as making a principled claim that prediction has
nothing to do with explanation.
In the course of thinking about these matters, I have stumbled on
an extraordinary website packed with simulations done by people at
the USGS in Menlo Park California. the page is
http://quake.usgs.gov/research/deformation/modeling/animations/. I
commend to you particularly, the simulations done on teh Anatolian
Fault in Turkey (BELOW the stuff on california) and ask you to
ponder whether the mix of simulatoin, explanation, and predicition
is appropriate here. I suggest you start at the top of the
Anatolian series and move from simulation to simulation using the
link provided at the bottom right of each simulation. Stress
buildup and stress release are represented by red and blue colors
respectively and the theory is one of stress propogation. I would
love to know where the colors come from i.e., how stress is
measured. If there is no independent measure of stress, then, as
in psychology, the notion of stress is just covert adhockery.
Please let me (us) know what you think.
Nick
Nicholas S. Thompson
Emeritus Professor of Psychology and Ethology,
Clark University ([EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>)
------------------------------------------------------------------------
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
