Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site watmath.UUCP
Path: utzoo!watmath!csc
From: csc@watmath.UUCP (Computer Sci Club)
Newsgroups: net.physics
Subject: Re: But why?  beCAUSE
Message-ID: <7654@watmath.UUCP>
Date: Wed, 2-May-84 17:38:42 EDT
Article-I.D.: watmath.7654
Posted: Wed May  2 17:38:42 1984
Date-Received: Thu, 3-May-84 19:22:52 EDT
References: <7451@decwrl.UUCP>
Organization: U of Waterloo, Ontario
Lines: 38

Just a couple of points.

1)  I very much like the idea of considering physics to be the construction
    of models to predict experimental results.  The success of a model to
    be judged by its applicability and accuracy.  Questions as to the
    "reality" of such models are interesting and possibly important but
    should be considered outside of Physics.  Certainy no one can survive
    without some kind of "metaphysic" (how does one define applicability
    and accuracy otherwise) but many such "metaphysics" can be consistant
    with the practice of physics as defined above.  Naturaly p implies q;
    q is obseved; therefore p; is a falacy.  However p implies q; q is observed;
    therefore p is a good model; is perfectly valid.  If r also implies q
    then r is also a good model.  If p and r agree on all predictions of
    experiments then questions as to the difference between p and r, and which 
    is more valid lie outside of physics.  This, I understand from limited 
    reading in the subject is the essence of the "Copenhagen" viewpoint
    championed by Bohr. (If I am wrong I am sure someone will correct me, this
    is the net  after all :-) )

2)  It is true that if we make a finite number of observations of a random
    process we cannot determine the underlying probability distribution.
    (Although sampling thoery says that we can make estimations which are
    very likely to be very close).  But we can produce a model which
    says that is we run experiment x, N times, we should get result x(1)
    n(1) times; x(2) n(2) ...  Then proceed to carry out the experiment and
    see if the model is any good.  Questions as to whether the outcome is
    "really" random are only meaningful within physics if phrased: do there
    exist deterministic models which also correctly predict experimental
    results.  (Experiments concerning the famous Bell inequality appear to
    show that a certain class of deterministic models do not correctly predict
    experimental results).  Questions as to what is "really" going on may
    be important to philosophers, and physisists as human beings but should
    not concern Physics.

                                                  William Hughes