Re: How to tell if a theory is a good one

From: Paul Draper (pdraper_at_yahoo.com)
Date: 10/05/04 

Date: 5 Oct 2004 10:01:19 -0700

"Ilja Schmelzer" <q6867901@mailstore.fernuni-hagen.de> wrote in message news:<cjrb38$i48$1@beech.fernuni-hagen.de>...
> "Paul Draper" <pdraper@yahoo.com> schrieb
> > But this isn't enough. There may be dozens of completely equivalent models
> > that adequately describe the same known phenomena, both qualitatively and
> > quantitatively. If this is all a theory does -- match evenly against an
> > existing model -- then it is no good. Here again is where many
> "alternative
> > interpretations" fall flat.
> >
> > 5. A good theory has to *quantitatively predict* an observable behavior
> that
> > has not yet been measured, and preferably a behavior that is at odds with
> > the predicted behavior of other proposed theories. If it fails to predict
> > anything new, then it must be dismissed as a theory, no matter how well 1,
> > 2, 3, and 4 are satisfied.
>
> False. In this case, SR would have been a worthless theory in
> comparison with the older Lorentz ether.
>
> It may be reasonable to use "interpretation"
> instead of "theory" to name the child. But that's all.
>
> Historical preference does not make the old interpretation
> better than the new one.

This is a worthwhile point, because the LET theory survives quite well
in a shoot-out with SR. Indeed, as Tom Roberts reported in an
excellent post that is often quoted, SR and LET predict exactly the
same phenomena *acting as isolated theories* (which I'll explain in a
minute), and as Roberts said, most of the arguments for SR being
favored over LET have to do with mathematical elegance or, perhaps
even more compellingly, because of the symmetry that it assigns to the
universe that LET does not.

However, in years since, SR has grown outside the domain of its own
predictions and is now incorporated as an underlying principle in
other theories, chiefly quantum field theory. The success of QFT is
then an indirect but just as compelling validation of SR as the direct
predictions are. In fact, I don't know that an attempt has been made
to rewrite a testable variant of QFT that incorporates LET rather than
SR.

In other words, sometimes it takes an application of two theories in
an even shoot-out to determine which is the winner.

>
> > In this sense, most of the string models and
> > spin-network models of the universe are NOT good theories yet, because
> they
> > fail to make a verifiable prediction. In this sense, even the Standard
> Model
> > with the Higgs boson is not a very good theory yet, because one of its
> > primary verifiable predictions (the Higgs boson) has not been confirmed
> yet.
>
> These two examples are also sufficient to prove you wrong.
> I don't give a penny for string theory, and I don't believe into
> the Higgs, but these are certainly physical theories.

Not in my mind. Certainly not the string model (note my term), which
has not yet generated a testable prediction that would distinguish it
from competing model. The "Standard Model" theory is an acceptable
theory insofar as the electroweak and QCD sectors have been highly,
highly successful, but the Higgs is such a key element of the Standard
Model that I would still hold it a bit askance until the Higgs is
found.

>
> > Alternative-theory-proposers: the onus is on you to make sure your model
> > satisfies ALL FIVE of the above requirements. If it does not, then the
> world
> > has a complete right to summarily dismiss your conjecture.
>
> It is nice to propose criteria for alternative theories which are
> not fulfilled by the leading (in terms of number of publications
> and man-years spend on it in professional theoretical physics)
> theory. That's named double standard.
>
> BTW, my theory (see gr-qc/0205035) fulfills all your requirements.
>
> Ilja

You are certainly right about the lemmings-style history of physics.
But that's politics, not objective science. It's a pity the two get
mixed. I have run into that kind of disappointment myself now and
again. Nevertheless, the fact that effort and attention are often
misplaced does not diminish the value of an established standard by
which the merits of a theory can be judged.

PD

Relevant Pages

  • Re: How to tell if a theory is a good one
    ... >> that adequately describe the same known phenomena, ... predictions are. ... In this sense, even the Standard ... > the Higgs, but these are certainly physical theories. ...
    (sci.physics.relativity)
  • Re: How to tell if a theory is a good one
    ... >> that adequately describe the same known phenomena, ... predictions are. ... In this sense, even the Standard ... > the Higgs, but these are certainly physical theories. ...
    (sci.physics)
  • Re: A whopping 50 percent... ???
    ... >> future and she is listened to by the wall stree casino analysts who are ... >> predictions often enough, they will actually happen. ... Unfortunately, too many people see as "industry standard" that which is, ...
    (comp.os.vms)
  • Re: How to tell if a theory is a good one
    ... >> Historical preference does not make the old interpretation ... of this theoretical type is certainly justified as part of science. ... but the Higgs is such a key element of the Standard ...
    (sci.physics.particle)
  • Re: How to tell if a theory is a good one
    ... >> Historical preference does not make the old interpretation ... of this theoretical type is certainly justified as part of science. ... but the Higgs is such a key element of the Standard ...
    (sci.physics)