ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] {Disarmed} Reality and Truth

To: patrick@xxxxxxxxxxx
Cc: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Mon, 07 May 2007 22:35:54 -0400
Message-id: <463FE20A.1080402@xxxxxxxxxxx>
Patrick,    (01)

The methods of testing a theory are a lot like debugging
a program.    (02)

 > Why isn't it sufficient to say that a theory makes correct
 > predictions in every case we have encountered? Isn't that enough?    (03)

If all the tests keep repeating a narrow range of possible cases,
they confirm the theory for that range of cases.  But it's better
to test a broad sampling of cases, especially extreme cases.    (04)

 > I am not sure what more you could ask of any theory, that is
 > that for every case encountered it made the correct prediction.    (05)

There's much more that one could ask for.  An unusual theory that
has no known relationship to other theories is not as convincing as
one that fills gaps that interconnect previously known special cases.    (06)

Another desirable feature is unexpected (and later confirmed)
predictions that are far outside the range of anything previously
encountered.    (07)

For example, Maxwell's famous equations for electromagnetism
related a wide range of previously known phenomena in four very
short equations.  Furthermore, it made predictions about previously
unknown wave-like phenomena, which were later discovered by Hertz.    (08)

Even more surprisingly, Maxwell's theory explained unrelated
phenomena, such as light, which nobody had expected to be
electromagnetic in nature.  Over the years, more and more
phenomena came under the electromagnetic umbrella, which gave
it more and more credibility.    (09)

 > First, "every conceivable input" isn't really possible. Testing
 > is always with a finite set of inputs.    (010)

For some systems, the number of possibilities is finite.  Even for
cases that have an infinite range of inputs -- say all voltages
from 0 to 100 -- a random sample that gives decent coverage of that
range would be sufficient (unless some unusual values -- such as a
sharp spike in output -- occurred at some point).    (011)

 > What claim is it, beyond accuracy (agreement of theory with
 > observations), that you want to make by saying something is
 > "part of the truth?"    (012)

If every prediction about a system X made by a theory Y turns out
to be true, that would imply that Y is at least a partially true
description of X.  And if Y is partially true about X, it would be
part of the truth about X.  (But there could be other unknown
aspects of X that might never be discovered.)    (013)

That isn't deep philosophy.  It's just English grammar.    (014)

John    (015)

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (016)

<Prev in Thread] Current Thread [Next in Thread>