From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!psych.toronto.edu!michael Thu Feb 20 15:21:16 EST 1992
Article 3785 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!psych.toronto.edu!michael
>From: michael@psych.toronto.edu (Michael Gemar)
Subject: Re: MUST Philosopy be a Waste of Time?
Message-ID: <1992Feb16.194200.13547@psych.toronto.edu>
Organization: Department of Psychology, University of Toronto
References: <409@tdatirv.UUCP> <1992Feb13.195506.23701@psych.toronto.edu> <416@tdatirv.UUCP>
Date: Sun, 16 Feb 1992 19:42:00 GMT

Stanley and I obviously have different views on this subject which may
very well be irreconcilable, so I'll try one more time, and suggest that
further discussion be taken to email.

In article <416@tdatirv.UUCP> sarima@tdatirv.UUCP (Stanley Friesen) writes:
>In article <1992Feb13.195506.23701@psych.toronto.edu> michael@psych.toronto.edu (Michael Gemar) writes:
>|In article <409@tdatirv.UUCP> sarima@tdatirv.UUCP (Stanley Friesen) writes:
>|>I would consider those as being more in the way of mathematics than pure
>|>philosophy.
>|
>|Huh?  Logic is not philosophy?!  Yow, stop the presses!!!!
>
>Certainly, logic is a *tool* *used* by philosophy, it is not itself philosophy.
>A tool is not the same as its use, a hammer is not carpentry.

Logic can be used as a tool, but it is also an area of philosophy itself.

>|>Besides, I only 'believe' them when I am doing classical bimodal logic,
>|>when I am reasoning using real-valued logic (also called fuzzy logic) they
>|>are *not* true.
>|
>|And who developed formalizations of fuzzy logic?  Scientists?
>
>Mathematicians and cognitive scientists trying to develop models for
>certain situations that classical logic did not handle well.
>
>The most any philosopher may have done is to provide some interpretation
>of what it means.

I must admit not knowing enough about the development of fuzzy logic to
disagree here, although I am greatly suspicious of this claim.

>|>Also, I find that classical logic has little *practical* value.  Most real
>|>problems are in finding the correct axioms or premises, not in the deductions
>|>from them.  Too many purely logical arguments are based on questionable axioms
>|>for them to be treated as more than suggestions of one possible reality.
>|
>|Be careful.  Axioms and premises are not completely interchangable.
>|And while it may be true that too many purely logical arguments are based
>|on questionable "axioms", too many purely scientific arguments are based
>|on questionable logic.
>
>In the long run scientific arguments on not resolved by logic at all, but
>rather by observation and testing.  Logic is *sometimes* useful in clarifying
>the implications of some observation, but that is all.  (Then those deduced
>implications become a 'theory' which is subject to resting to determine if it
>indeed corresponds to reality - the deductions are *not* taken as final
>per se).
Scientific arguments are not *always* resolved by observation and testing.
[insert Copernican astronomy example here]

Apart from this, the original claim is getting lost.  I am merely saying that
scientists use philosophically developed methods and approaches, including
logic, in doing their work.  I would not claim that all of science can
be done in an armchair - science *is* empirical.  But it is not *solely*
empirical, and much scientific progress is made through the (philosophical)
analysis of theories.

>My point about axioms and assumptions is that the only ones of any real value
>are those based on observation, or which follow directly from some definition.
>All others are immediately suspect.

This *in itself* is a philosophical position, like it or not.  "All truth
comes from observation" is a philosophical maxim, which itself needs
philosophical justification.

>
>I repeat logic is a *tool*, to be used when it is appropriate to do so,
>and which is dangerous when abused.  Logic *by* *itself* is of no value.
>And no matter what, the results of logical deduction *must* be subjected
>to observational testing to validate the axioms and assumptions.
>
>A purely logical argument is little more than a game.
>By choosing different reasonable sounding axiom systems you can
>prove almost anything.   This is *especially* true in philosophy
>where the basic concepts are not (yet) susceptible to observational
>validation.
>
>|>Probably, since in actual practice neither really contribute much to how
>|>science is actually done.
>|
>|What?!  Logical positivism had a *profound* impact on the way science
>|was done in its heyday.  Just check any history of science... 
>
>In its heyday, yes, but not so much now.

Then it serves as an example of how philosophy can impact science, even
if the particular movement cited is no longer influential. 

>And even then the *best*, most productive scientists were the ones that
>just got on with the job and didn't let little things like a philosophical
>concept constrain thier research program.

Who do you include amongst the "*best*" scientists of this time?  Certainly
the most influential psychologists and physicists were in the grips of
positivism...

>|Nonsense.  Many shamanistic religions also have methods that "work" within
>|the culture in which they exist.  It is only because we *philosophically*
>|refuse to admit the entities that they posit as having reality that
>|these systems are not labelled some form of "science".  We have alternative
>|accounts as to the way the universe works.   
>
>Bullpucky - they *cannot* produce repeatable, predictable results (except
>to a limited degree in areas of *human* behavior, where belief itself produces
>its own reality).
>
>Where they can produce repeatable results scientists eventually come to
>accept them.  To this day we use willow-bark extract to cure headaches,
>but most of us don't know it - we call it aspirin.  Most biologists today
>treat native people's identifications of animals and plants with a great
>deal of respect, to the extent of being willing to use them as a first
>aproximation for species determination.
>
>In short, where there *are* repeatable results that work *all* the time,
>they are accepted into science.  Science is how we got computers and
>a cure for pneumonia.  Can shamanistic religions match that kind of track
>record - no, not even close.

You miss my point.  We in the West have a materialist philosophy that does
not admit spirits, ghosts, and other such entities as valid terms in
explanations.  Other cultures do, and in such cultures they can "explain"
many phenomena that we can, using these entities.  What is needed is a
reason to say *why* these explanations are wrong.  Repeatability is *not*
a criteria because 1) many of their phenomena *are* repeatable, such as
using willow-bark extract to ease pain, and 2) their worldview *explains*
why repeatability doesn't work (hey, the gods don't always feel like answering
your prayers).

I do not question that science has produced great advances, and that it
has some kind of approximation to "truth."  That is not the question.  The
question is *why* we believe this.  If it is merely because it can *explain*
the world, other worldviews can also do this.  If it is merely because
we get repeatable phenomena, a) other worldviews can also get repeatability, and
b) repeatability is *not* the same as understanding. 

>|>Or you might say that most scientists have a theory of truth rather like:
>|>"If it walks like a duck, quacks like a duck, and looks like a duck, it must
>|>be a duck".  In short, if a model works, act like its true until it doesn't.
>|
>|Define "works" without using any implicit philosophical concepts.
>
>If it produces practical results.  It it allows one to make plans and rely
>on them working.  If it provides the capability to build something.
>
>In short, if it can be *used* *reliably*.

OK, so why should we take these criteria to yield "truth"?   Again, using
no philosophical concepts.

>|>|Does he wonder about his conceptual
>|>|distinctions (like between matter and non-matter)?
>|>
>|>Why is this even relevant to science?  What matters is observables.
>|
>|*This* is a philosophical position!
>
>I am not sure what you are getting at here.

That to claim that all that matters is observables is a philosophical
claim.  This position would deny the importance (or perhaps even reality)
of theoretical terms which cannot be observed.  


>|We're not talking about a *descriptive* account of science, but a *normative*
>|account.  If scientists reason poorly, or illogically, then they are doing
>|poor science.  Note that "reason" and "logic" are at least partially
>|defined by philosophers.
>
>I have noticed that most progress in science does not depend much on
>logic (though perhaps other types of reasoning, like contingency analysis
>or mental manipulation may often be inportant).


Nonsense.  The reason logic doesn't seem important is that it is used
ubiquitiously.  If a scientist said: "If a bird has a red breast, then
it's a robin, this bird has a red breast, so it's a sparrow" we'd be
aghast.  Any scientist that can't reason logically can't do science.
Experimentation is *based* on logic.
 
>|>   Actually, a large part of the scientific method (repeatability,
>|>independent verification and so on) is aimed at revealing this kind of error.
>|
>|You keep talking about "The Scientific Method."  Could you state what it
>|is?
>
>It is an approach to knowledge in which observation takes primacy over
>any and all theory.  It involves converting each problem into a *testable*
>form, devising a set of tests, performing them, and interpreting the results.

Then, as noted above, the decision to believe Copernican astronomy did
not rest on the scientific method.

>There are additional rules - the tests must be performable by any qualified
>operator, they must produce repeatable results, and they must be described
>along with the results.

Where have all these rules come from, if not from an analysis of what we
want science to achieve?  Are they merely "folklore" passed down from
generation to generation, "hey, *this* seems to work!"

>Actually a great deal of the practice of science is in the area of converting
>a question into a testable form.  This is where debates about definitions
>and such like are important.  The other large area of debate centers on
>interpretation of results - this usually leads to new tests being designed
>to differentiate between the competing interpretations.

Yes, Stanley, I know how science works, I practice it myself.  In the above
cases of "debates about definition" and "interpretation of results" philosophy
is constantly used to determine what are allowable definitions and what
are legitimate ways to interpret results.


>It is a very flexible 'method', since the only requirements are repeatability
>and verifiability, almost any test can be used - lab experiments, field
>observations, excavations and so on.
>
>|>The bottom line is,I trust a scientific result because the scientific approach
>|>has shown itself to be reliable in the past.
>|
>|You should read some Hume.
>
>What about Hume do you have in mind here.  I have almost certainly read
>some of his work while I was in college.  It was largely from reading the
>endless debates amoung various famous philosophers that I conclused that
>it was not a useful approach to knowledge (only to ideas).

I was referring to Hume's problem of induction.  

>|>I do not trust pure reason because it has so often lead to useless results.
>|>It is too easily mislead by incorrect asumptions or isolated false data.
>|>Science, because it is self-correcting, is not so gullible.
>|
>|Science without reason is witchcraft.
>
>But reason is only a *tool* in science, *pure* reason is reason *alone*,
>without any observations to back it up.
>
>I do not object to reason, only to *pure* reason.  Only to using reason as
>an end in itself.  Only to using reason by itself.

If this is your position, then we may not be terribly far apart.  Reason alone
cannot do science - empirical work *is* necessary.  But to believe that
all advances in knowledge come only through empirical testing it naive and
wrong.

As I suggested above, it might be better to take this to email, since the
topic is straying from the focus of the newsgroup.


- michael




