Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: Centrifuge (was: Just So Stories)



At 09:09 PM 11/3/00 -0500, Ludwik Kowalski wrote:
>> How do we decide what is and what is not an explanation?

> By whether it makes correct predictions!!!!!!!!!!!

Yes.

But a qualitaltive explanation could not be expected to produce
a quantitative prediction. Why should it provide answers to
all possible questions?

A theory need not answer all possible questions... but it should either
make reasonably accurate predictions, or keep silent. If it makes a
prediction, the prediction should be correct.

As an illustration, consider the Just So Story that says a wing produces
lift because it is curved on top and flat on the bottom:
-- It makes incorrect predictions about the effect of increasing or
decreasing the curviness.
-- It makes incorrect predictions about the feasibility of upside-down
flight.
-- It makes incorrect predictions about numerous other relevant physical
phenomena.

So it is with the zig-zag theory. The problem is not its
non-predictions; the problem is its false predictions. As I wrote previously:
> -- The zig-zag theory predicts the wrong dependence on temperature.
> -- The zig-zag theory predicts the wrong dependence on viscosity.
> -- The zig-zag theory predicts the wrong dependence on particle size.
> -- The zig-zag theory predicts the wrong dependence on particle mass.

And contrary to what I wrote previously, I now doubt that the zig-zag
theory correctly predicts that large particles settle out faster. In
kinetic theory, vertical velocity is independent of horizontal
velocity. (It's not like a car with a 55-mph speed limit that can either
go horizontal or vertical.) There's a separation of variables.

Meanwhile, taking a higher-level view, what is the goal? I thought it was
to give students some exercise using the centrifugal field formula. If so,
the question arises of why bring up the size-dependence at all? Why not
just consider particles with some uniform size? Why not just assert that
there is a terminal velocity proportional to net force?
-- The assertion is plausible, consistent with everyday experience with
viscous liquids.
-- The assertion can be justified as a phenomenological observed fact.
-- Any discussion of the detailed submicroscopic theory can be deferred
to a later chapter.

======================

The world is already oversupplied with Just So Stories being passed off as
"scientific explanations". The last thing we need is another one.

Recommendation: Be modest and honest. It is much better to say "small
particles settle more slowly, and I don't know why" than to say "small
particles settle more slowly because of thermal zig-zagging". Don't
pollute your brain with false explanations. Don't pollute your students'
brains.

> authors nearly always like what they compose

Good authors check their work!

Recommendation to all physlers, and students, and everybody else: Check
your work! If you devise a theory, check its predictions. Check as many
of them as you can, and make sure they are all true. Any theory worthy of
the name can be subjected to numerous tests:
-- Do the dimensional analysis and check the scaling properties.
-- Check the limiting cases.
-- Check the order of magnitude for reasonability.
-- Check against whatever experimental data is available.
-- Check the sign and magnitude of the dependence on the independent
variables.
-- Et cetera.

Check your work!