Well, maybe you have my hide on your barn door, but I'm not *quite*
convinced just yet.
You were probably right that I had read the Feynman excerpt too hastily,
although there are some oddities there in light of which I could dig in
for a bit longer. But in the interest of keeping the pedantry index on
this list at a reasonable level (well, at least not raising it) I'll let
it go.
HOWEVER, I'm not about to budge yet on my main contention. Just to
clarify, let me expand a bit, then get back to our direct discussion.
What I want to claim is that in any genuinely rich theory (e.g., not
merely phenomenological generalizations such as some of the classical gas
laws), there is a stipulative - or definitional, or human choice -
element. Such elements are not just arbitrary and purely armchair - they
are partially (but only partially) shaped by empirical considerations.
On the other hand, they are not simply *dictated* in any way by nature
(contrary to positivists, etc.). The presence of such stipulative,
definitional, choice elements does not in the slightest imply that the
containing theories are at risk of degenerating into closed definitional
loops (one of Feynman's suspicions, I think), nor does their presence
compromise, much less remove, the empirical character of the theories in
question - indeed, I tend to think that without such elements the
theories involved would not be able to make full empirical contact with
nature. But it does mean that at some point in the theory/nature link,
there will be a stipulative, definitional component - the theory will not
be some purely mechanical result of investigating nature.
Here's a (perhaps imperfect) example from another area, then I'll come
back to Newton.
The concept and proper definition of *mammal* is nowhere carved in stone
in nature. The accepted definition is partially a result of human
taxonomic decisions. There was nothing simply preventing us from
adopting a definition under which whales and platypi would turn out not
to be categorized (at the relevant level) with cows, etc. We would, of
course, end up with a somewhat different system of biological
categorization, theories, etc., but that is a different matter. (Such
other systems might not be as 'satisfactory' in some sense as our present
system, but that doesn't get us too far, given that the criteria,
relative weights of competing criteria, etc., for judging
satisfactoriness are themselves in part results of human decision,
choice, stipulation, and so forth.)
So *we* decided, which characteristics were and were not going to go into
our stipulative definition of mammal. The decision was not simply out of
thin air - lots of empirically-based considerations played roles here -
but it was not rigorously driven by nature either. Once that decision
was made (and I don't know the historical details), it was not, of
course, up to us what did or did not fall under the definition, whether
any nice theories could be wrapped around that categorization, etc. -
most of that was up to nature, and we had to engage in empirical digging
to find that out. Of course, that definition likely would never have
gotten off the ground had there not been some promising prior indications
- many of which had factored into the choice to begin with.
I want to claim that the same sort of thing can be found in physics - and
in particular in Newtonian physics and principles. There has been
historical commentary concerning just where the stipulative element
should be located - i.e., what should be taken as stipulative and what as
derivative - force, mass, etc. I don't really have any strong views
there - I said in my first post on this that I thought that we were
dealing with something definitional or at least some system of
interlocking definitions. But I do think that it is buried somewhere in
the system. That's my position.
In this connection you cited Cohen. I hadn't read that book for some
time (it is around 40 years old), but as I read him, he takes force as
esentially a historical given (in Newton), but introduces a definitional
stipulation concerning mass. You cite his suggested experiment involving
the F1/F2 = A1/A2 equation. That, as he presents it, involves a single
body having (presumably) a stable quantity of matter, and if the ratio
stays constant over a variety of Fs and As, the constant is 'the measure
of the amount of [the substance in question]' - the mass. Of course, in
order to complete the system, we need to be able to compare quantities of
different types of substances. We can get comparative 'mass'
measurements within each type of substance, but as yet we have no
guarantee that mass determined by F and A for one substance corresponds
to the 'mass' measurement for some different type of substance. Just as
the same quantity of heat might raise different amounts of different
substances to the same temperature, the same force (independently
determined) might produce the same acceleration in different quantities
of matter of different substances - whatever 'quantity of matter' turns
out to mean. How do we rule that out? An attractive attempt might be to
temporarily tie quantity of matter to weight, then later switch to mass.
But what is the justification - *independent of Newton*, since that's the
issue here - for doing that? Weight and quantity of matter were not
automatically linked historically ('levity' was a possibility in
Aristotelian physics, and phlogiston in some versions had negative weight
but not, so far as I know, a negative quantity of matter). What was the
solution? Here's Cohen's answer (p. 162):
"But for our aluminum blocks this same constant is also a measure of the
'quantity of matter' in the object, which is called its _mass_. We now
make precise the condition that two objects of different material - say
one of brass and the other of wood - shall have the same 'quantity of
matter': it is that they have the _same mass_ as determined by the
force-acceleration ration, or the _same inertia_." [his emphases]
That sounds awfully stipulative. We lay down the *condition* under which
they *shall* have the same quantity - matching numbers *as determined by*
the specified ratio. (As Cohen points out a bit earlier, Newton could,
given his system, explain why weight and matter correlated so nicely in
our neighborhood, but that is *given*, the system, which while justifying
it we aren't.)
(Re: Cohen. I'm not sure exactly how much weight (sorry) to put upon
his choice of words, but in his 1985 _Revolution in Science_, (p. 161)
Cohen credits Newton with "invention of the primary concept of physical
science (mass); invention of the concept and law of universal gravity
....". Since Kuhn, the term 'invention' has been a bit vexed, but anyway
Cohen does not use 'discovered', but credits Newton with *inventing* the
concept which in _Birth_ he cited (as I read it) as - by stipulation -
tying the system together across types of substance.)
Well, have I helped my cause any?
Del
Del Ratzsch voice: (616) 957-6415
Philosophy Department (616) 451-4301 (home)
Calvin College e-mail: dratzsch@legacy.calvin.edu
Grand Rapids, MI 49546 fax: (616) 957-8551