Re: causation in theory

[jonathan <jdh11@xxxxxxxxxxxx>]

A few responses.

David wrote that the "Post-modern" critique of science asserts that
"there is no scientific way of proving the truth of scientific
findings."
I hope this wasn't the point, because that would make it a regurgitation
of mainstream philosophy of science from Bacon, Huygens, and Newton on.
My claim was that some of the skeptism about causation despite a lack of
skepticism about laws was founded on the belief that for knowledge of
causation one needed proof, while for knowledge of (true, adequate) laws
one did not. I was trying to point out that assertions of causal
relations need no more "proof" than universal descriptive statements.
The demand is a relic of metaphysical thinking.

Harry remarked:
"The lawfulness of physical laws is not stripped because it was never
causality that made them lawful in the first place."

This is true, but unclear.  There is a distinction between empirical
laws (which do not indicate causal mechanisms and are mere "phenomenal"
laws...the classical gas law is an example) and explanatory laws (a
modern gas law which refers to the mean free path of molecules would be
an example, because it points to the mechanism by which volume varies
with pressure).  But if that is what you mean, it doesn't tell against
the existence of causal laws.

Or do you mean instead that there isn't some separate thing, the causal
relation, which we observe over and above our observations of
covariations?  That is of course true, and it is what was right about
Hume's attack.  Causation isn't a separate thing in the world.  But
nothing is wrong with Kant's response, either: it is WE who place causal
assumptions into the world. They are required by us to make inferences
about events.  It is exactly this Kantian kernel that many people have
taken up recently in arguing that when we make judgments about the
outcomes of interventions which change structural conditions, we must
use causal reasoning and not simply statistical reasoning. Statistics
tells us what is likely to happen if things stay pretty much the same as
they have been in the past.  Causal reasoning tells us what we should
expect to happen if we change some of those very things, based on other
things that don't change.  In most cases, a cause is an intervention
which alters what would have otherwise been the case.  this is why it is
so closely tied to counterfactuals and the idea of responsibility (if
only you hadn't insulted him, he would have invited us!)
Correspondingly, a causal generalization (I don't assume it's a law) is
one which remains invariant over some range of changes in the variables
it mentions, and which we usually regard as a means to change one
variable by bringing about or preventing another.

Harry also made some interesting remarks about what laws do:
"The lawfulness of laws is a matter of what they forbid not what they
cause. Under a set of laws, prediction is accomplished by combining a
causal explanation (making use of those laws) with appropriate boundary
conditions... Laws are still laws because of what they they say is
impossible."

What does a law forbid but not prescribe?  I think you mean to say that
simply given the true laws you cannot determine a course of events,
because different initial conditions lead to different end states.
That's fine, but that does not mean that a law cannot pick out an
invariant relationship.

We can achieve an invariant relationship between entities only when we
abstract from temporal change and from concrete events.  For example,
the gravitational law gives only the force on one mass at a given
instant due to another mass some distance away.  It does not determine
the total gravitational force on this basis, let alone the total force
or the acceleration of objects towards one another.  You are right that
for that you need to bring in other conditions (the presence of other
bodies and forces).

And yet, a strict law of physics does presume to say something that must
be true, namely this aspect of the total situation or total effect.  It
doesn't only rule things out.  But what it requires to be true are not
concrete sequences of events, and if that's all you meant, we agree.

I'll stop.

Jonathan Halvorson