Re: causation in theory

[jonathan <jdh11@xxxxxxxxxxxx>]

At the risk of getting too
philosophical, here are some
comments on the recent
discussion of causation:

Newton did indeed deny that
his laws identified causes, in
keeping with his slogan
"Hypotheses non fingo" (pardon
any misspelling). But he
didn't do this because he
thought his laws were merely
predictive in a statistical
sense.  He thought that they
could be exploited to
manipulate events, carrying
out the project set by Bacon
to first understand how we are
subject to nature in order to
dominate it.  In this sense,
perhaps the most important
sense, his laws can be called
causal laws.

I think (but am not
certain)that like a large
number of early modern natural
philosophers Newton required
of causation that one assume a
necessary connection between
cause and effect, and he
denied that we can have any
such knowledge. Only God
could. The reasons are
complicated, but many
philosophers of the age
believed it, including
anti-empiricists like
Malebranche, long before Hume
arrived and pretty much
eliminated even the theistic
solace.  This denial of causal
knowledge (at least as it
appears in the Occasionalists
and Hume) is just a species of
philosophical skepticism.  But
whatever we were sensibly
talking about when we spoke of
causes, it wasn't incorrigible
knowledge of future events or
of the necessary results of
states of affairs.

I think Harry had the right
take on it:  even if we want
to be 'hard-headed' types, we
can treat causal assertions in
a quasi-Popperian spirit: they
are falsifiable hypotheses
which allow us to make strong
predictions (stronger and
different from statistical
relationships which no longer
apply when structural
conditions change...and what
are structural conditions but
causal conditions!).  If the
predictions are right, great;
if they're wrong, back to the
drawing board.  But I don't
think this attitude has a
clear application when the
causal relations among the
phenomena change (that is,
when everything shifts and
nothing like the laws of
physics are present). All
regularities of action, of
course, have the systematic
problem that people react to
predictions in order to
realize or prevent them.  I
think scientism really must be
eradicated at the roots in
economics, but that's another
story.

Anyway, my first point was
that by denying causation and
insisting that the laws were
confirmed by their predictive
merits, Newton and others were
not asserting that the laws
were no more than statistical
correlations. They were
assumed to be rigid
regularities (i.e., LAWS)
which could be used to change
any fluid regularities (i.e.,
mere correlations) with the
proper tools, but which could
not themselves be changed.
There has been a lot of work
recently on the relationship
between causation and the
ideas of manipulation and
interference.  Judea Pearl's
'Causality' is very
interesting. An excellent
discussion appears in 'On the
Reliability of Economic
Models' edited by Daniel
Little (see the essays by
Woodward, Cartwright, and
Hoover).  Huw Price also has
written on causal asymmetries
and the symmetrical laws of
physics, and on the notion of
agency as the fundamental
causal notion (I think the
book is called Time's Arrow).

I hope this hasn't been too
much.

Jonathan Halvorson
Department of Philosophy
(PhD student, still!)
Columbia University