Re: Hume

From: Howard Engelskirchen (howarde@TWCNY.RR.COM)
Date: Fri Nov 14 2003 - 19:17:22 EST

The idea that Hume's position on causation has never been satisfactorily
challenged leaves out of account the last 30 or 40 years of the philosophy
of science.  About midcentury realist challenges to logical empiricism began
to take hold for the good and substantial reason that the propositions on
offer, including Hume's approach to causation, didn't correspond to what
scientists actually do.  Nancy Cartwright, for example, in "The Reality of
Causes in a World of Instrumental Laws," argues that although philosophers
have believed in laws and denied causes, explanatory practice in physics is
just the opposite -- we argue from effect to cause against a background of
our best available knowledge to find concrete causal processes.  The
inference is not to the best explanation but to the most probable cause.
Bhaskar's powerful critique of Hume in A Realist Theory of Science (1975)
was mentioned in a previous post.  Bhaskar argues that the Humean concept of
cause can't make sense of experimental activity -- the sequences of events
Hume's account presupposes are mostly produced in the closed system of a
laboratory, and the Humean can't give an account of what explanations hold
outside the experimental laboratory where such regularities rarely occur.
Moreover, reflection on experiment shows that it is not about sequences of
events at all.  Those are produced by the activity of the experimenter -- ie
the experimenter puts litmus paper in acid and turns the paper red -- but
the experimenter doesn't produce the causal process that caused the
transformation.  In other words in the Humean account causal processes are
confused with and collapsed into the sequences of events that permit us to
identify them.

Harre and Madden in Causal Powers (1975) also provide an important critique
of Hume. They show that Hume's argument depends on conflating logical
necessity -- something true in all possible worlds -- with natural
necessity -- a relation that holds between a generative structure and its
effects that can only be discovered a posteriori.  The fact that it is
logically possible for nature to change its course and for a stick of
dynamite to turn into a stone doesn't say anything to the point about
whether, given dynamite of the proper composition and structure, a stick of
dynamite will explode when detonated.  It is self contradictory to say that
the nature of dynamite is such that it both does and does not explain its
power to explode.

There's methodological significance here, as Harre and Madden emphasize by
recounting the story of Moliere's imaginary invalid:  opium puts people to
sleep because it has a "virtus dormitiva."  Put a Humean on the case:

"Following Hume's account of the empirical meaning of a causal statement,
the man who wanted to know more about the action of powerful particulars,
say soporifics, could proceed only by collecting more cases of similar
phenomena until he had enough to convince himself of the lawfulness of the
statement of concomitance -- for example, had formed the psychological habit
of expecting drowsiness to follow ingestion of opium."

It's been done in social science.

On the other hand, given a robust realist account of causation a scientist
"will not collect further statistics but begin[] an exhaustive chemical
anlaysis of the potent entity, trying to find out what sort of stuff it is,
i.e. what its nature is.  Then he tries to follow it through the body after
ingestion and tries to ascertain how it acts upon the central nervous
system, the higher centers, and so on.  He undertakes quite a different sort
of investigation from the strict Humean.  He does what scientists actually
do." (Causal Powers 91).

In other words, looking for a "virtus dormitiva" is better science, and, by
the way, which approach did Marx follow?

Richard Boyd is another contemporary philosopher of science who provides a
sustained and satisfying critique of Hume based on the practices of science:
"There is indeed considerable evidence that almost all the significant
features of the methodology of recent science rest ultimately upon knowledge
of unobservable causal powers and mechanisms (cites omitted), and thus that
the empiricist reservations about experimental knowledge of unobservable
causal powers and mechanisms are profoundly mistaken."  The basic idea is
that scientists actually do posit causal structures and processes and then,
relying on the best background theories available to them, design
experiments and explanations.  The success of the endeavor is good evidence
for the structures and processes proposed, whether unobservable or not.
((Boyd, "Observations, Explanatory Power, and Simplicity:  Toward a
Non-Humean Account.")

The Boyd and Cartwright articles can be found in Philosophy of Science
(1991) edited by Boyd, Gasper and Trout.

Value is an unobservable causal structure, and Marx, economics or social
science more generally without cause makes less sense of the world than we
can afford.  So getting past Hume's legacy is important for us.


----- Original Message -----
From: "Rakesh Bhandari" <rakeshb@STANFORD.EDU>
Sent: Friday, November 14, 2003 1:15 PM
Subject: Re: [OPE-L] Hume (just citations)

> Interesting discussion underway.
> On a Marxist theory of causality, see Anjan Chakrabarti and Stephen
> Transition and Development in India, pp. 31ff
> On Althusser's distinction between transitive and expressive causality,
> Robert Paul Resch Althusser and the Renewal of Marxist Social Theory.
> Rakesh

This archive was generated by hypermail 2.1.5 : Sun Nov 16 2003 - 00:00:01 EST