Re: [OPE] epistemological and scientific questions

From: Howard Engelskirchen <>
Date: Sun Dec 14 2008 - 19:37:46 EST

Hi Dave,

Perhaps the difference between us tracks different approaches to the
scientific method. I'm trying to insist that the projects of
science, including Marxism, are better understood as realist.

'Accurate' must be with respect to some standard. What is the
standard you appeal to? Is it how things stand in the world? Or is
it predictive power only? And are you concerned to give an accurate
account of underlying causal structures or mechanisms or only to give
an account of observable phenomena, ie the things we predict? I
don't mean to lock you into alternatives here -- perhaps you have an
altogether different approach to the thing that makes something
accurate when it is accurate.

But as it is, I'm not sure what actually you disagree with. I wanted
to suggest that the object of the scientific method was to give an
account of the causal structure of the world; we need to understand
to transform. I wanted to emphasize that the theoretical terms and
other representations we use, such as models, function well or badly
to refer to such real generative structures. You suggest we want to
use theoretical representations, such as models, to give an accurate
account of mechanisms that generate the data we observe. These seem
to be saying pretty much the same thing. But you want to reject the
idea that we can use 'true' to express the way representations we use
accurately account for the mechanisms to which we refer. What is the
problem you find here? Why is 'true' a category mistake? Is this
because truth applies only to the way we use logic and isn't a
helpful way of referring to the bridge we make between our concepts
and the world?

My concern with models was with the way economists, for example, can
abstract from this or that without regard to whether the way they
abstract remains consistent with or accesses more fundamental causal
determinations. Ptolemy showed we can enrich our models endlessly if
all we want to do is make sure we get our predictions right, and Marx
could be understood to insist that bourgeois models make Monsieur le
Capital and Madame la Terre pirouette to a Ptolemaic tune.


At 05:08 AM 12/14/2008, you wrote:
>on 2008-12-13 15:11 Howard Engelskirchen wrote:
>>I suspect neither of the formulations here, or in the post of
>>dogan's referred to, fully achieve what we're after.
>I am not sure what you are after, Howard. I was only trying to
>clarify the meaning of a "scientifically informed political
>position". To do so I also thought it is necessary to clarify how
>the scientific method works in general.
>>But I agree the idea of 'models' is tricky. We can think of the
>>double helix as providing a model for our understanding of dna, and
>>there is nothing wrong with that, but too easily, and especially in
>>political economy, models get expressed, as here, not as
>>corresponding to the way the world is, but as heuristic instruments
>>that guide prediction.
>No, I think your last distinction is incorrect. Scientific theories
>are not "corresponding" to the way the world is, but rather *model*
>the mechanisms that generate the empirical data we can observe, and
>therefore they *guide* predictions too. Recall that the Mendelian
>theory of the gene was formulated before the discovery of the double
>helix. However, the theory predicted, among other things, that there
>should be some physical entity passing from parent to offspring.
>This was subsequently confirmed.
>In other words, the theory modeled the mechanisms of evolution and
>reliably predicted several observations. Of course, the theory was
>made more accurate than Gregor Mendel's original ideas, as more data
>was gathered, new concepts were formulated and new tools were used.
>>On this approach, the model may be coherent, but it isn't true, and
>>we can leave out or introduce features according to model driven
>>(conceptual) imperatives.
>To say that the model --- i.e. a scientific theory --- is "true" is
>a category mistake. Is an airplane model "true"? No, it is more or
>less *accurate*.
>//Dave Z
>ope mailing list

ope mailing list
Received on Sun Dec 14 19:46:41 2008

This archive was generated by hypermail 2.1.8 : Wed Dec 31 2008 - 00:00:05 EST