File spoon-archives/bhaskar.archive/bhaskar_1996/96-05-20.182, message 62

Date: Fri, 2 Feb 1996 16:34:18 GMT
Subject: the importance of mechanisms

My apologies for entering this enjoyable discussion only
infrequently, and somewhat one-sidedly.

Some thoughts on the `metaphor' of mechanisms.

A mechanism is something we can understand: it is
composed of sub-mechanisms that interact in rule-governed
ways, producing emergent behaviours and new causal powers.

> The metaphor of mechanism was, indeed, invented in the modern
> political theory already in the 17th century, and it became ever
> more widesppread in the 18th century ...

Yes. The emergence of factory production and capitalism
was the `engine' that `fuelled' scientific materialism
and `clockwork' ideas of the universe and causation.

Since then, the mechanistic metaphor has been necessary
in science as it, quite simply, refers to processes that we
understand or think we could in principle understand.

And, of course, what scientists have in mind when they
use the word `mechanism' or `machine' changes with the prevailing
technological sophistication.

This century has seen new types of understanding that
enrich our concepts of mechanism and causation. For
example, I would say that Schrodinger's wave equation
is a mechanism (or rather a partial mathematical representation
of a mechanism), in spite of the fact that in its present
form it refutes regularity determinism.

Therefore, those that are uncomfortable with the term
`mechanism' are perhaps relying on an outdated 19th century
understanding of the term. In scientific circles there is
no self-consciousness about its use: when a phenomenon is
understood in terms of an underlying generative mechanism,
then it is entirely useful to think in terms of machines and
rule-governed causality (but not necessarily regularity determinism).

> It is my argument that there is no such "mechanism" as
> 'balance of power' or 'invisible hand'. There are unintended conse-
> quences of multiple, socially conditioned actions and there is
> interdependence of these actions, but they are not mechanical
> at all (to Colin: why use the term 'mechanism' if the intention is
> not to say that the production of effects is 'mechanical' at least
> in some sense of the term?; and is not the term 'machine' at the
> origin of both terms?).

`Unintended consequences', `multiple' and `interdependence' are
words that speak of choas, non-linear dynamics, complexity
and the like. And such types of system undoubtedly exist. However,
social mechanisms -- of which Adam Smith's `invisible hand' is
a very real example -- regulate social systems _in spite_ of
the multiple accidents, unpredictabilities, capriciousness
etc. of real life. One of the roles of a social science is
to discover these real and actual mechanisms that underly empirical
surface phenomena. In complex systems this is hard. From the
level of the individual the allocation and reallocation of the
social division of labour may appear an unintended consequence,
but from the level of the theory of those changes (a candidate
theory being, for example, Marx's `law of value') it is
mechanistic, i.e. it is understandable, rule-governed, perhaps
partially or wholly predictable.

> In RTS Bhaskar rejects regularity determinism ( a mechanistic
> concept, events of type A will always be followed by events of type B) but
> accepts ubiquity determinism (all events have a cause). In this respect his
> use of the term mechanism I admit might retrospectively be regretted, and I
> think Heikki has a point. But more importantly, given a Bhaskarian account
> of causality there is no way the term mechanism can be interpreted
> mechanistically, whatever its origins.

Ubiquity determinism falls out from ontological stratification.
An event can be multiply conditioned by many generative mechanisms;
therefore events of type A will not always be followed by events
of type B, IF a countervailing generative mechanism expresses its
causal powers. Ubiquity determinism can talk of mechanisms, and
_must_ be interpreted mechanistically, as long as we don't think
of `springs', `pulleys', `clocks' etc., but perhaps complex
computer systems, adaptive control systems, quantum computers
etc. These are all mechanisms, but of a different kind.

> Further, sometimes the metaphor of 'mechanism',
> even when understood in a critical realist manner, may well give rise
> to reificatory tendencies of thought. And, on the other hand, what
> would be the *good reasons* for using this metaphor?

Better the possibility of reification fallacies than the
possibility of `deus ex machina' causal explanations. Invoking
complexity arguments is a learned way of saying: we don't
know. Without good theories of social mechanisms we'll end up
with ideology.




Driftline Main Page


Display software: ArchTracker © Malgosia Askanas, 2000-2005