Identity, Structure, and Causal Representation in Scientific Models - Mechanism and Causality in Biology and Economics - page 49

Cartwright objects to Woodward defining direct cause by interventions that set a
variable to a value come what may precisely because, as we already observed, such
interventions alter the causal system (e.g., moving from Fig. 3.1 to Fig. 3.2 ), even
when the other causal arrows (and the equations to which they correspond) are left
intact. We should be concerned with the normal workings of a causal system, and
the workings of some other causal system are irrelevant to them (Cartwright 2007 ,
p. 107; Cartwright and Jones 1991 ).
Another of Cartwright's counterexamples to modularity - the operation of “a
well-made toaster” - helps to clarify the point:
The expansion of the sensor due to the heat produces a contact between the trip plate and
the sensor. This completes the circuit, allowing the solenoid to attract the catch, which
releases the lever. The lever moves forward and pushes the toast rack open.
I would say that the bolting of the lever causes the movement of the rack. It also causes a
break in the circuit. Where then is the special cause that affects only the movement of the
rack? Indeed, where is there space for it? The rack is bolted to the lever. The rack must
move exactly as the lever dictates. So long as the toaster stays intact and operates as it is
supposed to, the movement of the rack must be fixed by the movement of the lever to which
it is bolted.
Perhaps, though, we should take the movement of the lever to the rack as an additional
cause of the movement of the rack? In my opinion we should not. To do so is to mix up
causes that produce effects within the properly operating toaster with the facts responsible
for the toaster operating in the way it does; that is, to confuse the causal laws at work with
the reason those are the causal laws at work. (Cartwright 2007 , pp. 85-86)
What, we may ask, is proper operation? To ask such a question requires that we
can distinguish changes in its state that constitute its proper operation from changes
that undermine the proper operation or destroy the mechanism. Surely, such a
distinction is partly a matter of perspective and often driven by pragmatic
considerations. It requires that we be able to decide when the mechanism has
been so altered that it is effectively a new mechanism and when the mechanism
is preserved - that is, we need identity conditions for a causal system
(cf. Woodward 2003 , pp. 108-109).
Cartwright uses the toaster example to argue that the relevant interventions
operate only within a context of a preserved mechanism and that Woodward's
come-what-may interventions generally break the mechanism. While she does not
provide the necessary identity conditions, they are evident in the structural account.
Two mechanisms are causally the same when they have the same parameterization
(i.e., the same privileged set of variation-free parameters) and differ only in the
particular values that the individual parameters take within their admissible ranges.
In other words, two mechanisms are causally identical when they differ not in their
parameters or variables but in the token instantiations of their parameters and the
token consequences of those instantiations for the variables.
Another example illustrates both the pragmatic and the conceptual issue. In the
movie The African Queen , Charlie Allnut (Humphrey Bogart's character) runs a
steamboat. From time to time the pressure in the boiler of the steam engine gets
dangerously high. He hits a particular valve with a hammer, which frees up the
valve, and allows the steam to escape. Later in the film, as part of his general effort