Mechanisms and Laws: Clarifying the Debate - Mechanism and Causality in Biology and Economics - page 129

to package the variation into the antecedent of the conditional. In fact, however, one
finds that biologists typically characterize a mechanism in a particular strain of a
particular species (such as wild-type Drosophila ) and then recognize that there will
be subtle variations on that mechanism in other strains, mutants, and species. They
are not looking for general law statements that cover all of them but rather for sets
of prototypical models that stand in family resemblance relations to one another
(cf. Schaffner 1993 ). To push Bechtel and Abrahamsen's point one step further,
prototype models need not be general descriptions. Bechtel and Abrahamsen also
call their view an “exemplar” ( 2005 , p. 438) account, noting that models of
mechanisms often describe a particular, exemplary case. (Open a biology textbook
and look at some diagrams of mechanisms; more often than not, they are cartoons of
a single representative mechanism.) On such an exemplar view, generalization is
extrinsic to the mechanistic models (exemplars, prototypes); that is, the model need
not contain general statements or general representations at all. 5 Leuridan's insis-
tence that the model must contain such things is simply the imposition of a
philosophical prejudice onto actual scientific models that have the capacity rather
to surprise us if only we open our eyes to them. The generality of such a mechanistic
model is a matter of its scope of application and not something that must be
represented within the model itself. If one attempts to put the generality in the
model itself, to return to Bechtel and Abrahamsen's point, the model has difficulty
accommodating the variability characteristic of biological mechanisms.
Curiously, Leuridan fails to consider the possibility that most p-regularities are
stable and strong because they are produced or maintained by mechanisms (see,
e.g., Bechtel 2009 ; Craver 2007 ; Darden and Craver 2002 ; Glennan 2010 ; Steel
2008 ; Wimsatt 1998 ). Why might Morgan have expected the apparent exceptions to
Mendelian heredity he discovered in his lab to apply outside of the lab and in other
organisms? The simple answer is this: he expected the mechanisms of heredity
outside the lab and in other organisms to be more or less similar to the hereditary
mechanisms at work in his Drosophila . The p-laws of heredity are stable and strong
precisely because there is an underlying mechanism (e.g., involving crossing over
and replication of chromosomes) that explains them.
In his topic Across the Boundaries ( 2008 ), Daniel Steel builds on early
suggestions by Darden and Craver ( 2002 ) to develop an elaborate analysis of how
one can extrapolate scientific knowledge based on an understanding of the relevant
mechanisms. The idea behind his “comparative process tracing” (Steel 2008 , p. 85)
approach is simple and helpful: First, one uses a variety of strategies to learn about
the mechanism in the model organism. Second, one compares the mechanism in the
model organism to the mechanism of the extrapolational target at certain key
junctures. That is, one compares the two mechanisms at stages at which the
mechanisms are most likely to differ significantly from one another. The fewer
significant differences one discovers at these key points, the stronger is the basis for
5 This view fits with the semantic view of theories that Bechtel and Richardson embrace (cf. 2010 ,
p. 232).