Page 83 - MODES of EXPLANATION

Basic HTML Version

most realistic interpretation of the growth of scientific knowledge that is consistent
with both actual scientific practice and experimentally confirmed results.” (Shook,
2003)
This pragmatic approach to context, contingency, and context dependence is reflected in
positions such as that of Gordon Graham (1983), who contrasts historical explanation with
what he calls “theoretical explanation.” According to Graham, theoretical scientists are, in
contrast to historians, concerned with disclosing general patterns or with finding out and
explaining how things regularly work. Aviezer Tucker (2004) adopts a similar position. One
of his main theses is that what historians explain is token evidence (e.g., particular documents
or fossils) and token events (i.e., events that are “unique and unrepeatable”; Sober, 1988),
like the rise of Rome or the assassination of Kennedy. Contrary to the historical sciences,
says Tucker (2013), the “theoretical sciences” are not concerned with token evidence and
events, but rather with theoretical types of replicated evidence and repeated events. What
makes these ideas “pragmatic” is that they are explicit in articulating the differences between
generalities (explanations that apply regardless of contingencies or
ceteris paribus
) and the
individual, specific, contingent event.
Salmon (1984) had a similar thought:
“To offer an explanation … is to assemble a total set of relevant conditions for the
event to be explained, and to cite the probability of that event in the presence of these
conditions. This view of explanation, unlike the standard account of deductive and
inductive explanation, does not see an explanation as an argument showing that the
event has to be expected on the basis of the explanatory facts. The explanation is,
rather, a presentation of the conditions relevant to the occurrence of the event, and a
statement of the degree of probability of the event, given these conditions. That
degree of probability may be high, middling, or low, but whatever its size, it is an
index of the degree to which we would have been justified in expecting it. … To give
scientific explanations is to show how events and statistical regularities fit into the
causal network of the world.”
Notice Salmon’s emphasis on “conditions” and “fit” – context dependence – over the more
realist notion of “explanatory facts.” Other examples abound if one chooses to go looking for
them:
“‘Default realism’, the second preferred solution, which may appear nominalist,
rejects any universal and objective class of objects and proposes a world made of
individual, different objects – that is, every cat or iron atom differs from every other.
They are identified as unities both according to innate beliefs and principles … and by
pragmatic feedback or cultural and scientific models. Moreover, they are only
29