Another possibility is that (some of) the contemporary authors Wilson alludes to
are not implicitly quantifying over specific relations, but rather are asserting the
presence of agenericrelation of ground, and these authors are not taking a stand on
whether such a relation is in some way primitive or fundamental. The latter possi-
bility strikes me as more probable. But I have not surveyed the authors in question.^47
Let’s consider the following objection. Suppose that there is no small-g grounding
relation instantiated by some facts without those facts also instantiating the big-G
grounding relation, and that the big G-grounding relation is instantiated by some
facts only if the transitive closure of the disjunction of the small-g grounding
relations is also instantiated by those facts. Then grounding is necessarily coextensive
with the transitive closure of the disjunction of the small-g grounding relations. But it
doesn’t follow that they are identical. Why think that we have identified what
groundingisas opposed to merely what grounding is coextensive with?
This is a fair question, but recall the discussion of a similar question in section 7.4.
Perhaps the wrong conception of properties for considering cases of philosophical
reduction or identification is one that closely ties properties tomeaningsrather than
referents. But suppose we deny the numerical identity of grounding and the transitive
closure of the disjunction of the small-g grounding relations. The more interesting
question then is which is in some way prior. Given their necessary equivalence,
I think the grounding theorist should not say that grounding is metaphysically prior.
Consider any given grounding fact, such as the fact that F grounds G. Why does
F ground G? Perhaps F constitutes G. Or perhaps G is a disjunction of which F is a
disjunct. Or perhaps F constitutes some fact that is a disjunct of G. Any of these
possibilities would suffice to explain why F grounds G. Given the hypothesis of
necessary covariance, one of these possibilities does obtain and hence does suffice to
explain why F grounds G. Strictly, we needn’t say that, whenever grounding is
instantiated, its instantiation is grounded by the instantiation of the transitive closure
of the disjunction of the small-g grounding relations. But it also does not seem
plausible that any given instantiation of grounding grounds an instantiation of the
transitive closure of the disjunction of the small-g grounding relations. Rather, both
such facts would have a common ground, and this common ground is a fact in which
big-G grounding is not a constituent.
The notion of ground might be ideologically or methodologically primitive. But
I am unconvinced that it is metaphysically primitive, at least if we are thinking of
grounding as a relation instantiated solely by facts, and we are thinking that meta-
physical primitiveness should itself be understood in terms of grounding. On Schaffer’s
picture, for all I have said here, grounding might still be an ultimate relation, though
perhaps it too is grounded by each of the small-g grounding relations. And on the
(^47) Shamik Dasgupta in personal communication has informed me that his attitude towards grounding
is more or less this: there is a relation of grounding (provided that grounding talk is best regimented via a
predicate rather than an operator), it is highly general, it might be fundamental, but it also might not be.