potentially a weakness with the proposal. Alex Skiles has suggested to me in personal
communication a more moderate response. He writes,
Let’s suppose that there were such small-g grounding relations, R and R. A reasonable
expectation is that facts about relative fundamentality will systematically track the distribution
of R and R: in cases whereexactly oneis instantiated, relative fundamentality will always run
in the same direction; in cases wherebothare instantiated, one of the two will inevitably
“trump”the other insofar as the direction of relative fundamentality is concerned. So, even if
there were such relations as R and R*, it remains a reasonable expectation that a sort of
systematicity would be retained—a systematicity that (our judgments of) the facts about big-G
grounding would still track. One would then have to tweak other bits of your view a bit: in
particular, (i) what it takes for something to be a small-g grounding relation and (ii) which
relation“built up”from small-g grounding relations we should take big-G grounding to be
necessarily coextensive with. Regarding (i), the tweak could be to think of small-g grounding
relations as onlyindirectlyinducing differences in relative fundamentality rather than doing so
directly. Here’s one implementation of that thought. For any arbitraryxandy, let’s call a
complete small-g profile with respect to R 1 ,R 2 ,... any conjunctive property the conjuncts of
which specify, for eachi= 1, 2,... whether or notxbears Ritoyand vice versa. Someone could
just take these tobethe small-g grounding relations, rather than R 1 ,R 2 ,...;I’m not sure what
much would be at stake if one instead insisted on taking R 1 ,R 2 ,... to be them instead. Perhaps
one worry would be that this would undermine the ideological parsimony of your proposal,
since if we can’t analyze what it is to be a small-ggrounding relation in terms of its ability to
directlyinduce differences in relative fundamentality, what else could this be analyzed in terms
of instead? Perhaps though one could narrow down the class of complete small-gprofiles to
just those that uniformly induce differences in relative fundamentality, and then take the
small-g grounding relations to be whatever relations these complete small-g profiles are“built
up”from. Regarding (ii), and continuing with the proposal above, we could then take big-G
grounding to be necessarily coextensive with the disjunction of complete small-g profiles
rather than the disjunction of conjunctions of relations like constitution, determination, set
membership, etc.
Ifind Skiles’s suggestion here very intriguing and definitely worth keeping in mind
when assessing the tenability of the overall picture I am proposing.
Let us return to the discussion of Wilson (2014). I suspect that Wilson believes that
we antecedently think of what Wilson calls“small-g grounding relations”as inducing
metaphysical explanations, that is, they are the kind of relations one can appeal to
when giving a non-causal and non-teleological explanation of why something is the
case. For example, the fact that a statue has 100 kg of mass is explained by the fact
that a lump that has 100 kg of mass constitutes that statue. I also suspect she is
probably right about this, but since the notion of a metaphysical explanation and the
of which generates facts from universals and particulars and structural universals from unstructured
universals. If both such relations are small-g grounding relations, then we cannot say that small-g
grounding relations always induce relations of relative fundamentality. All the more reason then to deny
that mereological composition is a small-g grounding relation.