is in some way primitive. I will then discuss one split between proponents of ground,
namely, whether it is a topic-neutral relation or whether it is defined only on facts,
propositions, or other such entities with similar structure. One proponent of the
former view is Schaffer (2009), who thinks of grounding as a relation that relates
entities of any particular ontological category. Section 8.2 will focus on this view of
grounding. I will argue that ground on this conception can be reduced to either
degree of being or order of being. Accordingly, although talk of grounding in this
manner is perfectly acceptable, at the fundamental level, it is dispensable: it isn’t
needed for saying everything there is to be said about the world in the most
fundamental terms. In section 8.3, I will turn to the idea that ground is a relation
between facts. It is harder to give a reductive account of this notion of ground;
some accounts will be discussed, but perhaps none of them is wholly satisfactory. In
section 8.4, I will discuss the question of the unity of ground, which is a question that
pluralists about grounding face. Finally, in section 8.5, I will discuss whether ground-
ing monists have a reason to posit a further structuring feature, such as naturalness
or degrees of being.
Let’s start with the discussion of whether and in which sense or senses ground
might be primitive. We’ll distinguish ideological primitiveness, methodological
primitiveness, and metaphysical primitiveness.
To say that grounding is ideologically primitive is to say that we cannot give a
reductive definition or analysis of the notion of grounding.^2 This might be because in
principle no such reductive definition or analysis is possible. Or it might be merely
because we lack the conceptual or logical resources to state what is in principle
definable. I don’t want to get too bogged down in what is involved in successful
reductions, but a minimum requirement for a reductive account of grounding is that,
given such an account, any sentence in which a grounding expression occurs can be
replaced with a necessarily equivalent sentence in which no such expression occurs.
(Consider for comparison Lewis’s reductive account of modality, which provides a
recipe for replacing any sentence in which a modal expression occurs with a sentence
in which no modal notion occurs.)
To say that grounding is methodologically primitive is to say that it is dialectically
permissible to appeal to grounding in one’s metaphysical theories without attempting
to define or analyze this notion, regardless of whether it is ideologically primitive.^3 The
basic idea is that if grounding is methodologically primitive, then proponents of
ground do not have the burden of proof with respect to whether the notion of ground
is sufficiently intelligible to use in theorizing; rather, it is those who wish to cast doubt
(^2) Compare with Trogdon (2013b). Schaffer (2009: 364) and Rosen (2010) take grounding to be
ideologically primitive. 3
Vallicella (2002: 43–6) argues that the notion expressed by“in virtue of”is indispensable to
metaphysics. I take it that if a notion is indispensable, it is methodologically primitive.