• Metaphysics
• Mereology
• Identity
• Composition

It is sometimes stated that parts are identical to the whole they compose. Such thesis, known as “Composition As Identity”, comes in a strong variant, defended for the first time by Donald Baxter in 1988, according to which composition is numerical identity, and in a weak one, proposed by David K. Lewis in its “Parts of Classes” (1991), for which composition is merely analogous to identity.
Despite such difference, what is common to both variants of CAI is the intuition which their supporters rest on, that is the idea that parts and whole consist of the same portion of reality and that, as consequence, given an ontological commitment to the parts, a commitment to the whole does not seem a further commitment. Nonetheless, in the literature on CAI, the notion of portion of reality has been mostly used as a mere stopgap to back the plausibility of the thesis, without being given a clear characterization nor an effective role in carrying out the comparison between composition and identity. The aim of my thesis, then, is to fill this gap and outline an account of CAI wherein more space is left to portions of reality. I think, indeed, that such move might provide the debate about CAI with new responses to the objections and reduce the risk of a deadlock.
To do so, I will build a different meta-ontological framework based on the Finean notion of “interpretational modality”. The idea behind interpretational modality is that an absolutely unrestricted quantification is not possible, for the interpretation of the alleged unrestricted quantifier can always be expanded so as to include further objects in its domain, without that leading to an enrichment of our ontology, and so without that carrying any further ontological commitment. In general, we can read “It is interpretationally possible that A”, as ``there is a relevant possibility of extending the domain of quantification such that A is true''. So, it is not a traditional modality: the number of objects included in the domain of the quantifier does not vary with the different possible worlds. Rather, what varies is the interpretation of the quantifier with respect to the actual world. An interpretational possibility is, then, a possibility for the actual world.
In the light of this, I propose to re-interpret CAI within such framework. In particular, I claim that portions of reality are what exists under the non-expanded interpretation of the quantifier. On the contrary, parts and whole are picked out by an expanded interpretation of the quantifier and, in this sense, are possibile. Namely, given a fixed portion of reality which we refer to by a constant, there is a possibile expanded interpretation of the quantifier under which we can start speaking of the parts and the whole which correspond to that portion of reality. Moreover, by the nature of interpretational modality, parts and whole do not come with a further ontological commitment. Now, what is needed for some parts to be identical to the whole they compose is exactly what is needed for them to compose that whole: both parts and whole, indeed, have to be interpretational possibilities for the same actual portion of reality.