**3. From Tropes to Universals**

In BFO, "an entity is anything that exists. BFO assumes that entities can be divided into instances (your heart, my laptop) and universals or types (*heart*, *laptop*)" [8] (p. 6). In other words, for BFO, a type like *heart* is as real as the numerous hearts that instantiate it. The trope ontology is also realist in its stance. However, in the trope ontology your heart's relationship to your body has a distinct existence from Jane's heart's relationship with her body. In contrast, BFO views the property of your heart being a *part of* your body (at some time) as the exact same "part of" property as Jane's heart being *part of* her body7. Given this di fference between the trope ontology and BFO, we now face the question of how to match the di fferent accounts of reality that BFO and the trope ontology present.

There are two problems here: one is how to match tropes to general relations and, secondly, how to match certain particularized tropes to the universals of BFO. Note that we are just aiming for matching one ontology to another and not a justification of one ontology versus the other. However, a matching, if successful, will hopefully underpin a philosophy of reconciliation on these points.

Mereology, the theory of parts and wholes, provides a way to ge<sup>t</sup> from particular relations to general relations. That is, we could match general relations, such as *part of* used in BFO, to sums (collections) of all the particular tropes (such as your heart being part of your body and John's heart being part of his body) used in the trope ontology8. What matching does in this case is to acknowledge that in BFO there exists a *general* relation, but such a general relation does not exist in the trope ontology—and yet, wherever BFO's *part of* is applied, a *part of trope* can be understood to exist (from a tropist viewpoint). This is not an extensional equivalence, because for every *a being part of b*, BFO counts three things (*a*, *b* and *part of*) 9, whereas the trope ontology counts only one (*a being part of b*). However, we can match one such trope to a corresponding BFO structure of *a being part of b*, without having to compromise either ontology's philosophical principles (i.e., without having to commit to unacceptable entities in either ontology).

A similar use of mereological sums also provide a partial pathway to universals. The general extensional mereology discussed above allows for the existence of any collection of entities under some relation. For instance, in the trope ontology, objects would exist as the collection of individual objects, on the assumption that there is a particularized relation of " ... is an object" for every individual object. If we assume that tropes like "x is an object" exist, then the matching to BFO's universal *object* can proceed in the same way as the matching of relations like "part of". However, many such "relations" (e.g., BFO universals such as *object*, *process*, *quality*) do not exist in our sparse trope ontology. On the other hand, we can easily assert in the trope ontology that "x is part of objects", by using the existing parthood relations. What is missing for matching purposes is the assertion that there is a relation *in virtue of which* x is part of objects. This situation can be solved by simply asserting the existence of such relations. Note that we wo not necessarily add the relations as primitives to our ontology. Rather, for certain sums, we add an additional assertion that there exists a relation under which the individuals become a sum—without necessarily defining that relation in a formal sense. In essence, these are "placemarker" assertions. We'll use the following schema to assert the existence of such covering relations, nominating the functional term "universal" as a reserved term for this purpose:
