**6. Conclusions**

The trope ontology is oriented towards building up an ontology from individual cases, rather than focusing on universals. Universals, in the sense of entities that form collections under some covering relation, emerge in the trope ontology as entities that need to be talked about as collections and need to be distinguished for some reason from other entities. This reflects more of a "bottom up" approach than BFO. However, it also comprises an approach of developing ontologies of bounded scope. Limiting the ontological scope is something that the trope ontology has in common with BFO.

Matching universals in BFO to corresponding entities in the trope ontology is simple, as long as we formally recognize the existence of universals in the first place (i.e., as covering relations of mereological sums). From that point onwards, being a universal in the trope ontology is essentially a property that certain mereological sums have. If a universal can be wholly defined in terms of parthood (or possibly with additional relations that also exist in a trope ontology for a specific domain), then the universal can be defined in the trope ontology directly using the corresponding covering relation of a mereological sum. Alternatively, the existence of the universal can be asserted without formal definition of its trope relation, perhaps with annotation of the informal definition, as is the case with various universals in BFO.

The matching between the primitive causal relations in the trope ontology and *occurrences* in BFO is more complicated. What is available in the trope ontology is the ability to declare causal chains as entities, where those entities match BFO's occurrences. However, temporal measures like "time" are secondary inferences in the trope ontology. For example, a point in time would be derived from the length of a causal chain. Therefore, while causal chains can be classed as occurrences, some relations like "being part of X at time t" require specific assumptions about the measurement of time. Moreover, in the trope ontology a continuant is the mereological sum of all its parts, regardless of causal relations. This is not the case in BFO, because some mereological relations are only specified with explicit time arguments. As such, the matching of continuants that fall under causal relations will need to exclude causally dependent parts.

In general, the trope ontology can relatively simply match to universals in the Basic Formal Ontology, because universals have an extension that can be represented by corresponding individuals in the trope ontology, and because the trope ontology allows for mereological sums under some covering relation, where that covering relation matches the universal. However, since the trope ontology has causal relations that are not part of BFO, the matching of occurrence properties and relations is not as straightforward. With regard to occurrences, the trope ontology can be seen as an ontology of different granularity that can be matched with BFO, but where the matching requires assumptions about the nature of time.

**Funding:** This research received no external funding.

**Acknowledgments:** My thanks to Tim van Gelder for helpful review comments. My thanks also to the anonymous reviewers for the many insightful comments.

**Conflicts of Interest:** The authors declare no conflict of interest.
