Abstract
Classical mereotopology is sometimes thought to be represented by General Extensional Mereotopology with Closure Conditions (GEMTC). One reason typically given in favour of GEMTC is its relation to set-theoretic topology. However, the connection primitive in GEMTC lacks an obvious topological interpretation, and the alignment between GEMTC and topology varies across possible interpretations. This paper identifies, among several natural candidates, an interpretation that best aligns GEMTC with topology. Ten possible topological interpretations of mereotopological connection are examined, and for each, we identify (i) the conditions under which topological spaces can provide models of GEMTC, and (ii) the extent to which the definitions of GEMTC agree with their topological analogues in these models. It is observed that when connection is interpreted as the intersection of one set with the closure of another, the non-empty sets of any symmetric topology are a model of GEMTC, with agreement between several mereotopological notions and their topological analogues. This represents a stronger relation between GEMTC and topology than has thus far been observed in the literature. The results of the investigation also bear on issues like Peirce's puzzle, the possibility of external connection between regions, and our intuitive understanding of connection in terms of boundaries.