This is for two objects sharing a boundary. Using a negation of topo-overlaps in the definition would create an incoherence when defining 'true-part', since topo-overlaps is a composition of t-part and its inverse, thus t-part would actually be negated, while it is used positively in the definition of true-part. For this reason, 'common-part' from mereology is used in place of topo-overlaps. For naive external connection, see 'touches'. (AG) at the moment I am uncertain about the relevance of this issue for
general mereo-topology.
(<=> (Externally-Connected ?A ?B)
(And (Not (Common-Part ?A ?B))
(Exists (?C ?D)
(And (Boundary ?B ?D)
(Boundary ?A ?C)
(Connected ?C ?D)))
(Connected ?A ?B)))