Taking a look at how subgroups are defined in
set.mm (and submonoids, and I presume other sub-structures), I was surprised by what I found. It seems to be a mapping from a group to a (not necessarily proper) subset of the elements of the group. My (possibly naive) expectation was for SubGrp to be a relation on groups so that "A SubGrp B" means that A is a subgroup of B.