I don't think the paper cares, since it's about how the fields on the surface behave locally while the other pole needing to exist (and of course it could be split into multiple weaker ones, or smeared out, or whatever) is an invariant that applies to the object as a whole.