Why the verbal counting principles are constructed out of representations of small sets of individuals: A reply to Gallistel