Words which designate both themselves and their opposites can be found in many languages around the world. For the Arabic grammarians of the eighth and ninth centuries the existence of such a class of words was a concern as it was a potential hindrance to intelligibility and clarity of semantic interpretation. This paper seeks to engage one thread of that tradition, to reposition it in the context of more modern linguistic scholarship, and to expand on it implications for an understanding of language, all the while hopefully dispelling some of the more wanton psychological misconceptions it has engendered.