No universals? If there were linguistic universals, what is it they would be doing? Universals would not be mere playthings for typologists – call them “crosslinguistic descriptive generalisations” for respectability, but never mind what they mean. To have or not to have universals is no petty matter: at issue is the human mind and its history. Universals would be enforcing certain options, categorically or preferentially, and ruling out or disfavouring others as linguistic know-how is transmitted across generations and perhaps overhauled or impaired over the life spans of individuals. Everywhere and at all times, they would thus be superintending the construction of mental lexicons-andgrammars on the basis of the linguistic experience of the members of speech communities, in accordance with what the human brain, and those other parts of human bodies involved in the expression and perception of thought for purposes of communication, can and cannot do (well). Without universals, mental lexicons-and-grammars and the speech acts performed accordingly, and minds insofar as they are tied up with language, would be more diverse than they are and conceivably could be.1 Let me illustrate, chiefly, with adpositions and the ordering inside adpositional phrases, one of the structures figuring in Dunn, Greenhill, Levinson & Gray 2011, where the existence of universals has recently been contested, causing a similar stir as the more expansive universals bashing of Evans & Levinson (2009). No statistics or hard words are to follow here; but let’s try to be clear conceptually.