Correction by adversative and additive markers

Corrective uses of adversative markers like but, as in John isn’t going to Paris, but to Berlin, have proved rather difficult to capture in a unified theory of adversative markers, whereas corrective uses of additive markers, as in John is going to Berlin, and not to Paris, have been almost entirely ignored in theoretical semantics and pragmatics. These uses are taken under closer consideration in this paper, with special focus on the phenomenon I will refer to as (a)symmetric correction. I propose the following generalisation. Adversative markers are asymmetric in their corrective uses (e.g. the English but). That is, the first conjunct of but must be negated, while the second is positive. If the order of the negative and the positive conjunct is reversed, the corrective reading is not available for but, though it can be recovered if but is replaced by and or left out altogether. In contrast, additive markers are symmetric in this function. If a language standardly employs an additive marker to express correction (e.g. the Russian a), the order of the negative and the positive conjunct does not affect its corrective interpretation. The present paper develops a unified account of the semantics of but which accommodates its corrective uses and explains the above mentioned asymmetry. The proposed solution has non-trivial consequences for a general theory of additivity and adversativity, in particular, for the ongoing debate which function of but is the most basic, ‘denial of expectation’ or ‘formal contrast’.

► The paper gives a unified account of corrective and adversative uses of but.
► Describes the phenomenon of (a)symmetric correction,
► Explains corrective asymmetry as a consequence of argumentative asymmetry of but.
► Makes typological predictions wrt the (a)symmetry of correction markers recruited from the pool of adversative and additive markers.