Another view of the Gooniyandi “counterfactual” and its implications to the Van linden–Verstraete typology

Van linden and Verstraete (2008) present a compelling typological investigation of counterfactuals in simple clauses, showing that they are usually marked by both a modal element and a marker of tense. They propose that polarity reversal in the counterfactual represents a Generalised Conversational Implicature derived via the Gricean maxim of quantity, applied to a modalised past tense clause; in some languages this has grammaticalised, they suggest, into a coded rather than implicated meaning. In this note I attempt to do two things. First, I point to what seems to me to be a weakness of this pragmatic proposal: it fails to recognise the variety of meanings that can be coded by modal categories of the types employed in counterfactuals. This raises doubts about the universality of the suggested grammaticalisation pathway. Second, I examine one of the languages in their sample, Gooniyandi, and argue that the mono-clausal counterfactual is not double-marked in the way Van linden and Verstraete (2008) presume. It is double-marked paradigmatically rather than syntagmatically. I conclude with an alternative explanation that links counterfactuals to negative constructions, and may provide another motivation for their double marking in some languages.