Negative Concord in Afrikaans: Filling the typological gap

Many languages exhibit Negative Concord (NC), with multiple morphosyntactic instances of negation corresponding to one semantic negation. Traditionally, NC languages are distinguished as Strict and Non-strict (cf. Giannakidou 2000). In the former (e.g. Czech), multiple negative elements may or even must precede the finite verb, whereas in Non-strict NC languages, like Italian, only one negative element may precede the finite verb. In a recent analysis of NC (Zeijlstra 2004, 2008b), NC is analysed as an instance of syntactic agreement between one or more negative elements that are formally, but not semantically, negative and a single, potentially unrealized semantically negative operator. On this analysis, the difference between Strict and Non-strict NC languages reduces to the semantic value of the negative marker: in Strict NC languages, both negative indefinites and negative markers are semantically non-negative; in Non-strict NC languages, by contrast, only negative indefinites are semantically non-negative, negative markers being semantically negative. This analysis predicts the existence of a third type of NC language, namely one where negative indefinites are semantically negative, but negative markers are not. This paper demonstrates that a particular variety of Afrikaans (the standard) instantiates a language of exactly this type: while pairs of negative indefinites always yield a Double Negation reading in this variety, negative markers can be stacked incrementally without giving rise to a new negation.