EVEN and negative bias in questions revisited
Abstract
This paper discusses three Greek lexical items meaning EVEN in polar questions, focusing on the issue of negative bias. It is shown that negative bias does not arise with positive EVEN (a fact undermining the viability of proposals that rely on this option), or NPI EVEN, but only with what is identified as flexible-scale EVEN. The discussion has two implications. First, it provides an argument for lexical ambiguity of even. Second, it offers a novel way to reconcile the fact that strong polarity items (Pis) create negative bias, but any doesn't, by arguing that only strong Pis contain the flexible-scale EVEN.
