Terms like anything and anymore have long fascinated linguists. One of the central properties of interest is their apparent reliance on particular semantic contexts, such as negation. For example, there is a clear difference between the sentence “I didn’t see anything” and “I saw anything”, the latter of which is unattested and may be judged unacceptable by English users. Why would these terms be reliant on contexts such as negation in this way? Further complicating the picture, there is variation across dialects in the use and acceptability of the term anymore. For most English users, this term behaves similarly to anything in preferring to occur in contexts like negation. However, some English users in particular regions, such as Central PA, can use sentences like “Football is more popular than baseball anymore”, with anymore occurring in an apparently positive context—so-called “positive anymore”. In this talk, we use theoretical and experimental tools to explore the distribution of terms like anything and anymore, and to understand why anymore—a time adverbial phrase—displays this type of regional variation while anything does not. We further touch on broader issues related to the utility of formal modeling in understanding lexical semantics and semantic variation.