Paradigmatic predictions of the Scalar Model

The Scalar Model aims to explain what polarity items are and why they should exist. It also makes clear predictions about what sorts of constructions can be polarity sensitive and where they might be found in the lexicon. If all polarity items are scalar operators, it follows that polar sensitivity can only arise in scalar semantic domains and that all polarity items must profile an entity against a scalar base. Given the diversity of polarity items in both form and meaning and both within and across languages, this might seem an unlikely, or even quixotic, hypothesis. But scalar reasoning is itself a broad and abstract phenomenon, and given its pervasiveness in language and cognition and the ease with which almost anything can be construed against a set of alternatives, this is really a rather weak claim. However, if polarity items really are all conventional expressions of rhetorical affect – of emphatic and attenuating i-values – then they should arise in contexts where there are good pragmatic reasons for expressing such a rhetorical stance. Furthermore, if sensitivity really is an effect of frequently felt pragmatic needs, then there should be regular patterns in the types of expressions which become polarity sensitive and the types of functions these serve, both within and across languages. Thus, polarity sensitive constructions are likely to have sensitive synonyms (or near synonyms) within a language and sensitive counterparts in other languages.