The seamlessness of grammatical innovation: the case of be going to (revisited)

In this paper, I revisit the continued debate surrounding manner of grammatical innovation, i.e. whether it is abrupt or gradual. I show that the debate is complicated by different diagnostics for manner and argue that it is best understood in terms of degree of similarity (how similar the innovative use is to existing uses of a construction). However, even when adopting degree of similarity as a diagnostic, approaches differ with regards to how similar they find innovative and existing uses. The gradualness account argues that innovative uses are similar but distinct from existing uses, as they imply a new form-meaning pairing. A seamlessness account instead argues that innovative and existing uses are so similar that no new form-meaning pairing is required. I develop seamlessness into a theoretical position for semantic innovation in grammaticalization, which holds that grammatical innovations are maximally similar to existing uses i.e. they exhibit considerable conceptual overlap and the existing use is semantically underspecified. Seamlessness is empirically tested using semantic innovations in be going to as a case study.