Published online by Cambridge University Press: 01 January 2026
Grammaticalization research has led to important insights into the driving processes of innovation and propagation. Yet what has generally been lacking is a principled way of analyzing their interaction. Research into innovation focuses on the role of individual language users and tends to take a more qualitative approach, while propagation is typically studied in terms of the community grammar and tends to be more statistically driven. We propose an approach that bridges the two. Drawing on a much larger historical data set than is commonly done, our study shows how a highresolution analysis of semantic and morphosyntactic behavior can be married to statistics, resulting in a method that measures the degree of grammaticalization at the level of single attestations. We apply this method to the early grammaticalization of be going to INF, showing how a communal increase breaks down into different rates of change in the run-up to, the middle of, and right after conventionalization. Additionally, we trace lifespan change of individual authors longitudinally. While not robustly in evidence, there are hints of postadolescence reanalysis in the run-up generation, and of increased realization of innovative features in the middle generation.
First authorship is shared for this article. Both Peter Petré's and Freek Van de Velde's work has been carried out as part of a research professorship, both of them funded by the Special Research Fund (BOF) from the Flemish government. Peter Petré's work is part of the project ‘Mind-Bending Grammars’, funded by the ERC Horizon 2020 programme (Project ID 639008; https://www.uantwerpen.be/mind-bending-grammars/). Both institutions are hereby gratefully acknowledged. We would like to thank the editors David Willis and Megan Crowhurst, as well as the anonymous referees, for their generous and pertinent comments, from which we have much benefited in our revisions.