The main goal of the proposed project is to determine the extent to which individual speakers economize their mental grammar by reinterpreting local patterns as belonging to more widely applicable grammatical rules, as manifested in the further generalization of these local patterns. Conversely, analysing this will also lead to insight into which patterns in individuals behave as niches resisting such a realignment to more regularity. These goals tie in with and test psychological assumptions that abstract rules and concrete exemplars (instances accessible in memory) are simultaneously at work in grammatical cognition.
The case study that will be analysed to reach this goal is that of the introduction and diffusion of do as a grammatical word in Early Modern English (ca. 1550-1700). Before this period, do was largely limited to its lexical use, illustrated in (1). During this period, however, do was increasingly more often used in, among others, questions (2a) and negative statements (3a), as an alternative to the older structures (2b) and (3b).
(1) I already did my homework.
(2) a. Do you love me? b. I do not love you.
(3) a. Love you me? b. I love you not.
By the end of the 17th century do had become more or less obligatory in these uses. As such, it is one of the shibboleths of the English language, not only distinguishing it from other Germanic languages, but exploiting a grammatical option that is very rare across the worlds' languages.
The project focuses on the stages of the diffusion of do, where do was already established as an option, but much variation was to be found in the speech community. This poses the fundamental research question of which factors guide this variation at the level of individual speakers. It is examined if, and to what extent, the choice for do is primed by frequent use of modal auxiliaries (will, shall, may, must, can), which typically appear in similar syntactic environments. Together, they could be seen as syntactic markers of all non-neutral statements, i.e. statements that are not (non-emphatic) declaratives. Additionally, it is examined in what contexts patterns like (2b) and (3b) resist the insertion of do the longest. Together, these analyses are expected to considerably advance our knowledge of how much change is possible in abstract cognitive schemas such as grammar across the adult lifespan, and by what kinds of existing regularities this change is guided.