Constructional change across the lifespan of 20 early modern gentlemen

Many Construction Grammarians consider the development of linguistic structure (in acquisition as well in history) to be driven by usage (Diessel 2019). By extension, it is also held that an individual’s constructional inventory of abstract grammatical form-meaning pairings can be updated continually (Bergs 2012: 1637). Yet, while there are ample examples of constructional change across the lifespan of individuals, it has still proven difficult to find instances of categorical, ‘core’ grammatical change, where adults change from “one invariable and uniform grammatical system to another” (Meisel et al. 2017: 30; cf. Raumolin-Brunberg 2005: 47).

This study investigates “how much innovation and change is possible across the lifespan in the domain of syntax” (Petré & Anthonissen 2019). More specifically, this study tests the following hypotheses:

H1: The rate by which individuals use functionally equivalent (or ‘competing’) constructions across their lifespan can change.

H2: The extent to which and direction in which these usage-rates change across the lifespan varies between different individuals.

H3: Adults may ‘participate’ in the diffusion of the new variant to new grammatical contexts, and reach a more advanced stage of the development occurring at the community level.

The case study chosen to investigate H1-H3 is the diachronically unstable variation between the two competing forms illustrated in (1)-(2):

1. Idolatry consists in giving of that worship which is due to God, to that which is not God. (EMMA, Daniel Whitby, 1674)
2. the greatest part of Leviticus is imploy'd in giving Laws concerning Sacrifices (EMMA, Daniel Whitby, 1697)

The study presents an analysis of over 16,000 tokens (taken from the EMMA corpus (Petré et al. 2019) found in the writings of 21 authors, born between 1599 and 1640. Due to issues of quasi-separation (Kimball et al. 2018), the competing constructions in (1)-(2) are taken as the dependent variables in a series of multilevel Bayesian logistic regression model (brms). The results of this study confirm that individual usage-rates of competing construction indeed can (but need not) change across the individual’s lifespan (H1). Furthermore, changes in rates need not be consistently in the direction of the community trend (H2). Finally, while some constraint changes can be attested in the individuals' output, they appear to be very rare.