ABSTRACT
While variationist research has accrued an increasing body of evidence that our linguistic habits are more flexible than we would expect under a strict interpretation of the critical period (Sankoff and Blondeau 2007; Wagner and Sankoff 2011, inter alia), the complex conditions which drive linguistic (in)stability across the lifespan of the post-adolescent speaker are much less understood. Indeed, the notion of what exactly constitutes a change in speaker grammar is a contentious one. To date, the majority of panel studies operationalise shifts in proportional frequency as evidence of ongoing change in speaker’s linguistic habits. Only very few studies have probed the extent to which the constraint system that governs the use of individual variables can and does indeed change post-adolescence. Furthermore, the lion’s share of panel research has been conducted on the basis of phonetic change. The question whether later-life change is more likely at a particular level of grammar is still outstanding (MacKenzie 2017: 1). This chapter explores the ramifications of using different quantitative methods for modelling intra-speaker changes on the basis of variables situated at different levels of linguistic architecture (phonology versus morphosyntax). We draw on a small panel sample of six speakers recorded 42 years apart in the North East of England to explore: (i) the differences between heuristics employed for age-related linguistic (in)stability and (ii) the impact of linguistic architecture on the intra-speaker linguistic (in)stability.
