On the waning of forms – A corpus-based analysis of decline and loss in adjective amplification

This study takes a corpus-based approach to investigating the decline and near-loss of very as an adjective amplifier in spoken New Zealand English (NZE) based on The Wellington Corpus of Spoken New Zealand English. The paper analyzes the replacement of very as the dominant adjective amplifier across apparent time to provide insights into what factors correlate with and trigger the loss of formerly dominant variants from a variable context. While a wealth of studies has analyzed what causes innovative variants to become dominant, only relatively little attention has been placed on the process of waning. As a consequence, the expansion of innovative variants, e.g. the processes that underlie the expansion in use of really, is well understood; showing that the trajectory of change accompanying the increase of really follows a highly systematic and layered expansion in use. In contrast, questions as to whether the loss of very as a variant in spoken data is equally systematic remain unclear. The results of binomial mixed-effects regression models show that the decrease of very is remarkably uniform and does not parallel the highly systematic and step-wise trajectory of innovative incoming variants. This lack of an ordered heterogeneity that accompanies the retreat of very raises questions about the systematicity of loss as a linguistic phenomenon more broadly.