Predicting the Diffusion of Western Loanwords in Japanese Based on the Apparent-Time Hypothesis

Abstract

This paper aims to predict the diffusion of Western loanwords in contemporary Japanese based on the apparent-time hypothesis. An s-curve model is employed to predict the probability of the occurrence of 17 loanwords as opposed to their native equivalents.

Self-reports in previous questionnaire-based surveys show that loanwords are divided into two groups according to distribution by age. One shows a typical s-shaped curve, with the occurrence of loanwords monotonically increasing from older to younger generations. The other takes an s-shaped curve that ends in a drop in loanword occurrence among those in their late teens and/or twenties.

This delay in acquisition arguably reflects the age-specific domain of use of the loanwords in question. It is predicted that speakers acquire these as their ages increase. Consequently, both types of diffusion should be predictable with one monotonically increasing s-curve model. This raises the question of how the influence of the ‘acquisitional delay’ could be separated from the speakers' birth-year effect without excluding the data of the youngest generations. The problem is shown to be solved by including an ‘acquisition factor’ into the model as one of the independent variables.

The inclusion of this new variable, as a result, brings advantages in a more accurate prediction of change: it enables us to elicit the birth-year effect alone, and in turn, corrects underestimations of the rate of change.