The effect of age of acquisition: Partly frequency related, partly frequency independent

Abstract

A review of multitask investigations on the locus of the age-of-acquisition (AoA) effect in the English, Dutch, and French languages reveals two main findings. First, for most tasks there is near perfect correlation between the magnitude of the AoA effect and the magnitude of the frequency effect, even though the stimuli were selected so that both variables were orthogonal. This frequency-related AoA effect is as large as the frequency effect, despite the fact that the range of AoA values is more restricted than the range of frequency values. Second, a frequency-independent AoA effect is observed in object naming and word associate generation. Different explanations of the frequency-related and the frequency-independent AoA effects are reviewed and evaluated.

Notes

¹β-values or standardized regression coefficients are obtained when the regression analysis is done on the z-values of the dependent and the independent variables. In this way, the values are not influenced by differences in the variance of the independent variables (see Aron & Aron, 2003, pp. 114–124, for a good explanation).

²Notice that this exactly may be what the tachistoscopic presentation of Moore et al. induced: A need to covertly name the briefly presented stimulus, so that it could be kept in short-term memory.

³Interestingly, Levelt and colleagues never refer to their stimuli as polysemous words. They always refer to them as homophones and give examples of English heterographic homophones when they describe their manipulation (e.g., they say the words were of the type wee–we, moor–more, whereas in reality their stimuli were of the type bank, which can be translated in two different ways in Dutch, depending on whether it refers to a river or to money).