Subjective accent strength perceptions are not only a function of objective accent strength. Evidence from Netherlandic Standard Dutch

• Naïve listeners can recognize three regional accents of Netherlandic Standard Dutch.
• Naïve listeners agree in their subjective ratings of the strength of these accents.
• Accent strength can be objectively measured on the basis of vowel formants.
• Subjective ratings and objective measurements of accent strength strongly correlate.
• But: subjective accentedness ratings are co-determined by speaker prestige variables.

This paper compares subjective ratings and objective measurements of the strength of regional accents of Netherlandic Standard Dutch. In a first experiment, 21 respondents rated 80 short samples of spontaneous speech produced by male teachers of Dutch originating from four areas in The Netherlands; samples were rated in terms of the expected regional origin of the speaker and the strength of his accent. In a second experiment, we computed the phonetic distance between two accent benchmarks (the Randstad accent, the allegedly lightest accent of Netherlandic Standard Dutch, and the Randstad speakers rated as the mildest in the first experiment) and all other speakers’ accents. The subjective ratings were shown to correlate strongly with the phonetic measurements, but additional regression analysis revealed that subjective ratings of accent strength were also impacted by a number of speaker prestige variables. This indicates that the intrinsically quantitative process of accent strength perception (partly) depends on qualitative evaluations of speaker personality. Recommendations are made for a valid implementation of (native) accent strength in speaker evaluation experiments.