This paper represents a pilot investigation of accentuation patterns produced by advanced Dutch speakers of Italian as a second language (L2). Contrastive accent patterns within phrases were elicited in a semi-spontaneous dialogue entertained with a confederate native speaker of Italian. The aim of the analysis was to compare learners' contrastive accentual configurations induced by the confederate speaker's prime against those produced by Italian and Dutch natives in the same testing conditions. F0 and speech rate data were analysed by applying powerful data-driven techniques available in the Functional Data Analysis statistical framework. Results reveal different accentual configurations in L1 and L2 Italian in response to the confederate's prime. We conclude that learner's accentual patterns mirror those ones produced by their L1 control group (prosodic-transfer hypothesis), although the hypothesis of a transient priming effect on learners' choice of contrastive patterns cannot be completely ruled out.

Index Terms: prosodic transfer, (semi)spontaneous speech, priming, Function Data Analysis, Principal Component Analysis