Processing of linguisitic and musical dimensions contained in songs : effects of an expertise

Songs are universal in our lives and also offer an ecological material that links linguistic and musical information. It therefore provides a tool for observing the relationship between music and language.
To examine the specificity of language and music information processing in songs, we studied the influence of expertise (i.e. musicians or speech therapists) on the use of transitional probabilities (TPs) in learning an artificial language (AL) either spoken or sung. This type of implicit statistical learning is especially important in language development (Saffran, Aslin, & Newport, 1996).
We used the paradigm designed by Schön, Boyer, Moreno, Besson, Peretz and Kolinsky (2008). These authors have demonstrated the benefit of structural and motivational proprieties of music contained in the song in the learning process of AL in individuals with no expertise. Compared to a spoken AL, learning was enhanced for a singing AL, especially when the TPs of musical and linguistic materials were congruent.
Our results did not replicate those of this study in participants having no music expertise. However, an interaction occurred between the type of AL and expertise of the participants. Indeed, only music experts seemed to benefit from the congruence between the TPs of musical and linguistic materials in extracting TP-based linguistic units. On the opposite, experts in language seemed to be favoured by the incongruence between the TPs of linguistic and musical information.
These results are discussed in relation to the notions of the specificity of processing of lyrics and tunes in songs.