In recent years quite a lot of attention has been paid to the suprasegmental features of speech. In the field of second language acquisition, by contrast, the study of prosodic systems suffers from a considerable under-representation. Situated in the double theoretical framework of contrastive linguistics and interlanguage analysis, this study investigates the strategies underlying the distribution of pitch accents in L2 Dutch and French as well as the factors influencing them. The ‘Integrated Contrastive Model' used in this research involves four types of comparisons: Dutch (L1) - French (L1), Dutch (L1) - Dutch (L2), French (L1) - French (L2), Dutch (L2) - French (L2). Looking at Dutch and French L1 data, it appears that structural factors have a much stronger influence on the distribution of pitch accents in French than in Dutch, where their position in the utterance is mainly governed by semantic-pragmatic principles. Crucially, this contrast between the learners' L1 and L2 constitutes a major of interference in their inter¬language system. This is reflected by the use of a structurally motivated default pattern in L2 Dutch, whereas Dutch-speaking L2 learners of French have a clear preference for accent patterns reflecting the contextual news value of sentence elements. The systematic use of the French 'arc accentuel' in L2 Dutch results in a relatively high number of contextually infelicitous accent distributions (negative transfer), whereas the use of the pragmatic accent rules of Dutch in L2 French gives rise relatively few accentuation errors. This leads to a significant over- and underuse of some accent distributions instead. Besides those differences, there are also similarities between the two interlanguage varieties under investigation. In both cases, the relative correctness of the accentuation proves to be correlated with the overall quality of the segmentation of the utterance: the more accurate the learners' pausing strategies, the higher the probability that they also produce a contextually ...