Production of Phrasal Prominence in English and Bulgarian Bistra Andreeva 1 , Snezhina Dimitrova 2 and William J. Barry 1 1 Institute of Phonetics, Saarland University, Germany 2 Department of English and American Studies, University of Sofia andreeva / wbarry@coli.uni-saarland.de, snezhinad@hotmail.com Abstract We examine the production of different accentuation levels in English and Bulgarian on the one hand, and in the English speech of Bulgarian learners of English on the other hand, our basic presumption being that the control of accentuation is subject to language-specific strategies. An immediate implication is that this interferes with our production and perception of accentual prominence in another language. The acoustic-phonetic properties of words spoken with three different levels of accentuation (de-accented, pre-nuclear and nuclear accented in broad-focus and nuclear accented in narrow-focus) were examined in question-answer elicited sentences produced by English and Bulgarian speakers. For detailed comparison, iterative imitations (on the syllable “da”) were also produced. Normalized parameter values allow a comparative weighting of the properties employed in differentiating the three levels of accentuation. Clear differences were found between English and Bulgarian in the weighting hierarchy of the acoustic properties strategies. Key words: phrasal prominence, accentuation, English, Bulgarian 1. Introduction In the process of communication, we highlight the most important pieces of information in a number of ways. The purpose of such highlighting is to draw the listener’s attention to that part of the message which we, as speakers, consider to be the most important. Some of the ways to do this in speech are to make the relevant part longer, stronger, to say it on a higher, lower, or changing pitch, and to preserve the quality of its vowel. As a result, that part of the message is perceived as being more prominent than the rest of the utterance. The acoustic cues to greater or lesser perceptual prominence are all present in the speech signal at any time, and are known to co-vary to different degrees in both production and perception (Ladefoged 1967). At the same time, it has been claimed that these characteristics are utilized differently, with certain properties being exploited to a greater degree in some languages than in others. The prominence patterns typical of a given language are also closely related to speech rhythm, having a strong effect on the type of rhythmical organization in that language. Thus, Jakobson, Fant and Halle (1952) claim that the stress pattern of the native language affects listeners’ responses even to non-speech stimuli. For example, knocks produced at equal intervals, with every third knock louder, are said to be perceived by a Czech as groups beginning with the louder knock, and by a Frenchman as a group with the louder knock in final position, which reflects stress patterning typical of these two languages with fixed word stress. For languages with free stress, on the other hand, it is sometimes maintained that if stress is primarily signaled by duration, then it is perceived as final in a stress group, whereas if prominence is achieved by means of fundamental frequency variation or intensity, then it is heard as beginning the stress group (Lea 1977).