This book addresses students and researchers of both phonetics and phonology, and the semantics and pragmatics of discourse. It employs an autosegmental-metrical model of intonation to investigate the marking of aspects of information structure, concentrating on the Given-New dimension. It begins with an overview of the state of the art in the areas of intonation and information structure, and, since the term 'Givenness' has been used in the literature in diverging ways, provides a model of 'Givenness proper', focussing on the cognitive states of discourse referents, and how these states are expressed through the choice of words and their prosody. The empirical evidence provided here is based on German. It comprises the analysis of a read corpus and two perception experiments which show that the dichotomy of 'accented' versus 'uncaccented' corresponding to 'New' versus 'Given' information is inadequate. In fact, there is evidence that a range of pitch accent types (including deaccentuation) can be mapped onto the gradient scale of Givenness degrees, with the pitch height on the accented syllable being the determining factor.