Important for a high quality musical performance is a good insight in the way the composition you want to play, is constructed. However, since the beginning of the 20th century, composers develop more and more their own musical language, often even different from one piece to another. In the end you need almost as many models as there are compositions. That's why we want to go looking for a usable meta-model for musical analysis. This 'universal' analysis model should allow us - as a kind of an entrance model - to approach any composition without previous knowledge of the way it was conceived. Using some well defined procedures this model should be capable of indicating which of the existing analysis models could be most appropriate for a specific piece. We intend to use the newest insights in as diverent disciplines as linguistics, mathematics and artificial intelligence in order to get a deeper understanding of the essence of music as an art form.