Music analysis was exclusively conceived as the analysis of a score until computer science technology allowed us to fix in wave form images also the fleeting sounds of a performance. The first and most important principle in this context was to study the systematic deviations that each performer intuitively and more often unconsciously introduced in the graphic conventions of a score. For piano notation they particularly implied a great amount of nuances introduced in tempo and dynamics. Studies like these produced the tradition of computer expressive simulation in piano performance; its results, however, immediately introduced another even more important issue: why were deviations necessary? What were their aims? What was the proper meaning of “expressivity”? Musicological and cognitive sciences are trying to give answers to such questions. First of all a new historical branch of musicology arose in the last decades: the history of music performance, based on collections of a huge number of recordings and more recently of internet documents. But much more important were the studies on the very contents of expression: in music there are aspects, which can be scientifically studied, such as emotions, expectations, movements linked to performance and also to rhythmic and metrical structures, images linked to “intertextual” and “topical” conventions. The task of a performer, like that of an actor who reads a written text, is to highlight by means of particular expressive “deviations” and to give sense to the different aspects of implicit emotions, expectations, movements, images that moment by moment pass and flow in the time of the piece. The case of Liszt’s Réminiscences de Boccanegra is even more complex because of the historical, psychological and cultural relationships between two musicians that had different backgrounds and were active in different musical genres.