Musical semantics: Dimensions, processes, and neural correlates 115 Meaning emerging from large-scale relations Musical meaning can also emerge from large scale structural relations (such as rela-tions between phrases, parts, and movements of a work). This holds for all three classes of musical meaning (extra-musical, intra-musical, and musicogenic meaning). With regard to extra-musical meaning, one example is program music fea-turing a number of story components that are related to each other (e. g., symphonic poems such as Strauss’ Till Eulenspiegels lustige Streiche, Dukas' L’Apprenti sorcier, and many others).With regard to intra-musical meaning, one example is that the second theme in the sonata form has a different intra-musical meaning in the exposition than in the recapitulation due to the second theme being in the same key as the rst theme dur-ing the exposition, but not during the recapitulation. That is, the structural relation between rst and second theme, and thus the intra-musical meaning emerging from the second theme, differs between exposition and recapitulation. 8 Davies (1994) stated in this regard that “to understand the musical work is to understand how it is put together” (ibid., p. 48).Another example is that the last movement of a symphony, suite, cantata, etc. has a different meaning in relation to the previous movements compared to when heard in isolation, perhaps because “musical ideas t together – as complementary, or as variations, or as repetitions – so that there is a development or progress of ideas” (ibid., p. 368). With regard to the integration of meaningful information into a larger semantic context, a study by Krumhansl (1996) explored parallels between music and linguistic discourse. According to Krumhansl (1996) “music and discourse both consist of units that have well-de ned beginnings and ends. Topics [or musical ideas] are introduced and developed within these units, with various devices used to move the argument forward. Acoustic cues, such as pauses, pitch contour, dy-namic stress, and rhythmic patterning, serve to de ne these units and highlight cer-tain elements within them” (ibid., p.405).So far, however, such methods have rather investigated how listeners segment musical information. Krumhansl (1996) stated that beginnings of new segments are associated with the presence of “new musical ideas” (ibid., p.427), and that “musical ideas [are] marked by a variety of surface characteristics, such as changes in rhythmic and pitch patterns, register, and texture” (p.427). The speci cation of the semantic processing of such musical ideas remains to be speci ed.Further theoretical accounts With regard to other theoretical accounts on musical meaning, different theorists of-ten deal with different meaning dimensions when referring to musical meaning, and therefore it is not surprising that several theories are not compatible with each other, or that even within one theory different musical meaning dimensions are con-fused. For example, Hanslick’s argument that absolute music is not about expres-8 For a critical account on the perception of large-scale structures see Cook (1987).