Figure 3

(a) Scatter plot of size of vocabulary versus number of tokens (length) for the 76 composers in the corpus, who are grouped chronologically, as represented by the points of different color (one point is one composer). The year of each composer is the mean between the birth year plus 20 and the death year. Heaps’ law is given by the central straight line and the parallel lines denote one standard deviation \(\sigma _{c}\). Notice that when a limited chronological span is considered, the scattering is reduced. (b) Analog scatter plot for the 9489 individual pieces. Year of a piece is approximated to the year of its composer. The results of the fits are in Table 3 (in Appendix A)