The Difficulty of Measuring Musical Quality ( and Quantity ) * Commentary on R . W . Weisberg

In a recent article in Psychological Science, Weisberg (1994) examined the hypothesis that creative individuals suffering from manic-depressive disease are not only more productive during hypomanic phases (which they commonly are) but also produce works of higher quality than during normal or depressed periods. As his test case he took the composer Robert Schumann (1810-1856) who suffered from a bipolar affective disorder (Slater & Meyer, 1959; Ostwald, 1985; Jamison, 1993) and who left extensive records of his mood swings in letters and diaries. A plot of the number of Schumann's compositions by year of completion (Weisberg's Figure 1) reveals two periods of particularly intense activity: the years 1840 and 1849-1851, especially 1849. According to Slater and Meyer (1959), the years 1840, 1849, and 1851 coincide with hypomanic periods in Schumann's life. Years classified as mostly depressive periods, by contrast, show ~ry low productivity. This correlation between mood state and productivity seems robust and plausible, despite the crudity ofthe analysis. As Weisberg acknowledges, neither Schumann's periods of increased or reduced productivity nor his corresponding mood states necessarily extended over full calendar years, and there were other factors that impinged on his productivity, such as editorial work and periods of travel (see Ostwald, 1985). The measure of compositional quantity is also problematic. Weisberg's definition of "composition" (llot stated,.but following Slater & Meyer, 1959) is music published with a single opus number. However, works with different opus numbers may vary widely in size and complexity. InSchumann's case, they range from single short pieces such as the Toccata, op. 7, to extended works such asthe Three String Quartets, op. 41. Alternative measures of compositional quantity worth considering would be the number of movements of comparable length and complexity, or even the number of distinct musical themes ("ideas")) Defining musical quantity is nota trivial problem. There are also dangers in assuming that a composer's published works fully represent his creative output. Composers, like scientists,generally publish only the works that meet a certain self-imposed criterion of quality. However, that criterion may vary with the cycles of manicdepressive disease. IfSchumann lowered his criterion for what he deemed publishable during his hypomanic periods, thenthe data would be biased against Weisberg's hypothesis that average compositional.quality increases during these periods. The mosts~rious fla\V of Weisberg's research, however, is his measure of quality. To assess the relative quality ofSchumann's works in different years, Weisberg counted the number of recordings listed in two popular record guides. He placed the responsibility for this choke on an