Acoustic Constraints and Musical Consequences: Exploring Composers' Use of Cues for Musical Emotion

Abstract

Emotional communication in music is based in part on the use of pitch and timing, two cues effective in emotional speech. Corpus analyses of natural speech illustrate that happy utterances tend to be higher and faster than sad. Although manipulations altering melodies show that passages changed to be higher and faster sound happier, corpus analyses of unaltered music paralleling those of natural speech have proven challenging. This partly reflects the importance of modality (i.e., major/minor), a powerful musical cue whose use is decidedly imbalanced in Western music. This imbalance poses challenges for creating musical corpora analogous to existing speech corpora for purposes of analyzing emotion. However, a novel examination of music by Bach and Chopin balanced in modality illustrates that, consistent with predictions from speech, their major key (nominally "happy") pieces are approximately a major second higher and 29% faster than their minor key pieces (Poon and Schutz, 2015). Although this provides useful evidence for parallels in use of emotional cues between these domains, it raises questions about how composers "trade off" cue differentiation in music, suggesting interesting new potential research directions. This Focused Review places those results in a broader context, highlighting their connections with previous work on the natural use of cues for musical emotion. Together, these observational findings based on unaltered music-widely recognized for its artistic significance-complement previous experimental work systematically manipulating specific parameters. In doing so, they also provide a useful musical counterpart to fruitful studies of the acoustic cues for emotion found in natural speech.

Keywords: acoustics; communication; composers; corpus analysis; emotion; language; music.

Figure 1

Literature frequently played on the xylophone rarely exhibits minor keys. This sample contains frequent use of the minor mode for marimba, with 60% of the tonal pieces exhibiting minor keys (ignoring marimba solos with no tonal center). In contrast it contains few instances of minor key xylophone pieces, which appear only twice within this survey.

Figure 2

Major key pieces are higher and faster than minor key pieces. The 36 major key pieces exhibited faster rates of attack (left) and higher pitch heights (right) in major vs. minor keys. Data are displayed for the first eight measures of each piece. Error bars represent 1 standard error about the mean.

Figure 3

Trade-off of cues across the corpora. Bach's Preludes exhibited the largest timing difference between major and minor pieces, but the smallest pitch difference. Conversely, Chopin's Preludes exhibited the largest pitch difference but the smallest timing difference. Data are displayed for the first eight measures of each piece. Error bars represent 1 standard error about the mean.

Figure 4

Major key pieces are faster than minor across all editorial tempi assessed. Differences in recommended tempi for individual pieces varies considerably for Bach's Well Tempered Clavier. Despite these differences, the difference in major vs. minor key attack rate is consistent amongst each of the 7 sets of editorial tempi outlined by Willard Palmer.