In animal signaling theory, ethologists make the distinction between that which is intended to clearly convey information to the observer (a signal) and that which is not (a cue). When examining qualities associated with “sad” affect, Kraepelin (1899/1921) observed that the speech of depressed patients was slower, quieter, lower in overall frequency, more monotone, and more mumbled articulation. More recently, Erickson, et al. (2006) have included darker timbre. These acoustic features have also been observed in nominally “sad” music (Post & Huron, 2009; Turner & Huron, 2008; Huron, 2008; Huron, 2008; Schutz, et al., 2008). One plausible explanation for the association between acoustic features and the expression of sadness might simply be low physiological arousal. Low arousal is also linked with low acetylcholine levels which are then associated with the previously listed acoustic features (Huron, 2012). Additionally, Huron (2012) suggested that the expression of sadness is consistent with an ethological cue. If this is the case, then we can expect similar acoustic features to be associated with “sleepy” and “relaxed” music.
This thesis examines that conjecture with two studies, a production-perception study and a correlational study. The production-perception hypothesizes that participants will be unable to distinguish “sad” musical expressions from “sleepy” or “relaxed.” Participants generate stimuli by manipulating tempo, intensity, overall frequency, pitch
range, sustain, timbre, interval size, and modality for nine emotions, including the low-arousal states of sad, sleepy, relaxed. The directionality of the normalized scores for the low arousal states are as predicted. The perception study presents these participant-generated stimuli in addition to “average” stimuli for each emotion to a new group of participants. Their task is to identify which emotion the given stimulus best represents. This study is still on-going.
The second, correlational study consists of three parts and is ongoing. For all three parts, the dependent measure is syllable rate calculated using annotated audio recordings. In the first study, the syllable rates of randomly selected minor-mode songs are compared to year-matched major-mode songs from the Billboard Top 40 (Glenn Schellenberg’s list). If the minor mode is associated with sadness (Hevner, 1935; Post & Huron, 2009), those songs will exhibit a slower syllable rate than major mode songs. In the second study, the syllable rate of songs with lyrics rated by the author as expressing “sadness” is compared to those expressing “happiness.” The hypothesis is that songs deemed “sad” will have a slower average syllable rate than those deemed “happy.” Finally, the third study examines the relationship between the arousal and valence expressed by the lyrics and the syllable rate of the song. This time, sixteen participants will rate lyrics for negativeness, energy, sad, and happy. Now, the hypothesis is that that the level of perceived arousal expressed by the lyrics will be the biggest predictor of syllable rate rather than valence. Taken altogether, these studies will reveal more about the role of arousal in the expression of sadness