Comparing Word Affect and Tone Affect: Comment on Sun and Cuthbert 2017

Abstract

In the article "Emotion Painting: Lyric, affect, and musical relationships in a large lead-sheet corpus", Sun and Cuthbert (2017) explored the correlations between affect-carrying lyrics and musical features such as beat strength, pitch height, consonance, and mode. Several musical features did indeed turn out to be highly correlated with the affect of the lyrics. However, correlations between other features, particularly mode-related musical features and lyric affect, were either insignificant or even contradicted previous research. In the current commentary, it is argued that the difference between the musical features that show significant correlations and those that do not is that the former have a local musical effect whereas the latter tend to affect the mood of a whole phrase or piece, and that the way Sun and Cuthbert estimate lyric affect for sentences or song may not be appropriate. Furthermore, a few remarks are made about the way Sun and Cuthbert treat multi-syllable words and about some basic assumptions concerning the relation between music and lyrics in a song. Nevertheless, the authors are praised for their innovative and interesting work, while several alternative and additional analyses (for example with scale-degree qualia and syncopations) are proposed.

Popular music lyrics and the COVID-19 pandemic

“…Researchers might also consider the inclusion of how explicitly musical variables (such as dynamic range and melodic complexity) relate to turbulent socioeconomic conditions (North et al, 2018; Simonton, 1987). Since musical features can affect how sentences are interpreted (Schotanus, 2020), consideration of both lyrical and musical variables together would further increase our understanding of this topic.…”

“…While this allows objective analysis of a large set of lyrics, the approach is based on the use of dictionaries, and characteristics such as humor and metaphors are not understood by the software (North et al, 2018). Several studies have found connections between a tally of valence words in written text (e.g., Iliev et al, 2016) and the period in which they were popular; however, Schotanus (2020) suggested that song sentiments are not the sum of word sentiments and require complex computations or participant ratings of whole songs. Moreover, while none of the songs explicitly mentioned COVID-19 (or a pandemic in general), the current approach does not account for the idiosyncratic associations that some people may draw between a particular song and a particular social phenomenon.…”

Examining the Lyrical Content and Musical Features of a Crowd-Sourced, Australian Pandemic Playlist

“…However, the quantitative analyses still relied on computerized analysis of lyrical content and thus may not have fully captured differences between the three sets of songs. As commented on elsewhere, computerized analysis relying on word dictionaries does not capture metaphor and humor (North et al, 2018; Putter et al, 2022), and song sentiment can be more complex than sums of word sentiments (Schotanus, 2020). Future research may quantify differences in lyrics using computational approaches such as data mining and machine learning techniques to further examine these complexities (e.g., Anglada-Tort et al, 2021).…”

“…Thus, we quantitatively compared the lyrical content and musical features of the music within the pandemic playlist (as what people say they were or should be listening to in response to with that of music charting in 2020 and 2021 (reflecting the public's actual listening choices during COVID-19). In response to Putter et al's (2022) suggestion, we wanted to consider both musical features and lyrical content (given that both components can affect the interpretation of music [Schotanus, 2020]). The potentially contradictory predictions arising from two theoretical approaches previously discussed, absence of direct prior evidence, and idiosyncratic nature of the COVID pandemic, meant that we opted to approach the quantitative analyses as an exploratory study guided by the following two research questions: RQ2: How does the lyrical content of the pandemic playlist differ from charting popular music during the COVID-19 pandemic with respect to (a) interpersonal relationships and (b) valence?…”