Unearthing Tree Symbolism in Song: A Sentiment Analysis
How societies communicate about nature can shape the way that they interact with it. Messages contained in music are especially interesting to study because of the unique ability of sound and language to alter moods and/or induce physiological reactions. Research on cultural values in music is growing but studies on environmental themes are scarce despite pervasive natural symbolism in songs. Historically, most species of tree have gained a symbolic meaning in part based on their physical characteristics and the various ways they are used by humans (e.g., for construction or for medicine). The overall goal of this thesis was to understand the emotional sentiment associated with tree symbolism in English-language songs. To quantitatively investigate these associations, I assembled a corpus of 1335 songs that use common North American tree names in lyrics. Songs were categorized into two groups based on the evolutionary history of the tree used in lyrics. Trees are either angiosperms (typically flowering, fruiting, and deciduous) or gymnosperms (typically cone-producing and evergreen). I extracted lyrical sentiment (e.g., positive words) and musical qualities (e.g., tempo) of each song for analyses. Lyrically, I found that angiosperm songs were more likely to contain positive words and less likely to contain negative words than gymnosperm songs. Additionally, angiosperm songs were more likely to contain words of anticipation, joy, and trust, while gymnosperm songs were more likely to contain words of anger, fear, and sadness. Musically, gymnosperm songs had higher energy and tempo than angiosperm songs. Exploring these data further at other levels of taxonomy would likely provide higher resolution of thematic content. These results provide support for the idea that the sentiments we associate with trees are related to the tree’s evolutionary history which is important because our sentiments have the potential to affect how we connect to and interact with environments.
Bayliss_JREMThesis_07_20_21_FINAL.docx
3.33 MB
Microsoft Word XML
139fff4574780a19fa4340f4a95edc67
auto_convert.pdf
1.62 MB
Adobe PDF
b4912cfc82a83d3024b85545cbdf8f05