In this article, we apply the data science technique of sentiment analysis to the lyrics of Bob Dylan.
We can see Bob Dylan’s lexical diversity has increased over time.
Bob Dylan was the most prolific in the mid-1960s and the mid-1970s (with a trough in-between). No surprise there. There is also a surprise peak in 2009. This is because both Together Through Life and Christmas in the Heart came out that year. (Arguably, one might want to exclude Christmas in the Heart, but the sentiments were presumably his even if he didn’t write the words.)
We use two well-known sentiment lexicons:
Each has their advantages and disadvantages. Their results are sometimes inconsistent. Also, it’s arguable whether one should be doing automated sentiment analysis on poetry (or song lyrics). That sort of seems to be missing the point, somehow. Nevertheless, let’s look at some of the results.
It should also be noted that this type of analysis at the word level is a bit shallow and only skimming the surface. For example, in context a word can be negated. That’s not to mention irony or sarcasm in lyrics. However, I do think that unexpected or unorthodox findings can possibly lead to further human analysis and questioning.
I think it’s very interesting that Bob Dylan’s lyrics peaked in positivity in 1980. This was exactly in the middle of his “gospel period”.
Joy peaks in the 1980s (while quality of records dips) and sadness peaks in the 2010s.
Next we look at the relationship between the personal events in Bob Dylan’s life and the sentiments expressed in his songs at the time. Perhaps surprising is that his lyrics are the most positive during his “gospel period”. Perhaps less surprising is that they are also positive during his time of domesticity during the late 1960s.
Since Blood on the Tracks is perhaps Bob Dylan’s most autobiographical album (either that or it’s based on Chekhov short stories), let’s analyze the sentiments in its lyrics. Sadness is the predominant sentiment; not surprising for probably the best ‘break-up’ record of all time.
A look at the sentiment words in “Simple Twist of Fate”.
This article is highly indebted to the techniques described in “Tidy Sentiment Analysis in R” written by Debbie Liske.
My code is available in the bob-dylan-haikus project.