Things You Probably Didn’t Realize About Taylor Swift’s Songwriting

The words we use in daily life reflect what we are paying attention to, what we are thinking about, what we are trying to avoid, how we are feeling, and how we are organizing and analyzing our worlds.

-Tausczik & Pennebaker, 2010

Taylor Swift is an American singer-songwriter. Swift is known for her narrative songwriting, which often centers around her personal life. Although she has been pegged by the media as the “psycho serial dater girl”¹, Swift’s discography explores a range of subjects beyond just love and romance, including parent-child relationships, friendships, alienation, fame, and career ambitions. Swift famously writes or co-writes all of her songs; therefore, it follows that the words she uses in her lyrics should provide insight to her innermost thoughts and emotions. In this exploratory data analysis, I use computational text analysis on Taylor Swift’s corpus of lyrics to unlock those insights.

Motivation

Previous work done by Yla R. and Tausczik & James W. Pennebaker² has shown that our word choices, even ones as seemingly inconsequential as the pronouns we use, can reveal a lot about our psychological states. We can apply this idea to analyze the discography of one of the most influential artists of our time, Taylor Swift. In her Netflix documentary, Miss Americana, Swift acknowledges her place in the music industry, explaining that her niche that sets her apart from everybody is her unique storytelling abilities.

“I know that without me writing my own songs, I wouldn’t be here”

Anyone listening to Swift’s music will be able to feel the distinctly personal stories that she conveys in her lyrics. Of course, on the surface, Swift’s lyrical content is inherently confessional. She talks about some of the most intimate, private parts of her life (famously, her breakups). However, as demonstrated by research in the world of computational linguistics, there’s so much more waiting underneath the surface. Using cutting-edge computational text analysis methods, we can uncover what makes the mega pop star’s songwriting so compelling and explore how she has evolved as a writer over the course of her eight studio albums.

Dataset

I collected and prepared my own dataset for this project. I began by using Genius’s API to collect data for 134 Taylor Swift songs. This only includes songs released on Swift’s first eight studio albums (the data was compiled in October 2020, meaning folklore was the most recent album at the time). This includes deluxe tracks. However, I excluded remixes and demo tracks. For instance, the deluxe edition of Swift’s 2012 album Red includes the tracks State Of Grace and State Of Grace — Acoustic Version. Given that I am analyzing the lyrical content of each track and these two tracks are comprised of the same exact lyrics, the acoustic version is excluded from the dataset.

Once I had the lyrical data, I used LIWC2015 (gold standard in computerized text analysis) to calculate 73 different dimensions for each song. The way that the LIWC program works is fairly simple. It reads a given text (in this case, a Taylor Swift song) and counts the percentage of words that reflect different emotions, thinking styles, social concerns, and even parts of speech. For example, LIWC characterizes 4.72% of the words in Swift’s song Blank Space as exhibiting positive emotion. So, for the dimension posemo, Blank Space would have a score of 4.72. (The song would have scores for the other 72 dimensions as well.)

I manually added the album, track number, and release date of each song for organizational purposes.

The way Swift uses words like “I” and “ME!” matters more than you think.

Tracking attention can reveal a surprising wealth of information about a person’s priorities, intentions, and thoughts. Furthermore, we can use language to track people’s attention. Content words³ can explicitly reveal where individuals are focusing. This is intuitive: if a songwriter, like Taylor Swift, is thinking about love and romance, she will refer to them in her conversation or writing (i.e. her songs). We can also gather information regarding attentional focus from function words⁴, such as personal pronouns. One research paper found that people experiencing physical or emotional pain tend to have their attention drawn to themselves and subsequently use more first-person pronouns such as “I”, “me”, and “mine”⁵. Another study revealed suicidal poets use more first-person singular pronouns in their published works than non-suicidal poets⁶. This difference may show an attentional difference (i.e. more self-focus in response to emotional pain) or it may indicate a thinking pattern conducive to experiencing depression.

From the visualization below, we can see Swift uses the highest frequency of first person singular pronouns in her three latest albums. In fact, we see the sharpest increase in between the albums 1989 and reputation. The reputation album was released in 2017, one year after her very public feud with celebrities Kim Kardashian and Kanye West. In her documentary, Miss Americana, Swift opens up about this being a dark, isolating chapter in her life where she was more focused on herself in response to emotional distress. Therefore, it’s unsurprising that we see the same attentional focus, that is, an increase in first person singular pronouns, in her album reputation.

I was curious, though, why Lover ranked the highest on the first person singular dimension. Defined by Swift as “a love letter to love itself”, Lover celebrates the full spectrum of love, using light and bright tones that depart from the dark nature of its predecessor, reputation. Could it perhaps be the album’s controversial lead single, ME!, whose title is literally an exclamatory first person singular pronoun?⁷ I decided to create another visualization to uncover the 25 songs with the highest frequency of first person singular pronouns:

The songs on this list span all eight albums, but any experienced Swiftie would tell you: most of these are not just happy love songs. The songs from Lover are colored pink and most of the tracks that made the cut are consistent with the hypothesis that we see more self focus in response to emotional pain:

• Death By A Thousand Cuts illustrates a heart-wrenching breakup,
• The Archer delves into Swift’s propensity to self-doubt,
• The Man echoes Swift’s frustration with gender bias in the music industry and society,
• and Afterglow narrates Swift taking responsibility for struggles in her relationship.

It is important to note: while the frequency of first person singular pronouns psychologically correlates to emotional distress, the relationship is not causal. Songs like I Think He Knows and Lover, have high self attentional focus, but in these cases the focus stems from self-confidence and strong association with a romantic partner, respectively.

When artists, like Taylor Swift, write openly about their life and specifically their emotional turmoils, it signals a certain level of vulnerability. I’d argue Swift’s deeply personal lyrics that so many fans seem to connect with are what make her stand out as an artist.

Swift’s music is becoming more and more analytical.

Whereas personal pronouns can reveal the subject of attention, analyses of the tense of common verbs can provide insight to the temporal focus of attention. Studying attention also gives us a deeper understanding of how people are processing a situation or event.

I found a few interesting features in this visualization. The folklore album has the highest past focus and the lowest present focus among all the albums. Past focus typically points to more analytical language. In a personal essay explaining the album, Swift writes: “I found myself not only writing my own stories, but also writing about or from the perspective of people I’ve never met, people I’ve known, or those I wish I hadn’t.” Therefore, in this case, verb tenses could indicate increased psychological distance and a higher degree of resolution. This is in contrast to her earlier albums like Fearless and Speak Now where the narrative was almost exclusively driven by her own personal life. Moreover, the entirety of folklore was written in 2020 during Swift’s quarantining for the COVID-19 pandemic. Present focus psychologically correlates with “living in the here and now”, which can be a difficult feat in the midst of involuntary self-isolation.

The temporal focus of attention is intimately related to whether a text is more analytic or authentic, which LIWC measures using specialized summary variables. Per the LIWC website, “Each of the summary variables are algorithms made from various LIWC variables based on previous language research. The numbers are standardized scores that have been converted to percentiles (based on the area under a normal curve) ranging from 0 to 100.”

The authenticity variable seems to ebb and flow, but it appears the level of authenticity could be linked to genre. Taylor’s writing is most authentic when she’s producing country music, that is, albums like Fearless and Speak Now. We see a drop off after her transition to pop with albums like 1989, reputation, and Lover.

With regards to analytical thinking, we see an unmistakable increase in Taylor’s discography. There are numerous possible explanations for this trend, but perhaps the most obvious would be that Taylor has amassed more life experience, grown in maturity, and evolved as a songwriter. The Vox podcast, Switched On Pop, suggests that Swift is entering the “avant-garde” phase of her career, characterized by a style that is experimental and unorthodox (at least compared to what we’ve seen from Swift previously). For better or worse, Taylor is no longer the bright-eyed country starlet she once was.

Footnotes

¹ During a performance at the Grammy Museum, Swift explains her inspiration for the song Blank Space as a response to the media calling her a “psycho serial-dater girl”.

² Tausczik YR, Pennebaker JW. The Psychological Meaning of Words: LIWC and Computerized Text Analysis Methods. Journal of Language and Social Psychology. 2010;29(1):24–54. doi:10.1177/0261927X09351676

³ Content words are usually nouns, regular verbs, and many adjectives and adverbs. They convey the content of a communication. Consider the lyric from Swift’s song Blank Space:

’Cause darling, I’m a nightmare dressed like a daydream

The content words are “darling”, “nightmare”, “dressed”, and “daydream”.

⁴ Function words are words whose purpose is more to signal grammatical relationship than the lexical meaning of a sentence. In the lyric above, the style words are “I’m”, “a”, and “like”. Although we have almost 100,000 English words in our vocabulary, only about 500 (or 0.05%) are style words. Nevertheless, style words make up over 50% of the words we speak, hear, and read.

⁵ Rude, S. S., Gortner, E.-M., & Pennebaker, J. W. (2004). Language use of depressed and depression-vulnerable college students. Cognition and Emotion, 18(8), 1121–1133. https://doi.org/10.1080/02699930441000030

⁶ Stirman SW, Pennebaker JW. Word use in the poetry of suicidal and nonsuicidal poets. Psychosom Med. 2001 Jul-Aug;63(4):517–22. doi: 10.1097/00006842–200107000–00001. PMID: 11485104.

⁷ Fun tidbit: The song ME! doesn’t actually make an appearance on the list of Top 25 Songs Using 1st Person Singular Pronouns, due to issues related to computational text analysis. In the official lyrics of the song, we see words like “me-e-e” which mirror the way Taylor sings them. While this spelling reflects the catchiness of the hook, programs like LIWC cannot identify this as the word “me” and it is not counted as a first person singular pronoun.