During my recent recess from blogging, pop music lost two musical icons of the 1970s, and beyond: David Bowie and Glenn Frey. Some writers have been noting differences (musical and otherwise) between the two stars (with some including in their discussion the also-recently departed Lemmy Kilmister, as in this article). Here, I'd like to take a closer look at hit songs that Bowie and Frey had a hand in writing to see how such differences play out linguistically.

I began by looking at the songs of each artist that appeared in Billboard's year-end Top Ten charts. Both Bowie and Frey (with the Eagles) had one song apiece that made it to these charts, both in 1975. For Bowie, the song was "Fame," which was ranked seventh for the year by Billboard (and was co-written with John Lennon and Carlos Alomar).

For Frey, the song was "One of these Nights," which was ranked ninth for the year by Billboard (and was co-written with Don Henley). 

The following table shows the rank and frequency of the Top 20 words in both songs:

To be sure, the lexical differences between these two songs far outweigh the similarities. But let's first look at the latter. If you remember my earlier post on words that typically rank high in collections of pop lyrics, you will be able to spot two ways in which the wordlists for both songs deviate from such wordlists: 1) you and I do not appear at the top of the lists (in fact, I only occurs once in "Fame"), and 2) love appears in neither wordlist (in fact, the word doesn't occur in either song at all). Aside from these general observations on negative evidence, the similarities are rather scarce, but also unsurprising when they do occur: both lists include you, to, be/been, and a, function words that are bound to occur in any sample of English. Aside from these similarities, there are only differences, which are even more apparent when we take out the function words (and nonce words such as oooh) by applying a stoplist,* as illustrated in the table below: 

A comparison of content words used in the two songs reveals no words in common (babe and baby being the closest match, the latter being a mere oncer in "One of These Nights").

Clearly there are major lexical differences between the two songs, but are these indicative of general differences between the two writers? In an attempt to answer this question, I looked at each artist's remaining songs in Billboard's year-end Hot 100 from 1956 to the present. For Bowie, this search produced "Space Oddity" (1973), "Golden Years" (1976), "Let's Dance" (1983), and "China Girl" (1983), and for Frey, "Best of My Love" (1975), "Take It to the Limit" (1976), "New Kid in Town" (1977), "Hotel California" (1977), "Heartache Tonight" (1980), and "I Can't Tell You Why" (1980), all with the Eagles, and "You Belong to the City" (1985)** as a solo act.  

The following are the Top 10 words for the datasets:

Not surprisingly, the sets share several function words: you, I, my, the, to, and, in, and a. However, there are a couple of matters of interest, which can be seen when one compares the lists to the Top 10 words for the entire Billboard year-end Top Ten collection (1951-2015) presented below:

The first is that the list of words used most frequently by Bowie, despite him being, unarguably, the more unusual and less mainstream of the two artists, aligns more closely with the Billboard collection as a whole than Frey's does, particularly in his wordlist being topped by you and I. Another observation is that neither Bowie nor Frey share the affinity for the use of me in their Hot 100 songs than songwriters in the Top Ten collection do, as it ranks 35th in the Bowie dataset and 37th in Frey's.

The following table shows the wordlists once the function words were omitted via the same stoplist as used earlier, as well as the elimination of nonce words such as whop:

As the table shows, the sets share several content words, notably the very common words (at least in pop lyrics) baby and love, as well as dance, hear, little, and look, and the lemma, or variant forms of the same root word, life and living. It is also interesting to note that the words ranked 15th for each set, control and limit, can be used as synonyms for one another, both as nouns and verbs. 

It would be remiss not to mention that the table also shows that working with small datasets has its perils, especially when one is in a hurry to put something together in a timely manner. However, the comparison does indeed shed light on several differences in the foci of the two datasets. Bowie, for instance, uses color terms, such as gold, golden, and blue, while the only terms in the Frey set that come close to broaching this domain are dark and lights. Perhaps embracing his persona as Starman, Bowie uses several terms relating to the (celestial) heavens, including moonlight, angel, and stars, and if we allow our external knowledge of Bowie's catalog to inform our analysis, rather than just what is on the page before us, then the two-gram ground control as well. Frey, on the other hand, appears to keep his feet planted firmly on the ground, with references to town, city, place, street, California, hotel, and highway, important reference points in many a narrative in the genres of country music and Americana.

In short, a quick analysis of a selection of songs written or co-written by Bowie and Frey does suggest vast differences between the two artists, based on lexical use. Of course, more data would be helpful for supporting this view (or, as the case may be, to overturn it); thus, more songs by each artist should be added to their datasets if one wishes to draw any stronger conclusions. Additionally, the songs used in such an analysis could be vetted better (for instance, in the case of co-written songs, there might be a way to determine, say, through past interviews with various members of the songwriting or production team, liner notes, or even in some cases, the order that their name appears in in songwriting credits, how much bearing a writer in question had on the song). The information gleaned from such a process could then be used to better inform the decision-making that goes into whether a song is included in the analysis and how it is treated if it is. In addition, while the investigation of language as a bag of words is useful, an analysis that included the investigation of two- and three-word phrases as well could offer greater insight, in this case, into the styles of two of the great songwriters of our time.

Finally, let's hope that the rest of 2016 will not be as tough on our aging rock and roll stars, or for that matter, ourselves, as the first few weeks of the year have been. Peace.

------------------------------------------------------------------------------------------------------------------

*I tried to be as faithful to this stoplist as I could, despite my reticence to lose information through, at times, such drastic cutting. I did add a few words that were lemma of forms included on the list, such as don't and lately.

**Glenn Frey also recorded "The Heat Is On" as a solo artist in 1985, and it appeared at #19 in the Hot 100; however, it was actually written by Harold Faltermeyer and Keith Forsey.