Some Reharmonization Techniques for Popular Music: Melodic Skeletons, the Melodic-Harmonic Divorce, and Meta-Schemas

Trevor de Clercq

Introduction

If we subscribe to Bloom’s revised Taxonomy (Karthwohl 2002), our peak educational objective is to have students create. If we want students to understand harmony, that means getting them to write chord progressions. Melody harmonization has become a common way to facilitate this in modern workbooks (e.g., Laitz 2015; Kostka, Payne, and Almén 2018), usually within a chorale setting where each melodic note receives its own chord and chords change every beat. This environment differs from popular music, where chords typically change once per measure and support multiple melodic notes (de Clercq 2017). Moreover, recent monographs (Doll 2017; Temperley 2018) show that the harmonic language of popular music differs in many ways from classical music and even jazz. Melody harmonization in popular music thus presumably involves its own distinct approach. Surveying existing textbooks on popular music, however, it is difficult to find much advice on melody harmonization. Sometimes, a book primarily covers how to label chords, with little discussion of how these chords relate (Wyatt and Schroeder 1998). Other times, a book focuses more on analysis than composition (Snodgrass 2016). And other times, a book just drops the student in the deep end too quickly, asking them to “rewrite the hits” without much guidance (Kachulis 2004).

I thus present here some techniques for melody harmonization in popular music. Specifically, I discuss three tools, each of which addresses characteristic aspects of popular music: 1) melodic skeletons, to side-step rhythmic complexity in pop melodies; 2) the melodic-harmonic divorce, to address the lack of traditional coordination between melody and harmony, and 3) meta-schemas, to give directionality in the absence of traditional chord syntax. Technically speaking, I will discuss melody re-harmonization, since I use recordings of isolated vocal tracks (freely available on the internet) from well-known songs; I do so not only to focus the student on the harmonization task itself but also to replicate the real-world activity of reharmonization often employed in creating a cover version or remix (such as Pomplamoose’s reimagination of Lady Gaga’s “Telephone”).

As a bit of background, I introduce reharmonization assignments in my sophomore-level commercial musicianship course, which is an elective for songwriting and audio production majors. That said, any student with a basic command of triads and seventh chords on keyboard or guitar and some familiarity with the sound of popular music should be able to successfully participate in these types of exercises. In terms of deliverables, I have my students turn in videos of themselves playing along with the isolated vocal track, so that I can be sure they are hearing their reharmonization in context. The primary obstacle is that some students find it difficult to synchronize with the tempo of an isolated vocal track. To help, I sometimes provide the students with the vocal track in one channel (e.g., the left) and a click track in the other (e.g., the right). (This can be accomplished by aligning the isolated vocal track with the original recording in a digital audio workstation and then creating a tempo map of the song.) But the ability to follow a solo singer is a practical skill worth developing, so I see trying to play along with the vocal alone as a worthwhile side challenge to the task.

Melodic Skeletons

In order to reharmonize a song, we might first feel the need to fully transcribe the melody. A more efficient approach, though, is to create a streamlined version—what I call a “melodic skeleton.” Figure 1 (top row) shows a melodic skeleton for the first eight bars of “Hey Jude” by the Beatles (1968), which reduces the melody to one or two scale degrees per measure (with one cell equal to one bar). Creating a melodic skeleton is similar to Schenkerian analysis as typically taught to graduate music students, where non-essential tones are removed to reveal the underlying structure. But it does not take a graduate degree in music to create a melodic skeleton. The melodic skeleton simply identifies (somewhat subjectively, I admit) the most important melodic notes for each bar—typically those that are held for longer, occur on strong beats, or are chord tones in the original harmonization. In “Hey Jude,” for example, all the notes in my melodic skeleton are chord tones in the original song.

Melody:	3	2	4 1	3	6	5	5	1
Original Chords:	I	V	Vsus7	I	IV	I	V	I
Reharmonnization:	I	♭VII	ii IV	I	vi	iii	V Vsus7	I

Figure 1. Melodic skeleton of the first eight bars from “Hey Jude” with original chords and reharmonization. (Audio examples are embedded as links within the Figure.)

Because the melodic skeleton has only one or two notes per bar, the reharmonization task becomes greatly simplified. When creating my reharmonization (Figure 1, bottom row), I noticed that the first four bars of the melody can be conceptualized as a double-neighbor figure around scale-degree 3, so I mirrored this pattern in the bass. In the last four bars, I introduced some minor chords for variety and reintroduced the Vsus7 chord from the original harmonization.

In practice, reharmonization is inherently trial-and-error. As such, it can be a very interactive use of class time to have students suggest different options based on the melodic skeleton and then immediately try them out, discussing afterwards what did and did not work, and why. In general, the issue of whether a particular chord progression sounds “good” is more opinion than fact, and the reasons why are often debatable. I thus see my role more as a guide for ways of thinking about harmony rather than as a judge of what is right or wrong. For this reason, I usually grade these types of assignments as pass/fail, with a stated requirement that the student must use some particular technique or techniques (e.g., a secondary dominant, two instances of modal mixture).

The Melodic-Harmonic Divorce

In recent years, research on popular music has shown that melody and harmony often seem to operate independently, a situation dubbed the “melodic-harmonic divorce” (Temperley 2007; Nobile 2015). This feature of popular music can be employed in a reharmonization task. Consider, for example, the verse from “Someone like You” by Adele (2011). At the beginning of the song, the chords change every four beats (Figure 2, top row), but the melody (not shown) is rather static. The first few vocal phrases are almost identical, in fact, with Adele singing down the major pentatonic scale again and again. Even when Adele’s vocal melody moves away from this ostinato later in the verse, the melody remains pentatonic, still seeming to ignore the underlying chords.

Original:	I	iii64	vi	IV
Version 1:	I	V	♭VII	ii
Version 2:	I	♭VII	IV	iii
Version 3:	I	vi	ii	♭VII
Version 4:	I	ii	iii	IV
Version 5:	IV	I	V	vi
Version 6:	vi	IV	I	V
Version 7:	V	vi	IV	I

Figure 2. Harmonizations of the verse material for “Someone like You” (Adele, 2011).

As we think about possible reharmonizations for this song, we do not need to create a melodic skeleton. If we accept the melodic-harmonic divorce, any chord might work. For example, I like the sound of a ♭VII chord, so I can use ♭VII in, say, the third chord slot (Figure 2, Version 1); or as the second chord (Version 2); or as the last chord (Version 3); or maybe the harmonies simply ascend I–ii–iii–IV (Version 4). Moreover, I do not need to limit myself to harmonizations that begin on tonic. Note that the original chord progression can be seen as a variation on the I–V–vi–IV “Axis” progression (Richards 2017). Since this progression and its rotations are prevalent in popular music, I could harmonize the verse with the rotation starting on IV (Version 5), the rotation starting on vi (Version 6), or the rotation starting V (Version 7).

Ultimately, these seven reharmonizations are all viable. To me, the progressions that begin on tonic are somewhat better suited for a verse because they establish the tonality of the song more clearly. But there is nothing inherently wrong with any of these versions. Notice that in the second, third, and fourth chord slots, I have used every possible diatonic harmony from I to vi plus ♭VII. In other words, any of these common chords will work in any of these positions, showing the great compositional flexibility available to the student.

Meta-Schemas

Given that harmony in popular music does not need to coordinate tightly with the melody, and given that chords often do not follow classical syntax (de Clercq and Temperley 2011), students have many reharmonization options. This is a boon for creativity, but it can lead to directionless writing. As a solution, I introduce students to the idea of “meta-schemas” (see Doll 2017). In essence, a meta-schema is a scale-degree fragment, most often chromatic, that better explains the organization of a chord progression than a functional approach based on root motions. An example common to both rock and classical music is the linear descent from 1 to 5, which Doll refers as a “drooping” or “dropping” schema (depending on the chromatic content) and which others might refer to as a “lament bass” if in the lowest voice. When teaching reharmonization, I typically put these meta-schemas in the bass so that the chromatic motion can clearly be heard.

Consider the application of meta-schemas in the song “Let It Be” by the Beatles (1970). My melodic skeleton for the first eight bars of vocal material is shown in Figure 3 (top row), with the original chords below. In my first reharmonization (Figure 3, Version 1), I utilized a descending chromatic bass line, in this case from ♭7 down to 4 (the “dropping” [♭7–5] plus “slouching” [5–4] schemas). In my second harmonization (Version 2), I employed an ascending chromatic bass, now going from 3 up to 6 (the “swelling” [4–5] plus “stretching” [5–6] schemas). These chromatic bass lines were the primary factors driving my harmonization choices. To be clear, I do not think either reharmonization is revolutionary in terms of chord syntax, nor is either necessarily idiomatic to the harmonic practice of the Beatles. But both show a way of re-thinking the harmony of a song based more on voice leading—specifically, chromatic voice leading in the bass—and less on the expected succession probabilities for a particular chord function.

Melody:	5	5	3 2	1	3	2	1	.
Original Chords:	I	V	vi	IV	I	V	IV	I
Version 1:	I	♭VII	IV6	♭VI	I64	V65/V	IV iv	I
Version 2:	I	iii7	IV	V65/V	I64	V65/vi	vi V6	I

Figure 3. Melodic skeleton, original chords, and reharmonizations for “Let It Be” (Beatles, 1970) for the first eight bars of vocal material.

Conclusion

While reharmonizations of short passages may be adequate for introductory assignments, students should eventually grapple with an entire verse–prechorus–chorus module to get a sense of how harmony interacts with form. As a final illustration, therefore, let us tackle a complete verse-prechorus-chorus passage. I have chosen the song “Photograph” by Ed Sheeran (2014) because the relative simplicity of the original chord progression allows for many reharmonization possibilities (Figure 4).

Verse:	3	.	3 1	.	7 1	3 1	.	.	(x2)
Verse:	I	.	vi	.	V	.	IV	.	(x2)

Prechorus:	1	3	1	3 2	5	5	5	7
Prechorus:	vi	IV	I	V	vi	IV	I	V

Chorus:	3	3 4	3 2	3 4	3 2	3 4	1	1
Chorus:	I	.	V	.	vi	.	IV	.

Figure 4. Melodic skeleton and original chords (with audio here) for the first verse, prechorus, and chorus of “Photograph” (Ed Sheeran, 2014).

Figure 5 shows my reharmonization. Like many songs, the verse melody is mostly static and pentatonic. In the original verse, the bassline descends 1–6–5–4, but I exploit the melodic-harmonic divorce to write a bassline that is a step off, descending 1–7–6–5. I continue the diatonic bass descent into the prechorus, which now begins on IV instead of vi. From the melodic skeleton, I see in the third bar of the prechorus that I can move the bass down to ♭3 for a first inversion minor tonic chord, after which I use a chromatic bass ascent until the end of the prechorus. Exploiting the harmonic-melodic divorce again, I push the chromatic ascent through the last four bars of the prechorus, essentially ignoring scale-degree 5 in the melody, which we might think clashes with the V/V chord. In the chorus, I switch to a chromatic descent in the bass. Note that the #IV half-dim7 chord at the end of the chorus arises primarily from voice-leading, supporting scale-degree 1 in the melody while sharing common tones with the chords before and after it.

Verse:	3	.	3 1	.	7 1	3 1	.	.	(x2)
Verse:	I	.	I42	.	vi7	.	V	.	(x2)

Prechorus:	1	3	1	3 2	5	5	5	7
Prechorus:	IV	I6	i6	I6	IV	V65/V	I64	V65/vi

Chorus:	3	3 4	3 2	3 4	3 2	3 4	1	1	IV iv
Chorus:	vi	.	V65/vi	.	I64	.	♯IV^ø⁷	.	IV iv

Figure 5. Melodic skeleton and reharmonization (with audio here) for the first verse, prechorus, and chorus of “Photograph” (Ed Sheeran, 2014).

I do not expect all students to create reharmonizations at this skill level. Often, though, students will stumble onto fairly sophisticated harmonic techniques just by thinking about chord possibilities and using their ears. I remember one student, for example, coming up with the same reharmonization as shown in the chorus of Figure 5 and then fluently explaining (as I have them do) his reharmonization strategy even though (as he admitted) he was not entirely sure what Roman numerals were most appropriate. To be frank, I am not entirely sure what Roman numerals are most appropriate in this passage either. I am sure, though, that having students reharmonize melodies—in particular, recordings of isolated vocal melodies—is a great way to encourage students to use their ears to better understand harmony in popular music.

References

de Clercq, Trevor. 2017. “Interactions Between Harmony and Form in a Corpus of Rock Music.” Journal of Music Theory 61 (2): 143–170.

de Clercq, Trevor and David Temperley. 2011. “A Corpus Analysis of Rock Harmony.” Popular Music 30 (1): 47–70.

Doll, Christopher. 2017. Hearing Harmony: Toward a Tonal Theory for the Rock Era. Ann Arbor, MI: University of Michigan Press.

Kachulis, Jimmy. 2004. The Songwriter’s Workshop: Harmony. Boston, MA: Berklee Press.

Karthwohl, David. 2002. “A Revision of Bloom’s Taxonomy: An Overview.” Theory into Practice 41 (4): 212–218.

Kostka, Stefan, Dorothy Payne, and Byron Almén. 2018. Workbook for Tonal Harmony with an Introduction to Post-Tonal Music, 8th ed. New York: NY, McGraw-Hill.

Laitz, Steven. 2015. Writing and Analysis Workbook to Accompany The Complete Musician, 4th ed. Oxford: Oxford University Press.

Nobile, Drew. 2015. “Counterpoint in Rock Music: Unpacking the ‘Melodic-Harmonic Divorce’.” Music Theory Spectrum 37 (2): 189–203.

Richards, Mark. 2017. “Tonal Ambiguity in Popular Music’s Axis Progressions.” Music Theory Online 23 (3).

Snodgrass, Jennifer. 2016. Contemporary Musicianship: Analysis and the Artist. Oxford: Oxford University Press.

Temperley, David. 2007. “The Melodic-Harmonic ‘Divorce’ in Rock.” Popular Music 26 (2): 323–42.

———. 2018. The Musical Language of Rock. Oxford: Oxford University Press.

Wyatt, Keith and Carl Schroeder. 1998. Harmony and Theory: A Comprehensive Source for All Musicians. Milwaukee, WI: Hal Leonard Corporation.