A little while back, we talked about pitches,
notes and scales. Admittedly, scales, or at least scale fragments,
are frequent components in popular melodies, but generally we want more
than that in our music.
We often think of this improvisational approach as a modern phenomenon.
The truth is that it was an important component in operas and concertos
until the 19th c., and has most likely always been a part of popular
Melodies usually include both rhythmic and tonal aspects. Some might be
written with a loosely defined structure. In such cases, performers
might modify either component for effect. One particularly finds this in
styles, like jazz, in which the audience is fully expecting such
modifications as aspects of proper performance.1 I will talk about
rhythmic aspects of melody briefly, but most of what I have to say is
about the tonal properties, so we will talk about those first.
It should be obvious that tonal movement takes place in intervals of
various varieties, including repetition of the same note (often with
some rhythmic variation). The most common movement is step-wise (up or
down one note in the current scale, technically called 'conjunct
motion'). The following familiar melody fragment is entirely step-wise.
Usually we want to hear some more variety. In fact, you probably
remember that Ode to Joy starts to include other intervals very soon.
Most pieces include varieties of intervals, even though seconds are the
most common. This famous piece, written just a few years earlier by John
Stafford Smith,2 has some conjunct motion, but is primarily
'disjunctive' (plenty of skips).
Surprised by the author? Francis Scott Key famously wrote the poem
("Defence of Fort McHenry") during the war of 1812. It was
later set to this pre-existing popular tune (this was common practice;
very few lyricists were also musicians of any stature, and it is
unlikely that its becoming a song was ever Key's intention). The music
had been written (ironically, in England) about 30 years earlier.
Both varieties create interest by varying the combination of rhythm,
range, and types of movement; but they have other similarities, which
may be clearer as we move along.
Most melodic phrases have a 'contour'—a generally recognizable shape:
usually up, down, or curved (down, then up, or the reverse). A larger
melody will combine these in different ways to create connection and
diversity, so when we talk about contour, we are almost always
discussing a single phrase within the larger melodic structure. To talk
about it at a larger level would be either too complex or grossly
over-simplified. Here is a brief melody with a series of clear upward
contours, even though they slip back down a step at the end of each
Of course, taken together, the entire verse has a perfectly curved
contour, starting and ending on the same note.
For obvious reasons, this is less true when the vocalists are trying to
show off their impressive range. Mariah Carey comes to mind, although I
don't know whether this was ever a consideration in her choice of songs
(although I suspect so).
If other notes from the tonic chord are sounding simultaneously, and
especially if the lowest sounding note (the bass note) is the tonic
itself, a fifth in the melody is usually a satisfactory resting note. We
will see this again when we look a little deeper at Ode to Joy, below.
Sometimes the degree of rest is affected by musical context—how it
got there. More on this later when we talk about chord movement.
By no means should step-wise resolution be considered a rule. I
am simply observing that it is the most common. The more jumps there are
in the melody, the more normal a jump to resolution will sound; a
predominantly step-wise melody will often sound a little odd if it jumps
to resolution (although there are plenty of exceptions).
There is an observable preference in post-'80s pop music for downward
moving contours. No rule—just a trend. There is also an observable
tendency to limit melodic range in the primary vocal part.3
Melody and chords
One of the most common approaches to melody is to follow the outline of
the current chord (which is what is going on at the beginning of "The
Star-Spangled Banner"). This technique is called an arpeggio.
Even when a chord is not being explicitly stated in the melody, though,
it is normal for the melody to focus on the chord notes, even though
others may be present. When melodies include notes that are not part of
the underlying chord structure, those notes are called
non-harmonic tones (the second note in "Frère Jacques" is an
Successful tonal melodies draw their strength from the internal tensions
and releases created by each note's relationship to the tonic. Some
intervals are more 'at rest' than others. The tonic of the key is the
most at rest, with the other notes in the key's tonic triad reasonably
so, although the fifth often wants further resolve.4 They
create a sense of completion, but the other notes (2, 4, 6, &
7) are in various degrees of tension and leave the listener
wishing for resolution. That resolution consists of movement to one of
the resting notes, and that, in turn, appears to be most satisfying if
it comes from step-wise movement.5
'Ah! vous dirais-je, Maman', better known in the English speaking world as 'Twinkle, Twinkle Little Star,' or 'The Alphabet Song.' (a)
Of course, it's not just folk tunes that go this way. Taylor Swift does
this in the end of "Fearless", as does
Third Eye Blind in "Jumper", and about a
gazillion other songs (give or take a couple).
Larger melodies are often built from brief musical ideas. Consider
Beethoven's fifth symphony. The first movement starts with this simple
musical statement that ends up getting restated and developed throughout
Listen to Australian Chamber Orchestra performing this brief fragment
from the beginning of Beethoven's Fifth symphony, to hear how he
develops this theme within the first half minute. A brief musical idea
like this is often called a motif (or motive), and
occurs occasionally in pop as well (consider the opening sax part in
Gerry Rafferty's "Baker Street"). Most musical ideas are a little longer, but it is
not the length of the basic idea, but rather, how it is
developed-restated, reused, and modified-that makes the larger piece
Slightly longer musical moments (like one line of a song) are called 'phrases.'
Repetition is an important part of what holds musical pieces together.
It creates a familiarity that listeners appreciate. Obviously, this can
be as simple as just doing it again. This is the approach taken in
"Frère Jacques" (over and over). And just in case you were thinking that
this repetition stuff was just for orchestras and folk tunes, take a
second to remember Mark Ronson & Bruno Mars doing "Uptown Funk".
More often, however, you will hear similar phrases with slight
variation. Consider what happens in the second phrase of "Ode to Joy."
They are the same melody except that the first phrase ends , while the second ends . Since this is in the key of
C, you will notice that the first ending does not resolve on
a resting note (c=1, e=3, or
g=5). The second does—on the tonic, in fact.
The result is that at the end of the first phrase, you feel like you
need to keep going, while at the end of the second, well, you could stop
there (even though we know the melody actually continues). In a pair
like this, the first phrase is called the antecedent (= 'goes
before'), and the second is called the consequent (= 'follows
closely'). The opening lines of "Falling Slowly" by
Glen Hansard (Once) do this sort of thing as well, although much more
You can also see a version of this technique in the first couple of
phrases of "Sound of Silence," above, with the variation that the second
phrase is moved down a step, and lacks the resolve. There are no rules,
of course, only certain common practices that generations of experience
have found to be pleasing to the ear.
As Ode to Joy continues, we get a very different phrase:
This is just for contrast, and it will go back and repeat the second
line next with its resolve. But there is something worth noticing here.
This phrase also has an incomplete feel, which sets up the next line;
but look at the last note. It is a g=5. Isn't that
one of the notes I have been calling a 'resting note'? Well, yes, but as
I mentioned earlier, the 5 is a special case. It is part of the
tonic chord, but by itself it implies the 5 chord, which
always wants to resolve. We could make this restful by adding other
chord notes from the tonic (polyphonically) which would make it a clear
resolve, but we don't want to do that here because its instability is
setting up the resolution in the next phrase.
Putting chords with melody
Also, if you are writing for a band or choral group it is a frequent
technique to move all the voices, such that at any given moment you can
identify a chord from the current voicing. Although in these cases,
there is usually an implied larger chord pattern.
Quicker chord changes may be part of the arrangement, particularly in
rock, but they are usually creating texture rather than actually
supporting the melody.
Of course, if you have a minor key song, you may have to consider that
this may be a harmonic or melodic minor. In any case, part of what makes
Radiohead's "Creep," interesting is that they ignored this rule—of course,
they weren't doing country [See my
discussion of Creep].
The first two notes don't make up an entire measure. They are called a
pickup. Normally, the time value of the pickup notes are removed from
the last measure. Happy Birthday does this as well. I am counting this
as the first measure. Some people do not, and would say that measure 1
begins with the b.
In fact, this is an example of a 'passing tone.' These are a variety of
non-harmonic tones you get when you are moving conjunctly from one chord
tone to another. We will discuss these sorts of notes later.
We have already seen how a melody is tied to the chords in a piece of
music. If you only have the melody, how do you decide what chords go
with it? Occasionally you will find a melody that simply follows the
notes of a chord (arpeggio), in which case how you add chords is usually
Most melodies are not so obvious, though, so you have to look at the
predominant notes in an area. One of the most basic mistakes people can
make when first writing chords for a melody (or the reverse) is to
assume that there needs to be a one to one correspondence between the
two. By this I mean that folks will assume that every note in a melody
corresponds to a particular chord in the arrangement. This can
be the case, of course; if you have a particularly slow tune, you may
like the sound.6 But most of the time in popular music,
functional chord changes do not take place more that once or twice in a
measure,7 and usually the melody is moving a good deal more
frequently. In any case, the rhythm instrument (guitar, piano, etc.) is
just playing the straight chords underneath these extra notes. As we go
on, we will see some common approaches to adding other notes, but most
of the time the important rule is what sounds right.
When you are considering chord possibilities, try to stick with the
chords in the key-chords that would not require any accidentals relative
to key signature.8 In fact, most popular songs, especially
country songs, like to major on the majors (1, 4,
& 5), so go there first, but don't be afraid of minor chords.
They usually make songs more interesting.
If you have a melody like we see at the beginning, you simply look and
see what chords are suggested by the notes. You can see that the first
two measures9 are made up entirely of notes from a
B♭ chord. The first exception, in fact, is the third
note in the third measure, which is a c. We could be bothered
by it, and go back to evaluate the first three measures as a
B♭7, but this is unnecessary. As noted earlier,
we usually just refer to notes like these as 'non-harmonic
tones.'10 They make the music more interesting by adding
movement and tension. We can ignore those out-of-arpeggio notes, though,
when we are identifying chords.
But at the end of the next measure we get a challenge. The
e♮ isn't part of any chord in this key (which is why it
needs an accidental). In fact it takes us out of key temporarily, so you
just have to play around to find the chord. You know there are a limited
number of chords with e♮ in them (E,
A [and their minor versions], C, and
C♯m). You try them all. C works best (we
will understand why later). You can extend it to include the whole
measure by making it a C7 (the d is just a
non-harmonic tone), or limit the C to include only the
e♮, in which case the b and the
d are still part of the B♭.
At this point, we find ourselves not having any arpeggios to help. If we
stay in 2-flats, the fs in the next measure could work with
B♭, Dm, or F. Coming after the
C we just had, the F sounds best (also keep in
mind that songs like this are often made up of 1,
4, & 5 chords which in B♭ would
be B♭, E♭, & F). The next
measure doesn't give us an arpeggio, but we have a b♭
and a d with a passing tone between, so we are going to guess
that we are back in B♭. The a's in the next
measure suggest another F chord, and the g in the
middle could make it an F7. The last two measures give us a
clear B♭ arpeggio, which also brings us to a
resolution for the phrase. Here is what we've got: