Harmony and Voice Leading

Harmony is chords treated as motion rather than as a list — where a progression is heading and what makes it want to get there. The chords lesson handed you the seven chords a key contains and the numerals for writing them down. This one is about why they pull, and about the levers that let you steer that pull instead of taking whatever the shapes give you.

The usual mistake is treating a chord chart as the harmony of a song. It is not. A chart lists destinations and leaves the route blank — which is why the same four symbols handed to two bands produce two different records. Players also inherit a set of voice-leading rules as though they were physics. They are not, and knowing whose rules they are is what lets you break them on purpose.

Every Chord Has a Job, and the Chart Never Prints It

Function is a chord's relationship to the home note, and there are three jobs going. Tonic is rest. Dominant is tension aimed at rest — V and vii°, the chords carrying the leading tone. Between them sits pre-dominant, ii and IV, which walks you off home and sets the dominant up. The other two, iii and vi, each hold two notes in common with I, so they extend the tonic rather than leave it. The default order runs tonic, tonic extension, pre-dominant, dominant, tonic, and both familiar departures have names: IV falling back to I, and V landing on vi.

None of that is visible on a chart. An A minor in the key of C does two different jobs depending on where it sits — after the tonic it holds the ground, and on the way to G it sets up a cadence. Play it identically in both places and one will sound wrong for reasons nobody at the rehearsal can name. Function also tells you what is interchangeable: chords sharing a job can stand in for one another, and that is the basis of every substitution below.

Where the idea came from governs how much authority to give it. Rameau published the treatise it grows out of in 1722, and Riemann formalised the three functions in 1893 — both writing about music that already existed. This is a description of what people were doing, and it fits some repertoires far better than others.

Your Ear Is Following Voices, Not Chords

A chord change is not one block swapped for another. It is several lines moving at once, and hearing tracks lines — which is why a melody survives being played on any instrument you like. Voice leading is how those lines get from one chord to the next. It reduces to three habits: hold any note both chords contain, move the rest the shortest distance available, and let lines travel in opposite directions where you can.

Common tones are the cheap win. C major and F major share a C, so a keyboard player who leaves that C alone has two notes left to move and both go by a step. The same player reaching for a root-position F moves every finger, and the change announces itself instead of passing under the vocal. In a band the voices are spread across people — bass, two guitars, keys, a vocal — so nobody is watching the connections. That is usually why an arrangement sounds lumpy while every individual part is correct.

Move every voice the same direction by the same distance and something else happens: the lines stop being separate and fuse into one thickened line. The power chord from the intervals lesson is exactly this, root and fifth sliding up the neck locked together — heard as one fat note rather than two notes in harmony, which is why it survives distortion. That fusion is the mechanism behind the rules in the next section.

The Voice-Leading Rules Are One Period's House Style

Three rules get handed down as though they came off a stone tablet: no parallel fifths or octaves between two voices, resolve the leading tone upward, prefer contrary motion. The reason given for the first is independence — parallel perfect intervals were held to destroy the separateness of the parts. The prohibition is old and specific. It was written down around 1300 and had settled into consistent practice by roughly 1450. From there it ran through the common practice period, about 1650 to 1900, and into counterpoint exercises still being set today.

Read that reason again, because it describes a goal rather than an acoustic fact. Parallel fifths do not sound bad. They sound like fewer voices, which is only a problem if independent voices were what you wanted. The earliest notated polyphony in Europe was built out of the very parallel fifths and octaves later forbidden — the rule arrived to prohibit something people had been doing for centuries.

It has been broken deliberately ever since. Chopin used consecutive fifths as a feature. Rock rhythm guitar is nothing but parallel fifths — folk polyphony in Georgia and Iceland harmonises in parallel by design, and by the twentieth century the prohibition had stopped being relevant to most new music. Leading-tone resolution is just as local: Mixolydian has a flat seventh and no leading tone to resolve, as the keys lesson pointed out. Learn the rules, because working through them trains you to hear separate lines where you used to hear blocks. Then stop obeying them the moment fusion is the sound you want.

A Secondary Dominant Aims the Pull at Something That Isn't Home

If you write songs, this is the advanced device to learn first. Take any major or minor chord in your key, treat it for one chord's length as though it were home, and play its dominant in front of it. In C major the chord built on D is normally D minor. Raise its third one semitone to F sharp and you have D7, the V of G. That F sharp sits a semitone below G — a leading tone borrowed for a chord that is not the tonic. One note from outside the key is the entire cost.

It gets written as a numeral over a numeral. V/V is said out loud as "five of five", V/ii as "five of two", and V7/vi is a dominant seventh aimed at the sixth degree. Applied dominant is the other name for the device. Every major and minor chord in the key takes one; the diminished chord is the exception — tonicising it is not worth the trouble. V/IV is the odd member, because the tonic triad is already major, so what you add there is a flat seventh that leaves your home chord leaning on IV. Beethoven opened his First Symphony on exactly that chord, in the wrong key on purpose.

The keys lesson owns the line between this and changing key. The short version: a secondary dominant leaves the key for one chord and is back before you register the trip — which is why it costs a listener nothing to follow. Chain them and the effect compounds, each dominant resolving into the next one down the circle. "Sweet Georgia Brown", written by Ben Bernie and Maceo Pinkard in 1925, spends a sixteen-bar section on VI7–II7–V7–I, three dominants handing off before anything diatonic happens. Arlo Guthrie's "Alice's Restaurant" runs the same chain in 1967.

Tritone Substitution Is One Tritone With Two Possible Roots

The chords lesson already located the tritone inside a dominant seventh: in G7 the third is B and the seventh is F, six semitones apart, each a semitone from somewhere stable. What it had no reason to mention is that the pair is not the property of one chord. Build a dominant seventh on D♭, a tritone away, and you get D♭, F, A♭ and C♭ — and C♭ is B under your fingers. The same two notes are back with their jobs exchanged, F now the third and B now the seventh.

The tritone is the part that resolves, and it resolves the same way whatever root sits underneath it — so D♭7 lands on C exactly as G7 does. This is not a trick somebody licensed. It is one resolution wearing a different bass note, and the bass is what you are buying: a root falling a fifth into the tonic becomes a root sliding down one semitone into it. A ii–V–I becomes Dm7, D♭7, Cmaj7, and the bass walks D, D flat, C.

Two of the four notes do not survive the swap. G and D leave, D♭ and A♭ arrive — so a melody note that sat comfortably over G7 can be sour over D♭7, and checking it before you commit is not optional. One consolation is built into the arithmetic: the old root returns as the sharp eleventh of the new chord, which is why the substitute is normally voiced with that note in it.

What to Change First When You Reharmonise

Reharmonising means keeping a melody and changing what sits underneath it. These moves are ordered by what they cost.

• Mark the jobs before the names — write tonic, pre-dominant or dominant over each bar. That separates decoration from whatever is holding the phrase up.
• Substitute inside a function first — iii or vi for I, ii for IV. Two notes of three survive, so the melody usually still fits untouched.
• Put a secondary dominant in front of anything you want to arrive — take a chord that currently just appears, and precede it with its own dominant seventh.
• Leave the diminished chord alone — every other chord in the key takes an applied dominant, and forcing that one sounds like an error.
• Tritone-sub a V when the bass should step — a root falling a fifth becomes a root sliding one semitone into the target, which is usually the point.
• Check the melody against every substitution — you kept two notes and replaced two, and the melody was probably written over one you removed.
• Move voices, not chords — hold what both chords share and shift the rest the shortest distance. That beats finding a cleverer chord.
• Break the parallel rules on purpose — parallel motion fuses voices into a single line, which is right any time you want weight instead of counterpoint.
• Reharmonise the repeat, not the first pass — a listener needs the plain version in memory before the altered one reads as a change.