Cadences

A cadence is a place, not a pair of chords

Nearly every explanation you'll read says a cadence is two chords. That's the useful shorthand, but it isn't true, and believing it will trip you up. A cadence is a point of arrival that punctuates the end of a phrase. Four things decide how final it feels:

• Where it sits. It has to end a phrase. The same two chords mid-phrase are just chords.
• The chords. Necessary, but not sufficient.
• The melody. The note on top changes everything — that alone is the difference between the two strengths of authentic cadence.
• The beat. Landing on a strong beat sounds conclusive; the same chords on a weak beat don't.

The second demo in the bench above makes this concrete: it plays the same IV–I twice, once as a phrase ending and once buried mid-phrase. Only one of them is a cadence. Nothing about the chords changed — only where they sat.

This also explains a phrase you'll meet later: post-cadential. An outro vamp or a tagged-on "amen" comes after the music has already closed, so it decorates an ending rather than making one. Same reason a looping vamp like the Andalusian Am–G–F–E isn't really a cadence — it never ends, it just goes round again.

The four you actually need

Everything else is a variation on these. Roman numerals just number the chords of the key — capital = major, lowercase = minor — and the number system covers them properly if they're new.

• Authentic — V–I — the full stop. The strongest ending in tonal music. Everything else is defined by how it differs from this one.
• Half — ends on V — the comma. Not an unfinished V–I: it's an arrival on V. Worth knowing that the V of a half cadence is normally a plain triad in root position, not a V7 — on guitar, end on E rather than E7 and you'll hear why.
• Deceptive — V–vi — the swerve. V promises home and delivers the minor vi instead. In a minor key the swerve lands on the major VI— V–VI, brighter and more cinematic than you'd expect. (It's written VI, not ♭VI: a minor key already has the flattened 6th.)
• Plagal — IV–I — the "amen". Softer, and genuinely argued about; see below.

Authentic cadences come in two strengths, and this is the distinction exam boards care about. A perfect authentic cadence needs all three: V and I both in root position, and the tonic note in the top voice. Soften any one of those — an inversion, or the 3rd or 5th on top instead — and it's an imperfect authentic cadence: the same two chords, a less conclusive landing. Press the bench's "Same chords, different ending?" button and it plays them one after the other — listen to the top note only; the chords underneath never change. There's also a rarer flavour where vii°⁶ stands in for V: it has the leading tone but no root motion underneath it, so it's always an imperfect authentic cadence, never a perfect one.

One footnote, since Caplin comes up below: he would say an inverted tonic isn't a soft cadence but no cadence at all. Most North American textbooks are more relaxed and file it as an IAC.

If you learned "imperfect cadence" somewhere else

One genuine landmine: the word imperfect means two different things on either side of the Atlantic, and neither camp mentions the other.

This page uses the US / North American names throughout, because that's what most study material uses. If you were taught the British set, read "imperfect" as "half".

Why V–I sounds final: two fingers, one fret each

The mechanism is visible on the neck if you strip the chords back to a shell. Play 3 2 3 x x x — that's G on the low E, B on the A string 2nd fret, and F on the D string 3rd fret. Those last two are the notes that make the chord itch:

• B — the leading tone, a half-step below home.
• F — the chord's 7th: the note that turns G into G7, the one that makes it sound like something is about to happen.

Watch out for one thing, because there are two different sevens here: the leading tone is the 7th degree of the scale and it rises; the 7th of the chord is a different note and it falls. They are the two protagonists of this cadence and they move in opposite directions.

On adjacent strings like this they form a true tritone, the most unstable interval in tonal music, and it wants to contract. Now play x 3 2 x x x: the A-string finger goes up one fret (B→C) and the D-string finger goes down one fret (F→E). Two fingers, one fret each, in opposite directions — that tritone (some books call it a diminished 5th) squeezed into a major 3rd. That squeeze is the entire mechanism.

The same thing happens in the full G7 (320001) → C (x32010), just spread across the neck: both B's rise a fret (A string 2→3, B string 0→1), the F drops a fret (high E 1→0), the open G is a common tone held over, and the D moves up a whole step to E. It's a complete re-grip rather than two fingers — which is exactly why the shell version is worth playing first. It's also why every cadence in the bench lists its leading tone and its 7th underneath.

Now play F (133211) → C instead. There's a half-step in there (F down to E), but no leading tone at all— nothing pulls upward into home. And here's the deeper problem, the one you can see on the fretboard: F major already contains a C. The chord that's supposed to be arriving at home is already sitting on it.

The plagal argument — worth knowing you're in a fight

That last observation is the reason theorists genuinely disagree about whether IV–I is a cadence at all. The argument, made most influentially by William Caplin, runs: a cadence has to confirm the key, and IV can't — it contains no leading tone to resolve, and the tonic is already sitting inside it. Look at classical repertoire and the IV–I you find is nearly always a tag after the phrase already closed with a real V–I. On that reading, plagal motion prolongs the tonic rather than cadencing on it — which is exactly how several modern textbooks now file it.

And in the music you actually play? The opposite. In rock, gospel, folk and pop, IV–I is one of the most common ways a phrase genuinely ends — often with no V in sight for the whole song. So the honest answer is: IV–I is a cadence in your music, and arguably wasn't one in Mozart's. The reason for the fight is the fretboard fact above, not snobbery. Worth saying plainly: this is an influential position, not a settled one — Caplin scopes his claim to the eighteenth and early nineteenth centuries, and genuine plagal closes are common in later Romantic and modal repertoire.

Two variations worth stealing. Minor plagal (iv–I) borrows the minor iv from the minor key that starts on the same note: in C major the iv comes from C minor, so it is Fm → C. That bittersweet lift is unmistakable, and the reason it aches is the ♭6 (A♭) sliding down into the 5th (G). And the double plagal (♭VII–IV–I) stacks two fourth-falls in a row: in C that's B♭–F–C, the classic rock outro chant. Both are presets in the bench.

Minor keys: the one-finger cadence

This is the single most useful thing on this page for a guitarist. Natural minor has no leading tone — its 7th sits a whole step below home — so the chord on the 5th degree comes out minor (v), and a v–i ending doesn't resolve so much as stop.

Raise that 7th by a half-step and you manufacture a leading tone; the v becomes a major V, and the ending snaps shut. In A minor, that is literally one finger:

• Em (022000) → Am — v–i. Diatonic, modal, unresolved.
• E (022100) → Am — V–i. One extra finger puts G♯ in the chord: the leading tone.

Hit the bench's "One finger, in a minor key" button and the two versions play back to back. Eight seconds of A/B teaches this better than any paragraph — including this one. (That borrowed G♯ is where harmonic minor comes from: the scale exists because the cadence demanded it.)

The flamenco ending

Minor keys have their own half cadence: iv⁶–V, the Phrygian half cadence. In A minor the bass falls a half-step — F down to E — while a higher note rises into that same E: the two arrive at one note from opposite directions, the pincer move again. You already know the sound: it's where the descent Am–G–F–E lands. Strictly the F in that progression is a root-position VI rather than a true iv⁶, but the half-step bass slide into E is identical. On guitar the strict version is worth playing once: Dm/F (1x0231) → E (022100), where the F on the low E string 1st fret slides down onto the open E.

And one honest caveat: a flamenco player doesn't hear that final E as the V of A minor at all. They hear it as home— the tonic of E Phrygian dominant. Calling it a half cadence is the classical reading of a progression that, in its own tradition, has already arrived. That's the real reason it sounds Spanish rather than merely unfinished.

Where they turn up when you're actually playing

The half cadence deserves special mention for songwriters: it is the most useful and least discussed of the four. Ending a section on V is how you make a listener need the next section. Nothing else creates that particular forward lean.

One decoration worth naming: the cadential 6/4

You'll meet this on any theory paper, and you already play it: it is Gsus4 → G7 → C. Books write it I⁶₄–V–I, but despite that "I" it is dominant harmony the whole way through — a V that arrives late. It's the suspension sound, doing cadence duty.

Train it by ear

Cadences are one of the few pieces of theory that's genuinely easier to hear than to read — the four types sound nothing alike once you've met them. The tool has a cadence ear-training drill that plays one and asks which it was, and the Academy teaches it as a lesson with a comprehension check. From there, the natural next steps are the chords of a key (where V and IV come from) and secondary dominants(how to aim a cadence at a chord that isn't home).

Sources & further reading

The plagal argument above is William Caplin's. If you want it first-hand: "The Classical Cadence: Conceptions and Misconceptions", Journal of the American Musicological Society 57/1 (2004), expanded in Cadence: A Study of Closure in Tonal Music (Oxford University Press, 2024). For the modern textbook treatment of plagal motion as prolongation rather than cadence, see Open Music Theory. Everything else on this page — the explanations, the fretboard demos and the bench — is our own.