So I've begun writing several orchestral pieces (forgive me if I use incorrect names, I didn't study orchestrastion or composition beyond my GCSEs).

When writing a piece, you often change the key part way through to inject some flavour and keep the piece interesting. Should you return to the original key when ending the piece?

I've found that it sounds finished regardless of whether I go back to the original key or stay in the new key to the end. Are there any hard and fast "rules" that say you should or shouldn't return to the original?


Music that deals in key relationship as a structural device - which could be considered a definition of the Classical style - almost inevitably returns to the original key (or possibly the tonic or relative major if it started in a minor key). Yes, of course you can ignore this 'rule'. I suspect, however, that you will grow into appreciating the reason for it.

Do you know the song 'Unforgettable' by Nat King Cole? It ends in the subdominant key. If you choose to start in G, it will end in C. It's a great song, but for me the key structure is a negative feature.

  • Unforgettable is a tad awkward to go round two or three times, in that it sounds like it changes key for the next time round, which in a way it does.
    – Tim
    Oct 11 '17 at 12:55
  • Yup. I wondered if there was a lost verse in the same key as the end of the chorus, but apparently not. He went where the music took him, but it wasn't an altogether satisfactory journey! It's interesting to see what arrangers have done with it. Nat's recording has an interesting series of modulations, one into an instrumental half-chorus another into the final vocal. Have a listen. Oct 11 '17 at 15:55
  • I have, many times - and played it in many guises. Still can't work out exactly where the mod occurs, though...
    – Tim
    Oct 11 '17 at 16:32
  • Sometimes it's hard to define just when you stop visiting and actually set up house! Maybe only in the very last few bars in this case. But it's pretty clear where it starts and where it ends. Oct 11 '17 at 16:44

The answer as Lawrence and Tim have stated is that music prior to the Twentieth Century most often did. However, beginning roughly at the start of the 20th Century, the harmonies and progressions matured to the point that any relatively strong cadence will sound completed. By using other musical elements such as tempo, dynamics, etc., you even get away with a dissonant cadence, such as at the end of Ravel's 'Bolero'. Bach's continually rising canon ends one whole step up on each iteration.

If you are writing music in the style of the common practice period, the final cadence for the piece as a whole should end in the key it began on in general. However, if you have musical reasons for ending on a half cadence, dissonance, a chord in a foreign key, etc, and it sounds convincing, then go ahead and do it.

I would strongly suggest reading the chapters on Tonality and Cadences in Leon Dallin's book "Twentieth Century Composition" 3rd edition, 1974 published by WM. C. Brown.


If it's modulated a good few bars before the end, then playing that I harmony will sound like it's arrived back at home, even though, technically, it hasn't. It's what sounds best that is the best. If it's just modulated somewhere away from the original key shortly before the end, then it'll sound better to either prolong that ending in the new key, or return to a key that sounds final. I say 'a key', as that may indeed not be the original at all.

So, to sum up, whatever chord sounds like it signifies the end in that the piece has come to rest, permanently, rather than temporarily, will be the most convincing, be it the original key or the one it last modulated to. No hard and fast rule. As usual, ears are best judges.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.