The reason a happy ending changes the moral

The same eleven years of struggle, sacrifice and lost relationships read as vindication if the venture succeeds and as warning if it fails. Not one event differs between the two tellings. The reader is given the last chapter and reinterprets the previous ten.

What everyone sees

Endings are treated as the final instalment, one more piece of the sequence, weighted about the same as the others. They are not weighted the same. Retrospective evaluation is dominated by end states, so the ending is not a part of the story so much as the instruction for reading the rest of it.

What is actually happening

Kahneman, Fredrickson, Schreiber and Redelmeier demonstrated the general mechanism outside narrative entirely: participants preferred a longer aversive episode with a milder ending to a shorter one that ended at peak intensity, showing that remembered evaluation tracks peak and final moments rather than integrating over duration. Bruner’s account of narrative as a mode of thought supplies the interpretive layer — events acquire meaning from their position in a plot rather than from their properties, so a change in the plot’s resolution reassigns the significance of its parts. The two combine into an editorial fact: whoever controls the last scene controls the reading of every earlier one, without touching a single earlier scene.

Why it stays hidden

The reassignment hides because it happens in memory rather than on the page. A reader who is asked can recount the events accurately, and would deny that the ending changed them. What changed is which events count as central and which as incidental — and centrality is not something readers can inspect in their own recall.