The other night I assembled an Aviation cocktail for the first time. It’s a drink my friend, Jared, likes, so I texted him a picture of my first attempt. It’s made with gin, maraschino liqueur, crème de violette and lemon juice. I’d followed a recipe out of a bar book. When I commented that it didn’t taste as good as I remembered, he suggested his own recipe, which differed from the book very slightly. Specifically, he suggested 1/2 an ounce of lemon juice rather than 3/4 as the book. That was the only difference.

I tried it, and that tiny change made a major difference in the taste. And not in the way I had expected. The drink tasted slightly less sweet with a bit less lemon. Two of the ingredients are very sweet, whereas lemon juice is generally more tart, so I don’t know if it was just a contrast change in the mix or what, but the tiny adjustment made a big improvement.

I’ve been struggling with the revision of my novel, The Trickster Apocalypse for a while. After working on the first draft for a long time, regularly reading chapters to my monthly Writers’ group, I had revised and assembled the whole thing, and gotten three people to agree to read it all the way through. There had been some common comments from all of them regarding some frustration with the protagonists or inconsistencies in their characterization.

So I’ve been re-reading and revising. I recently shared two new scenes and one heavily revised one with two of the readers and my group, and there was a consensus that these little revisions changed their perception of the main plot and one subplot significantly.

I’ve described the novel as “a light fantasy in an epic fantasy wrapper using anthropomorphic tropes to tell how reluctant and unlikely heroes try to avert a prophesied apocalypse.” As a light fantasy, certain things happen in the story because they’re funny. It was easy, especially when I was working on the first draft and reading it to others in a serialized fashion, to pepper in jokes throughout. People laughed when they read the scenes, so that seemed like a good thing, right?

But when someone read the whole thing in un-serialized circumstances, a couple of the jokes late in the book subverted the emotional arc of at least one of the protagonists. It’s not that I can’t have jokes late in the book, but I can’t show one character’s emotional journey from reluctant to get involved to taking a stand in a big showdown if I keep showing ways that he is trying to dodge responsibility. A scene that would have been funny and in character in chapter four doesn’t work in chapter seventeen, after the reader has watched the character start growing beyond that.

The scene is funny, which is good in a light fantasy, but any scene in a novel needs to either advance the plot, establish or resolve a conflict, illuminate a character, show how a character has changed, reveal new information to the reader, or hit an important emotional beat. It’s not that the scene has to go, but every scene, particularly jokes about the protagonist’s character need to move the things forward, not back.

I should have realized during an earlier revision phase a couple of these developments were actually throwbacks to earlier versions of the character. I didn’t in part because in a couple of cases the jokes were working so well, I wanted to keep them in. When people repeat the classic writing advice to “kill your darlings,” it isn’t because your favorite lines or sequences are always bad, it’s because that sometimes, because some bit is a favorite, it blinds us from noticing that it’s wrong for this scene or this stage of the story.

Removing a single misplaced joke can change the taste of the entire tale.