Show Don't Tell: How to Find It in Your Own Manuscript

Everybody knows "show don't tell." It's the first piece of craft advice any fiction writer hears, usually within the first week of taking writing seriously. It gets repeated so often it starts to sound like background noise.

And that's the problem. You KNOW the rule. You nod along when someone explains it. You can spot blatant telling in someone else's work without breaking a sweat. But sitting in front of your own manuscript at 6 AM, coffee going cold, staring at chapter nine for the third time this week... can you find it in YOUR prose?

Probably not. And that's not because you're bad at this. It's because telling is sneaky in your own work. You know what you MEANT, so you read what you meant instead of what's on the page.

What "telling" actually looks like (it's not what you think)

Most writers think telling means writing "she was angry" instead of showing the anger through action. And yeah, that's telling. But it's the obvious kind. The kind you'd catch on a first pass.

Craft editors identify a much subtler category called "telling after showing." This is where you DO the good work of showing a character's emotional state through action, dialogue, body language... and then you tack on the explanation anyway.

Here's what that looks like.

Telling after showing (the sneaky kind):

She slammed the folder on the desk and turned to the window, her jaw tight. She was furious.

You already showed the fury. The slammed folder, the tight jaw, the turned back. "She was furious" doesn't add information. It tells the reader something they already understood, and in doing so, it says: I don't trust you to get it.

Just the showing:

She slammed the folder on the desk and turned to the window, her jaw tight.

That's it. The reader lands the emotion themselves, and it hits harder because THEY did the interpretive work. Developmental editors call this "the most insidious" form of telling because writers feel like they're showing (they are!) and then accidentally undercut it with a summary.

The four patterns to search for

The challenge is building a reliable method for catching telling in your own work. Not a vague "read it out loud and feel for it" approach. Something specific. Something repeatable at chapter forty without your eyes glazing over.

Here are the four patterns that actually work as a show don't tell checker for your own prose.

1. Emotion words after action beats

Search your manuscript for emotion labels that appear within a sentence or two of physical action. Words like: angry, sad, nervous, excited, terrified, frustrated, relieved, jealous, ashamed.

If you've got a character clenching their fists and THEN you write "he was angry," you're telling after showing. The action beat did the work. The label is a crutch.

This is the single most common telling pattern in fiction manuscripts. Developmental editing guides devote entire chapters to it because it's that pervasive.

2. "Felt" and "realized" as filters

These two words are what craft instructors call "filter words." They insert the author between the reader and the experience.

Filtered (telling):

She felt the cold wind cut through her jacket.

Unfiltered (showing):

The wind cut through her jacket.

See the difference? "She felt" tells you about the character's experience from the outside. Dropping it puts the reader inside the experience directly.

Search your manuscript for "felt," "realized," "noticed," "watched," "saw," "heard," "wondered," and "thought." Not every instance is a problem (sometimes you genuinely need them for clarity). But each one is worth a second look, because most of the time the sentence works better without the filter.

3. Paragraphs of internal summary between scenes

This one's harder to search for mechanically, but it's worth knowing the shape of it. After a scene with dialogue and action, you'll sometimes find a paragraph (or three) where the viewpoint character reviews what just happened and explains what it means.

After he left, she sat in the kitchen thinking about everything Marcus had said. He clearly didn't trust her judgment anymore. Their partnership was falling apart, and she wasn't sure she could fix it.

That's the author providing an interpretation guide. If the scene worked, the reader already knows Marcus doesn't trust her judgment. If the scene DIDN'T work, the summary paragraph won't save it.

The standard editorial advice here is blunt: cut the interior summary and see if the scene still communicates. If it does, the summary was telling. If it doesn't, the scene needs work, not a footnote.

4. Adverbs on dialogue tags

"He said angrily." "She whispered nervously." "He asked suspiciously."

These are emotional stage directions. They tell the reader how to interpret dialogue that should be doing the interpretive work itself. If "I don't want to talk about it" needs "she said coldly" to convey the coldness, the dialogue probably needs rewriting, not an adverb.

This is maybe the most searchable pattern of the four. Open your manuscript, search for "ly " near quotation marks. You'll find them.

(One FirstReader beta tester found fourteen in a single chapter. Fourteen. It was humbling for all of us.)

How to actually do the audit

Knowing the patterns is one thing. Sitting down and working through 80,000 words is another. Here's what actually works.

Go chapter by chapter, not all at once. You can't audit an entire manuscript for telling in a single session. You'll go blind to it by chapter five. One chapter per sitting, maybe two if they're short.

Use your word processor's search function. Start with the mechanical patterns. Search "felt," search "realized," search "ly " near dialogue. Mark every hit. Then go back and decide which ones to cut. This is grunt work, but it's the kind of grunt work that actually moves the needle.

Read the dialogue scenes with ONLY the dialogue visible. Copy a scene's dialogue into a blank document, stripped of all tags and beats. Does the conversation make sense? Can you tell who's speaking? Can you feel the emotional temperature? If yes, your showing is working. If no, the scene has a telling dependency, and that's worth knowing.

Get an alpha reader to look specifically for telling. Not a general "what did you think?" pass. A specific ask: "flag every place where you feel like I'm explaining something I already showed." Alpha readers are gold for this because they haven't read your draft forty times. Their eyes are fresh. They'll catch the telling-after-showing that you've gone completely blind to.

Where this advice doesn't apply

Worth being honest about the limits. Not all telling is bad. Every serious craft guide says this explicitly.

Transitional passages sometimes NEED telling. "Three weeks later, she was back in New York" is telling, and it's fine. You don't need to show every elapsed time period through scene work.

Genre conventions matter too. Thrillers tell more than literary fiction. Romance readers expect some interior emotional access that technically counts as telling. Knowing your genre's tolerance is part of the craft.

The goal isn't zero telling. The goal is INTENTIONAL telling. Telling because the story needs it, not telling because you didn't notice you were doing it.

Tools that can help

The manual audit works. But it takes a long time, and it gets harder to sustain attention across a full manuscript. A few things can help.

Your word processor's search function handles the mechanical stuff (filter words, adverbs on dialogue tags). That's a solid first pass and it costs you nothing but time.

A good alpha reader catches the interpretive stuff, the telling-after-showing that no search function will find. If you've got someone willing to do that kind of read, that relationship is worth investing in.

And if you want a tool that checks your manuscript against established craft principles, including show vs. tell, FirstReader does this at the chapter level across your entire manuscript. It's rules-based, not statistical. Every finding traces back to a named craft principle. It flags the specific passages and tells you which principle applies, so you're not guessing at WHY something got flagged. It functions as an alpha reader that applies consistent craft analysis to every chapter without losing focus at page 200.

If you're curious whether it would catch things in your prose, you can try a free chapter analysis. One chapter, no credit card.

(And if you're skeptical about whether AI can do this kind of craft analysis at all... good. Healthy skepticism is the right starting posture. That piece goes into the honest answer.)

The real test

Show don't tell isn't a rule you learn once and then execute perfectly forever. It's a calibration you keep adjusting. Every manuscript has telling in it. Every single one. The question is whether you can find it before your readers do.

The four-pattern search I described above won't catch everything. But it'll catch the structural, repeatable telling that accounts for most of the problem. And once you start seeing the patterns in your own work, you can't unsee them.

Which is annoying, honestly. But also the point.

If you found this useful, I'd love to hear about it. And if you've got your own techniques for catching telling in your own prose, drop a comment. I'm always looking to sharpen this stuff.