Behind the Edit - September 13, 2026

When ‘Said’ Isn’t Boring: The Dialogue Tag Debate, Settled with Data and Examples

When ‘Said’ Isn’t Boring: The Dialogue Tag Debate, Settled with Data and Examples

Open any fiction forum and you will find the argument in full swing. One writer posts a paragraph with “she exclaimed,” “he growled,” “they murmured darkly” — and gets told, politely or otherwise, to just use “said.” Another posts a page of clean, unadorned “said” and “asked” and is told the prose is monotonous. Both pieces of feedback can be correct. And that is exactly the problem: the dialogue tag debate is not really about which word to use. It is about understanding what these small grammatical workhorses are actually doing on the page.

This article cuts through the noise. We will look at what readability studies tell us, what professional editors actually recommend, how bestselling authors navigate the choice, and give you a practical, data-informed framework for making the right call in your own manuscript.

What Are Dialogue Tags — and Why Do They Matter?

A dialogue tag is any phrase that attributes spoken words to a character: the “she said,” the “he asked,” the “they whispered” that anchors dialogue to a speaker. Simple as they seem, they perform several jobs at once. They tell readers who is speaking; they can carry tonal information when chosen carefully; they function as rhythm markers, giving the eye a beat to rest on between lines of speech; and they can be replaced entirely by action beats, which ground the scene in physical space.

The stakes are real. Research into reading fluency consistently shows that common, predictable words are processed faster; they function almost subliminally. This is the core of the “said” argument. But it is also only half the picture. Overused or misjudged dialogue tags can pull a reader clean out of the story, which is the very opposite of their purpose.

The Case for ‘Said’: What the Data Actually Shows

The argument for “said” as the default dialogue tag is not aesthetic snobbery, it has a legitimate cognitive basis. Analyses of top-selling literary fiction consistently find that roughly 85% of all dialogue tags are “said” or “asked.” Eye-tracking studies show that fluent readers fixate on “said” for less than 70 milliseconds on average, processing it in the same near-automatic way they process function words like “the” and “and.” High-frequency words are processed up to four times faster than rare synonyms, which means unusual tags actively slow comprehension.

Finish your manuscript in 12 weeks, guaranteed suport, every step. 

This is the invisibility effect that champions of “said” point to. Bestseller analyses of Cormac McCarthy, Kazuo Ishiguro, and Sally Rooney find these authors using “said” as their near-exclusive tag, sometimes dispensing with tags altogether for pages at a time. The reader’s imagination fills in tone from the context and the dialogue itself. Writing teachers from Elmore Leonard to Stephen King have publicly championed “said” for precisely this reason: the word performs its attribution job without interrupting narrative flow.

The road to hell is paved with adverbs, and most dialogue tag sins are really adverb problems in disguise.

When ‘Said’ Genuinely Falls Short

Here is what the “always use said” camp tends to understate: there are genuine situations where a more specific tag does meaningful work that neither “said” nor an action beat can efficiently deliver. Vocal quality that carries narrative weight is one such case. A character who “rasped” or “croaked” may be establishing a physical condition the reader needs to understand, laryngitis, terror, old age, and a single well-chosen tag delivers that information more efficiently than a separate sentence of description.

Genre convention matters too. Thriller and crime fiction readers often expect a slightly denser tag vocabulary; the stripped-back minimalism of literary fiction is a stylistic register, not a universal rule. Children’s fiction and middle-grade actively employ a wider range of tags because the emotional signposting serves developing readers. And in a negotiation scene, a character who “stated flatly” versus one who “ventured” signals status and confidence the dialogue alone may not convey. Context, form, and audience all inform the right choice.

The Hierarchy of Dialogue Tags: A Practical Framework

Rather than “use said” or “never use said,” experienced editors tend to think in tiers. Tier one contains the invisible tags — “said,” “asked,” “replied” — which can be used freely because they cost the reader nothing. Tier two holds light-touch alternatives like “whispered,” “called,” and “continued,” which are acceptable when they carry genuinely useful information about volume or continuity of speech. Tier three — “murmured,” “snapped,” “laughed” — earns its place only when the emotional quality of the speech cannot be conveyed through the words themselves.

Tier four, the riskiest category, includes tags like “expostulated” or “opined.” These almost always read better when replaced by an action beat or when the dialogue itself is rewritten to carry the weight. Action beats — “He set down his glass. ‘I know what you did.’” — are not dialogue tags at all, but they are frequently superior: they ground the scene spatially, slow the pace, and add physical texture. The guiding principle is straightforward: if the dialogue itself communicates the emotion clearly, the tag is doing redundant work. If the tag is doing work the dialogue cannot do, it may be worth keeping.

Explore Resources for Indie Authors

Explore expert resources designed to help indie authors plan, revise, and publish with confidence, from editing guidance to industry insights you can actually use.

 

Explore Here

Common Dialogue Tag Mistakes (and How to Fix Them)

Most dialogue tag problems fall into a small number of recurring patterns. The first is the said-bookism: replacing “said” with fancy synonyms on every line — “averred,” “voiced,” “conveyed” — under the mistaken belief that variety signals quality. The second is the adverb crutch: “she said softly,” “he said angrily.” When the adverb is doing work the dialogue itself should do, the solution is to rewrite the speech, not to hunt for a better modifier.

A third common error is the physical impossibility: “he laughed,” “she smiled,” “they shrugged.” Characters cannot literally laugh or shrug words. These are action beats masquerading as tags and they require a full stop and restructuring. The redundant emotion tag is the fourth pattern: “‘I hate you!’ she said angrily.” The exclamation mark and the word choice already convey the emotion; “angrily” is noise. Finally, tag stacking in two-character scenes — labelling every single line after the speakers have already been established — signals distrust of the reader. After three or four exchanges, trust the alternation to carry itself.

I can’t believe you did that,” she hissed angrily.
“It wasn’t my fault,” he retorted defensively.

“I can’t believe you did that.” She turned away from him.
“It wasn’t my fault.” He crossed his arms.

What Bestselling Authors Actually Do

One of the most instructive exercises any writer can do is audit dialogue attribution in published fiction they admire. Literary fiction — Ishiguro, Rooney, Munro — tends toward near-total reliance on “said” and action beats, with long stretches of untagged dialogue between two speakers. Commercial thriller writers like Grisham and Patterson use a slightly wider vocabulary but still heavily favour “said”; the exception is when pace demands a beat with more information density.

Contemporary literary fantasy has largely moved toward the minimalist end, even as older genre fantasy was more permissive with creative tags. Dialogue-heavy screenplays use no tags at all, relying on speaker labels and action lines, a useful reminder that “said” is already a compromise, not the purest form of dialogue presentation. The pattern that emerges across all these forms is the same: the writer’s register is consistent, chosen deliberately, and serves the story’s pacing rather than signalling effort to the reader.

The Verdict: Rules, Exceptions, and Your Own Voice

The dialogue tag debate is not one debate. It is at least three: a debate about cognitive processing, a debate about genre convention, and a debate about authorial voice. Understanding which argument applies to your situation dissolves most of the confusion.

Default to “said” and “asked” — they are invisible, efficient, and never wrong. Replace tags entirely with action beats when you want to add physical or spatial information to the scene. Use Tier 3 and Tier 4 tags sparingly, only when the specific vocal quality is narratively significant and cannot be conveyed through the speech itself or an action beat.

Audit your adverbs: if you find “softly,” “harshly,” “nervously” attached to tags, that is a signal to rewrite the dialogue, not to hunt for a better adverb. Read your dialogue aloud; the ear catches tag problems that the eye slides past. The best dialogue tag is the one the reader does not notice. That is usually “said.” But a great writer knows when a single well-chosen departure from that default does work nothing else can, and uses it without apology.

Go back to your current manuscript. Highlight every dialogue tag that is not “said” or “asked.” For each one, ask: does this carry information the dialogue itself cannot? If the answer is no, cut it. If you find yourself cutting nearly all of them, the data suggests you are making your prose stronger.

Not Sure If Your Dialogue Tags Are Working?

Get clear, expert feedback on your manuscript and find out exactly what to fix.

 

Enquire Now

Found this helpful? Let us know in the comments!

Leave a comments