Colloquialism in Literature: What Writers Change

Last month I was reading Huck Finn out loud — first time in years. And something caught my attention. When Twain writes "I says" instead of "I say," it's not just a spelling quirk. It's a whole different voice. But I realized most people confuse spelling with dialect. They're not the same thing at all.

A novelist who wants to show a character speaking non-standard English has exactly three things on the page they can change: the spelling, the grammar, or the words. Twain changed all three. Hurston changed all three. Most novelists working today change one or two. The one they usually skip is the spelling.

The reason: these three layers don't cost the same. Spelling is loud and loaded. Grammar does more work than people give it credit for. Word choice is quiet and ages best. I'll show you what each one actually does, what it costs, and why most contemporary writers choose to leave spelling alone.

If you're here for the terms themselves — colloquialism versus slang, idiom, and jargon — I've covered that separately in Colloquialism vs Slang vs Idiom vs Jargon. And if you want spoken examples rather than literary ones, there's a guide to everyday colloquialisms. This one is about what writers actually do with dialect on the page.

The loudest layer isn't the most useful

Two very different techniques get filed under "dialect spelling." Confusing them is the most common mistake in this whole subject.

Phonetic respelling is straightforward. Gonna for going to, fust for first, tuh for to — these tell you a sound actually changed. The reader's ear is being instructed.

Eye dialect does nothing of the kind. Wimmin, sez, wuz, git, enuff — read them aloud next to their standard spellings and you'll produce identical sounds.

The linguist George Philip Krapp named this in 1925 in The English Language in America. His description is still the sharpest one I've found: the convention violated, he wrote, is one of the eyes, not of the ear.

So what is eye dialect doing, if not recording speech? Krapp's answer is uncomfortable and, I think, correct. The misspelling is a signal passed from author to reader over the character's head — a shared glance that places both of them above the person speaking. It marks the speaker as uneducated by exploiting the fact that English spelling is chaotic enough to make any word look wrong.

That's the layer to handle carefully. It isn't evidence about pronunciation. It's a judgment about standing.

I should be fair here: not all respelling works this way. When Hurston writes Ah for I in Their Eyes Were Watching God, that's not eye dialect. It's a real feature of Southern speech — the flattening of the vowel in I and prize that linguists call monophthongization. Same with tuh and 'im. She was recording something. When Irvine Welsh writes tae and whae in Trainspotting, he's doing the same job for Scots.

The cost of that job is measurable. Trainspotting was published with a glossary. A book that needs a glossary in its own language has spent something to get what it got. Welsh decided the trade was worth it. Most novelists don't.

Grammar carries the voice

Strip out every respelling and you can still hear the voice. The grammar is doing most of it.

Huck Finn's narration runs on nonstandard subject-verb agreement, double negatives, and clause structures no schoolbook would allow. The wages is the same. He couldn't stand it no longer. None of that requires a single altered spelling. Hurston's characters do the same. So does Holden Caulfield, in a milder register, with sentences that keep adding clauses because he can't be bothered to stop.

This layer travels better than spelling. It survives being read aloud. It survives translation. It doesn't ask the reader to decode anything. It also happens to be where regional English is genuinely systematic — Southern English has structures like might could and I got me a truck that are rule-governed and specific. A writer who marks grammar is drawing on real patterns. A writer who marks only spelling is drawing on the reader's prejudices.

I don't have a neat statistic for how much of the voice comes from grammar versus vocabulary. Neither does anyone else, honestly. But if you've ever read dialect-heavy fiction, you know the feeling: you can hear a character even with all the spelling normalized. That's grammar at work.

Word choice is the quietest, and the most interesting

Then there's plain vocabulary. No spelling touched. No grammar broken.

Huck says he's got the fantods. Holden says someone's a phony. Neither word is misspelled. Neither sentence is ungrammatical. Both tell you exactly who's talking and roughly when.

Salinger's case is the interesting one: Holden is an expensively educated boy who could speak formally and refuses to. His informality isn't a limitation. It's a position. Spelling can show what a character can't manage. Only word choice can show what they won't.

The far end of this layer is invention. Anthony Burgess built an entire lexicon for A Clockwork Orange — a slang called Nadsat that runs Russian roots into English syntax. No respelling. No broken grammar. Just an unfamiliar dictionary.

What Twain actually said

Every discussion of literary dialect eventually cites the "Explanatory" note at the front of Adventures of Huckleberry Finn. Twain lists the dialects he's using and insists the distinctions weren't guesswork but done "painstakingly, and with the trustworthy guidance and support of personal familiarity."

It's a good defense. It's been quoted approvingly for over a century. It also doesn't survive much checking.

David Carkeet, who studied Twain's dialects closely, concluded that Twain's recall was imperfect. Huck's own dialect isn't as consistent as the claim suggests — plausibly because Twain wrote much of the novel years after starting it and didn't hold every choice in his head across the gap. Critics have argued since publication that the seven dialects are hard to tell apart on the page.

I'd take one thing from that. A writer's statement about their own technique is a frame, not evidence. The Explanatory worked as a frame — it's still working, since everyone still quotes it — but it isn't a measurement of what's in the book.

Worth noting how contested it was at the time, too. When the Concord public library committee rejected the novel in 1885, the dialect was part of the stated objection: rough, ignorant language, systematic bad grammar. What reads as literary daring now read as a lapse in hygiene then.

One technique, two verdicts

Hurston is the case that shows how unstable all of this is.

Their Eyes Were Watching God came out in 1937. Richard Wright reviewed it savagely in New Masses, charging that Hurston was carrying on a "minstrel technique" — writing Black characters as quaint entertainment for white readers, and sidestepping the political subjects he thought the moment demanded. Several Black critics of the period agreed. The dialect writing was central to the objection.

The counterweight: Hurston wasn't guessing. She was a trained anthropologist who had done fieldwork in Southern Black communities. Her transcriptions rest on more direct observation than most of her critics had.

The novel went out of print anyway. Alice Walker's rediscovery in the mid-1970s brought it back. Henry Louis Gates Jr. later described it as a "speakerly text" — a book built to be heard, with the vernacular as the whole architecture rather than decoration on top.

Same orthography. Opposite verdicts, fifty years apart.

I don't think you can settle that from the page. There's no textual feature that distinguishes faithful rendering from caricature. The difference lives in who wrote it, what they knew, who was reading, and what the surrounding culture was doing with such portrayals at the time. That's an unsatisfying answer for a craft question. It's the honest one.

What most writers do now

Contemporary fiction mostly keeps spelling near standard and lets grammar and word choice carry the voice. Some of that is inherited caution about the Hurston problem. Some is just that heavy respelling slows readers down for a return that turns out to be small.

I think it's the right trade. The two quieter layers were always doing more of the work anyway.

Here's a test you can actually run, and it's Krapp's. Before you respell a word, read the respelling and the standard spelling aloud. If they sound the same, the spelling isn't telling anyone how the character talks. It's telling them what to think of him.