You wrote three sentences. Then you read them back and they looked curt, so you added a softening line. Then the softening line looked defensive, so you added context. Then the context looked like an excuse, so you added a joke. Now it's four paragraphs long, you've apologized twice for something that isn't your fault, and you're hovering over send with the same low hum in your chest you get before a doctor's appointment.

Here's what almost nobody tells you: the problem isn't that you're insecure, or verbose, or bad at writing. The problem is that you are having a conversation with someone who never nods. And your brain, which was built across a few hundred thousand years to talk to faces, keeps waiting for a signal that will never arrive — so it keeps talking.

Every conversation runs on receipts

When you speak to someone in person, they are working just as hard as you are. They're nodding. Saying mm-hm. Tilting their head at the word that didn't land. Laughing a half-beat early because they've already guessed the punchline. Linguists call these backchannels, and psychologists Herbert Clark and Susan Brennan folded them into a broader theory called conversational grounding: the idea that understanding isn't something a speaker transmits, it's something two people build together, in real time, with constant tiny receipts.

Grounding is why you can say to your sister, "So the thing with Mum," and she says, "Yeah," and you both know exactly which thing. Fourteen words did the work of four hundred. You weren't being efficient. You were being fed. Every micro-signal she gave you told you: keep going, that's enough, I have it, you don't need to say more.

Clark and Brennan also observed that grounding costs shift depending on the medium. Face-to-face is cheap — feedback is instant and free. Writing is expensive. Your reader isn't there. There is no mm-hm. There is nothing to tell you the sentence landed.

So you do the only thing an ungrounded speaker can do. You keep speaking, hoping that somewhere in the excess, the receipt will arrive.

Over-explaining is a search, not a flaw

Watch what actually happens in your head when you pad an email. You write Could we push Thursday to Friday? and then some part of you — the part that has been reading rooms since you were four years old — asks: did that sound annoyed?

In a room, you'd already know. You'd have seen the face. Instead, you get silence, and into that silence your brain pours the most useful thing it has: simulation. You imagine the reader. And here's the cruel part — the imagined reader is not neutral. Under uncertainty, we don't guess the average outcome; we brace for the bad one. Your simulated reader is faintly irritated. Slightly disappointed. A little suspicious of you.

So you write to that person. You add the reason. Then the reason behind the reason. You apologize preemptively. You attach a small biography of your week. And when you're done, the email doesn't read as thorough. It reads as anxious — because it is a transcript of anxiety, and readers are very good at hearing that.

The deep irony: over-explaining is your social intelligence working perfectly, in a medium that starves it. It's a thermostat in a room with no thermometer, cranking the heat higher and higher because it can never read the temperature.

Why you don't do this out loud

Notice that you almost never over-explain when you talk. Ask yourself the last time you left a voice note that spiraled into four apologetic paragraphs. It happens, but rarely. Say the same request aloud — hey, can we push Thursday to Friday? — and you'd stop there. It sounds fine. It is fine.

Part of this is prosody: spoken language carries warmth in the signal itself. Pitch, pace, a slight upward lilt at the end. Those are all doing the reassurance work that, in text, you have to do with extra words. Strip the voice out and the warmth goes with it, so you rebuild it manually, sentence by sentence, out of hedges and apologies and just wanteds.

But part of it is simpler and stranger. Speech is committed. Once a sentence leaves your mouth, it's gone; you move forward. Writing is infinitely revisable, and revisability invites rumination. The cursor sits there blinking, offering you unlimited chances to imagine being misread. Take one of them and you'll take twenty.

This is the practical insight underneath everything: the fastest way to write like a person who isn't afraid is to say it before you write it. Not as a trick. Because speech is the one mode where your brain stops waiting for a nod it isn't going to get, and simply finishes the thought.

The test that ends the spiral

Before you send anything you've rewritten more than twice, read it out loud as though the person were sitting across from you.

You will hear it immediately. Nobody says, Sorry for the delay, I just wanted to quickly circle back and check whether, if it's not too much trouble, we might possibly be able to — you'd never get through it. Out loud, padding is unmistakable. It sounds like someone backing out of a room.

And when you cut it down to what you'd actually say — Sorry, slow reply. Can we do Friday? — a small alarm goes off. Too blunt. Sit with that alarm for a second, because it is the whole thing. That is not blunt. That is what warmth sounds like without padding. Padding isn't warmth. It's fear wearing warmth's coat.

Your next moves

  • Say the message out loud first, then write what you said. Before typing any email that has made you nervous, speak the request to an empty room in one breath, as if to a friend. Whatever you said is the draft. Write that.
  • Run the apology audit. Search your last twenty sent messages for sorry, just, and I hope this isn't. Count them. Note how many mark an actual wrong you committed. The rest are backchannel substitutes — receipts you were writing to yourself.
  • Delete the last paragraph before you send. In an over-explained message, the final paragraph is almost always the anxiety, not the content. Cut it. Reread. It will be better nine times in ten.
  • Set a one-question rule for hard emails. If a message contains a request, it gets exactly one sentence of context and one question mark. Everything else goes in a follow-up, if it's needed. It usually isn't.
  • Move the grounding to a medium that provides it. If a message has taken you more than ten minutes, it was never a message. Say "Can I call you for two minutes?" and let their voice give you the nods your document can't.

The nod you can give yourself

Somewhere in this, there's a gentler idea than write shorter emails. Over-explaining isn't a character defect to be beaten out of you. It's what a socially attentive person does when the social feedback is switched off. You are not too much. You have been talking, patiently and at length, into a room where nobody was allowed to say I understand.

So stop writing into the silence. Speak first, and let the writing be a record of the sentence you already finished.

That's the small thing Quill was built for: you speak into any app, on-device and private, and clean text appears — in the shape your mouth gave it, before the second-guessing sets in. If you want it kinder or crisper afterward, one tap rewrites it. But most of the time you'll find you don't need to. The version you said out loud was already the version you meant.

If you'd like to hear what your writing sounds like when it stops apologizing, try Quill.