Spoken vs. Written Language
Spoken language and written language are different registers. Writing a script in prose and then reading or reciting it aloud tends to sound unnatural — because it wasn't generated the way speech is generated.
Written language is optimized for the eye: longer sentences, more subordinate clauses, denser vocabulary, no filler, no repetition. Spoken language is optimized for the ear: shorter phrases, natural rhythm, repetition for emphasis, the kind of word choices that come out when you're thinking aloud.
When someone writes a presentation script the way they'd write an essay and then tries to deliver it, the audience hears prose being recited — not a person talking to them.
The Dictate-First Method
Harvard professor Dan Gilbert's approach: speak the talk into a recorder before writing anything down. Use the transcript as the first draft of the script.
The result is language that sounds natural when delivered because it was generated naturally. The editing pass shapes and tightens it, but the register — the rhythm, the vocabulary, the sentence length — stays true to how the speaker actually talks.
When Written Language Works on Stage
The distinction isn't absolute. Some talks are deliberately literary — Amanda Gorman's spoken word poetry at a TED-Ed event is an example where the written register is itself the point. The audience understands they are hearing a crafted piece of writing, and the elevated language creates its own effect.
The rule: written language works on stage when the audience is meant to experience it as writing. When the goal is to feel like a person talking directly to you, spoken language wins.
Delivery Register
Choosing spoken-language words is only half the problem. Julian Treasure separates what you say from how you say it: prosody, pace, pitch, and the other controls in the vocal delivery toolbox layer meaning on top of the script. A natural-sounding script delivered in monotone still fails the ear.
Connections
- memorizing-a-talk — the dictate-first method produces the raw material that gets memorized; solving register and memorization together
- active-listening — the listener's ear processes differently from the reader's eye; spoken language is designed for that mode
- prosody — the melodic layer that written text cannot carry without a speaker
- vocal-delivery-toolbox — how delivery choices shape meaning beyond word choice