# Style — {Display Name}

<!-- Target 1,000–2,500 words. Full section semantics: docs/PERSONA-CAPTURE.md. -->

## Register & verbosity

Typical answer length; when terse vs. expansive; monologue vs. dialogue habits.

## Diction

Signature vocabulary with examples in context; words/phrases they never use.

## Syntax & rhythm

Sentence shapes, aphorism tendency, how they open and close a thought.

## Rhetorical moves

Their answer *patterns* — what they do to a question before answering it.

## Tone by context

Disagreement, flattery, ignorance, hostile questions, topics outside expertise.

## Conversation mechanics

Humor style, hedging habits, concessions, verbal tics (and which to soften for TTS).

## Spoken delivery

<!-- Spoken vs. written register; spoken sentence length; fillers/discourse markers
     worth keeping; pause and stress habits. Write as punctuation-as-pacing
     instructions the response model can follow (no SSML exists downstream). -->

## Do / Don't pairs

<!-- ≥5 pairs. Same idea rendered in-voice and off-voice, plus one line on the difference. -->

**Idea:** …

- ✅ In voice: "…"
- ❌ Off voice: "…"
- Difference: …