Voice Analysis

Voice & Tone Analysis

type: utility mode: generative triggers:

• "capture writing voice"
• "analyze writing style"
• "write like me"
• "voice guide"
• "tone analysis"
• "style fingerprint"

Purpose

Extract and document a writer's distinctive voice patterns for consistent reproduction. Creates a "voice guide" that enables authentic writing that sounds like the source, not a generic approximation.

Core Principle

Capture spirit, not just mechanics. The goal is writing that makes the source say "yes, that's me" not "I guess that's accurate."

Phase 1: Sample Collection

Gather 5-10 Examples from Each Category

Peak Voice - Writing they identify as "most them"

Off-Voice - Writing that doesn't represent them well

Different Contexts:

• Technical/instructional content
• Persuasive/argumentative pieces
• Narrative/storytelling
• Casual communication (emails, messages)
• Formal communication
• Emotional/vulnerable content

Self-Report Prompts

Rewrite Exercise: Ask: "Rewrite this neutral paragraph in your voice:"

"The new policy will be implemented next month. It includes several changes to current procedures. Employees should review documentation and submit questions by the deadline."

Rule Breaking: "What writing 'rules' do you consistently ignore? Why?"

Pet Peeves: "What writing choices immediately signal something wasn't written by you?"

Evolution: "How has your writing changed in 5 years? What stayed constant?"

Phase 2: Linguistic Analysis

Sentence Level

Paragraph Architecture

Punctuation Signature

Phase 3: Lexical Fingerprinting

Word Choice Matrix

Register Analysis

• Consistent register (formal/informal throughout)
• Deliberate register mixing (formal content, casual asides)
• Context-dependent shifting (formal for X, casual for Y)

Recurring Constructions

List phrases/patterns appearing 3+ times:

Phase 4: Conceptual DNA

Metaphor Mapping

Reference Pool

• Cultural touchstones: (movies, books, memes, history...)
• Time period: (contemporary, 90s, classic...)
• Accessibility level: (mainstream, niche, insider)
• Domains drawn from: (sports, cooking, science...)

Reasoning Patterns

Rate 1-5 for prevalence:

• Analogical reasoning (like X, therefore Y)
• First principles (from basics up)
• Empirical evidence (data, studies)
• Personal anecdote (I experienced...)
• Hypotheticals (imagine if...)
• Socratic questioning (but what if...?)

Phase 5: Emotional Texture

Enthusiasm Spectrum

Criticism Styles

Vulnerability Patterns

• Admission phrases: "I'll admit...", "honestly..."
• Uncertainty markers: "I think...", "not sure but..."
• Personal revelation style: Direct? Buried in humor? Rare?

Phase 6: Reader Dynamics

Positioning

The writer positions as:

• Expert/teacher (I know, let me explain)
• Peer/collaborator (we're figuring this out together)
• Student/learner (I'm working through this)
• Challenger/provocateur (conventional wisdom is wrong)
• Guide/facilitator (here's how to navigate)

Assumed Context

• Shared knowledge level: Assumes expertise? Explains basics?
• Cultural assumptions: In-group references? Universal?
• Relationship warmth: Distant professional? Familiar?

Interactive Patterns

• Questions per 1000 words: ___
• Direct address frequency ("you"): ___
• Imperative usage (commands): ___
• Inclusive language ("we/us"): ___

Phase 7: Voice Guide Synthesis

Core Voice Statement

In 2-3 sentences, capture the essence:

The Rules That Matter Most

Always:

Never:

Usually, unless:

Sentence Construction Guide

• Preferred length:
• Variety pattern:
• Opening moves:
• Power positions: (where key info lands)

Word Selection Principles

• Go-to words for [concept]:
• Banned words/phrases:
• Register rules:

Structural Signatures

• Paragraph rhythm:
• Transition style:
• Information architecture:

Emotional Register

• Default tone:
• Excitement expression:
• Criticism approach:
• Vulnerability threshold:

The Litmus Test

A piece captures this voice when: 1. 2. 3.

Red Flags

Definitely NOT this voice when: 1. 2. 3.

Phase 8: Validation

Before finalizing the voice guide:

• Can identify the author in a blind test?
• Guided writing feels authentic, not performative?
• Patterns are descriptive, not prescriptive?
• Captures spirit, not just mechanics?
• Source would say "yes, that's me"?

Quick Reference Template

In Every Piece

The Heart of the Voice

[Single paragraph essence]

Emergency Voice Recovery

When writing has gone generic, add: 1. 2. 3.

Usage Notes

For AI Writing

Once the voice guide is complete, include relevant sections in the prompt to guide generation toward authentic voice reproduction.

For Self-Analysis

Writers can use this framework to understand their own voice, identify what makes their writing distinctive, and consciously apply those patterns.

For Editing

Use the voice guide as a checklist when editing to ensure consistency and authenticity.

Anti-Patterns

1. Mechanics Over Spirit

Pattern: Cataloging every linguistic feature without understanding what makes the voice feel distinctive. Why it fails: A perfect inventory of word frequencies and sentence lengths can produce writing that's technically accurate but feels like a parody. Voice is gestalt, not components. Fix: Start from "what makes this voice feel like this?" Work backward to mechanics. The inventory serves understanding; understanding doesn't emerge from inventory alone.

2. Single-Context Capture

Pattern: Analyzing voice from one type of writing, then applying it to all contexts. Why it fails: Writers shift voice across contexts. Technical writing voice differs from casual email voice. Capturing one context and forcing it everywhere creates uncanny artifacts. Fix: Sample across contexts. Map how voice shifts. Include context-switching rules in the voice guide. Understand which elements are constant vs. context-dependent.

3. Frequency as Rule

Pattern: If they use em-dashes 8% of the time, the voice guide prescribes 8% em-dash usage. Why it fails: Frequency is a statistical average, not a style rule. Forced frequency creates awkward placement. Natural writers don't count punctuation. Fix: Understand when they use em-dashes, not how often. "Uses em-dashes for dramatic interjections, rarely for lists" is actionable. "8% em-dashes" is not.

4. Imitation Artifacts

Pattern: Voice-guided writing that feels like someone doing an impression—technically accurate but overperformed. Why it fails: Distinctive features become tics when isolated. Real voice balances distinctive and neutral. Guides that catalog only distinctive features produce caricature. Fix: Include neutral baseline alongside distinctive features. Most sentences should sound natural, with distinctive features emerging at appropriate moments, not constantly.

5. Frozen Voice

Pattern: Treating the voice guide as permanent, not updating as the writer evolves. Why it fails: Writers change. A voice guide from 2020 may not fit 2025 writing. Using outdated guides produces writing that feels like an old version of the person. Fix: Note the capture date. Plan periodic updates. Include the writer's own reflections on how their voice has evolved. Treat the guide as living documentation.

Integration

Inbound (feeds into this skill)

Outbound (this skill enables)

Complementary