Message Corpus
About this corpus
This is a fictional corpus of 273 text messages between two co-parents — David Mercer and Sarah Chen — regarding their 9-year-old daughter Lily. The messages span January through August 2025 and were designed to illustrate the full arc of coercive control patterns in post-separation communication.
Browse the messages below. Click any message to see it in context with surrounding messages. Use the sidebar filters to narrow by sender, date range, or keyword. Click a Fontes tactic in the sidebar to see only messages classified with that tactic.
Framework Loading
Why this step matters
Every AI tool you might use probably knows something about coercive control. But it may not know what you want it to know, or how you want it to analyze.
The AI performs dramatically better when you give it the analytical framework before you give it any data. Without this step, you get generic conflict analysis. With it, you get CC-informed pattern detection.
This prompt loads three frameworks simultaneously:
Evan Stark
Coercive control as a course of conduct — not isolated incidents. The pattern across time reveals control.
Lisa Aronson Fontes
10 Tactics taxonomy — Isolation, Degradation, Gaslighting, Rules & Restrictions, Withholding, Economic Abuse, Institutional/Legal Abuse. The classification system used throughout this tool.
Emma Katz
3 post-separation patterns — dangerous behavior, seemingly caring behavior masking abuse (the most critical for text analysis), and omnipresent control.
Tip: Modern AI platforms can process an entire book's worth of text at once. You can feed the AI academic papers or legal briefs alongside the messages if you want it to apply a particular analysis.
Click Show Prompt above to see the exact framework loading prompt, then copy it for your own use.
Pattern Detection
What this analysis does
Pattern Detection applies Fontes's 10 Tactics taxonomy to every message in the corpus. Instead of reading messages one at a time, the AI identifies patterns across the full body of communication — connecting messages weeks or months apart that belong to the same tactic category.
A single demanding message is co-parenting friction. A pattern of unilateral demands, conditional cooperation, and information withholding across months is coercive control. This analysis tells the difference.
Asymmetry Analysis
What this analysis does
Asymmetry Analysis compares the communication registers of both parties. It answers the question: who sets terms, and who accommodates?
In coercive control, the controller sounds reasonable in any individual message. The asymmetry only becomes visible when you measure it across the entire corpus — who initiates, who demands, who references authority, who adjusts. The AI quantifies what no human can hold in their head.
Escalation Timeline
What this analysis does
The Escalation Timeline tracks how tactics change and intensify over time. Coercive control is not static — it typically escalates when the controlling party faces resistance or perceives loss of control.
Each dot represents a classified message, color-coded by Fontes tactic. Click any dot to see the message in context. Use the tactic filters to isolate specific patterns, and toggle phase labels to see the AI's assessment of escalation stages.
Legal Mapping
What this analysis does
Legal Mapping takes the patterns identified in earlier rounds and organizes the evidence against specific legal standards. For each statutory element, the AI cites specific messages and explains how the cumulative pattern — not any individual message — meets the standard.
Select a framework below to see how the evidence maps. Click any element card to see the supporting messages. The analysis also anticipates and rebuts common counter-arguments.
Export Report
Generate DOCX Report
Export all analysis sections as a formatted Word document suitable for court filings or case documentation.