Argument Graph Discussion Platform

Overview

The Argument Graph Discussion Platform is an experiment in turning chronological discussion into a reusable knowledge structure. Claims, counterclaims, rebuttals, and evidence are treated as independent nodes, then connected through explicit relationships.

Can online discussion be accumulated as structured knowledge instead of remaining a simple conversation log?

---

Problem

Traditional forum discussions push context downward as new comments arrive. Important issues become scattered across reply chains, making it difficult to see which claims were challenged, which evidence remains useful, and which questions are still unresolved.

Structure

This project treats discussion as an argument graph instead of a linear timeline. Claims, responses, contradictions, rebuttals, and evidence are stored as separate record units, and each relationship receives a defined type. Readers can follow the debate by moving through the local structure around a claim.

Claim A ──supports──> Claim B
Claim C ──rebuts────> Claim A
Evidence D ──grounds──> Claim C

Key Idea

A debate record should not disappear as a finished conversation log. It can become an accumulated knowledge asset that supports repeated issues, previous rebuttal paths, and later summarization or evaluation.

• Claims remain as independent nodes.
• Counterclaims and rebuttals are connected as relationship edges.
• Evidence becomes a reusable record for repeated discussions.

discussion
↓
claim extraction
↓
argument graph
↓
knowledge structure

Feature Direction

The initial direction includes claim registration, response linking, contradiction markers, rebuttal tracking, and evidence node accumulation. Later work can add issue-focused views, knowledge map exploration, LLM-assisted summaries, and argument quality scoring.

Technical Notes

The core model is a statement node and a relationship edge. Nodes hold the statement and its explanation, while edges express support, rebuttal, contradiction, or extension. The interface should prioritize graph-local exploration and issue filters over a plain comment list.

Future Possibilities

LLM support can help summarize distributed claims, detect unresolved issues, and expand the debate record into a knowledge map without replacing human judgment.