Exhibits/debate/how it works
The Debate Mapper (DM) does the following:
- enforces discipline in the discussion by requiring every comment ("debate point") to include certain information:
- which other point it is responding to
- whether it is intended to support or refute that other point
- calculates whether the central claim (the "root" point) of the debate has been refuted or not (this is the complicated part)
There is a slightly more technical explanation on HTYP: InstaGov/Debate Mapper
- Each comment in the debate ("debate point" or just "point" for short) is either a root claim or a response.
- Each response point is either a supporting ("pro"), dissenting ("con"), or neutral/informal ("inf") claim.
- The root claim and its responses, and their responses, etc. are all mapped into a tree (directed acyclic graph).
The slightly complicated part is the rules by which responses affect their "parent" points. Any comment may be considered "refuted" depending on the state of its responses; the rules are applied recursively, starting at the "twigs" (points with no responses) and working inward.
There is some room for experimentation to determine which sets of rules work best; the rules we are currently using are:
- If a point has no active refutations, then it is considered "active" ("provisionally true", or "true" for short).
- A single active refutation ("con" point) causes the parent to be considered "false".
- Note that if a con point is inactive, then it does not count against its parent.
- If all of a parent's pro-points have been refuted (deactivated), then the parent is considered false.
There needs to be a way to make a point dependent on a point that already has another parent, or that is a root point.
In practical use, there will need to be a method to prevent trolling, gaming the rules, etc. Peer-to-Peer Accountability Enforcement seems like a promising solution.
- allow users to vote on the "leaf" points (agree/disagree), making it possible to map out different belief-systems.
- allow users to vote on non-leaf points, making it possible to see where people are unwilling to accept the existing evidence.