Structuring the claim graph (Wiki vs. DAG vs. DCG)

After much trial and error, I have given up on the prospect of organizing claims into a graph with sufficient formal structure such that Bayesian reasoning can be performed mechanically. There are simply too many kinds of claims, which decompose in too many different ways, for that to work robustly beyond toy examples within a single frame of analysis. To do Bayesian reasoning on any particular claim, one has to have priors across a very large, in many cases practically infinite, set of claims, many of which are likely to be normative or otherwise unfalsifiable.

So I have settled on allowing any claim to be a parent of any other claim, and for the relationship between claims to be presented informally in Arguments, which are text objects belonging to claims. Claims can have multiple arguments for and against. Those arguments often contain links to claims with their own pages.

This leaves open how to define the edges between claims and how to determine which claims matter most.