I would start with the concept of adding the number of reasons to agree (#As) and subtracting the number of reasons to disagree (Ds). But what to do with reasons to agree with reasons to agree (AA) and so on (AD, DA, DD)? I would suggest this framework: each argument has its own unique Argument Score (AS). It also gets Linkage Scores (LS) to each conclusion. LS are in %. You multiply AS*LS, to determine strength given to each conclusion by a argument through a linkage.
A unique score (US) should be used to remove duplication. People would suggest arguments are saying "essentially the same thing". Then people would support or weaken this belief. If an argument has a low unique score between 2 other arguments, and they are all used to support the same conclusion, they would each give 1/3 (or 33%) of their individual score. 2 ideas, each with 50% US, would each give 50% of their points.