Four Rules of Inference for Ranking Argumentation

I think any scholar working in Optimality Theory (Prince and Smolensky 1993, and much subsequent work) would agree that rigorous language-particular analysis within this framework should provide solid arguments for the ranking of constraints. It is sometimes underappreciated just what is involved in developing such arguments. I will suggest that arguments from ordinary human reasoning are (probably) not generally as reliable as arguments produced by algorithm. A sample algorithm of the appropriate type, which uses four rules of inference, is laid out here. I also suggest that algorithmic ranking argumentation has some potential theoretical as well as practical consequences.