On the Expressivity of Optimality Theory versus Ordered Rewrite Rules

I prove that there are phonological patterns which are expressible by ordered rewrite rules but not by any Optimality Theoretic (OT) grammar whose constraint set contains only markedness constraints and single–state faithfulness constraints, i.e. faithfulness constraints that assign violation marks to pairs of single input–output segments in correspondence, with no reference to other segments in the input or output. The intention is to capture formally the widespread intuition that certain opaque patterns, which are expressible by ordered rewrite rules, are problematic for classic, or traditional, OT.