|Abstract:||This paper investigates the stress system of Central Rarámuri (CR), an interesting case of morphologically conditioned phonology that provides a testing ground for the predictions of Root Controlled Accent (RCA) Hypothesis (Alderete 1997, 1999, 2001) and Cophonologies. The typological and theoretical relevance of this system stems from the following factors: a) it is a mixed system (i.e. partially lexical, partially rhythmically assigned); b) it is left anchored, and restricted to the first three syllables (an extremely uncommon pattern that has been only reported to occur in Icua Tupi (Tupi), Terena (Tupi), Wishram and Chinook (Penutian) (Kager 1993)); c) the interaction of (a) and (b) yields markedness reversals. The Raramuri case challenges previous accounts of mixed stress systems, such as RCA, which rank indexed faithfulness with respect to a fixed markedness hierarchy. I examine the implementation of the RCA to the CR mixed stress system (through undominated Prosodic Faithfulness to the Root), and argue that compound stress and third-syllable stress generate a markedness reversal, contrary to what is predicted by the RCA. Through examination of a large body of original data collected through field research, I argue that the cophonology approach is superior to the RCA in delivering the correct empirical generalizations.