|Abstract:||This paper tries to capture variations of vowel harmony in Conflation Theory (de Lacy (2002, 2004, 2006)). Yamane-Tanaka (2006) assumes Gitksan harmony has variations across speakers: translaryngeal and transguttural. These two types cannot be captured in Conflation Theory unless the composite set of Place features in a markedness category is properly revised. The proposal here is to divide the markedness constraint prohibiting dorsal (= *K) into two distinct constraints (= *K and *Q). This solution captures laryngeal transparency and guttural transparency only with the reranking of conflated constraints, without stipulating a language-specific markedness reversal. This approach limits variations among sets of the so-called transparent consonants.