This dissertation proposes a novel phonological framework, directional Harmonic Serialism, that synthesizes constraint-based, rule-based, and formal language theoretic approaches to phonology. I illustrate its advantages in the domains of feature spreading, quantity-insensitive footing, and autosegmental phonology. Specifically, I demonstrate that across these disparate domains, directional Harmonic Serialism makes empirical predictions that more tightly model natural language phonology than alternative theories and that it does so using fewer theoretical mechanisms. At a high level, the theory outperforms alternatives using a simpler, more restricted toolkit.