I propose a reformulation of realism, as the pursuit of ontological plausibility in our systems of knowledge. This is dubbed plausibility realism, for convenience of reference. Plausibility realism is non-empiricist, in the sense that it uses ontological plausibility as an independent criterion from empirical adequacy in evaluating systems of knowledge. Ontological plausibility is conceived as a precondition for intelligibility, nor for Truth; therefore, the function of plausibilty realism is to facilitate the kind of understanding that is not reducible to mere description or prediction. Difficulties in making objective judgements of ontological plausibility can be ameliorated if we adhere to the most basic ontological principles. The workings of plausibility realism are illustrated through a detailed discussion of how one ontological principle, which I call the principle of single value, can be employed with great effect. Throughout the paper the discussion draws on concrete examples from the history of science.