<p> Supposing they exist, what work are value properties supposed to do? What difference do they make? What is the difference between a world in which they exist and a world in which they do not? </p><p> One obvious answer invokes the claim that evaluative properties make a causal difference. While this is an interesting topic, it is well-covered elsewhere by Gilbert Harman and Nicholas Sturgeon. But there are other possibilities put forth by moral realists that are independent of the question of causal explanation. In my dissertation, I examine a number of alternative possible jobs that value properties are thought to fulfill. </p><p> 1) <i>Reference and supervenience</i> - Some argue that evaluative properties serve as the <i>referents</i> of evaluative <i> predicates,</i> or as the <i>extension</i> of supervening evaluative concepts. I consider arguments from McDowell and others to the effect that our ability to correctly sort evaluative cases into the correct categories requires the mapping of these concepts onto evaluative properties. My arguments show that these considerations alone cannot support evaluative realism, as there are alternative accounts of evaluative language that do not require separate value properties. For instance, a semantics grounded in conceptual-role can adequately account for the ability to think with and use evaluative concepts but nevertheless have natural properties serve as the extension of these concepts. </p><p> 2) <i>Resemblance</i> - One might think that, as Russ Shafer-Landau and David Brink argue, the resemblance of items belonging to the same evaluative category needs to be explained given the manifest differences in their <i> non-evaluative</i> properties. Stealing candy from a baby, cheating on one's spouse, and refusing to tell the police where a perpetrator is hiding all belong to the same moral category (the category of wrong actions), but they share little in common from the view of physically manifested behavior. I offer two alternative methods for explaining evaluative categorization that do not require accepting the existence of distinctly evaluative properties, thus showing the inference that distinct value properties are necessary to explain resemblance to be unwarranted. I claim that the way we <i>think </i> about value is enough to ensure correct categorization &ndash; there need not be some further existent to explain this. </p><p> 3) <i>Qualitative Character</i> - Last, I consider the view that evaluative properties possess a distinctive and irreducible <i> qualitative character.</i> I address the purported qualitative natures that value properties are thought to possess and argue that understood in one way, we would have justification for accepting that they exist. This interpretation has it that evaluative qualities are literally perceptible &ndash; their qualitative characters are of the same general sort as the properties <i> redness</i> or <i>pain.</i> I argue, that there is no need to posit distinct value qualia, at least not if qualia are necessarily representational, since we can have the same phenomenology of value whether or not we are directly perceiving an evaluative episode &ndash; we can have the same phenomenology just by considering or imagining the relevant episode. I offer a model of value perception which captures this important point. </p><p> Though my arguments might appear to push one toward anti-realism, they are all compatible with the truth of (suitably qualified versions of) 1), 2), and 3) after all is said and done. My goal is not to undermine arguments for evaluative realism, but I do intend to show that there is no master argument for it; any argument for realism must delve into thorny and often distinct metaphysical questions. Furthermore, I emphasize the role that metaphysical preconceptions and their implications play in many debates in value theory and the need to be clear and consistent with regard to these implications. </p>

<p> In the thesis, I provide a defense of the moral error theory. In the first section, I solve the formulation problem, which is the puzzle related to properly formulating the moral error theory without self-defeat. In the second section, I defend the claim that moral concepts commit their users to irreducibly normative reasons. Finally, in the third and last section, I defend the reduction argument against new criticisms and conclude that there are no irreducibly normative reasons.</p><p>