This paper addresses Peter Singer's claim that cognitive ability can function as a universal criterion for measuring moral worth. I argue that Singer fails to adequately represent cognitive capacity as the object of moral knowledge at stake in his theory. He thus fails to put forth credible knowledge claims, which undermines both the trustworthiness of his moral theories and the morality of the actions called for by these theories. I situate Singer's methods within feminist critiques of moral reasoning and moral epistemology, and argue that Singer's methods are problematic for moral reasoning because they abstract from their object valuable contextual features. I further develop this claim by showing the importance of embodiment for the construal of objects of moral knowledge. Finally, I develop the moral and scholarly implications of this critique. By showing that the abstract, universal methods of reasoning Singer employs cannot credibly construe the objects of ethical inquiry, I call into question the validity of these methods as a means to moral knowledge in general. Furthermore, since moral reasoning takes place within an embodied moral landscape, it is itself a moral enterprise. Singer's moral reasoning, and ours, must be held accountable for its knowledge claims as well as its concrete effects in the world.