In the Appendix to the Transcendental Dialectic of the Critique of Pure Reason, Kant argues that even though we can have no knowledge concerning the objects of ideas of reason – the rational soul, the world as a whole, God – these ideas nonetheless contribute, positively, to human epistemic enterprises: they serve as heuristic guides to empirical scientific investigation. This function of the ideas of reason is glossed by Kant in the form of regulative principles that both prompt further, ever-on-going empirical investigation, and direct it so that it will generate properly scientific knowledge of nature, i.e., a system of empirical laws. In the Critique of Judgment, Kant appears to reallocate most of these regulative functions, and the corresponding regulative principles, to a newly identified cognitive faculty, that of reflective judgment: our activity of engaging with given empirical particulars so as to form concepts of them, to find ever more specific and also more general empirical laws, and thereby to attain to more systematic natural science. One might wonder, then, what role the ideas of reason play on this re-considered Kantian position, or, more pointedly, in what sense (if any) are the ideas of reason – ideas of objects that are definitively beyond the empirical — necessary or even at all useful for empirical scientific investigation? Kant does not address this question, but in this paper I shall attempt to think through what his position might be by attending to the Appendix in light of his later changes in position. I shall suggest that on Kant’s view, the ideas function as “objective correlatives” for the norms governing sought-after empirical knowledge, i.e., are images of objects that have the structure of nature as we attempt to know it scientifically, and therefore function as goals for empirical investigation. They can continue to play this supplemental role even after Kant formulates his theory of reflective judgment. The specifically beyond-empirical status of the ideas endows them with a further, important, critical or limiting function, however: they indicate the insufficiency both of current empirical knowledge – hence their motivational role as goal-setting for empirical investigation – and indeed of empirical knowledge of nature altogether. For the ideas give a glimpse of what an intuitive, holistic grasp of the ultimate essences of things might be – by contrast to the actual accomplishments of empirical knowledge, namely (at best) a knowledge of laws, which govern merely the relations among some of the external properties of things. I end the paper by asking whether the objective-correlative function of the ideas can in fact be reconciled with this critical function: does it make sense to hold that empirical scientific investigation aims to approximate to a goal that requires knowledge of an entirely different kind, to be truly satisfied?