In this paper, I argue that in spite of suggestions to the contrary, Merleau-Ponty defends a positive account of the kind of abstract thought involved in mathematics and natural science. More specifically, drawing on both the Phenomenology of Perception and his later writings, I show that, for Merleau-Ponty, abstract thought and perception stand in the two-way relation of foundation, according to which abstract thought makes what we perceive explicit and determinate, and what we perceive is made to appear by abstract thought. I claim that, on Merleau-Ponty's view, although this process can sometimes lead to falsification, it can also be carried out in such a manner that allows mathematics and natural science to articulate what we perceive in a way that is non-distortive and in keeping with the demands of perception itself.