In one of his working notes Merleau-Ponty remarked that the role of scientific facts is not to replace philosophical reflection, but to “provide us with variants of our experience that awaken us to what it contains”. My aim in this paper is to clarify Merleau-Ponty’s corresponding view that empirical sciences describe facts that are not accessible in imaginative variation, but of which phenomenology must, and can, avail itself. Phenomenology thus understood strives neither to provide a transcendental foundation of science nor does it generate insights that could be simply “applied” in empirical studies, but explores a variable “human-scale” window into experience that remains indispensable for sciences. To substantiate this idea, I draw on Merleau-Ponty’s explicit methodological reflections and his usage of scientific studies (e.g., cases of neuropsychological pathologies) to clarify in particular his view of the transcendental-empirical distinction and the eidetic variation. I contrast my account to several analogical attempts from current secondary literature.