Entertain, for a moment, the possibility that reason evolved to facilitate the use of tools. If you take this point as a given—and I cannot “prove” to you that this idea is true; I can only ask you to entertain it—it would appear that the tool-using animal is the same one who engages in metaphysical ideation, and that the animal that engages in scientific study is a different (and shrewder) beast entirely.
Which in turn makes “scientism” seem less like science, and more like just another religion. The way the tool-using animal thinks when, say, cracking a nut with a rock, begins with the cracked nut in mind; the rock is something harnessed to accomplish that end. Religions begin with some metaphysical assertion or another and harness reasons why that assertion is true. Similarly, rationalism begins with a specific conclusion in mind: Not that all religions are false, though it is often spun that way; nor even that its assertion of their falsehood is true; rather that reason is the only means by which to come to conclusions about Being—implicitly invalidating imagination, storytelling, emotion, and irrationality, which are the standbys of traditional religions. (A fine distinction, but not quite chop-logic.) Both in rationalism and religion, the thinking runs as follows:
This is why doctors and researchers are always very skeptical of studies in which the investigators and participants aren’t blinded, and why every researcher knows that drug trials funded by a given pharmaceutical company (which wishes to prove the efficacy of its product), aren’t all that reliable. Data gets fudged and statistics get massaged to prove desired conclusions all the time, and not always intentionally; reason starts getting used as a tool again, a means to an end.
Scientific inquiry, in any decently-set-up experiment, forces itself into a state of agnosis regarding the conclusions of its own study. This dispassionately curious agnosis is critical in preserving the integrity of scientific method. It does not use reason, observation, and data analysis as tools to attain a result decided upon at the outset; that would be bad science. This willed agnosis, literally not-knowing, is understood to be a temporary, deliberate state of mind (and distinct from agnosticism, which is a metaphysical school of thought).
Such dispassionately curious agnosis can be preserved much more easily in the study of physical phenomena. In religious thought, there is little to analyse, only abstract terms, stories old and new, myths, images, and parables to bandy about; reason is relegated to its conclusion-proving role, not its result-discovering role; and there is little to observe, save our own like-dislike responses to ideas and states of mind. I observe myself disliking no state of mind so much as absolute certainty. If I ask myself why this is, I realize I treasure agnosis as a spinning arrow pointed at the truth.