The one point R. E. Warner made that I do agree with is that an intuitive process is not as flexible or amenable to change as an inductive process. However, Warner didn't demonstrate that Mrgan doesn't understand the principles behind the adjustments he'd make to the icons.
In fact, Mrgan did elucidate reasons that the grid was being used the wrong way. He asserted that the grid should be used as a bounding box, not the size of the actual icon and he specifically showed how this works with "pointy" (concave, complex shapes) vs. full shapes, providing an actual example that showed side-by-side how his tweak looked compared to a full-sized grid-based icon.
The concept of visual balance and optical adjustment is related to trompe-l'oeil and perspective drawing. The conscious knowledge that human vision introduces artifacts was known at least as far back as ancient Greece, because they deliberately designed subtle curves and other proportional changes into their temples to avoid perceptual distortions.
There are some common optical illusions that make things look "wrong" unless they are adjusted for how our brains process things. Some designers may not be able to explain the exact rules behind their adjustments, but many probably do know exactly why they need to make subtle changes, even if they don't know the details of the neuropsychology that necessitates them.
I would bet cash that if you gave a group of designers an assignment to adjust the size and spacing of icons that had to be based on a standard set of proportions, most if not all of them would come up with very close to the same solution. Whether they arrive at the answer through explicit training and understanding of human perception, or an intuitive understanding and experience in making things "look right" doesn't really matter for the final product, only for the process.