How we best construct parameter correlations depends on what we want to learn. Projections of the fundamental plane correlations separate galaxy classes; e.g., bulges+ellipticals from disks+Sphs (Figure 4). So they teach us about differences in formation processes. In contrast, it is possible to construct parameter correlations that make most or all galaxy types look continuous. These encode less information about galaxy formation. E.g., in a projection of the structural parameter correlations that encodes mass-to-light ratio, the difference between ellipticals, spheroidals, and even irregulars largely disappears (Bender, Burstein, & Faber 1992). Zaritsky (2015) regards this as progress – as replacing correlations that are flawed with ones that capture some inherent simplicity. That simplicity is real. But it is insensitive to the power that other correlations clearly have to tell us things about galaxy formation.
I therefore disagree, not with Zaritsky's operational results but with his motives. If you look at the fundamental plane face-on, it contains lots of information. If you look at it edge-on, then it looks simple. This may feel like a discovery. But it just means that you are looking at a projection that hides the information content in the parameter plane. Other combinations of parameters make still more types of objects looks continuous and indistinguishable. But this means that we learn still less, not more, about their nature and origin. The simple correlations are not uninteresting, but the ones that teach us the most are the ones that correctly identify differences that turn out to have causes within formation physics.