By "that trait" I believe he's referring to that particular way of getting to bilateral symmetry, not that an organism that doesn't do it like that can't be bil

By "that trait" I believe he's referring to that particular way of getting to bilateral symmetry, not that an organism that doesn't do it like that can't be bilaterally symmetric. As an example, think of what it would take to add a third leg to a human being. Now we have two and one leg is a mirror image of the other. Those would presumably be the outer legs, how's the central leg going to look like? Is it going to be a copy of either the right or left leg, or some combination with 9 toes (4 outer, big toe in the middle) etc.? Carrying that sort of information in the DNA is going to be harder than following the default template of "make this here and mirror it", which is how most structures are encoded in DNA. Furthermore I think the reason the paper's abstract mentions phylogenetics (but I don't know, I can't access the full thing) is to say something to the effect that if an organism doesn't have any hands or legs, adding them in pairs of two is inherently simpler than adding three, since the rest of the body structure is already mostly following that template. Once you start going down that road it's hard to "go back", which is why modern animals carry a lot of structures that are only the way they are today because they evolved from some initial past state.