Right, so this is the subject of a paper that I have in process on the applications of some of the newer material in conceptual analysis to this kind of work in ethics. Suppose we consider an account where there are multiple criteria but the various criteria divide individuals in a more fine-grained way. That seems to give us what you're talking about. We have criteria (a) being self-aware and (b) having a desire to continue living as associated conditions. We can say that neither of the individuals meet the conjunction of condition (a) and (b), but they are still morally different in virtue of meeting the categories differently. (This is a perfectly fine move; it's just not compatible with the one-criterion account I model above.) What I have in mind is actually something a little bit more sophisticated. Suppose that we have a set of criteria, call them (w), (x), (y), (z). Some individual has a right to life just in case it meets three or more conditions. We might wind up with a set of very fine grained groupings of individuals with moral status (there would be 5 such groups) and individuals without moral status (there would be 11 such groups). I suspect that when Warren talks about characteristics "generally associated with moral status," this is the kind of structure she has in mind. Are there moral differences between the various groups? Perhaps the 5/11 split is the only one that matters to the right to life, but other configurations matter to other rights associated with pain, sociality, etc.