A Journal of Philosophy, Applied to the Real World

For Hierarchy in Animal Ethics

Yale University


In my forthcoming book, How to Count Animals, More or Less (based on my 2016 Uehiro Lectures in Practical Ethics), I argue for a hierarchical approach to animal ethics according to which animals have moral standing but nonetheless have a lower moral status than people have. This essay is an overview of that book, drawing primarily from selections from its beginning and end, aiming both to give a feel for the overall project and to indicate the general shape of the hierarchical position that I defend there. In this essay, I contrast the hierarchical approach with its most important rival (which holds that people and animals have the very same moral status), sketch the main idea behind one central argument for hierarchy, and briefly review three potentially troubling implications of the hierarchical view. I close with a discussion of a promising possible solution to the most worrisome of the three objections.



One of the most striking developments in moral philosophy over the last half century has been the remarkable explosion in the discussion of animal ethics, that part of moral philosophy that deals with our moral obligations toward (nonhuman) animals. It would of course be an exaggeration, but only a mild one, to say that fifty years ago philosophical discussion of the treatment of animals was virtually nonexistent. The topic suffered from something close to complete neglect. On the rare occasion when a moral philosopher had something to say about animals, it was largely a matter of admitting—albeit only in passing—that it was wrong to be cruel to them, that the gratuitous infliction of pain was morally problematic. And then, for the most part, the matter was typically left at that.

Fifty years later the pendulum has swung the other way. Animal ethics is now a well entrenched subdiscipline within the field of moral philosophy as a whole. There is an ever growing cascade of books and articles devoted to the subject, a constant stream of journals and conferences. What’s more, it seems to me that a particular philosophical position within animal ethics has emerged as well.

I hesitate to say that it is the dominant view. I doubt if there is enough consensus in the philosophical literature on animal ethics to have much of anything substantive lay claim to a title like that. But it does seem to me that many theorists are drawn to some version of the view I have in mind.

Here’s the basic idea. According to this view, otherwise similar harms or benefits for people and animals count equally from the moral point of view. “Pain is pain,” as the point is sometimes put.1 In this sense, animals and people can be said to have the same moral status. To be sure, there are important differences between people and other animals, including differences in terms of which goods and which bads are likely to be at stake in any given case. These, in turn, can make it morally appropriate to treat people and animals differently. But that’s not because animals somehow count less than people do, from the moral point of view. On the contrary, similar goods (or similar bads) are to be treated the same, regardless of whose interests are at stake. That is to say, in and of itself it matters not at all whether we are talking about the interests of a person or the interests of an animal. Similar interests are to be given equal weight in our moral deliberation, regardless of whether we are dealing with a person or an animal. Strictly speaking, everyone has the same moral status.

For obvious reasons, it would be natural to call this position egalitarianism. It assigns the same weight to the interests of animals and of people. It gives the same moral status to both, considering neither group higher or lower than the other.

But for still other reasons, equally obvious, it would be potentially misleading to call the position in question egalitarianism, for the label is already in use as the name for views that hold that equality has moral significance in its own right (for example, that there is value in the equal distribution of welfare). Using the term “egalitarianism” for the first sort of view as well would only invite needless confusion. So we’ll need another name for the position I am trying to describe.

Accordingly, I propose to call the view in question unitarianism, since it holds that there is only one kind of moral status—a status shared by both people and animals. The name is far from ideal, I suppose, but I cannot think of a better one, and if nothing else it has the advantage that “unitarianism” is not already the name of any sort of prominent position in moral philosophy.

Unsurprisingly, unitarians differ from one another in all sorts of ways. For it is one thing to say that all of us—people and animals alike—have the same moral status. It is quite another thing to spell out what that status involves, just how it is that we are morally required to treat one another. Thus there can be, for example, unitarian utilitarians, instructing us to bring about the greatest balance of pleasure over pain. As unitarians, such utilitarians simply remind us to count the pleasures and pains of animals every bit as much as the pleasures and pains of people. And there can be unitarian deontologists as well, instructing us, say, to avoid harming the innocent (even if the results of harming them would be better overall). As unitarians, such deontologists remind us to avoid harming innocent animals, just as we are to avoid harming innocent people. In short, almost all of the sundry debates within normative ethics remain contentious and unresolved even if we embrace unitarianism. In and of itself, unitarianism doesn’t tell us how to treat people or animals; it only tells us that the same fundamental rights extend to all.

There is a lot to be said in favor of unitarianism. But one advantage should be apparent from the start. If we accept unitarianism then it is reasonably easy to see how to extend our moral theory so that it covers not only people but animals as well.

The truth, of course, is that in the past almost all of our moral theorizing has been limited to thinking about people. And while, as I have just observed, the debates in normative ethics are far from resolved, it does seem fair to say that for the most part what we have been working our way towards, by means of these debates, is a moral theory that would accurately tell us about the obligations that people have toward people. So even if you have such a normative theory worked out to your own satisfaction, strictly speaking you still face the question of how to extend or generalize that theory so that it covers animals as well. Unitarianism provides a simple and straightforward answer to that question: our interactions with animals are governed by the very same set of principles that govern our interactions with people (as spelled out by your favorite moral theory). Armed with a normative theory adequate for dealing with people, there is no further work to be done.

In contrast to the unitarian approach to animal ethics, it seems to me that common sense embraces, rather, a hierarchical approach, where animals count, but count in a lesser way. On this alternative view, people have a higher moral status than animals do. There are still restrictions on how we are to treat animals, but these are not the very same restrictions that govern our treatment of people. People have rights that animals lack, or have stronger rights, or perhaps a person’s interests count for more than (or count in different ways from) an animal’s.

Admittedly, one should probably hesitate before making confident assertions about common sense in this area. Some people apparently believe that animals don’t have any sort of moral standing at all; they are merely one more resource to be used as we see fit.

I suppose there is a sense in which a view like this—where animals lack moral standing altogether—could still be described as hierarchical, since people clearly have a higher status on this account than animals do. But similarly, there is a sense in which such a view could instead be called unitarian, since it holds that there is indeed only a single moral status (that had by people). But as I intend to use the terms, at any rate, neither label applies to those who simply deny the moral standing of animals. As I intend to use the terms, both unitarians and hierarchy theorists agree that animals do indeed count, morally speaking; animals have moral standing. Unitarians and hierarchy theorists differ only in terms of whether animals have the very same moral status as people or a lower one. Accordingly, if enough people believe that animals don’t count morally at all, then it would be a mistake to claim that common sense embraces a hierarchical approach.

I suspect that most people reject the extreme claim that only people count. What I take to be the common view, rather, is that animals do indeed count morally, but they simply do not count in the very same way that people do. Animals count for less.

Of course, here too, there remains tremendous room for disagreement. In addition to the familiar debates from normative ethics about the details of our obligations toward people, questions about the appropriate extension of our normative theory (so that it covers animals too) now become pressing and difficult. After all, it is one thing to say that animals count, but in a lesser way. It is quite another thing to spell out exactly how they count, what it really means to say they count in a lesser way. If the interests of animals are not to be counted in precisely the same fashion as the interests of people, how then are they to be counted? Although it does seem to me to be true that common sense accepts a hierarchical approach (or, at a minimum, it is true that a lot of people accept something like that idea), I don’t think there is anything close to a clear understanding of what the lesser standing of animals entails.

In my book, How to Count Animals, More or Less, I argue for a hierarchical approach to animal ethics. Given what I have just said, then, at best I can only partially claim the mantle of common sense. I do think that many readers will find my central thesis—that the right approach to animal ethics is a hierarchical one—to be fairly obvious, hardly worth arguing for. But at the same time, if I am right that there is nothing like a consensus about what the lower status of animals comes to, then I imagine that the various specifics that I discuss remain controversial.

Accordingly, at various places in the book I take some initial steps toward trying to develop a moral theory that is appropriately sensitive to differences in moral status. I explore, for example, what might be involved in extending some common distributive principles (such as egalitarianism, or a priority view) to animals, while taking into account the fact that animals count for less than people do, with some animals counting still less than others. Similarly, I ask what certain deontological principles or rights might look like—such as the right not to be harmed, or the right to self-defense—once we modify them so as to reflect the various differences in status that we find between people and animals, or among animals.

However, it is probably best to admit that in my book I don’t actually develop a detailed hierarchical theory. At best, I offer a sketch of what a theory like that might be like. In fact, truth be told, in many places—really, in most places—all I do is try to point out how desperately far we currently are from having an adequate moral theory when it comes to the treatment of animals. Unlike the unitarians, who think it a relatively trivial matter to extend moral theory to cover animals, I find myself thinking that we remain very much in the dark about how best to do that. I can only say how to count animals more or less.

I do however want to emphasize one further point. Although I defend a hierarchical approach to animal ethics, I do so with considerable misgivings, for I am afraid that some may come away thinking that my aim to is to defend an approach that would justify much or all of our current treatment of animals. After all, it seems reasonable to suggest that it is part of the commonly accepted view that our treatment of animals is, in the main (even if not in all specifics), morally acceptable; and I have already suggested that the common view is a hierarchical one. So in defending hierarchy, aren’t I defending—in broad strokes, at least, if not with regard to every detail—our current treatment of animals?

But nothing like this is remotely the case. Our treatment of animals is a moral horror of unspeakable proportions, staggering the imagination. Absolutely nothing that I say here is intended to offer any sort of justification for the myriad appalling and utterly unacceptable ways in which we mistreat, abuse, and torture animals.

In this regard the unitarians have an easier time of it. No one would be tempted for even a moment to suggest that we already treat animals in anything like the way that morality requires us to treat people. So unitarians are very well positioned to condemn current practices for the moral monstrosities that they are.

But that doesn’t make unitarianism the truth. On the contrary, it seems to me to be true both that animals count for less than people and yet, for all that, that they still count sufficiently that there is simply no justification whatsoever for anything close to current practices. It may be less straightforward to condemn our abuse of animals once one embraces a hierarchical view, but it is still important to do so.

Having said that, however, I should nonetheless warn the reader that the requisite arguments for the unjustifiability of our treatment of animals will not be found in my book. To work out those arguments with care one first needs to articulate in detail the appropriate hierarchical normative theory; and as I have already suggested, it seems clear to me that we are very far indeed from having anything like that. My book is intended as a contribution to the attempt to produce the relevant hierarchical theory. But the truth is, it throws out far more questions than it answers.


A hierarchical approach to normative ethics emerges rather naturally from two plausible thoughts. First, the various features that underlie moral standing come in degrees so that some individuals have these features to a greater extent than others do (or in more developed or more sophisticated forms). Second, absent some special explanation for why things should be otherwise, we would expect that those who do have those features to a greater extent would, accordingly, count more from the moral point of view. When we put these two thoughts together they constitute what is to my mind a rather compelling (if abstract) argument for hierarchy.

Some of the implications of this line of thought seem to me completely congenial. On the one hand, many animals clearly do have some of the features that ground moral standing, so these animals count, morally speaking. Indeed, it is plausible to think that they count for far more than we ordinarily recognize. (Certainly they count for far, far more than one would think, given the appalling ways we normally treat them.) But at the same time, I think it is also clear that animals have fewer of the relevant features than people have (or they have them to a lesser degree), so that animals count for less than people. All of which is just to say: there are different degrees of moral status, and people have a higher status than that had by animals. What’s more, and this is a third plausible implication of this basic line of thought, since animals themselves vary, one to the next, in terms of their possession of the relevant features, some animals have a higher moral status than others.

Now it sometimes happens in philosophy that an abstract argument that seems otherwise persuasive has implications that are hard to accept. One then faces the difficult question of whether to accept the argument and its troubling implications or instead somehow resist the argument (and thus avoid the implications), by abandoning some initially plausible premise. Happily, we don’t face this dilemma with regard to the abstract argument for hierarchy that I just rehearsed. For as I argue at length in the book, these implications are plausible in their own right. Indeed, the hierarchical approach allows us to avoid various unattractive or absurd conclusions with which we might otherwise be saddled. Overall, then, the hierarchical approach is rendered even more plausible by virtue of its implications.

But that’s not to say that there are no cases at all where hierarchy leads to results that may be surprising or even intuitively difficult to accept. On the contrary, I suspect that no position that we could adopt on the issues surrounding animal ethics will be a completely comfortable one for us to embrace. Accordingly, our aim should be to think through the comparative strengths and weaknesses of the various alternative views and tentatively accept the view that seems to offer the most attractive position overall—all the while recognizing that even the view that does best in terms of this comparative assessment will have at least some implications that may trouble us.

In my book I identify three such implications of hierarchy, ones that might reasonably give us pause. First, just as ordinary adult human persons have a higher moral status than animals do, anyone who accepts hierarchy must be open to the possibility of “superior beings,” beings with a status even higher than our own. I know that this is a possibility that some would like to disallow, but speaking personally I think it shouldn’t really bother us. The possibility of such beings may be a humbling one; but it’s not, I think, unacceptable.

More worrisome, second, is the existence of “marginal cases,” humans with such severe cognitive impairments that they simply fail to count as persons at all. Given their restricted and diminished psychological capacities, it seems inevitable that their moral status will be lower than our own. I argue that an appeal to something I call modal personhood—the fact that such severely impaired humans could have been people—can reduce the bite of this implication somewhat (giving marginal humans a higher status than animals that are their psychological peers); but I would not want to deny that most of us would find it difficult—initially, at least—to accept the thought that the severely impaired count for less. Nonetheless, since the alternatives (such as unitarianism) seem to me even less plausible, this does seem to me to be the view that we should, upon reflection, accept as well.

But that still leaves a third troubling implication, which I call “the problem of normal variation.” It is one thing to accept that those with capacities far beyond ours (superior beings) or far below ours (marginal cases) would have a status different from our own. It is quite another thing to accept the claim that since even ordinary adult human persons differ, one to the next, in terms of their various psychological capacities, they must also differ, one to the next, in terms of their moral status, with some of us having (slightly) lower status than others.

Even this last implication will not seem unacceptable to everyone. Viewed against the tremendous range of relevant capacities that we find among creatures with moral standing (think of the minimal agency and—perhaps—sentience of a fly, as compared to the incomparably rich cognitive and emotional lives of persons), the differences found among ordinary adult humans will be relatively trivial. For all practical purposes, then, these differences may be ones that are sufficiently limited that we are quite justified in simply disregarding them. On such a view, there may indeed be minor differences in moral status from one person to the next, but these differences will justifiably play no role in our moral deliberations.

As I say, some may find this last response sufficiently plausible and reassuring to put the concern to rest. Indeed, for the most part, that’s true in my own case as well. But I know that many others will remain uncomfortable (to say the least) at the thought that there may be genuine, even if small, differences in our moral statuses. These people will understandably wonder whether there is a more robust answer to the problem of normal variation.

A response capable of giving all of us (that is, all normal adult humans) the very same moral status would need to somehow overcome the presumption stated above that variations in the features underlying moral standing should result in corresponding variations in moral status. The claim would have to be, rather, that certain minor variations in psychological capacities (the kinds of differences we see displayed among ordinary persons) actually make no difference to status at all. Instead of status increasing steadily with increases in the relevant capacities, status would have to grow less smoothly than this, remaining flat or constant over the range of variation we find among normal humans.

Of course, if there were only one moral status—if unitarianism were true—we would have this result automatically. For if there is only one status, not only does the normal variation in the relevant capacities that we find among ordinary persons make no difference to status, no variations in those capacities make any difference at all. Under unitarianism, after all, as long as one has moral standing of any sort, one’s status is exactly the same as everyone else’s, no matter how much it might be the case that other individuals have the relevant capacities to a greater or to a lesser extent. In effect, under unitarianism status would remain flat or constant across the board—for all beings with moral standing whatsoever—not just for ordinary adult persons.

But as I argue in my book, we have ample reason to reject unitarianism. So a plausible view here is going to have to be a more complicated one than that. Another alternative, I suppose, would be a view according to which moral status can indeed vary and does indeed increase with greater capacities (as we move up the animal kingdom)—until we reach the range of capacities found among normal adult humans, at which point status stops increasing. (Conceivably, it might start up again once we get past the normal range of human capacities, so that superior beings might still have a status higher than our own.) The problem with a view like this, of course, is that it is difficult to see what could be so special about the particular range of capacities that normal humans happen to display, such that here and here alone small variations in capacity make no difference to status, while at other levels they do.

More plausible, I suspect, would be a view according to which this sort of feature—where certain variations in capacities make no difference to status—is found repeatedly, not just when it comes to normal human capacities. Perhaps status is regularly flat or constant over a given range of variation in capacities. But instead of imagining that all levels of capacity elicit the very same status (which is, after all, the unitarian position), we can suppose that there are a number of such ranges, where each such range elicits a different (constant) status.

If a view like this can be defended, it would hold out the possibility of solving the problem of normal variation. It would also, obviously, have implications for what we should say about the status of animals as well. Instead of holding—as we might otherwise do—that whenever two animals differ in their capacities they differ in moral status as well, we will instead think of the animal kingdom as being divided into an appropriate number of broad categories, where each such category is assigned a single corresponding status. No doubt dogs will still have a higher status than fish, say, but minor variations in capacities among fish (for example) will not ground even minor differences in the resulting statuses.

Note that a position like this needn’t be seen as incompatible with the second premise of the abstract argument I sketched at the start of this section. As I intended that claim, at any rate, it wasn’t an assertion that status must go up with increasing capacities. Rather, it was describing what we should expect to happen in the absence of some special mechanism capable of producing a different effect. That is, I meant only to suggest that status will increase with capacities unless there is some special mechanism or force that prevents this from happening. Accordingly, those attracted to the idea that minor differences in relevant psychological capacities normally make no difference to moral status need only insist that there is such a mechanism, something that “overrides” the fact that there is a range of capacities that holds among normal adult humans (or, for that matter, among fish), thus blocking the effect—increasing status—that we would otherwise expect.

What might such a mechanism look like? I believe that a promising answer emerges if we adopt a “realistic” approach to evaluating moral principles, that is, if candidate rules are evaluated (by the relevant foundational machinery) under realistic assumptions about the epistemic and motivational limitations of actual moral agents.

The details will differ, of course, depending on the specifics of one’s foundational theory (whether rule consequentialism, contractarianism, ideal observer theory, or what have you). But whatever one’s foundational theory, one can adopt a version of the theory which is sensitive to what we are actually like, one which uses realistic assumptions about our cognitive and motivational limitations, when evaluating potential moral rules. Theories like this share a commitment to practical realism. And I believe that if one accepts practical realism then one is likely to end up with a normative theory that includes only a few different levels of moral status, since more complicated approaches to status will be poor choices for moral agents with the kinds of limitations that we actually have. Arguably, then, despite the differences in psychological capacities that we really do find among people, all of us may nonetheless have the very same moral status.


I have suggested that theories that incorporate practical realism may end up with a coarse-grained approach to hierarchy, one where instead of having a continuum of possible levels for moral status there will only be a fixed number of such levels (so that even individuals who differ in terms of their psychological capacities may have the same status). Is there more that we can say about this? More that we can predict about the hierarchical theory that will emerge?

I do have a few conjectures, but it is worth emphasizing the point that they are indeed only conjectures. Precisely because practical realism attends to empirical facts about the kinds of rules that we are well suited or ill suited to try to act upon, it is to a significant extent an empirical question what rules will emerge as optimal from any given foundational theory. So the conjectures I offer remain just that—mere conjectures—until the relevant empirical work is done.

Similarly, it is important to bear in mind that the precise rules that emerge will also depend on the particular foundational theory that you embrace. Since I do not try to develop any of those theories in my book, at best all I can point to are a few broad features that, I suspect, will be common to those accounts.

Nonetheless, it does seem possible to make a few predictions. The most important of these is of course the very claim I have already emphasized, that theories that embrace practical realism will generate only a few different levels of status. I’m not prepared to offer an exact number, but it seems likely that there won’t be more than a half dozen or so. (Perhaps, if we ever do encounter superior beings, we may need to add a few more.) My thought here is that much more than this would already involve us in a larger number of levels than we can readily call to mind.

It isn’t so much that we cannot readily think to ourselves that there are ten or twenty—or even more—levels of status. It is, after all, easy enough to think that some creatures have a status represented by the value 1, others by .9, others still by .8, and so on, down to .1. (Or, for that matter, 1, .99, .98, .97, and so on.) The worry, rather, is that the more categories there are, the more difficult it becomes to identify who goes where, and this difficulty radically increases the chances of making a mistake.

Perhaps, then, we will have only a handful of categories: one for persons, and another few dividing up the animal kingdom into large groups with roughly similar capacities. Perhaps those covering animals will simply include one level for extremely intelligent animals (that is, animals with fairly developed psychological capacities), one for moderately intelligent animals, and one final level for minimally intelligent animals. Or perhaps the divisions will be a bit narrower than that, with one or two more levels than this.

I hesitate to actually attempt to demarcate the relevant divisions, since I know too little about the actual capacities of different animals, but if only for the sake of illustrating the kind of approach I have in mind, let me just suggest that there might be one level for the most intelligent animals, those closest to being full-blown persons (like dolphins, whales, squid, or great apes), another for highly intelligent animals (like dogs, pigs, parrots), another for “midlevel” animals (rabbits, cows, squirrels), still another for “lower” animals (other birds, fish, reptiles), and one last level for the very lowest animals, with the least developed psychological capacities (such as insects and spiders). Again, the point here is not to claim that these are the relevant divisions, but only to suggest that the actual divisions will be comparably broad and inclusive, with relatively clear indicators of which types of animals fall into which groups.

The idea here would be to have not only a relatively small number of groupings, but also a relatively easy way to assign a given animal to its relevant group. After all, it would hardly be feasible to expect us to undertake a detailed investigation of a given animal’s specific psychological capacities each time we were going to interact with one. This makes it almost inevitable that in normal circumstances we will assign a given animal on the basis of its species (or, more likely still, on the basis of even larger, more general biological categories).

An approach like this would be similar to what we see when a rule consequentialist favors a rule prohibiting doing harm. Although harming an innocent isn’t always the act with the worst consequences, there is nonetheless a broad correlation in place, so that normally we do better to simply avoid such acts rather than trying to directly calculate the possible consequences of our actions. Because of this, despite its imperfections, a coarse-grained rule like “don’t harm the innocent” still earns its place as part of the optimal set of moral rules. Similarly, then, even though assigning status on the basis of species (or family, order, or even class) will not always correctly identify a given animal’s overall level of psychological capacities, there should nonetheless be broad correlations, so that ordinarily we will do better to assign on the basis of broad biological groupings, rather than trying to directly determine the given animal’s capacities. (Indeed, given the near universal ignorance and underestimation of the cognitive and emotional capacities of animals, left to our own devices most of us would routinely place animals at far too low a level if we had to estimate capacities directly.)

Does this mean that the view that emerges will reject an individualistic approach to status (where one’s status turns on one’s individual properties, rather than on the various biological groups to which one belongs)? To some extent, perhaps, but not completely. For despite our rather limited ability to size up the psychological capacities of individual animals, it would be silly to suggest that we are altogether incapable of doing this, under any circumstances whatsoever. Consider the fanciful example of a golden retriever who has been given a supervitamin and has now miraculously been turned into a person (while still remaining, nonetheless, a dog). We might well be able to recognize that this particular dog has psychological capacities far exceeding that of ordinary dogs—that this dog, unlike ordinary dogs, is a person. (Imagine, for example, that the dog begins to discuss with us its plans for next summer, or that we translate the poetry it has been writing!) In such extraordinary cases it will presumably be appropriate to give the dog the very same status as we ourselves have, despite the fact that it remains canine. Similarly (if a bit less extreme), if a snake, say, were to display psychological capacities at the level of, for example, a cat, then once we recognized that fact it would be appropriate to give it a correspondingly higher status.

Cases like this would be somewhat similar to the rule consequentialist’s recognition that despite the suitability of a general prohibition against harming the innocent, the right not to be harmed should have a threshold. For even though we are not normally very good at identifying cases where slightly more good would be done by doing harm, when the amount of good at stake is great enough our judgment to the effect that we are now in an exceptional case becomes much more reliable. Accordingly, a rule that permits doing harm in such extraordinary cases is actually optimal. Similarly, then, while the optimal rule for assigning status (given practical realism) is likely to tell us that in ordinary circumstances we are to do this on the basis of biological classification, it is also likely to recognize exceptions, unusual cases where we should assign a given individual a higher status—or, for that matter, a lower status—than that to be given to its biological peers.

An especially important example where we would probably be justified in departing from the ordinary practice of assigning status on the basis of biological classification might be that of marginal cases, humans who are so severely impaired as to fall short of being a person. Just as it should be possible to recognize animals who tremendously exceed the psychological capacities typical of their kind, and just as optimal rules for assigning status should allow for more individualized assignments in such cases, it also seems clearly possible to recognize humans who fall tremendously short of the psychological capacities typical of normal adult humans and who should thus be assigned a lower status than the rest of us have. And tragically, as we also know, such cases—unlike our imaginary example of the golden retriever who is a person—are all too real.

Where then should we expect marginal cases to be placed? What status will they be given? To say that they will have a lower status is not yet to say on what level they will be placed. In this connection it is worth recalling the role that modal personhood (the fact that one could have been a person) may play in raising one’s status. Since we do seem capable of recognizing cases that involve a significant degree of modal personhood, there is no reason to assume that practical realism will direct us to simply disregard this feature altogether. Furthermore, it is also worth bearing in mind the fact if we adopt practical realism then it is important to attend not only to our epistemic limitations but also to our motivational ones. So if it should turn out—as might well be the case—that we are simply incapable of being motivated to conform to a set of rules that would direct us to treat impaired humans as no better than their psychological peers, then we should anticipate that the severely impaired will have a higher status than we might otherwise have expected (based on their actual capacities alone).

Suppose, then, that on the optimal classificatory scheme that emerges from practical realism there is a group for animals that fall just short of being persons, or that are persons but only in a more limited way than normal adult humans. It is conceivable that marginal cases may belong here as well (despite having less developed capacities than the various animals that are assigned to this level). Alternatively, it might turn out that practical realism will direct us to distinguish between more and less severe instances of marginal cases. (Perhaps those who fail to be persons but who nonetheless manifest a reasonable degree of agency will be assigned a higher status than those who are so impaired that they display little or no agency at all.) Indeed, given the likelihood that we are not particularly effective at identifying impaired humans who fall just short of being persons, it could easily turn out that the optimal rules will tell us to place even the significantly impaired at the same level as ordinary persons. Perhaps only the most extreme and unmistakable cases of severe cognitive impairment will be assigned a lower moral status at all.

I am not prepared to choose among these various alternatives; too much turns on empirical matters I don’t feel at all confident about. (I am particularly uncertain as to what motivational limitations there might be concerning how we are able to treat impaired humans.) But the point remains that practical realism could play a significant role in determining the moral status of the severely impaired, and this might well result in assigning them a higher status than we might otherwise think warranted.

In broad outline, then, the kind of view that I think most likely to emerge if one accepts practical realism looks like this: there will be only a small number of levels of status, and in ordinary circumstances individuals will have their status assigned on the basis of species or other, broader, biological classifications, though in exceptional cases—where the individual clearly has significantly higher or lower capacities than is normal for creatures of their kind—appropriate adjustments (up or down) to status are to be made. Clearly, any number of details remain to be worked out, but it does seem to me that this kind of limited hierarchy (as we might call it), is the most likely implication of practical realism.

For reasons that I have suggested, I think it unlikely that there will be all that many different levels of status in such a system, probably no more than four, five, or six, or so. But in principle, of course, the number could be higher than that, if we, as moral agents, have rather higher epistemic abilities than I am currently inclined to give us credit for. Still, even if the number of levels were twice or three times what I am imagining (and I very much doubt it could be much higher than that), it is extremely unlikely that more than one of these levels applies to normal adult humans. So however the details get worked out, it does seem reasonable to expect that an appeal to practical realism will solve the problem of normal variation. In effect, a suitable form of limited hierarchy will have only one level “to spare” to cover the full range of ordinary human capacities.

It might be claimed, however, that I have actually been too liberal in describing the number of levels that will emerge from practical realism. I have suggested there may be as many as six or so, but conceivably, of course, the number might be even smaller than that. Since it is, after all, an empirical matter (at least in part) how many levels are optimal, anyone who embraces practical realism must be open to the possibility that given our actual epistemic limitations it is a mistake to generate a system with any distinctions in moral status at all. That is to say, it might be that the optimal set of rules, given practical realism, would involve only one single status, with the very same status being assigned to all creatures with any moral standing at all. This is, of course, the unitarian position. So in principle, at least, we can see how one possible argument for unitarianism might emerge out of an appeal to practical realism.

Less extreme, though still more limited than what I have suggested, would be a view according to which the optimal set of rules would recognize only two different levels of moral status, one for persons and another for animals. On such a view, all animals would have the very same moral status, though people would nonetheless have a higher one. Conceivably, a version of what I call restricted deontology might be defended along lines like this; that’s a view according to which people have deontological rights, but animals, in contrast, are to be treated in keeping with consequentialism. Alternatively, it might be possible to defend, instead, a view where animals were granted deontological standing as well, but of a weaker sort than the deontological standing granted to persons.

As I say, views of these more extreme types do seem possible, and in principle, at least, someone who appeals to practical realism should be prepared to entertain their possibility. But that’s not to say that it seems especially likely that views of these more extreme sorts will emerge. On the contrary, it seems to me that these views go too far in their pessimism about our epistemic abilities. At least, that’s the case if they are defended (as I have just been considering the possibility of doing) by means of an appeal to practical realism.

To be sure, anyone who embraces practical realism must take full cognizance of our epistemic limitations. And I have of course been emphasizing the thought that if we do this we are led to a more limited form of hierarchy, one with at most a small number of levels of moral status. But practical realism requires not only that we be realistic about our limits, it also requires that we be realistic about our abilities. So should the suggestion be made that we are altogether incapable of successfully drawing any distinctions at all (not even one between persons and animals) or at most one such distinction (precisely that between persons and animals), then it seems to me that any such assessment of our epistemic abilities is unduly crimped. It would run afoul of practical realism to have too many levels of moral status. But it would also run afoul of that very same idea to have too few.

Of course, it must be conceded that if even a limited form of hierarchy is to be adequately defended one must eventually show just how and why one’s favored foundational theory supports a hierarchical approach to status in the first place. Although I argue that practical realism—if one accepts it—puts pressure on hierarchical views to guarantee that there are neither too many nor too few levels, I don’t try to argue at all for the claim that it is in fact a hierarchical view (of some sort) that will emerge from the most plausible foundational theories. Or rather, a bit more accurately, I don’t to do that directly. For I do argue at various places in my book that only a hierarchical normative theory will yield intuitively acceptable answers on a number of different moral issues. So I presume that any acceptable foundational theory will, in fact, not only grant animals moral standing, it will also generate hierarchy of some sort. Still, in How to Count Animals, More or Less I don’t try to explore exactly how and why such hierarchy would emerge from any given foundational view. That is work for another occasion. The more limited point I am trying to make here is that given that an adequate foundational view will in fact generate hierarchy at the normative level, if we also embrace practical realism then the result will be a limited hierarchy of the sort I have been describing.


As I have already remarked, in broad outlines the view I arrive at bears a striking resemblance to what may well be the commonsense view. For I imagine that most people would concede that animals count morally, and they would certainly also insist that animals count less than people do. Similarly, I take it to be something like the commonsense view that although animals count, they don’t all count in the same way: some matter more, morally, than others do. Furthermore, I imagine that most people either accept or come close to accepting something very much like a limited hierarchy view, according to which there are at most only a small number of morally relevant divisions within the animal kingdom, with higher animals (for example, dogs, monkeys and whales) counting for more than other animals (such as chickens, rabbits, and mice) who in turn count for still more than the rest (like fish, perhaps, or maybe insects).

But this very similarity brings us back to a concern I expressed at the outset of this overview, that my position in How to Count Animals, More or Less will be misconstrued and taken to be a defense of something like current attitudes and practices toward animals. After all, people count more than animals do, right?

I hope it is clear that any such interpretation would be a gross misunderstanding of my actual view. Animals count for less than people do, but they count for far, far more than we ordinarily acknowledge.

The day may come when it will be common to look back on mankind’s long history of abuse of animals and recognize it as the disgrace and horror that it is. But that day is not yet upon us. Conceivably, then, given the widespread mistreatment and disregard for animal interests that continues to this very day—indeed, given the innumerable ways in which abuse of animals runs almost unnoticed through countless aspects of human life—it may well be the case that the most pressing task for moral philosophy with regard to animals is to establish that they really do count morally, and that they count for a tremendously great deal more than we seem ready to acknowledge (given the horrific ways we actually treat them).

Crucial though it is, that is not the task I have undertaken in my book. For it seems to me that no such project can be successful unless it is undertaken in full acknowledgment of another essential fact about animals—the fact that although they do count morally, they count for less than people do. The moral theory with regard to animals that we need to be defending is indeed a hierarchical one; and until that fact is more widely recognized in the philosophical literature, I suspect that many of our efforts to secure decent and just treatment for animals will be doomed to failure.

Perhaps that is an overly pessimistic assessment. I hope so. But even if so, at the very least it seems obvious to me that our understanding of ethics—not just animal ethics, but all of ethics—will be confused and incomplete until such time as the significance of status is properly taken into account.


1.For example, Peter Singer, Animal Liberation, updated ed. (HarperCollins, 2009), p. 20; and David DeGrazia, Taking Animals Seriously (Cambridge, 1996), p. 234.