There’s an approach to a number of different domains in ethics, which we can call “specificationism”, that is seldom explicitly discussed but that I think warrants greater attention. Easily the most famous example of specificationism is found in the theory of rights and is owed to Thomson’s “A Defense of Abortion”, where she argues that no one has a right not to be killed simpliciter, only a right not to be killed unjustly. The basic idea is that the content of the right not to be killed – or indeed any right if we wish to generalize – must be specified to reflect what may and may not be done to the right-holder. The right not to be killed may clearly prohibit killing in some uncontroversial cases, but according to this picture it is rife with qualifications that reflect the many instances where one may be permissibly killed. This is a thumbnail sketch of a specified conception of rights, but the theory of rights is only one of the domains where the strategy of specification is employed.
Scanlon, for example, employs the strategy with respect to principles in What We Owe to Each Other. Regarding the principle of fidelity to promises, Scanlon maintains, “We are not morally required to keep a promise no matter what. The clearest part of the principle is this: the fact that keeping a promise would be inconvenient or disadvantageous is not normally a sufficient reason for breaking it, but “normally” here covers many qualifications.” (199) In other words, the seemingly simple principle to keep promises masks a complicated, highly contextualized set of duties and permissions that are revealed as one specifies the principle.
Obligation, too, is a candidate for specification. In The Second-Person Standpoint, for example, Darwall argues that “if someone were able to establish that she did in fact have good and sufficient reason for a putative violation of a moral obligation, then it seems she has accounted or answered for herself and shown she did no wrong.” (98) Moral obligations, then, are subject to qualification by Darwall’s lights – if there is good reason to do that which is normally wrong, then doing it isn’t wrong all-things-considered.
What all of these examples of specificationism have in common, I think, is a commitment to the primacy of basic if complicated normative reasons in moral reasoning. It is the particular reasons that one has in a particular situation –not rights, principles, or obligations – that do the real normative work. These normative reasons are more fundamental or primitive than rights, principles, or obligations on this picture because the contents of rights, principles, and obligations are provided by the interaction of the various normative reasons that one has in a particular situation. On this view, to appeal to a bona fide right, principle, or obligation is conclusory. Fundamentally, one argues towards rights, principles, or obligations and not from them.
For my part, I have long been attracted to specificationism generally, and to a specified conception of rights in particular. But it is worth noting that Thomson subsequently rejected the specified conception of rights that “A Defense of Abortion” relied on in “Self-Defense and Rights”. One of the chief reasons why she rejected specificationism was that it rendered rights explanatorily impotent. After all, if one has (only) a right not to be killed unjustly, then determining that the right has been transgressed is to conclude that the transgressing action is wrong for reasons that have nothing to do with its being the violation of a right. This is just an implication of the fact that one does not argue from rights according to specificationism. That some action transgresses a right cannot be a potential explanation of the action’s wrongness. Clearly, Thomson couldn’t live with this feature of the specified conception of rights because she believes that rights have explanatory power. She believes that rights must be among the potential reasons why what is wrong is, in fact, wrong. (This commitment led her famously to distinguish between a right violation and a right infringement, a distinction that allowed her to hold that rights have a content that does not depend on what it is permissible and impermissible to do to the right-holder.)
Now, Thomson’s worry about specified rights can be restated as a general worry about specificationism in any moral domain. Whether it is a right, a principle, or an obligation, as soon as one specifies the object, that object would seem to lack explanatory power – one cannot explain what is wrong about some action by appeal to a specified right, a specified principle, or a specified obligation. For the content of that right, principle, or obligation is precisely what is at issue.
Is Thomson’s rejection of the specified conception of rights well-founded? Is specificationism across all of the moral domains tenable? Is the case for it stronger or weaker depending on the moral domain at issue? I am inclined to go primitive across the board here and defend specificationism generally. I’d be curious to know what you think.
I think that first we need to distinguish between two things that really are independent from one another. I worry that the term specificationism fudges them.
The first issue you discuss is whether rights, obligations and principles are global or hedged. The second issue is whether violating rights, obligations or principles is constitutive of the wrongness of actions.
These issues do not go hand in hand. So, in the quote you give from Scanlon, he does seem to say that principles are not global in their scope. But, elsewhere, he seems to be the first to say that violating a principle no-one can reasonably reject *is* what is wrong in the actions we consider to be moral wrong. In the contractualist framework, the principles do play a clear explanatory role with regard to wrongness and they also ground our duties and obligations. But, this has no effect on the fact that the content of the principles is not that they always apply.
For what is worth, with regards to the second issue, I think Thomson’s early view was much better. As explanatory wrong-makers, rights, obligations, or duties seem to have little role to play. The reason for this, for me, is the kind of intuitions that come up in Williams’s one thought too many arguments. The first-order considerations that its killing, or cruel, or painful, and so on seem to be sufficient to make actions wrong and also what moral agents should care about. I would be worried if someone was not concerned about these features of the actions but rather about whether it would violate rights and duties. That sort of person seems to be missing something fundamental about morality. This is not to say that notions like rights, duties and principles cannot play a useful role in pointing our attention to the reasons there are.
Hi John
Very interesting. I have a question.
It seems that the specificity of rights is a matter of degree. A right against being killed unjustly is more specific than a right against being killed; but the latter is in turn more specific than, say, a right against being either killed or allowed to die. It seems, then, that a thoroughgoing non-specificationist would have to hold that there is exactly one right, e.g. a right against being killed, or allowed to die, or insulted, or deceived, etc.
Is that right?
Regarding your question, Campbell, I think that you’re on to something. Of course Thomson would balk if told that, in virtue of rejecting specificationism about rights, she is committed to holding that there is only one right. She spends a good deal of The Realm of Rights enumerating many different rights, after all. But non-specified theories of rights nevertheless seem arbitrary in an important sense: they have no account of what distinguishes the reasons that determine the content of a given right from the reasons which determine when those rights are either (permissibly) infringed or (impermissibly) violated.
Consider how this plays out in Feinberg’s famous case, where a hiker must break into another’s cabin to seek shelter from a storm. There is widespread consensus that the hiker would be justified in breaking into the cabin under such circumstances. The more specific claim advanced by proponents of what we can call the general conception of (property) rights is that the hiker would be justified in breaking into the cabin out of necessity even though the cabin owner has a right that the hiker not do so. Now, to say that the break-in is justifiable is to say that the hiker has sound reasons for contravening the cabin owner’s right in the cabin. This is what proponents of the general conception of rights are committed to. It follows that the right is only a partial summary of, or an intermediate conclusion about, the many practical reasons that bear on the hiker’s action in that case. According to the general conception of rights, in other words, the content of the cabin owner’s property right is fixed prior to the determination of the justifiability of the hiker’s break-in, and then a separate inquiry is made into the justifiability of transgressing that right, which will take into account any context-sensitive reasons – like the circumstances of necessity. I believe that this bifurcation is artificial. By separating these steps, the general conception of rights reifies rights, and erroneously invests special moral significance in an intermediate conclusion about what it is permissible to do instead of in the final conclusion about what it is permissible to do. Specificationism, on the other hand, invests moral significance only in the result of all of the relevant reasons being taken into account.
I’m happy, Jussi, to have a fellow-traveller on the question of specified rights. I too prefer Thomson’s earlier position. Regarding your first point, having to do with Scanlon, I appreciate that clarification. Although he doesn’t accept a buck-passing account of moral wrongness, Scanlon does maintain that it is the “more concrete considerations” that make an act one which could not be subsumed by any principle that no one could reasonably reject, and so it seems to me possible to make that case that it is those considerations, which I take to be primitive normative reasons, that really do the work in his moral theory. But this may simply be me projecting my own preferred view onto Scanlon’s theory.
John
Very interesting post. I tend to think that obligations, rights etc,. are context dependant and that if the parameters that define the context changes thru no fault of my own I am ‘excused,’ to use Austin’s notion, from fulfilling the obligation or recognizing the right. For example, a person may have a general right not to be harmed, but if he is trying to harm me unjustly the I do not have to recognize his general right not to be harmed. Generally, non-contextually, I have a general obligation to keep promises justly made, and to recognize the rights of others. However, these can be overridden (excused) within a specific context if the parameters wherein the specific promise was made, or the right recognized have changed thru no fault of my own.
John (Alexander) — it seems to me that one isn’t merely excused for, say, killing another in the circumstances that you stipulate. Rather, when one kills in self-defense in such a case, one is justified. It seems to me that justifications but not excuses help to define or specify rights/obligations/whatever. After all, it makes perfect sense on a specificationist picture to say that one wronged another but was excused for doing so (say, due to duress or something), but it really doesn’t make sense, at least according to specificationism, to say that one wronged another but was justified in doing so. That, I take it, is Darwall’s Point (to adopt Darwall’s parlance).
Hi John (O) –
Nice post, with an interesting connection to particularism.
Let’s say we agree that rights and obligations are, as you put it, “conclusory.” They are, as W. D. Ross would have put it, “toti-resultant.” Now, what about the “reasons,” as you call them, that these normative conclusions reflect? Are they not also subject to specification—not, indeed, so as to convert them into ATC or automatically overriding status, but in the kind of degreed way Campbell has in mind.
I am myself happy to go with Ross and take considerations of obligation to be among the reasons, in this broad sense; but let’s suppose we go with Jussi and stick with killing, cruelty, pain, etc. I would think that if we are to work up from these basic reasons without the intermediating help of something like a welfare-maximizing principle, we’d need to recognize that these reasons, as reasons, need specifying. Pain is a wrong-making factor, unless it is inflicted as justified punishment, in which case it is a right-making factor.
One could, here, follow this kind of thinking one-step further, and hold, with Dancy, that such valence-switching will continue, and show us that there are no general reasons, either. Sticking with the set of reasons, however, I have found Lance & Little’s recent work on moral defaults convincing in showing that general reasons (or wrong-making principles) have an explanatory role, despite being defeasible in the kinds of ways Dancy’s examples demonstrate. If this is right, specifying these reasons at the right level of generality will be important to their well serving an explanatory role.
Henry,
I wonder about this:
“I am myself happy to go with Ross and take considerations of obligation to be among the reasons”.
Do think that there is textual evidence in Ross for the idea that that something is a duty is a reason to do something? That that it is a pf duty is some reason to do it, and that it is a duty property a sufficient one?
The Rossians I know seem to hold the opposite interpretation. They usually begin from Ross’s story of how we come to learn about the duties. We recognise that certain considerations (i.e., reasons) are morally salient in situations. We then make a generalisation that those considerations have some salience whenever faced with – this is just what the talk of the duties picks out or encapsulates.
Henry,
Thanks very much for your comment, which I found illuminating. The link between specificationism and particularism is one that I’m interested in exploring in greater depth than I have — I’m drawn to particularism for many of the same reasons I’m drawn to specificationism. Its worth noting, I think, that in “Specifying Absolute Rights” (Arizona Law Review, 1995), Russ Shafer-Landau maintains that the specified conception of rights’ compatibility with particularism is one of the strengths of that conception of rights. But I will have to think more about Lance and Little’s take on particularism. I do want to preserve the explanatory power of primitive normative reasons — after all, something has to have explanatory power. I just don’t know whether Lance and Little’s approach is the only way to do that.
John,
I’ve been trying to think of an argument for extreme non-specificationism, i.e. the view that there is only one right. Here’s what I’ve come up with.
Let R be the set of all rights, and let v be a function on R such that for any right x, v(x) is the set of all actions, actual or merely possible, that are violations of x. For example, if x is a right against being killed, then v(x) is the set of all killings.
The argument has three premises:
1. ∀x,y∈R[v(x) ⊆ v(y) ⇒ x = y]
2. ∀x,y∈R[∃z∈R[v(z) = v(x) ∪ v(y)]]
3. R ≠ ∅
(I hope the symbols display correctly. The preview looks fine in my browser.)
Premise 1 rules out redundant rights. It says there cannot be two rights such that the violations of one are a subset of the violations of the other. For example, there cannot be both a right against being killed, and a right against being killed with a screwdriver. If we have the former, then the latter is redundant.
Premise 2 rules out arbitrary restrictions on the composition of rights. It says that for any two rights, there is a third right that is the “composed” of the first two. For example, if there’s a right not be killed with a screwdriver and a right not to be killed with a hammer, then there’s a right not to be killed with either a screwdriver or a hammer. It would be arbitrary to accept the former two rights, yet reject the latter.
Premise 3 requires that there be at least one right.
Together, these three premises imply that R is a singleton; there is exactly one right.
What do you think?
By the way, what’s wrong with saying there’s only one right?
John,
I’m sure Lance & Little’s way of preserving the explanatory power of primitive normative reasons is not the only way to go, but it does have advantages. In my 1990 article on specification in PAPA, I used the relatively primitive device of supposing that all of the norms were “generally speaking” norms rather than universal generalizations. I took this to be important to explaining the usefulness of specification, for otherwise the more specific norms would be redundant in just the way that Campbell’s Premise 1 is intended to finger and rule out. Conversely, of course, if the norms are thought of as being logically loose in this way, or in the Lance/Little defeasibility way, or some other way, then that Premise 1 is undercut.
Jussi,
I would have thought that the Rossian story of learning that you sketch could be continued as follows: starting from their particular observations (and habituated responses?), individuals come to learn that certain moral considerations are salient or relevant whenever they arise. In this matter, too, Ross was a good Aristotelian, or so I’ve always thought. Of course, a p.f. duty is a reason of a distinct kind, one that will indicate a duty if it is not overridden. But like normative reasons more generally, it is a consideration bearing on what ought to be done that is subject to being overridden.
Campbell,
I’m not sure that I accept the possibility of constructing a unified violations function [v(z)] is an adequate criterion of whether there is only one right. I’m thinking that rights, like prima facie duties, tend to conflict with one another. When they contingently do, it is a possible theoretical choice to say that in that circumstance, anything one does violates one of them, and so violates the compound, singleton right you’ll construct. But the fact that, in some circumstances, the set of rights generates a conflict such that violation is unavoidable is some evidence of plurality.
More generally, while there is much attraction in saying that all-things-considered rightness is one, I don’t see the attraction in thinking that there’s only one right. Rights, such as the right not to be killed, are among the main considerations we deploy to figure out what we ought to do. They also seem to go towards explaining what ought to be done, and may further play a constitutive role. The items that play these roles are many. While mitigating conflict among them is an important theoretical desideratum, an equally important check against this systematizing ambition is fidelity to what, on reflection, seem to us sound considerations.
Campbell and Henry,
In order to make sense of the interest theory/will theory debate about the foundation of rights, it seems to me that we must accept that there are considerations more basic than rights and that rights are indeed based upon those more basic considerations. On an interest theory of rights, for example, a right is based upon some interest or value (or several) that is sufficiently important that it imposes duties on people to respect the right-holder in certain ways.
The reason, then, that I don’t believe that there is only one right is that I accept value pluralism; for every value, there is potentially a right (and this is to say nothing of what we get when values are mixed, etc.). But for a related reason, I’m also hesitant to accord rights the explanatory role that Henry believes they have. I’m drawn to the view that rights pass the buck to the more basic considerations upon which rights are founded — on an interest theory, that would be to the value(s) that grounds the right. I accept that rights play a role in our actual dialogue about right and wrong and whatnot, but I think that rights are likely place-holders in that role. What is doing the work at a philosophically deep level, what is playing an explanatory role, by my lights, are the more basic reasons upon which rights are founded. I should add, for what its worth (not much!) that I’m not at all sure that these considerations apply to principles or obligations, because its not clear to me that either are based on more basic considerations in the way that rights are.
Thanks, all, for the discussion. I’ve found this very helpful!
Henry,
The following seems a plausible definition of conflict between rights: a right X conflicts with a right Y in a situation S iff every action available in S violates either X or Y. On this definition, a right might conflict with itself; this would happen in a situation where every available action violates the same right. Suppose, for example, you’re in a situation where you can save one person only by killing another. We might describe this as a conflict between two rights: the right to be saved, and the right not to be killed. Or we might say there’s just one right, the right to be saved and not killed, which is in conflict with itself.
You write:
Perhaps these many items are not many rights, but many aspects of one right.
John,
What’s wrong with saying there’s only one value? It seems that for any pluralistic theory, we can construct a monistic counterpart theory, by combining the plurality of values into one. So what’s wrong with the monistic counterpart of the pluralistic theory you favour?
Campbell,
I guess I simply don’t believe that it is in fact possible to construct a monistic counterpart to a pluralistic theory of value without loss. I realize that that just restates the thesis of value pluralism, to which I already pledged my allegiance: namely, that there are innumerable, irreducibly distinct values. But we may just be at a stalemate on this one.
John,
Fair enough. I don’t really have a firm view about the number of values, partly because I’m not sure how to count them. But I was interested in why you think the number can’t be one.
Let me suggest another argument against specificationism.
If we want to know the correct specificity for a theory of rights, we need to compare theories that, as far as possible, differ in specificity alone.
For example, consider these two theories:
T1. There is a right not to be killed.
T2. There is a right not to be killed or tickled.
The more specific theory, T1, is more plausible. But that’s no reason to favour specificationism. T1 and T2 differ in another important way. They differ with respect to which actions count as violations of rights; in particular, only T2 holds that tickling violate rights. The anti-specificationist might claim that it is this latter difference, rather than the difference in specificity, which explains why T1 is more plausible than T2.
Consider, then, a third theory:
T3. There is a right not to be killed painfully, and another right not to be killed painlessly.
T3 differs from T1 only in specificity. They count all the same actions, i.e. all killings, as violations of rights. (I assume all killings are either painful or painless.) But whereas T1 says there is only one less specific right, T3 says there are two more specific rights. So this is a better test case for specificity. Here it seems the less specific theory, T1, is preferable. So this supports anti-specificationism.