INTRODUCTION
Fairness is a concept that pervades our moral and political discourse. But what is it for something to be fair or unfair – what is our best understanding of the concept? Perhaps the best-known answer to this question is provided by John Broome: fairness is the proportional satisfaction of claims.Footnote 1
Broome's account of fairness has been subject to criticism. In particular, its focus on proportionality and pure comparativeness (and as such its refusal to countenance absolute satisfaction as relevant to fairness)Footnote 2 and its focus on claims to the exclusion of side constraintsFootnote 3 have been scrutinized. In this article I question whether the focus on claims is appropriate (but not in the sense that some other things, like side constraints, should be considered as well). My doubts about claims as the relevant focus for a theory of fairness rest on the logical relationship this implies between claims and fairness – in order for fairness to be the proportionate satisfaction of claims, fairness must respond to claims, and to claims only, and claims must not respond to, or be generated by, fairness. Fairness's job, Broome implies, is to handle and mediate between conflicting claims. But if this picture of fairness and claims were correct, the following would be true: fairness could play no role in determining the correct distribution of claims; fairness could not be used to evaluate the distribution of claims; we could not allow fairness to guide us in distributing claims (or unowed goods); we would never have a claim to be treated fairly (and fairness could not be the source of a claim); and we would not be wronged when treated unfairly. I find these entailments implausible, and as such doubt that the proportional satisfaction of claims is the right way to understand fairness. At the very least, I think that the relationship between fairness and claims is more complex than Broome suggests and the ways in which it is more complex are not those that Broome imagines may be the case.
BROOME'S THEORY OF FAIRNESS
Broome's theory of fairness is relatively straightforward – fairness requires that ‘claims should be satisfied in proportion to their strength’.Footnote 4 Broome also makes the following claims, which I will not question.
(1) Fairness is a pro tanto value – fairness is not everything, and we should not always do what is fair.Footnote 5
(2) The importance of fairness varies with context.Footnote 6
(3) Fairness is a strictly comparative concept.Footnote 7
Obviously if we are to understand this theory of fairness, we must understand what Broome means by ‘claims’. Claims, Broome explains, are things that are owed to us and can have a variety of grounds, such as desert, need and agreements.Footnote 8 However, Broome recognizes that not all claims necessarily require proportional satisfaction. In particular, Broome differentiates the kinds of claims that fairness must mediate between (which he calls ‘fairness-claims’) and ‘side constraints’. If there is a side constraint reason in play (and Broome is sceptical that such reasons exist), then that is the end of the distributive story – the person must have it and fairness doesn't come into it. In making this differentiation, Broome provides further information about fairness-claims: such claims generate pro tanto reasons for the claim-holder to have the good in question.Footnote 9
Note, therefore, that one's having a fairness-claim to something is doubly removed from the judgement that it is (all things considered) right that one should have it. (Hereafter, following Broome, I will refer to ‘claims’, meaning fairness-claims, and ‘side constraints’). First, claims are themselves pro tanto. We must weigh claims against other claims, and we may not be able to satisfy all claims. It is fairness's job to mediate between competing claims. Second, fairness itself is not all that matters, it is also a pro tanto reason for someone to have something, so even if fairness favours your claim being satisfied, other reasons may need to be considered alongside, and may even completely trump, fairness, meaning that it would be all things considered wrong to satisfy your claim.
To sum up, then, Broome identifies fairness as the proportionate satisfaction of claims. Claims are special reasons for us to have something: they are things that are owed to us. Broome differentiates his ‘fairness-claims’ from side constraints, as fairness-claims are pro tanto, but he also speculates that these fairness-claims may exhaust the category of claims altogether.Footnote 10 At the very least, I hope to show that this speculation is false, regardless of whether side constraints exist or not. In other words, the relationship between fairness and claims is more complex than Broome imagines and not because of the potential complication he considers (side constraints).
Broome gives persuasive reasons for focusing on claims. He states that:
Certainly, fairness is only concerned with claims, and not with other reasons. Suppose there is some reason why a person should have a good, but she has no claim to it. Then if she does not get the good, that may be wrong, but she suffers no unfairness. It cannot be unfair to deny her a good she had no claim to in the first place.Footnote 11
Broome has offered this example as evidence for this stance:
if I threaten to destroy a city if I do not get a drug, this is a reason for giving the drug to me, but not a claim I have to it. Fairness is concerned with claims only, not with other reasons. If I have no claim to a resource, and I get no share of it, then no unfairness has been done me, even if there are good reasons why I should have got a share.Footnote 12
It would indeed seem ludicrous to suggest that the threat to destroy the city had any bearing on who should get the drug as a matter of fairness. This seems to bolster Broome's thought that fairness's job is to mediate between competing claims.
SOME CONSEQUENCES OF THE FOCUS ON CLAIMS
Despite these seemingly good reasons for restricting the domain of fairness to claims, I think there are reasons to doubt that this is the correct understanding of fairness. To say that fairness is the proportional satisfaction of claims makes a sharp distinction between fairness and claims and says that claims exist independently of, and logically prior to, questions of fairness – claims are the subject of fairness. When we think about what would be fair, we need to know what claims people have.Footnote 13 Fairness responds to claims, which means that claims cannot respond to fairness, so what claims people have to a good cannot depend on it being fair that they have it. This would be dangerously circular. In addition, for questions of fairness even to arise, people must have conflicting claims to a good. Training the focus of fairness exclusively on existing claims, and the logically prior nature of claims to fairness, has the following consequences, which I think tell against understanding the relationship between fairness and claims in this way.
Fairness has nothing to say about how non-voluntary claims are or should be distributed
Some claims just exist – other people don't need to do anything for us to get them. Other claims are distributed to us by other agents, through promises, contracts or other voluntary acts by others. A claim based on need is an example of a claim that does not arise through voluntary acts of others – nobody gives me that claim, I just have it. I will call these non-voluntary claims, or NV-claims for short. An example of a claim that is generated by the voluntary acts of others is the claim generated by your promising to pay me £5 tomorrow. In doing so, you give me a claim to £5 tomorrow. I will call these voluntary claims, or V-claims for short.
In this article, I will largely focus on V-claims and the distribution of unowed (i.e. non-claimable) goods. But, first, I want to consider briefly what Broomean fairness can say about those claims that we just have, NV-claims. Imagine somebody says that white people's needs are more important than those of black people, and therefore black people's needs give rise to weaker claims than those generated by the needs of white people. If Broome's theory were correct, and fairness merely responded to claims, we could not say that the posited distribution of NV-claims is incorrect because it is unfair. Fairness's role is to mediate between conflicting claims, not to tell us who has what claims. We could say that the posited distribution of claims was incorrect, and indeed morally repugnant, but fairness could play no direct role in our indignation. We could also say that any attempt to satisfy these claims proportionately would lead to unfairness. Therefore, epistemically, we may reach the judgement that the posited distribution of NV-claims was incorrect through the concept of fairness, but the logical relationship is they key point here. Any distribution of goods based on the posited distribution of claims would be unfair because that is the incorrect distribution of claims. The distribution of claims would not be incorrect because it was unfair. Fairness cannot, on Broome's theory, ground or explain the correct distribution of NV-claims.
I don't think this first consequence is a huge issue for Broome's theory – we are left with resources to condemn the posited racist distribution of NV-claims, but some may find it odd that they are disallowed the concept of fairness to do so.
Fairness cannot be used to evaluate the distribution of claims
In addition to being unable to condemn posited incorrect distributions of NV-claims, Broomean fairness could not assist us in evaluating actual (i.e. correct) distributions of claims. We could not say, for example, that we recognize that the distribution of claims looks a certain way, and the fact that it looks that way is in one way regrettable, because it is unfair. Imagine, for example, that someone has an NV-claim based on need because she has been left destitute by a hurricane. According to Broome's theory, we could say that fairness requires this claim to be satisfied in proportion with other claims, but we could not say that it is unfair that this person is in need, and thus that she has this claim. If fairness simply responds to existing claims, it cannot condemn or praise the circumstances that lead to claims (unless those circumstances come about through disproportionate satisfaction of claims). Some will find this example untroubling. A hurricane, they may think, cannot be fair or unfair – only agents can act fairly or unfairly. Others will hold a telic understanding of fairness, and think that states of affairs not brought about by human (in)action can be praised or condemned by fairness. But even those who think of fairness as purely concerning the interactions of agents may be troubled by our inability to use fairness to evaluate the distribution of claims. For example, it seems plausible that the fact that Kerry's parents leave her a lot of money generates one type of (V-)claim to that money. Therefore, Kerry may be owed the money in a way that others, through no fault or choice of their own, are not. It seems plausible to me that we could recognize Kerry's claim, while also declaring the distribution of claims to be unfair (and thus in one way regrettable). On Broome's theory, this is simply not possible – fairness's job is to mediate between claims as we find them, not to help us work out what we think of them, or in what ways that distribution may be thought regrettable.
The problem identified here is one of evaluation – regardless of what we must do, we cannot say that the actual distribution of claims is in one way regrettable because unfair. This concern can be put aside if we restrict fairness's domain to the normative.
Fairness has nothing to say about how we should distribute V-claims, nor unowed goods
Not all claims are non-voluntary. There are some claims that are created by our voluntary actions, such as when we promise things. We also give away things that nobody has any claim to. The problems identified thus far have focused on things that Broome's theory does not allow us to say, or evaluations that fairness cannot help us make. However, when distributing V-claims, or giving away unowed goods, we must decide how to act, and thus the distribution of V-claims and unowed goods is a matter for a normative theory. Yet if claims are logically prior to fairness, and fairness is only concerned with sorting out how to respond to existing claims, then fairness can play no role in how we ought to distribute V-claims and unowed goods. This is, I think, more troubling for Broome's account.
Consider a divorced father who lives separately from his two daughters. He is a good father, and has fulfilled all of his daughters’ claims over him and his time. Despite being a good father, he has, as divorced fathers sometimes do, purchased a two-seater sports car. This means that he can only take one daughter out at a time. The father decides to spend more time with his daughters and decides to take one to the zoo. Recall that he has already fulfilled their claims on him, so neither has an existing claim to a trip to the zoo – if he took nobody to the zoo, nobody would be wronged. Imagine that the father either calls up and promises to take the elder daughter to the zoo (distributes a V-claim to her) or turns up in his car and takes the elder daughter out to the zoo without warning (distributes an unowed good to her). This, to me, seems unfair – the elder daughter has had a trip to the zoo while the younger daughter has not. This may be a justifiable unfairness – after all, the alternative was that nobody got a trip to the zoo – but it nevertheless seems unfair. Yet on Broome's account, fairness can have nothing to say about this situation – there was no prior claim, and fairness cannot say anything about how we should distribute (or create) claims, it is only concerned with how we satisfy claims that exist independently of fairness considerations. The father has either distributed a V-claim, or has given away an unowed good. On what basis can the younger daughter claim that she has been treated unfairly? She has no claim, and, on Broome's account, fairness is only interested in claims – not in how they are distributed or in unowed (and thus unclaimable) goods. In order to make a fairness-based complaint, the younger daughter has to have a claim, and she does not have one.
Now look at the situation from the father's point of view. He wants to act fairly, and so wonders how he should distribute claims to zoo trips, or zoo trips (to which nobody has a claim). Broomean fairness has nothing to say about this – it tells us how to treat claims, not how to create or distribute them, nor how to distribute non-owed goods.
The father turns up again in his sports car the following Saturday. It is obvious from the point of view of fairness what he should do – he should take the younger daughter to the zoo. Yet on Broome's account, not only has the father no reason to correct his unfairness, since he was not unfair, he has no reason of fairness to take the second daughter out, or indeed not to take the elder daughter out a second time. Imagine the father turns up the second week and beckons the elder daughter into the car for her second trip to the zoo. Intuitively, this is now definitely unfair. Before, he could only take one, so either there was no unfairness, or a potentially justifiable unfairness. But now the father has the opportunity to even things up, but in fact exacerbates the inequality. What, however, could the second daughter say, with Broome's theory of fairness at her fingertips? The father has fulfilled all NV-claims (which, let's assume, are equal between the daughters). He now has some unowed goods or claims to distribute, and fairness can say nothing about how he should distribute them. This is as true when he decides whom to take on the second zoo trip as it was when he was deciding whom to take on the first.
You might be tempted to think that whilst neither daughter had a claim to the zoo trip in the first week, the distributing of the trip to the elder daughter generated a claim to any future zoo trips to the younger daughter. This would make the father's behaviour in week two unfair, since the younger daughter would have a claim and the elder would not, and therefore fairness would demand he take the younger daughter. But it is hard to see how such a claim could be generated. One natural response would be to say that the younger daughter has a claim to be taken to the zoo on the basis that that is what would be fair. This would involve claiming that claims themselves are sensitive to, or generated by, considerations of fairness, or that there are claims to be treated fairly. However, such things would be impossible on Broome's account, as I will show in the next section. Another response, which would, again, allow us to end up saying that she was being treated unfairly, would be to say that the younger daughter has a claim grounded in another comparative value, such as equality. I consider this response, and what it means for both Broome's theories of fairness and equality, in the next section.
Alternatively, it may be that there is some other way to condemn the father's behaviour, other than branding it unfair. But I cannot think of a more apt one, and avoiding calling it unfair in order to respect the posited relationship between claims and fairness would be to twist our understanding of fairness to fit with Broome's theory, rather than searching for an understanding which best fits our intuitive sense of it. That we want to call the father's conduct unfair but cannot with the resources of Broome's theory is a reason to doubt that Broome's theory has captured what we understand fairness to be about.
Nobody has a claim to be treated fairly, or claims generated by or based on fairness. Fairness cannot be owed
It may be thought that in week two the younger daughter has a claim to be taken to the zoo, based on the fact that her sister was taken last week. Yet it is hard to see how such a claim could have been generated. If it was an NV-claim, such as one based on need, she would have had it the week before (and we are assuming that the father has fulfilled all NV-claims). And it cannot be a distributed V-claim since the father has not distributed any claims to her.
The only way that I can see that a claim for you could be generated by a V-claim being distributed to some other, or by some other being given an unowed good, is if you had some pre-existing NV-claim to be given an NV-claim that matches all V-claims and/or unowed goods given to the other (perhaps restricted to this particular distributor). The existence of such a claim, however, would be problematic for Broome's theory in one of two ways. Either the claim is a claim to be treated fairly (or otherwise grounded in fairness) or it is not, and each of these possibilities seems to create problems for the idea that fairness is the proportional satisfaction of claims, especially if (as Broome speculates) all claims are fairness-claims, or if (as Broome implies) side constraints are the only other possible type of claims. Endorsing the existence of the claims considered here means endorsing the existence of types of claims which Broome does not consider, making the relationship between fairness and claims considerably more complex.
A claim to be treated fairly (a) cannot do the work we need it to here and (b) is impossible on Broome's theory anyway. Regarding (a), establishing a claim to fair treatment would do no good to the younger daughter, as it would just say that she has a claim to having her (fairness-)claims treated proportionately with those of her sister. But that is the only relevant claim she would have here, and without any other claims to be satisfied proportionately, fairness still cannot explain why she should be taken to the zoo. Since she has no other claims, the claim to fairness has no bite. In other words, in order for a claim to be treated fairly to do any work here, we must be able to show an unfairness, and we still do not have the resources, from within Broome's theory, to show that there has been an unfairness.
More problematically for Broome's theory, both in terms of whether it can assist our hard-done-by younger sister and for the theory's plausibility more generally, is (b): if fairness is the proportional satisfaction of claims (and the only types of claim are fairness-claims and, perhaps, side constraints), then there can be no such thing as a claim to fairness. If claims are the subject of fairness, fairness cannot be the subject of claims. What could such a claim amount to? It is the claim to have our claims satisfied proportionately. In order to make sense of this claim, we will need other claims. And fairness will then have to mediate between the claim to fairness, and those other claims. Which means a fair distribution must take into account (amongst other things) a fair distribution. This becomes (viciously) circular.
Broome recognizes that the importance of fairness varies with context. But his theory as it stands cannot recognize that sometimes fairness is important enough such that persons have a claim to be treated fairly – we owe it to them to be fair. A family seems like the kind of context where this will be the case. Not being able to recognize claims to fairness also has the consequence, discussed below, that we do not wrong people when we treat them unfairly.
The only way for Broome to escape this is to endorse the existence of a type of claim that is neither a fairness-claim nor a side constraint.Footnote 14 There is no acknowledgement of this kind of claim in Broome's theory: as we have seen, the only type of non-fairness claim that Broome even contemplates are side constraints, and he is sceptical even of those, speculating that fairness-claims may exhaust the category of claims. The kind of claim imagined here, a claim to be treated fairly, is neither a fairness-claim (the kind of claim that fairness must take into account) for the circularity reasons just mentioned, nor a side constraint, since side constraints have a veto whilst, as Broome recognizes, fairness does not. Fairness, on Broome's theory, is not the end of the distributive story, while side constraints are.
Is recognizing this further type of claims – claims to be treated fairly – an easy way out for Broome? First, it doesn't solve our problem here: the younger daughter still needs the resources to say that she has been treated unfairly, and a claim to Broomean fairness cannot provide those. Second, to do so would be to acknowledge that fairness-claims are definitely not the only type of claims. Third, this response seriously complicates the story Broome tells about fairness and claims and in doing so considerably lessens our ability to get a handle on Broome's theory of fairness. Broome's story about fairness and claims is initially quite simple. Claims are owed and fairness is the proportional satisfaction of claims. We can understand this quite easily. He then introduces the complication of side constraints, and distinguishes ‘fairness-claims’ from those. ‘Consequently,’ he says, ‘I cannot pretend to have defined claims independently of the notion of fairness . . . The subclass of claims I am talking about is partly identified by the way they work.’Footnote 15 However, it is simple enough to get a handle on the fairness-claims/side constraints distinction: we can understand the difference between a pro tanto claim and a claim that is a veto, as they work differently. But since a claim to be treated fairly is like a fairness-claim in being pro tanto, we can't apply the pro tanto/veto distinction to help us understand which claims are which. In addition, unlike side constraints, these claims are heavily bound up with fairness. The story at the moment is that claims are either inputs to fairness (fairness-claims) or completely unrelated to fairness (side constraints, if they exist). These new claims are, unlike side constraints, outputs of considerations of fairness – they are heavily bound up with the concept of fairness, making differentiating them from, and understanding their relationship with, other kinds of claims a more difficult task. These complications amplify a problem with Broome's account as it stands, identified by Brad Hooker: ‘sometimes Broome's account of fairness seems to be the thesis that fairness requires proportional satisfaction of reasons that are to be satisfied proportionally’.Footnote 16
If the claim to have whatever another has had is not a claim based on fairness, then we can only say that fairness requires that the father take the younger daughter in the second week if she now has a claim based on the elder daughter's previous zoo trip. However, we would need to know what, if not fairness, this claim could be based on. But for any given alternative to fairness as the basis for the younger daughter's claim two questions will then arise. First, we can ask whether ascribing a claim on such a basis fits our assessment of the situation. Second, we can ask whether generating claims on this basis properly fits in with Broome's (or our) understanding of the nature of claims, the role of fairness and the value of equality.
Let's ask the first question first: does ascribing the younger daughter a claim to be taken to the zoo on a basis other than fairness fit with our assessment of the situation? We would have to say that the younger daughter has a claim to be taken to the zoo, and that this claim is grounded in some other value (like equality) or is generated by some higher-order or meta-claim to be given that which her sister is given (or equivalent). However, the claim that the younger sister has can be neither a claim to be treated fairly nor a claim based on or related to fairness – it is just one of the claims that fairness must take into account. The idea that the second daughter's claim is not one of fairness doesn't seem the best explanation of what is going on here. It seems to me that the father should take the younger daughter out the second week because that would be the fair thing to do, and she should be, indeed has a claim to be, treated fairly. Yet on this response to the zoo example, the reasoning runs in the opposite direction: it is fair to take her because she has a claim to equal treatment, rather than having a claim to equal treatment because that would be fair.
We can now turn to the second question: does generating claims in such a way fit with our understanding, or Broome's understanding, of the nature of claims, the role of fairness or the value of equality? One possible response to the zoo example is to say that the second daughter (now) has a claim to a zoo trip which is generated by considerations of equality. However, in order to recognize that a claim can be generated by the value of equality, Broome would have to acknowledge three things. First, that claims themselves can be comparative, or generated by comparative concerns. Second, that it is therefore not for fairness alone to work out comparative morality – claims themselves do some of this work. Third, that equality has value and moral standing independently of fairness.
Broome does not, so far as I am aware, explicitly deny either of the first two (that claims can be generated by comparative concerns or that fairness alone does not work out comparative morality), but he does not explicitly endorse them, and they rather cut against the tenor of his theory and his understandings of the respective roles of fairness and claims. The theory appears to be one in which we assess our claims to some good independently of comparative concerns, through looking at our direct relationship to the good (Do I need it? Do I deserve it? Have I earned it? Did someone give it to me?). Comparative concerns are then taken care of by fairness. Certainly, none of the potential grounds of claims mentioned by Broome are comparative.Footnote 17 Yet in order to recognize this newly generated claim to a zoo trip, Broome would have to acknowledge that claims can have comparative grounds. Broome claims that ‘fairness is concerned only with how well each person's claim is satisfied compared with how well other people's are satisfied. It is concerned only with relative satisfaction, not absolute satisfaction’Footnote 18 and that ‘the particular business of fairness is to mediate between the conflicting claims of different people’.Footnote 19 So, fairness is exclusively concerned with comparisons. But if we could have comparative claims of the sort considered here, then we would have to admit that comparisons are not exclusively fairness's business – claims themselves also respond to comparisons, and form part of comparative ethics.
In addition, this way of explaining how the younger daughter is treated unfairly – one that states that she has a NV-claim grounded in the combination of the value of equality and her sister's prior (unowed) zoo trip – seems to contradict Broome's understanding of the relationship between fairness and equality. On the view under consideration here, the value of equality generates claims, which in turn need to be considered in treating people fairly. So equality is a value independent of fairness, and one which we can make use of in working out what claims people have, and thus what is fair. But Broome sees the relationship the other way round. For Broome, the value of equality is explained by fairness – fairness tells us why and in what way equality is valuable.Footnote 20
We do not wrong anyone when we treat them unfairly
As I noted at the start of the article, Broome clearly sees fairness as a normative value (a reason for action), which it surely is, albeit a pro tanto one. In describing claims, Broome states that there are three kinds of moral reason: teleological, side constraints and claims.Footnote 21 Broome never, however, says what kind of reason fairness is. He does say that unfairness is harmful.Footnote 22 Now, we know that since fairness is a pro tanto consideration, it cannot be a side constraint. And, as we saw in the previous section, it cannot be a claim, since a claim to fairness would be nonsensical on Broome's theory (as it stands), which requires a one-way relationship between claims and fairness: fairness responds to claims and so claims do not respond to fairness. Therefore, if Broome's taxonomy of reasons is correct, fairness must be teleological, and the harm must be of that sort.Footnote 23
I am not sure that Broome can help himself to the claim that unfairness harms.Footnote 24 His claim appears to be that since fairness is concerned with claims, and claims are held by individuals, then unfairness harms the individual: not having one's claims satisfied as much as another with the same strength of claim is a harm.Footnote 25 I don't see that this follows. Not having one's claim satisfied may well be a harmFootnote 26 – but that says nothing about why having your claims satisfied less than some other is (in itself) a harm, which is what Broome requires to be true. Indeed, for Broome, the fairest thing will often be to satisfy no claims, and it seems to me that this is to be maximally but equally harmful.
Regardless, even if unfairness harms, I don't think that Broome, given the current taxonomy of reasons he uses, can claim that we are wronged when treated unfairly. This is because of the type of reason fairness must be for Broome. Since fairness cannot be a side constraint, and it cannot be a claim, then it cannot be owed. And since it cannot be owed, we do not wrong those who do not receive it, since we did not owe it to them. It might be wrong, on Broome's account, not to treat your children fairly, since fairness is a reason, and acting rightly will mean taking account of, and balancing correctly, all relevant reasons, but you do not wrong your children in so doing – you do not fail to give them something that they are owed. This would seem to make treating someone unfairly (or not according fairness its due weight) the same kind of wrong as not giving the drug to the person who threatens to destroy the city – it's the wrong thing to do, but it does not wrong the individual you did not give it to. This, I think, conflicts seriously with our understanding of the nature and value of fairness.Footnote 27
