PEA Soup is pleased to introduce our newest PEA Soup Book Review, as Patrick Tomlin (Warwick) takes a look at Larry Temkin’s book Inequality.
This continues PEA Soup’s book review series, where philosophers introduce a book that influenced them or that they found important, and provide some analysis of that work.
Over to Patrick.
Patrick Tomlin on
Larry S. Temkin, Inequality (New York: Oxford University Press, 1993).
For PEA Soup
1.
I first read this book as a graduate student. I loved it. There were two aspects of it which I was particularly struck by at the time, and, on revisiting it, I was struck by the same two aspects all over again. I cannot hope to do justice to the richness, complexity, and inventiveness of Temkin’s book in this short piece. But I would like to articulate the two aspects of it I found so exciting as a graduate student, and also, by way of doing a little philosophy, introduce the complexity of equality that Temkin both illuminates and revels in.
2.
The two elements of the book I found so striking concern the very type of project or enterprise that Temkin is engaged in here. The first is the overall approach. Very crudely, let’s distinguish between positive projects (‘think this!’) and negative projects (‘don’t think this!’ or ‘don’t think this alongside this!’) This book fits into neither of these categories. It is a third type of project—it is primarily an exploration. You have the sense that for Temkin the philosophical excitement is not necessarily in coming up with the right answer, and getting people to agree, but discovering the puzzle, and helping people to think about what the right solution could be.
I think it is this aspect of this book (and other books written in a similar style) that, as a graduate student, I found most inspiring. Philosophical contributions do not have to come from having all the answers. They can also come from discovering a puzzle, or presenting a range of positions or options that people have not developed, even if you are not quite sure which is the right one.
3.
The other striking feature about this book is its topic—inequality, rather than equality. Philosophical work on equality tends to concern whether equality is (in itself) valuable, and what valuable equality looks like—in what way is it good or just for us to be equal? It is, in other words, about the ideal of equality. This book, however, is not about a philosophical ideal, it is about the absence of an ideal, and about which situations of inequality are worse than others. Temkin gives two reasons for this focus on the negative: ‘First, it can be of little practical consequence that one regards inequality as bad—as many do—unless one is generally able to determine if one situation’s inequality is worse than another’s. Second, and more fundamental, I believe it is only by addressing this question that one can begin to understand the nature and complexity of the notion of inequality’ (p. 3)
I think Temkin somewhat understates, here, the practical importance of the question he is asking. Most egalitarians (or distributive egalitarians at any rate) are pluralists. Equality may be good, and inequality bad, but other things matter too. Equality and other values must be ‘traded off’ against one another. Once we have this pluralist picture in mind, it really matters what makes one unequal distribution better or worse than another, for once we get into trading off equality against other values, we’re unlikely to get a perfectly egalitarian distribution. But trading can only occur if there is a clear sense of what you’re ‘giving up’. We can’t know which trade-offs are all-things-considered best if we don’t understand the costs involved to the different values.
4.
This is not an easy book to summarise, because it has no overarching single thesis that it is arguing for. It is made up of different puzzles and different solutions to, or things we might think about, those puzzles. In closing—and as a way of demonstrating the type of question and the type of approach the book takes—I want to discuss just one argument from the book (or more accurately one place where Temkin can’t find an argument).
(I should note here that I am responding here directly to Temkin’s book, and not to the significant literature which has responded to and/or been inspired by it).
We need to start with Temkin’s ‘Sequence’. This sequence concerns what ‘balance’ of better off and worse off individuals is best. (The best balance is to have everyone equally well off, but we assume that is not possible). So Temkin asks us to imagine 999 distributions between 1000 people, where people can either be better off or worse off (i.e., there are only two groups, there is a fixed population size, and the only changes are changes in the size of the groups). So we get this sequence:
The Sequence:
Better off Worse off
D1 999 1
D2 998 2
D3 997 3
…
D500 500 500
…
D999 1 999
D1 is clearly the best in terms of overall utility. But Temkin is not interested in that question. Temkin identifies five possible positions concerning The Sequence. The first is that as The Sequence progresses, things get worse and worse regarding inequality (I’ll drop the clarification ‘regarding equality’ from hereon in). The second is that things get better and better. The third is that things first get worse, and then get better (so the distributions roughly in the middle are the worst). The fourth is that things first get better, and then get worse (so the distributions roughly in the middle are the best). The fifth is that all the distributions are as bad (or as good) as one another. (Temkin ultimately (p. 297) endorses the worse-then-better view.)
Temkin concentrates on our ‘complaints’ about inequality. He identifies two different ways of deciding who has a complaint, and three different ways of identifying the size or content of their complaint. He then identifies several ways in which complaints can be used to form a picture of the overall badness of an inequality.
One thing I found surprising on revisiting Temkin’s discussion of The Sequence is that Temkin claims that there is nothing plausible to be said for the fourth view—the claim that things get better, and then get worse through The Sequence (such that the distributions roughly in the middle are the best). This is surprising because Temkin, as I have emphasised in the above, is so brilliant at finding what may be said in favour of a huge range of positions. So in the spirit of Temkin, I wanted to try to articulate some thoughts on what may be said in favour of the better-then-worse worse.
Let me start with observing that Temkin finds things to say for both the ‘things are getting better’ view, and the ‘things are getting worse’ view. In articulating the pull of both of these positions, Temkin focuses on the ‘bad extreme’ distribution, and notes ways in which it is particularly bad. (As a side note, it is interesting that we might find these extreme distributions (D1 and D999) particularly bad, since they are so close to being wholly equal. But as Temkin observes, the question of which distribution is more unequal is not the same question as which distribution is worse regarding inequality).
In the better-and-better view, the bad distribution is D1, where almost everyone is better off. Temkin notes that in that distribution ‘the whole burden’ of inequality is concentrated on a single person, and this may seem especially unfair. He goes on to give various technical reasons (around the types of complaints he articulates) to back up this basic thought, but the main point is that there are plausible reasons for thinking D1 especially bad.
In the worse-and-worse view, the bad distribution is D999, where almost everyone is worse off. Temkin notes that these later distributions bring to mind medieval Europe, with a chosen few living extremely well, and the vast majority living extremely poorly. There, again, seems to be something especially invidious about this distribution. Again, Temkin is able to show that there are plausible ideas about complaints and the way in which they operate that would explain this view.
My point here is simple. If it isn’t merely the case that we might find ‘things to say’ about the better-and-better view and the worse-and-worse view, but that both latch on to things that genuinely matter about inequality (that is, if we see them not as competing answers to how to view The Sequence, but as two elements of a pluralistic account of the badness of inequality) might we not find a plausible view that views both extremes as particularly bad, and therefore the middle distributions (those around D500) as not so bad?
It may be thought that if we endorse whatever underlies the worse-and-worse view and whatever underlies the better-and-better view, they will cancel each other out, and we’ll end up viewing all 999 distributions as equally bad. But this is so only if they are equally weighted and are both linear (that is that every distribution is precisely equal in terms of how much better or worse it is than the previous distribution on both measures). But Temkin’s vivid descriptions of the worries about the extreme ends of The Sequence might lead us to question whether this is the case. When it comes to our concern about the concentration of complaints[1], we might think that the gap between D1 and D2 is much larger than the gap between D499 and D500. Indeed, if the problem with D1 is that one person ‘bears the burden’ of inequality alone, it seems quite plausible that this concern is halved by the move to D2, whereas things are only mildly improved by the move from D499 to D500.
This, then, is one reason to favour the better-then-worse account – the extremes are both particularly bad. But now let me try to say something in favour of the middle distributions. In Appendix A, Temkin considers whether there is anything to be said in favour of distributions like D500 in which the unequal groups are of equal size. He concludes that there is not. This equality is, he says, ‘purely mathematical’ and is not ‘itself relevant to how people fare relative to others’ (p. 311). I wonder if this is right. To see this, I want to think about relations of equality.[2]
Temkin phrases everything in terms of complaints against inequality. He briefly considers whether relations of equality are independently important, but concludes that they are basically the absence of complaints of inequality (pp. 313-314). I think this questionable. Consider this example:
Better Off Worse Off
D1* Anna Bella
D2* Anna Bella; Charlie
D3* Anna; David Bella; Charlie
D4* Anna Bella; Charlie; David
Start with D1*. There is only inequality here. Now imagine we move to D2*. According to Temkin, things get worse regarding inequality (ch 7). In D1*, one person has a complaint about inequality (Bella). In D2*, two people have such a complaint. But I think there is something to be said in favour of D2* over D1*, and indeed, I think it is overall better regarding inequality. This is for two reasons. First, there are now relations of equality—before we had only inequality, now we have some equality. Second, those relations of equality belong to the worse off—Bella no longer bears the burden of inequality alone. Imagine you are a parent, and these are your children. Under D1*, there is nothing positive you can say about equality to Bella. Under D2*, you can acknowledge that things are still not fair between Bella and Anna (this may not be your fault), but you can point out that things are fair between Bella and Charlie. In this sense, things have improved. It may be thought that this ‘value’ of the egalitarian relationship can be explained by the absence of inequality between Bella and Charlie. But that would be something to be said in favour of D2* over some distribution in which Charlie was also better off than Bella, it’s not something that can be said in favour of D2* over D1*.
If relations of equality matter, then we may think that we should maximize them. This, however, would not speak in favour of the better-then-worse view of The Sequence, or in favour of D500 being the best distribution, because D1 and D999 both have many relations of equality (999 people enjoying a total of 499500 relations of equality). Moreover, in D999, these relations of equality are all held by the worse off. But these two distributions are distributions in which some people have no relations of equality. Moreover, in distributions close to these extremes, the relations of equality are very unequally distributed (for example, in D2, two people have one relation of equality, while 998 people have 997 relations of equality each).
Consider being at D2* and moving to either D3* or D4*. D4* has three relations of equality, whilst D3* has only two. So if relations of equality matter, D4* might seem to be better. But there does seem something better about D3*—because everyone is equal to someone, and they are all equally equal to someone. This is why I think it is a mistake for Temkin to say that there is nothing of egalitarian worth in the groups being of equal size. This doesn’t in itself matter—the size of the groups themselves have no egalitarian value—but it might matter because it reflects a situation in which people enjoy an equal number of relations of equality. D500 represents a situation in which all 1000 people are equal to 499 other people. There are fewer relations of equality than in D1 and D999, but they are equally distributed. (Temkin rejects the idea that we should care about the distribution of complaints at pp. 200-209). (We should at least also care about the number of relations of equality, otherwise a situation in which nobody was equal to anybody would be just as good as D500).
I don’t know what the right answer to The Sequence is. But Temkin’s book is brilliant at exploring what might be said in favour of all sorts of views. I think there is something that can be said in favour of the better-then-worse view.
[1] For an excellent account of how to take account of this concern, and so a defence of the better-and-better view, see Julia Mosquera Ramil, ‘An Egalitarian Argument Against Reducing Deprivation,’ Ethical Theory and Moral Practice 20 (2017): 957-968.
[2] I was first prompted to think about whether relations of equality had independent significance when supervising Julia Mosquera Ramil’s PhD thesis. Mosquera Ramil uses relations of equality to ‘offset’ complaints against inequality in ‘An Egalitarian Argument Against Reducing Deprivation’ (drawing on work by Gustaf Arrhenius).
Thank you for this review, I really enjoyed reading it and could feel the graduate student excitement shining through :)!
You say, towards the end and when exploring how a worse — better — worse view might be defended:
“This is why I think it is a mistake for Temkin to say that there is nothing of egalitarian worth in the groups being of equal size. This doesn’t in itself matter—the size of the groups themselves have no egalitarian value—but it might matter because it reflects a situation in which people enjoy an equal number of relations of equality.”
I wonder whether it is correct to say that “the size of the groups themselves have no egalitarian value.” Maybe you are right that the sizes themselves don’t have EGALITARIAN value. But if we grant the assumption that “equality cannot be achieved” (for whatever reason), it seems to me there is some political value to the relative group sizes that might well favour roughly equally large groups. The idea is that if one group is a clear minority, it is easier to view that group as ignorable or as less essential to seek cooperation and agreement with. If we have the aristocracy and a large peasant class, the peasants might one day decide to fight their oppression simply by getting rid of the aristocracy. And if we have a minority of poor individuals, it is easy for the large majority to just ignore them. A balance in group sizes, by contrast, seems to make it more salient and imperative to seek cooperative terms with the other group and to regard it as “different but (in some sense) equal”, partly because of a similarity in size.
Thanks Susanne.
What I meant by ‘the sizes of the groups themselves have no egalitarian value’ was that it isn’t that ‘unequal groups of equal size’ are the target of equality, or provide some good-making egalitarian feature of the distribution. Instead, the egalitarian goodness to be found in the 500-500 distribution is the (equal number of) relations of equality that *individuals* enjoy *within* the groups.
The value you point toward of having groups of equal size are either instrumental to equality (though it is a large peasant class, not an equal peasant class that seems beneficial there), or the benefits are not (as you recognise) not CLEARLY egalitarian benefits.
However, your comment about egalitarian relations between groups did make me think about whether the benefits might be egalitarian in some sense. When I was writing the short piece above, I did think a bit about the relationship between Temkin’s egalitarianism and the ‘relational’ egalitarianism of Anderson, Sheffler, Scanlon etc. This was because in positing that there was egalitarian value in us having equal numbers of ‘relationships of equality’, I seemed to be turning Temkin’s clearly distributive theory into a relational one. In addition, I wondered whether the intuitions driving the idea that there was something better about the 500-500 distribution (intuitions which many people share) were driven by relational rather than distributive concerns. (Though I do not think these things are fully separable–we can, as the above shows, be concerned about the distribution of relationships).
What your comment suggests, I think, is that we should also be concerned about the relationships not only within groups but between them. And you seem to suggest that the badness of relations of inequality between groups/individuals within the groups will be softened when the groups are equal. There is possibly something to this, though of course empirical questions abound. Philosophically, this also creates a metric headache. If the metric of equality is something like ‘general wellbeing’, then shouldn’t the good or bad effects of inequality already be built into the numbers/distribution? That is, if it’s better for people to be equal sized groups (because they get more respect, for example), then we haven’t kept all-else-equal, as the worse off are better off in the 500-500 distribution than they are in the 999-1 distribution.
(But this is a more general worry about Temkin’s methodology. Temkin likes to present an inequality, and then ask how bad it is, and sometimes he seems to be asking how bad it is FOR THE UNEQUAL (‘complaints of inequality’). But if having a complaint is bad for you, or you can only have a complaint about something that is bad for you, then the numbers that are used to demonstrate the inequality cannot be the whole story about wellbeing.)
Thanks for this review (it sounds like this should be on my eventual, post grad-student reading list) and your ‘new piece of the puzzle’.
While I get that the sequence aims to abstract so we are dealing with as few variables as possible, to me, it seems to matter at least how much (if not in what way) they are better or worse off. Take the medieval Europe example: isn’t this objectionable on the basis that most were extremely poor while some were extremely rich cf. a hypothetical where most were reasonably off and some a bit richer.
In the latter intuitively I think I might be drawn towards the quarters (ie D250 or D750 (rather than my feudal punishment for treason)) if your ‘second-order’ equality (equally) compensates for the first.
Patrick Tomlin offers an interesting and thought-provoking criticism of one of Larry’s key arguments in Inequality. What he says raises many interesting points of discussion, but in the following I want to look at just the main point of criticism he makes. Even more specifically, I want to examine the positive suggestion he makes as a proposed solution to that weakness.
First, though, I want to start with something on which both Patrick and I agree. Thirty odd years on, Inequality is still the brilliant and defining work it was when it came out. It is packed full of arguments, in a way that is rarely matched in works of moral and political philosophy these days.
Patrick focuses his criticism to one of the claims Larry makes regarding the famous Sequence. The Sequence is of course key to Larry’s approach to evaluating the badness of inequality:
Patrick targets Larry’s claim that one of the potential accounts about the Sequence, namely that moving from A to B to C, things get better and then worse, is too hastily dismissed. Specifically, he takes issue with Larry’s claim that unlike other interpretations of the Sequence, the better-then-worse view has nothing to recommend it. He then goes on to offer precisely such reasoning, to explain and justify why we might think there is something to be said for that view. I am not so much interested in who is right, Larry or Patrick, about that particular point. What I am interested in is the reasoning that Patrick provides to back his assessment, and what they may imply for egalitarianism.
The crux of the matter is the following. In evaluating how outcomes fare in terms of the value of equality, Larry, Patrick correctly observes, only looks at the badness of inequality (measured in terms of complaints). But suppose, he says, we also looked at the goodness of equality (measured in terms of pairwise relations of equality). That would question many of the value judgments made by Larry in his book, including regarding the Sequence. ‘If relations of equality matter, then we may think that we should maximize them’, he goes on to suggest.
Patrick goes on to provide a nice example that captures the logic and intuitive force of adding pairwise relations of equality to the measurement of how good or bad things are with regard to equality (I won’t go into the details of the example). In my brief comments below I want to offer some thoughts about Patrick’s suggestion, and what it may amount to.
Let me start with a methodological point. There are two nuances to the claim that is the focus of Patrick’s criticism, or at least two interpretations of his claim. One is that there is absolutely nothing to be said for the view that in the Sequence things get better and then get worse. Let’s call this the Strong claim. If that is what Larry intended, then I am with Patrick that this is too strong, and in all probability false, and precisely for the reasons offered by Patrick. That claim, if indeed Larry meant it like that, is indeed quite hard to defend. It is quite hard to find a view about equality that has nothing to recommend it. Indeed, even mere mathematical elegance might serve as a reason.
There is therefore another, more plausible, interpretation about the truth of the Better-then-Worse view, namely that while there is something to recommend it, nevertheless, all-things-considered it proves false. One way to make this more precise is to say that while the view may capture some aspect of the Sequence, overall such a view encounters more costs than benefits, in terms of explanatory power.
In my remarks below I shall exclusively focus on the latter, Weak claim. That is, I will offer some thoughts about whether or not it makes sense for egalitarians to incorporate a positive measure (the number of equal relations) to their negative measure (how many pairwise relations of inequality there are, and the size of their gaps).
First though, here is a point of context. As Patrick acknowledges, the suggestion that we might want to add the positive value of equal relations to balance the negative value of unequal ones is nothing new, and as far as I know is owed first to Gustaf Arrhenius. Since then, there hasn’t been much discussion of this, but there is some. (Gustaf Arrhenius, ‘Inequality and Population Change’, Shlomi Segall, ‘Why We Should be Negative about Positive Egalitarianism’, Utilitas 31 (2019), 414-430; Gustaf Arrhenius and Julia Mosquera, ‘Positive Egalitarianism Reconsidered’, Utilitas 34 (2022), 19-38.)
I said that in support of his Positive egalitarian stance Patrick offers a nice example involving the comparison of four different outcomes. Now, he does not stress this, but there is at least one crucial difference between his example and Larry’s Sequence. While the Sequence concerned a fixed number of individuals across all one thousand distributions, Patrick’s example concerns what in Population Ethics is called Different Number case. This is not an illegitimate move, of course (far from it), but it is one that carries important implications (more on which below), and therefore the difference better be flagged.
Ok, with all this as background let me offer just two reasons (I believe there are more) to think that adding the value of pairwise relations of equality to the overall goodness of an outcome (henceforth Positive Egalitarianism) might not be such a great idea. Each of the reasons picks out a different feature distinguishing Positive from Negative Egalitarianism (the standard Temkin view that the value in terms of equality is measured by only looking at pairwise relations of inequality).
The first reason trades on the following difference between Positive Egalitarianism (PE) and Negative Egalitarianism (NE). NE looks, at minimum, at two factors. One, it counts the numbers of pairwise relations, and two, it measures at the gap in whatever it is that the currency of egalitarianism consists in between members of an unequal pairwise relation. This is just a fancy way of saying that for (negative) egalitarians the gap between myself and Patrick does not count in the same way as the gap between Patrick and Elon Musk, say. That’s obvious. Not so with PE. For an obvious reason it only counts instances; with no need to measure gaps. This of course makes this measurement a lot simpler. But this feature also potentially comes at a cost.
To see this consider a version of the Sequence. Suppose a thousand individuals are living quite goods lives, and perfectly equally so. We therefore have almost half a million relations of equality. Now suppose we add a second of pleasure to the life of the first person, two such seconds to the second person, and so forth. The result is a population with not one single relation of equality. Measured in terms of the goodness of equality (in distinct from the badness of inequality) this constitutes quite the drop in value. But it seems rather counterintuitive that such a minute difference in welfare distribution would result in such a dramatic change to how we evaluate how good or bad things are with regard to equality.
Patrick, or anyone else defending PE, may offer the following ready reply. Rather than looking at equality in this fine-grained way, we could instead talk of ‘rough equality’. All thousand individuals, we could say, still stand in a relation of rough equality. This makes a lot of sense. More precisely, for all practical purposes, looking at rough equality is very sensible, we must concede. But it does come at a price in terms of parsimony. To see this, suppose we stretched the example from one second to one day. Each person is now living one extra happy day, compared to the other. The gap between the first and last person is now almost three good extra years; surely non-negligible. Patrick might then say that rough equality implies that we may group each 30 individuals, say, into a cluster, thereby generating quite a few relations of equality (within each of the 30 odd batches). This may solve one problem, but we can see that the overall result is rather ad hoc. More importantly, this discussion fleshes out that what might be doing the intuitive work behind PE is some sort of ideal of solidarity, or communality of fate. That, in itself is certainly a fine ideal. But one of the problems of anchoring the badness of inequality with that ideal is that it potentially breaks apart the second we move away from laser-perfect equality.
The first reason to doubt PE traded on the fact that unlike NE, it does not incorporate gaps into its measurement. The second reason to doubt PE, a far more damning one in my view, trades on another difference between PE and NE, namely that PE gives us reasons to bring people into existence. This stands in stark difference to NE, which never gives us reasons to bring people into the world (quite the opposite, more on which in a second). Patrick does not discuss this explicitly, but it is an obvious implication of his view. “If relations of equality matter, then we may think that we should maximize them.” (p. 6). The surest way of maximizing relations of equality is of course to bring into existence more and more people who stand equally to one another. I, for one, don’t think this is counteractive in itself. But anyone who is still enamored by the Repugnant Conclusion (I am not), might want to think twice. Paraphrasing Jan Narveson’s famous quip, on PE we are not only in favour of making people equal, but also of making equal people.
As hinted, this is not my second objection to Patrick’s PE. My objection, rather, is the following (picking on a comment by Ingmar Persson, I forget where). Since PE gives us reasons to bring in new people (provided they lead equal lives), then presumably this also applies in cases where these people live lives that are so horrible that they are worth not living. Consider a world with no people in it (just for simplicity). If we could bring two, or a thousand, or a billion individuals all living hellish lives PE tells us to do so, so long as they lead equally hellish lives. We already saw that PE generates a Repugnant Conclusion; now we can add a Sadistic Conclusion to that list. (For more, see Shlomi Segall, The Future of Equality (Oxford University Press, 2025), ch. 3.)
Just to polish things off, let me stress that NE does not give us reason to bring into existence miserable individuals, for the simple reason that NE never gives us reasons to bring people into existence at all. In fact, as Larry mentions in Inequality, if anything NE is vulnerable to the exact opposite problem, namely what he called the ‘Shrinking World’ problem. If you subscribe to Larry’s view (as I do) that the badness of inequality consists (among other things) in the number of pairwise relations of inequality, then it follows that proportional shrinking of an unequal population makes things better with regard to inequality. Indeed, it makes things better all the way to when a population is shrunk to one individual.. To paraphrase a fascist slogan currently popular in Israel, ‘No people? No inequality!’. Is this a problem? Only if you thought equality is the only guide to policy we should appeal to. But thankfully neither Larry nor myself, nor anyone I know, think that.
Let me finish by mentioning a potential reply to the Sadistic Conclusion to PE. The objection, the reply goes, deals with Different Numbers cases. And such cases, it is noted, are notoriously difficult for everyone. Alluding to such cases is no measurement to a view’s plausibility. As you will have guessed, I don’t find this reply particularly moving, but regardless it does not seem one that Patrick would entertain. And that is because his own objection to Larry uses, as its star evidence, an example that is itself one of Different Numbers (top of p. 6).
Patrick’s discussion highlights interesting aspects of Inequality. While I am not convinced by the solutions he proposes, I am convinced that the questions he raises are important and enduring.
Fascinating stuff, and outside what I work on. I read inequality years ago, and now want to re-read it. I had something of a similar response to Shlomi, without knowing the literature at all. I don’t have very strong egalitarian intuitions but I can just about get myself to believe that it is bad for X that X is worse off than Y. This is a relational bad, but a bad for a person. I find it harder to believe that it is good for X that X is equally well off to Y, and as Shlomi points out such a good would not be scalar but absolute. And I find it even harder to believe that is good as such, but good for no one, that X is equally well off to Y. Patrick, I wonder if you had anything to say in support of the idea that equality is good (for a person, or impersonally), or whether there is some way of supporting the conclusion without relying on it.
In response to the Shrinking World problem that Shlomi mentions, I find it hard to believe that there is anything good about having fewer people so that there is less inequality, and I didn’t find the pluralist response that Shlomi offered convincing. But there is another response for negative egalitarians: it is always better for a X to share the world with more people, but it diminishes the value for X of sharing the world with a person, Y, that X is worse off than Y. Inequality, on this view, diminishes a good rather than introducing an independent bad. On that view, if X has 50, introducing Y with 70 makes things better for X than not introducing Y, but it is in one way worse for X than introducing Y with 50. That reply might already be in the literature, of course; I’d be interested to know. I find it better than the pluralist response.
Thanks James. I think you’re right that when we have an objection or worry about a distribution it can be hard to know what drives it. Larry’s sequence is very vague, but then concrete examples (like feudalism) perhaps put more meat on the bones than we’d like, in that they introduce specific factors (like social inequality, as I have been discussing with Susanne. What is clear is that Larry always asks us to focus on how bad things are *in terms of inequality*, and not more generally, or overall. So, the way to try to find out whether the pattern of distribution is bad, or the particular variant, is to think of lots of examples of inequality that follow that pattern, but varying the parameters. That way we should find out whether it is absolute poverty, or relative positions, that trouble us.
Thanks Victor, just a quick reply to your second point. You offer the view according to which “inequality diminishes a good”. You suggest that this avoids the Shrinking World problem, because on your view “it is always better for X to share the world with other people”. Here is my question: would this work also for negative welfare? Suppose there is a person at minus 10, and we contemplate bringing into existence another person at minus 5. Here it is not the case that inequality ‘diminishes a good’. If anything it worsens a bad. More troublingly, it does not seem that “it is always better to share the world with more people”. Consider: first person as before at minus 10, and now we contemplate bringing another person into existence also at minus 10 (which recall is worth not living). It seems very counterintuitive that we ought to add this second person.
Would love to hear more about why you are convinced by the Pluralist strategy, but perhaps at some other point, so as not to steal the spotlight from Patrick.
Shlomi, thank you for this fantastically detailed and thought-provoking response. I can see that, if I want to pursue these thoughts further, I will have some reading to do!
So, we have differentiated between egalitarian reasons and non-egalitarian reasons. We now have a distinction between positive and negative egalitarian reasons. However, I am not fully clear on whether PE (in your post) is: A. a position that incorporates (but is not reducible to) the claim that relations of equality are good; B. a set of egalitarian reasons that are not necessarily the only egalitarian reasons; or C. a position in which there is no negative egalitarianism (so unequal relations just are the absence of positive egalitarian relations). The important question is whether PE and NE are opponents, or can be combined.
To explain my confusion, you introduce PE like so (with my emphasis): ‘But suppose, he says, we *also* looked at the goodness of equality (measured in terms of pairwise relations of equality)’ and it concerns ‘*adding* the value of pairwise relations of equality to the overall goodness of an outcome’
This suggests that PE is A (above). But the criticisms of PE appear to rely on PE not being able to differentiate between different levels of inequality (i.e., view B or C). For example:
‘This is just a fancy way of saying that for (negative) egalitarians the gap between myself and Patrick does not count in the same way as the gap between Patrick and Elon Musk, say. That’s obvious. Not so with PE. For an obvious reason it only counts instances; with no need to measure gaps. This of course makes this measurement a lot simpler.’
But PEs only don’t need to measure gaps if they are C-like PEs. Or perhaps PE here just means a set of reasons (B). But if so then PE can be combined with NE and avoid this implication that the size of gaps doesn’t matter. In other words, Shlomi seems to introduce PE as a position that incorporates positive value for relations of equality, but I think he criticises it as a view that excludes NE. I didn’t mean it as such (and I’ll say more on this in the next post).
Now onto Shlomi’s first example. I wonder, as Shlomi predicts, whether our egalitarianism should be sensitive to single added seconds of life. But once we make the changes more significant, as Shlomi does, all theories are going to agree that we have introduced some egalitarian badness. I think there is a badness in the inequality (as Shlomi does) and I wonder if there isn’t also a badness in the absence of equality. Now, I agree with Shlomi that what stands behind this is something like a relational egalitarian thought. But that doesn’t make it a non-egalitarian thought, necessarily.
Now, onto Shlomi’s second worry. Shlomi is right that PE *gives us reason* to create people, and gives us reason to create people with very bad lives (though I had a similar thought to Victor’s about NE — maybe the value of equal relationships only ‘kicks in’ once people exist). But even if Shlomi were right that PE has this implication, why can’t the PE avail themselves of a very similar reponse to the one that Shlomi gives on behalf of NE’s troubles with population ethics? Namely, equality isn’t the only thing that matters.
More generally, Shlomi outlines two interpretations of Larry’s ‘nothing to be said’ claim (and so two interpretations of what it is to ‘say something’ on behalf of a position on the Sequence. The first is that there can be ‘something to be said’ for a view if we can find something at all plausible in its favour. That, I admit, was my aim, and that is the position that I took Larry to reject in Inequality: p. 47: ‘By now it may seem that there are bound to be several plausible positions supporting the judgment that the Sequence first gets better than gets worse. However, if there are such elements, I am not aware of them.’ However, I suppose Larry does (in Appendix A) *consider* a position, so he doesn’t mean a merely apparent something to be said. Whatever Larry’s own view, my aim was to say something plausible about the better then worse position. And, as Shlomi (fairly) notes, in doing so, I set myself a comparatively low bar! Though I was trying to write in the spirit of Larry, which is to say we try to find all the different angles we might take on a puzzle, and think them through one by one.
The second interpretation of there being ‘nothing to be said’ for the view is that it overall proves false. So if there is something to be said, it has to at least be a contender for the truth.
I want to introduce a third. Larry is very good at separating out our all-things-considered judgements from our egalitarian judgements. A certain point in the sequence may be best from the point of view of equality without being best overall. But Larry also recognises that equality is pluralistic. So a certain point in the sequence might be best from the point of view of AN ASPECT OF EQUALITY without being the best from the point of view of equality as a whole, or best overall. It might be that there is ‘something to be said’ in favour of the better-than-worse view in the sense that it correctly captures an aspect of equality.
I’ll just finish my response to Shlomi by thanking him again. These are just initial thoughts/responses and I will need to think more about his comments.
Hi Victor.
Thanks for this. Is there something good about being equal to others? It seems at least plausible to me. It may, as I have stressed, be a sort of relational egalitarian thought. In order to test the claim, it won’t do to compare inegalitarian relationships with non-egalitarian relationships, because all egalitarians can agree that the former are better than the latter.
So, I think we need to consider different number cases, as Shlomi highlights. Now, there’s obviously a lot going on in different number cases, but we should try to isolate the egalitarian elements. In my case above D1*, D2* etc, it seems to me that there is a way in which D2* is better than D1*. We have introduced some fairness to the world, while before there was only unfairness. Perhaps it isn’t good in and of itself, but offsets the badness of the inequality in D1*. Perhaps our reasons for caring about relations of equality are not egalitarian all the way down but concern the goodness of people’s lives. But maybe that’s ok.
Your comment also made me think about something I was going to put in my original post, but for space reasons didn’t. I have always found Larry’s work on equality/inequality a little difficult on the precise ‘location’ of the badness of inequality. How do we square an approach which asks us isolate and measure the complaints against inequality that the worse off in particular possess versus (or the best off do not possess) with a defence of equality as an impersonal value?
Shlomi: On your comment to Victor about not stealing the spotlight…. I am in favour of an equal distribution of the spotlight!
Obvious mistake in my comment to Victor: ‘it won’t do to compare inegalitarian relationships with non-egalitarian relationships’ — one of those should say egalitarian relationships.
Hi Shlomi,
Happy to qualify the qualified view I outline to people with lives worth living, or people with good lives. It is always better to share the world with more people with good lives. We might also think that it is not worse for me to be worse off than people with lives with negative value. If I’m at -50 there is no bad inequality where another person is created at -10. To put this in emotional terms, their existence is just a cause of regret, not a cause of envy, even though I would rather be in their shoes than mine.
My reason for rejecting the pluralist response is that it suggests that there is something disvaluable about a new person coming into existence who is better off than me, because then I am worse off than them. I don’t feel the force of that at all – it seems a bit dismal to react negatively to the person’s creation, when compared with their non-existence, even in one way, because they are better off than me. That being said, as I suggested at the outset, I don’t have very strong egalitarian convictions, and I’m tempted to think that it doesn’t even diminish the value for me of the creation of a person with a good life that I’m worse off than them.
Finally, just to mention an even weaker egalitarian view – it is always better for me to share a world with a person whose life goes well, and the better that person’s life goes, the better for me. But how much better that is for me depends on whether I am worse off than them. Suppose Y will be created with 100. This is always good for X, and always better for X than Y being created with 90. But how much better it is that Y is created with 100 rather than 90 depends on how well off X is. If X is on 100, this is better for X than if X is on 90. Of all the egalitarian views, this is the one I’m most tempted by, perhaps because I’m not much of an egalitarian!
I take it that this kind of view suggests that there is no egalitarian reason to prefer moving away from D1 along the spectrum. Egalitarianism never provides a reason, even an outweighed reason, for this. I find that attractive.
Hi Patrick, I guess I think this. When we see inequality, we feel bad for the victims of inequality, not bad about the inequality itself.
Now consider: X has 100, Y has 90. Then Y is given 10.
We now feel good for Y, and perhaps good that the inequality has been removed. But we don’t feel good for X that they are now equally placed with Y. It’s Y who has benefited (both in egalitarian and non-egalitarian ways), not X. Or, if we feel good for X, it is just in the fact that they are now not contributing to an injustice or unfairness to Y, not in that there is some good of equality that we have achieved that is good for them. But, again, I just find it hard to believe that egalitarianism is cashed out in impersonal rather than personal terms. So that leads me to think that the disvalue of inequality is located in the disadvantaged person, and equality is good only because it eliminates inequality, and not in itself.
Victor,
Regarding your reply that you are happy to restrict your view to positive life: I’d be apprehensive about taking such a move. As Nils Holtug says on a somewhat similar matter: “it’s like saying the equality is valuable, but only on Wednesdays…”
When we search for an account of the badness of inequality, we should aim to find one that applies always and everywhere. Otherwise, it’s a bit ad hoc.
A brief reply to Shlomi’s first argument against PE.
There’s a continuum of degrees of inequality running down from perfect equality to, in principle, vast inequality. Larry’s view puts the zero-value point on this continuum at the top, i.e. relations of perfect equality have zero value and ones of even the tiniest inequality have negative value. A plausible version of PE won’t say that only relations of perfect equality have positive value; it will move the zero-value point some distance down the continuum, maybe so relations where one person is 10% better off than another have zero value and only ones where someone is more than 10% better off have negative value. Then a relation where one person is 5% better off than another will still positive value with respect to equality, though less such value than if they were exactly equally well off. And Shlomi’s example where some members of an exactly equal population are given tiny additional benefits will still be one with positive egalitarian value, though less than when the equality was perfect. This won’t be because of any notion of “rough equality”; the view as so formulated measures gaps and degrees of equality/inequality exactly as precisely as Larry’s pure NE view does. It just does so after moving the zero-value point on the continuum down from its top to a point somewhat below that.
In reply to both Tom and Patrick’s earlier comment:
Starting with Patrick: you are absolutely right that a plausible version of PE will be a combination of it with NE. I think this is Gustaf’s proposal. So the view is: an outcome’s value in terms of equality is a function of the negative value of inequalities PLUS the positive value of equalities.
Now over to Tom: your suggestion is interesting. But it’s not what’s on offer as far as I know. Put simply, it’s not PE; it’s something else. That of course doesn’t make it wrong. But it is does mean it needs some justification or a new story. And here is the puzzle: if we are already counting inequalities, why add a positive value, if it does not attach to equalities (the way PE suggest)? Isn’t this simply double accounting?
Hi Shlomi
I didn’t think the view was ad hoc. I suggested that the role of inequality is to temper the good for a person of sharing a world with another person. You pointed out that it isn’t always good to share a world with another person – where that person has an overall very bad life. If the significance of inequality is to temper a good, though, it makes a difference only where that good exists. So it doesn’t have significance with very bad lives.
Now, you might think that if X being worse off than Y makes it less good that X shares the world with Y where Ys life is good; then, X being worse off than Y makes it worse that X shares the world with Y where Y’s life is bad.
This is a possible view, and I don’t find it completely implausible. But I don’t think it follows. There are plenty of asymmetries between the good and the bad in personal value (which is what I’m focusing on). Consider agent-regret type cases:
X and Y have involuntary muscle spasms. One of them strikes V killing them.
It is bad for X that X kills V, even though involuntarily. If X doesn’t know who has killed V, they have good reason to hope that it was Y and not them.
Compare
X and Y have an involuntary muscle spasm. One of them prevents V from being killed.
It is not good forX that X has prevented V from being killed. If X doesn’t know who has saved V, they have no good reason to hope that it was them rather than Y.
This is just one instance where the good and the bad are asymmetrical in personal value. There are plenty of others I think. So I’m not too worried about the restriction being ad hoc.
In reply to Shlomi,
I don’t see where there’s double-counting: every relation, of equality or inequality, is given one value. And why assume, if there’s a zero-value point on the equality/inequality continuum, it has to come at the top? For an analogy, in the theory of virtue and vice, at least as I think of them, a very weak appropriately oriented attitude, e.g. a tiny bit of compassion for another person’s horrendous suffering, is evil, not good. That’s because, most plausibly, the zero-value point on the virtue/vice scale isn’t total indifference to suffering — that’s positively bad, and a form of callousness — but some more than tiny compassion.
And re what’s “on offer”: as I told Gustaf (whom I happen to be seeing in Toronto tomorrow) a while ago, the idea that relations of equality should have positive value, and an argument for that based on population cases, was made in a Calgary MA thesis by Risa Kawchuk Dennis McKerlie and I supervised in the 1990s or maybe earlier. So my sense of what the PE view says comes from that and the discussions we had then.
I agree with a lot of what you say here Victor… and it takes us somewhat off topic… but have we NO reason have saved someone’s life, even involuntarily!?
I think none at all Patrick, but I don’t even need that view; at least, there is a much stronger reason to hope that I haven’t involuntarily killed than to hope that I have involuntarily saved.
I still owe Patrick a reply for his rebuttal of my two arguments against PE, and hopefully I can do so in a way that will respond to Tom’s latest reply as well.
My first criticism of PE was the “mere addition of a second of pleasure” (let’s call it). Patrick replies that the objection only holds for the narrow understanding of PE as representing all the value of equality, whereas the truth (as I concede) is that PE is meant as an addition, not a substitute to PE. I don’t think this gets rid of my first objection. Here is why. Suppose we give a score to the blissful perfectly equal scenario (here I begin also my reply to Tom). On the scale of inequality the state of affairs gets the score of Zero, which is the best score you can get for the badness of inequality. On the score of the Goodness of Equality you get a very high score on account of the half million relations of perfect equality. Now you add the tiny benefits: one second of pleasure to the first person, two seconds for the second person, etc. What’s the score now (on the view that combines NE with PE)? On NE the score is no longer zero, but it is also not much worse, since the gaps are tiny. On the PE component, however, there is a huge drop of value. From half a million relations of equality we drop to zero number of relations of equality. The overall score then drops dramatically. And that, I argued, is counterintuitive given how little the welfare distribution changed. The combination of PE with NE does not rescue PE from the first objection.
On to the second objection: Patrick’s first thought is that “maybe the value of equal relationships only ‘kicks in’ once people exist”. This is a legitimate move. If I was a PE advocate I would be reluctant to take it, however. The reason is that this would represent quite the retreat from Patrick’s initial suggestion that we should “maximize relations of equality”. Because if you are only committed to maximising relations of equality contingent on people already existing, you basically piggyback your PE on NE… You want to create relations of equality only when this gets rid of inequality.. There is nothing wrong about that, but then it is questionable that PE is doing any work. More importantly, it would restrict PE to Same Number cases, which contravenes Patrick’s key example against Larry (a Different Numbers case).
But the main reply Patrick provides to my Sadistic Conclusion to PE is that “why can’t the PE simply resort to value pluralism the way NE does” (say, in reply to Levelling Down or Shrinking World). This is fair point (I have discussed this line of thought in writing in the past), and I cannot do it full justice here. I will just give the gist of why I don’t think it works. Consider Shrinking World again. The anti-egalitarian says: “your view is so stupid, it basically says that you give the perfect score, in terms of the badness of inequality, when there is only Robinson Crusoe”. In reply, the NE says: “I never said shrinking the population all the way to one is the best outcome all things considered; I only said there is little, or actually none, of the bad stuff I hate, namely inequality”. So you admit that NE gives us a reason to shrink the population?? Yes, I do. If inequality is all we cared for, and if the only way to rid our society of inequality is to render the population infertile, then yep. Fortunately, I don’t think the badness of inequality is ALL we care about.
OK, replies the PE, but why can’t I also say that the positive value of equality gives us a reason to bring new people into existence, and yes, even when that implies bringing pairs of people with lives that are so wretched they are worth ending? Because, replies the NE, unlike the Robinson case, where there is SOMETHING to be said for shrinking the population (we get rid of inequality!!), in your case, my PE friend, there is NOTHING to recommend this course of action. And the fact that you resort to “the value of equality only kick in when people already exist” attests to that. There IS value to equality, it is value that piggybacks on the absence of inequality…. There is nothing to recommend bringing people into horrible existence just because they would lead it in communality of equal relations.
I know this is tentative, and intuitive-heavy, but there goes..
To Shlomi,
I don’t see how your “reply to Tom” is one. On the version of PE I described a two-person relation of perfect equality has, say, +10 units of value, If one of the two gets a tiny additional pleasure, the value of the resulting relation is, say +9.9, i.e. smaller but still positive. That’s exactly the same change as on a pure NE view, where perfect equality has value 0 and the result of adding the tiny pleasure is, say -0.1. Given that the only difference between the view I described and a pure NE one is where the zero-value-point on a common scale comes, the change in value resulting from any addition or subtraction of pleasure is, on the two views, the same. If it’s small on pure NE, it’s small on the mixed PE/NE view I described.
In reply to Tom:
To truly assess your version of PE we need more information about how it operates. Does it apply to Different Numbers? Does it apply to negative welfare? If so, how does it avoid the Sadistic conclusion.
Also – what’s the justification? Patrick’s (Gustaf’s) maybe ultimately unworkable but at least it has a compelling story: there is something good about being in a relationship of equality. But what’s good about one person being 5% better off than another, as you seem to suggest? Suppose all the men in Sweden get 5% more than every women, in terms of wages. Would anyone put a positive value on this?
Or suppose we have in Canada a population of 30 million living in perfect equality. On your model, creating a thin layer of hundred thousand individuals leading lives that are 5% better is a move for the better. Are you willing to stomach that?
In short, we need to more about your proposal and how it operates in order to be able to evaluate it.
Thanks Shlomi….
This is all fantastic stuff! I really like Tom’s creative proposal, and I really like your objections to it. Oh dear!
I wonder if some view that combines the insights of Tom’s version of PE and your NE is possible. The thing I like about Tom’s view is that the positive value of equality is not all-or-nothing, and that’s where the bite of your first objection comes from. It seems to me that if there is positive value in relations of equality, it is where equality is indeed warranted, or valuable, and there is something good about being treated with equal concern and respect where this is warranted. So, Robinson Crusoe is not treated unfairly, but he is also not treated fairly. The same is true in my D1* above… nobody is being treated fairly, but then two people ARE being treated fairly vis-a-vis one another in D2*. I doubt that the positive value of this requires PRECISE fairness (which is what Tom’s view seems to get right). So could it be that the NE aspect of the view trends toward 0 for perfect equality, but the PE aspect of the view begins to pick up positive value before perfect equality? (I really am just thinking out loud here).
Shlomi, you are of course quite right that I cannot have my population ethics cake and eat it too. I can’t rely on population cases and then restrict my view to same number cases! That was a mistake. But I could restrict, I think in a non ad hoc way, the value of PE to cases in which people have lives worth living.
I also wonder if the right view might be one in which equality is negative but relations of equality offset the badness of relations of inequality. That is a version of what goes on in the move from D1* to D2*. (it is also how Julia Mosquera Ramil tries to account for the value of PE). NE says D1* and D2* are equally bad in inequality terms. I find that counter-intuitive.
Two final points… Shlomi…you have said several times that I say we should maximise relations of equality. I don’t quite say this. I say we should care both about the number and their distribution (we have to care about the number, other no relations of equality is equally valuable to zero). Part of what I think might be attractive about the 500-500 distribution is that everyone enjoys equal numbers of relations of equality.
And you also say ‘this is all very intuition-heavy’. In this kind of debate, I don’t see how we can avoid that.
Thanks Patrick, this is great stuff!
Three quick points:
1. I like the idea of placing a positive value on “being treated fairly”. This gets you into John Broome territory (fairness gets a positive value that is then factored into how good things are for individuals), which I am less an expert on. It also triggers a concern for what Johann Frick called “conditional value”. Compare: Broken promises represent negative value. But do fulfilled promises represent positive value? If they do, that means we should generate more and more promises… but that seems wrong. PE might suffer from a similar problem.
2. In your initial post you wrote (and I quote): “If relations of equality matter, then we may think that we should maximize them.” Is this not your view?
3. When I say “this is all very intuition-heavy” I was referring to myself!… It was not a reflection of your position. So we agree on this score.
Thanks Shlomi.
On 2. To clarify….I did say that, but then I went on to say….’This, however, would not speak in favour of the better-then-worse view of The Sequence, or in favour of D500 being the best distribution, because D1 and D999 both have many relations of equality (999 people enjoying a total of 499500 relations of equality). Moreover, in D999, these relations of equality are all held by the worse off. But these two distributions are distributions in which some people have no relations of equality. Moreover, in distributions close to these extremes, the relations of equality are very unequally distributed (for example, in D2, two people have one relation of equality, while 998 people have 997 relations of equality each).’ I think, if we care about relations of equality, we need to care about the number AND distribution. So all else equal we maximise.
On 3. Yes, I realise you were talking about yourself, but I don’t think it is a negative of or strike against your position or post (as you perhaps implied)… We can come up with various competing models of equality and inequality, but in choosing among them intuitions will surely have an important role to play. I don’t see any other way.
Hi Patrick,
Many thanks for readvertising Larry’s still splendid volume and prompting a thought-provoking discussion. I agree that an appeal to intuition is unavoidable in responding to the Sequence but thought it might be worth paying more attention to intuitions that express indifference to the relative size of the more advantaged and less advantaged group as well as intuitions at a higher level of generality.
As the discussion so far perhaps suggests, there may be some bias against saying ‘Who cares?’ in public philosophical conversations. But when I’ve discussed the Sequence with others in private ‘Meh?’ has been a common reaction. Quite a few philosophers start with what may be the fifth equivalence response, which denies that size variation in the groups of unequally advantaged individuals makes any difference to the value of the various outcomes.
These Sequence Skeptics, as we might label them, often begin by emphasizing that they share many of our egalitarian intuitions in other contexts, and then assuming that, like those intuitions, the relevant intuitions about the Sequence are supposed to concern our reasons for action or aspiration. So, if one partition of individuals into unequally advantageous positions involves less objectionable inequality than another partition, then the partition must, in at least one respect, be more choice or hope worthy than the other. But then the Skeptics insist that the Sequence is such an unusually austere context that those reason-involving intuitions lack their normal justifications and just shrug their shoulders when asked to rank the outcomes.
At this point, the Skeptics say that they are assuming that completely different rigidly designated individuals exist in the various outcomes comprising the Sequence. So, moving from one unequal outcome to another unequal outcome involves replacing some lives with other lives rather than improving anyone’s life. Moreover, they assume, it is consistent with the Sequences that everyone involved leads a sufficiently good life free of various non-comparative harms; perhaps everyone even has a life at least as good for the individual leading it as the life they would have in any other possible world. For those who occupy the disadvantaged position, then, only non-existence can protect them from being less advantaged than others. Moreover, their relative disadvantage is not explained by others’ misconduct and might have no impact whatsoever on their interactions with each other. If any of the thousand participants in the Sequence have a ‘complaint’ against their disadvantage, then it’s against the universe they inhabit and its ‘treatment’ of them rather than other agents. And they may not even be aware of their relative position in the distribution of benefits and burdens, let alone act or feel in ways that involves that position having any impact on their lives.
Having noted that the inequalities present in the Sequence have these features, the Skeptics emphasize how dissimilar those inequalities are to the inequalities disfiguring the real world. Unlike many actual inequalities, their presence is not symptomatic of any moral failure and their reduction need not improve the quality of anyone’s life or influence anyone’s interactions with others. They concede that strictly speaking the Sequence might still exhibit more, or less, objectionable forms of inequality even if the variations are disconnected from the factors present elsewhere: admittedly, they haven’t shown that connection is essential for inequality to matter. But even so the Skeptics maintain their attitude of indifference to the variations in group size present in the Sequence. I suspect they would be similarly indifferent to the promotion or absence of what Patrick, and others, call ‘relations of equality’. When those relations play so little role in enhancing or diminishing individuals’ lives it seems a challenge to convince the Skeptic that they matter. Does anyone know how Larry’s book responds? (I must return to it!) Or has anyone had more success than me in producing arguments to undermine the Skeptics’ indifference?
Hi Andrew
Thank you for this. While it is obvious from the above that I do not have the ‘meh’ reaction, it is interesting that some have this reaction. That said, your interlocutors seem to have put a lot of conditions onto Larry’s Sequence in order to reach their meh reaction. Some of these may have been stipulated by Larry, though I am not sure if all of them are. Moreover, we can remove Larry’s stipulation and think about a nearby puzzle or Sequence that isn’t strictly speaking Larry’s but is a close cousin.
For example, I am not sure that Larry specifies that different individuals exist in each of the distributions. It is true that on Larry’s own telic, impersonal view of equality (assumed, not argued for, in Inequality) that we should be indifferent between whether the people existing in D1-D1000 are the same people or different people (insofar as we rank these outcomes in terms of equality/inequality). But I don’t think that the questions the Sequence poses are only for those that hold this particular understanding of equality.
To find ourselves needing to ask the questions that the Sequence poses, all that has to be the case is that we have some situation in which (a) equality is a relevant value; (b) it is impossible to achieve perfect equality (or perfect equality is clearly undesirable or impermissible); (c) a range of distributions between better off and worse off are available. Which, from the point of view of equality, should we choose. Maybe the distribution is some kind of educational opportunity (which should be equally distributed on Rawls’s fair equality of opportunity for example). Maybe perfect equality is impossible to achieve without abolishing the family, and we have good reason to keep the family. We now want to know, is it better for a small number to have the better opportunity, better for equal numbers to have the better opportunity, or better for a small number to have the lesser opportunity.
Or, we might ask about class inequality: have things improved regarding class inequality when a small percentage of working class people own their own businesses, or become professionals? (I am reminded here of Jerry Cohen’s locked room example, and the class inequality issues he ties it to).
As Larry says in the book, the boxes just represent cases we can describe with words. These abstract examples can be filled in in many different ways. Maybe the way we fill in the example will affect the answer we think is right. But then I think that tells us something interesting.
It can be hard to have intuitions about abstract pictures of boxes. But the Sequence cottons on to real choices, and even if different versions of the Sequence have different answers, we should think about Sequence-like questions, I think.
Lots of interesting stuff here, is there a deadline for this discussion? Cheers G (in Toronto)
Hi Patrick,
Thanks for those very helpful remarks. They made me realise that I should have emphasized that the ‘Meh’ reaction applies to the outcomes in the Sequence and not to Larry’s book. For me, the reaction summed up the fifth equivalence view, which denies those outcomes are better or worse than one another. As such, it’s fully available to people, like you and me, who are convinced that we can learn a lot from Larry’s book and so want to maintain interest in it. To adopt the reaction to the outcomes whilst maintaining our estimate of the book you just have to think that the lessons to be learned are – as so often happens in philosophy – not quite the ones the author anticipated.
For example, as I mentioned, many Sequence Skeptics who affirm equivalence have firm egalitarian convictions in other contexts. They think the Sequence is worth studying for a negative reason: it shows the implausibility of egalitarian theories that overlook factors relevant when explaining why inequality matters. Incidentally, many of these philosophers are thoroughly at home with highly abstract arguments and have firm convictions about how to rank boxes under various descriptions. The worry I raised was that they express confident indifference when I offer what I assumed was the appropriate abstract description of the outcomes, and I fail to undermine their indifference.
Here you rightly mention the possibility of some discrepancy between my description and the one supplied in Larry’s book. I’ll certainly have to investigate that possibility when I return to the text. You also raise the possibility that we could learn something from the Sequence by adjusting parts of my description even if it was accurate, and you mention, more specifically, a revision to my compositional assumptions.
Suppose then that the various outcomes in the Sequence are not composed of different individuals but, as you suggest, the very same ones in each outcome. I agree that this is an interesting variation and in part because the revision is likely to make some of us change our ranking of the outcomes.
In my experience, however, once we stipulate that very same individuals exist in D1 and D2, people then realize the change from D1 to D2 involves (1) loss for the former member of the more advantaged group, who now descends into the less advantaged group, but (2) no gains whatsoever for the individual remaining in the less advantaged group. As a result, it now becomes even more difficult to convince them that equality provides any sound reasons to choose D2 over D1. Instead many egalitarians insist that D1 is clearly better than D2 and that there’s no countervailing reason to decrease inequality by levelling down from D1 to D2. For these levelling up egalitarians, as Paula Casal calls them, reducing inequality, or promoting egalitarian relations, is only sometimes rather than always choice worthy; it ceases to be choice worthy when it involves only losses for individuals, and no gains for anyone, or inefficiency in the economists’ sense. Some, who assume egalitarianism expresses equal concern for each individuals’ life and that such concern counts against inefficiency, may even argue that equality favours D1 over D2.
To conclude, I worry then that you’re more likely to provoke opposition to your own preference for the ‘first better, then worse’ view of the Sequence if you revise the description of the outcome in the way you mention. For people like me, who affirm Pareto in full compliance theory and axiology, the first view then becomes more compelling and the Sequence is downhill all the way. Of course, there’s no more ingeniously effective advocate of levelling down egalitarianism than Larry, and so perhaps my opposition to inefficient equalising is mistaken. But I suspect Larry may have avoided describing the Sequence in the way you mention in order to make it less difficult to say something in favour of the move from D1 to D2 and to reduce his discussion of the Sequence being more reliant on the defence of levelling down supplied later in his book.
Many thanks again – I look forward to further discussion in person, in Warwick or Barcelona!