One proposal that easily comes to mind when faced with the RepugnantConclusion is to reject total utilitarianism in favor of a principleprescribing that the average well-being per life in apopulation is maximized. Average utilitarianism and totalutilitarianism are extensionally equivalent in the sense that theygive the same moral ranking in cases where the compared populationsare of the same size. They may differ significantly indifferent-number cases, however. This is easily seen when the averageapproach is adopted in the comparison of the population outcomes inFigure 1 and Figure 2. That the average well-being is all that mattersimplies that no loss in average well-being can be compensated for by again in total well-being. Thus in Figure 1, A is preferable to Z,i.e. the Repugnant Conclusion is avoided. In Figure 2, A is preferableto A+, i.e. the Mere Addition Paradox is blocked at the very firststep.
Despite these advantages, average utilitarianism has not obtained muchacceptance in the philosophical literature. This is due to the factthat the principle has implications generally regarded as highlycounterintuitive. For instance, the principle implies that for anypopulation consisting of very good lives there is a better populationconsisting of just one person leading a life at a slightly higherlevel of well-being (Parfit 1984 chapter 19). More dramatically, theprinciple also implies that for a population consisting of just oneperson leading a life at a very negative level of well-being, e.g., alife of constant torture, there is another population which is bettereven though it contains millions of lives at just a slightly lessnegative level of well-being (Parfit 1984). That total well-beingshould not matter when we are considering lives worth ending is hardto accept. Moreover, average utilitarianism has implications verysimilar to the Repugnant Conclusion (see Sikora 1975; Anglin1977).
Another way of modifying total utilitarianism is to introduce acritical level (Kavka 1982, Parfit 1984; Locke 1984; Blackorby etal. 1997, 2004; Broome 2004). As is the case with the variablevalue view, a critical level view is based on a split betweenprudential and moral value. However, in this case the split is evenmore radical. The idea is that a person's life only contributespositively to an outcome if the quality of the person's life is abovea positive critical level. Below this level the moral value of a lifeis neutral (Kavka 1982) or negative (Blackorby et al. 1997,2004) even though the person leads a life worth living. In itssimplest form, the critical level view is a modified version of totalutilitarianism. The contributive value of a person's life is herwelfare minus a positive critical level. The value of a population iscalculated by summing these differences for all individuals in thepopulation. The critical level view thereby succeeds in avoiding theRepugnant Conclusion. Assuming the critical level is higher than thelow quality of the lives led in population Z, the Repugnant Conclusionis blocked: Z now constitutes an outcome (much) worse than A.
An attempt to produce a compromise between a total principle and anaverage principle is provided by a variable value principle. The ideabehind this sort of view is that the value of adding worthwhile livesto a population varies with the number of already existing lives insuch a way that it has more value when the number of these lives issmall than when it is large. Taking this view, Noah's obligation toprocreate after the Deluge was much stronger than our obligationtoday. Variable value principles are sometimes called“compromise theories” since they can be said to be acompromise between total and average utilitarianism. With smallpopulations enjoying high welfare, a variable value principle behaveslike total utilitarianism and assigns most of the value to the totalsum of welfare. For large populations with low welfare, the principlemimics average utilitarianism and assigns most of the value to averagewelfare. This property is what makes it possible to avoid theRepugnant Conclusion. If the value of extra lives decreasesasymptotically, then there exists an upper limit to the total value ofa population (exactly as the sum of the infinite series 1+1/2+1/4+… has the upper limit 2). As long as population A in Fig. 1 has atotal value above this limit this population will be preferable nomatter how populous the low quality population Z is (Hurka 1983, Ng1989; Sider 1991).
The Repugnant Conclusion may seem to rely on a certain view aboutwelfare. It treats welfare on a single additive scale, where lownumbers, if added to themselves often enough, must become larger thanany initial larger number. This is like the Archimedean property ofthe real numbers: For any positive numbers x and y,there is a natural number n such that nx is greaterthan y. Any finite number of lives in population A cantherefore be outweighed by a sufficiently large number of lives in Zbecause the gain in the quantity of lower values outweighs the loss ofcertain higher values. However, some theorists – includingParfit – have suggested that this axiological assumption ismistaken. (Parfit 1986; Griffin 1986; Rachels 2001; Crisp 1988, 1992;Glover 1977; Edwards 1979; Lemos 1993; Skorupski 1999; Klint Jensen1996; Portmore 1999; Riley 1993, 1999). The idea is that one type ofgood A can be superior to another type of good B, in the sense thatany amount of A is better than any amount of B or that some amount ofA is better than any amount of B. Suppose that what happens as we movedown the alphabet from the high-quality population A to thelow-quality population Z is that the best things in life are graduallylost. For instance, as Parfit has suggested, the first step from A toB involves the loss of Mozart's music; in the move from B to C Haydn'smusic is lost; in the move to D Venice is destroyed; and so on downthe alphabet. All that is left in the final move to Z is “muzakand potatoes”. The claim is that the lives in the beginning ofthe sequence involve goods that are superior to the goods involved inthe lives at the end of the sequence. The loss of the most worthwhilethings in life cannot be compensated for by any gain in the quantityof muzak and potatoes. Consequently, whatever the number of people inpopulation Z, there will be less well-being in this world as comparedto population A and thus the Repugnant Conclusion is blocked.
It has been pointed out, however, that Feldman's reasoning involves aquestionable interpretation of the ceteris paribus clause inthe repugnant conclusion (Arrhenius 2003b). He implicitly assumes thatthe ceteris paribus clause is satisfied whenever the peoplein the compared populations have the same desert level. A moreplausible reading of the ceteris paribus clause is that it issatisfied if and only if the compared populations are (roughly)equally good in regard to other axiologically relevant aspects apartfrom welfare. Consequently, what we are looking for are cases wherethe compared populations are, in some sense, equally good in regard todesert. Given this interpretation of the ceteris paribusclause, justicism does imply the repugnant conclusion. If A and Z areequally good in regard to desert, then justicism ranks Z as betterthan A since Z has greater total well-being.
The above discussion of the Repugnant Conclusion has presumed thatthere are possible lives with a very high quality of life. If thereare no such lives, then the Repugnant Conclusion would be neutralisedby making it an empty truth. If there are no possible lives with avery high quality of life, then the Repugnant Conclusion is vacuouslytrue, since the antecedent – “[f]or any possiblepopulation … with a very high quality of life” – isfalse of every possible population. Consequently, if there are nopossible lives with very high welfare, then total utilitarianism wouldonly imply the Repugnant Conclusion in a trivial and uninterestingsense. Christoph Fehige (1998) has suggested, in the tradition ofSchopenhauer, that there are no lives with a positive well-being. Inhis account, only frustrated preferences count, and they countnegatively, whereas satisfaction of preferences has no positivevalue. Since it is possible that a population with very high welfareonly involves lives with complete preference satisfaction, theRepugnant Conclusion is avoided since any population with completepreference satisfaction is better than any population involving atleast one frustrated preference. More troubling, however, is that thisview implies a strong version of the Reversed Repugnant Conclusion: Apopulation with very high positive welfare can be worse than an emptypopulation. Since most lives with very high welfare can be assumed tohave at least one frustrated preference, such lives are worse thannon-existence according to Fehige's theory. Moreover, a theory aboutwelfare that denies the possibility of lives worth living is quitecounter-intuitive (Ryberg 1996a). It implies, for example, that a lifeof one year with complete preference satisfaction has the same welfareas a completely fulfilled life of a hundred years, and has higherwelfare than a life of a hundred years with all preferences but onesatisfied (Arrhenius 2000b).
Several philosophers have taken up the challenge from Parfit and haveengaged in considerations as to what would constitute a satisfactorypopulation ethics. Some theorists have done so by accepting theframework Parfit has set for the investigation; that is, they havecarried on the work of developing an ethical theory that succeeds inmeeting Parfit's minimal requirements for adequacy. Other theoristshave been less willing to accept the scenery as set by Parfit and havemore or less radically challenged his requirements for a plausibletheory. The Repugnant Conclusion has been placed at the centre of thisdiscussion. The suggestions in the literature on how to deal with theRepugnant Conclusion can roughly be divided into eight categories: (1)introducing new ways of aggregating welfare into a measure of value;(2) questioning the way we can compare and measure welfare; (3)counting welfare differently depending on temporal or modal features;(4) revising the notion of a life worth living; (5) rejectingtransitivity; (6) appeal to other values; (7) accepting theimpossibility of a satisfactory population ethics; (8) accepting theRepugnant Conclusion.
It might be tempting for people who have little sympathy withutilitarian thought to try to set the problems raised by the RepugnantConclusion to one side, thinking that it constitutes a problem onlyfor utilitarians. However, most people tend to believe that we havesome obligation to make the world a better place, at least ifwe can do so without violating any deontological constraints, and at anot too high cost to ourselves. Clearly all who think along theselines, even without being utilitarians, are faced with the problem ofthe Repugnant Conclusion. We can assume that other values andconsiderations are not decisive for the choice between populations Aand Z in Fig. 1 (e.g., promises, rights). The Repugnant Conclusion isa problem for all moral theories which holds that welfare at leastmatters when all other things are equal.
Leaving the mentioned Non-identity case aside, there are otherarguments establishing that the Repugnant Conclusion is not easilyavoided. Parfit has developed an argument to this effect. Consider thethree population scenarios indicated in Fig. 2.