EGALITARIAN JUSTICE AND CHOSEN BRUTE TASTE* Knight, C. (2009) Egalitarian justice and valuational judgment. Journal of Moral Philosophy, 6 (4). pp. 482-498. ISSN 1740-4681 http://eprints.gla.ac.uk/49956/ Deposited on: 22 March 2011 Enlighten – Research publications by members of the University of Glasgow http://eprints.gla.ac.uk http://eprints.gla.ac.uk/view/author/3829.html http://eprints.gla.ac.uk/view/journal_volume/Journal_of_Moral_Philosophy.html http://eprints.gla.ac.uk/view/journal_volume/Journal_of_Moral_Philosophy.html Egalitarian Justice and Valuational Judgment CARL KNIGHT* Adam Smith Research Foundation and Department of Politics University of Glasgow Glasgow, UK Contemporary discussions of egalitarian justice have often focused on the issue of expensive taste. G. A. Cohen has recently abandoned the view that all chosen disadvantages are non-compensable, now maintaining that chosen expensive judgmental tastes – those endorsed by valuational judgment – are compensable as it is unreasonable to expect persons not to develop them. But chosen expensive brute taste – the main type of non-compensable expensive taste on the new scheme – cannot be described in such a way that there is a normative difference between it and chosen expensive judgmental taste. As there are related problems with denying compensation for the other kind of expensive taste that might remain non- compensable, Cohen’s position on taste appears to be either implausible or virtually indistinguishable from that of equality of welfare. However, compensation for valuational judgment-based expensive taste might be justified on grounds of responsibility. Keywords: Choice, G. A. Cohen, Equality, Justice, Responsibility, Expensive Taste. I. Introduction Consider the following scenario: 1. (a) Steve holds £10. (b) Harry holds £10. 2. (a) Steve’s taste is such that he would obtain 1 unit of welfare (hereafter: unit) from consuming a hamburger. (b) Steve’s taste is such that he would obtain 10 units from consuming a steak. (c) Harry’s taste is such that he would obtain 10 units from consuming a hamburger. d) Harry’s taste is such that he would obtain 1 unit from consuming a steak. e) No change in the tastes of Steve or Harry is now possible. 3. (a) A hamburger now costs £2. (b) A steak now costs £10. 4. Steve would now prefer to have Harry’s taste. 5. (a) Prior to t (=a point in the past) Steve’s taste was that of Harry’s. (b) Harry has not deliberately cultivated his taste, nor could he ever have eliminated it. (c) Steve deliberately cultivated his present taste at t. Which of these facts is relevant to egalitarian distributive justice? A wealth egalitarian replies that 1 alone is relevant, and that the existing distribution is perfectly just. A welfare egalitarian begs to differ: on her account, 2 and 3 also come into play. She views the distribution as unjust, on the grounds that Steve is only able to secure 10 units with his funds, whereas Harry is able to secure 50 units with his. Steve has expensive taste, in the sense that he is only able to secure below average welfare with average income. The welfare egalitarian views Steve’s expensive taste as a disadvantage that entitles him to compensation. In his highly influential article, ‘What is Equality?’, Ronald Dworkin criticized the welfare egalitarian’s stance on ‘champagne tastes’ as counterintuitive.[1] His favoured theory of equality of resources did not, however, agree with equality of wealth’s stance that 1 is the only relevant fact. Dworkin suggested that 4 is also relevant, for on his scheme those with dispreferred tastes (or ‘cravings’) are entitled to compensation. Another position was proposed in response to Dworkin in G. A. Cohen’s well-known paper ‘On the Currency of Egalitarian Justice’.[2] Cohen urged that compensation for expensive tastes did not appear counterintuitive where those tastes were unchosen; on the contrary, such compensation was mandated by justice. On this account, which may be referred to as equality of opportunity for welfare,[3] 1, 2, 3 and 5 all appear to be relevant. Cohen has recently modified his account, suggesting that some tastes that are chosen are nevertheless compensable.[4] The present article argues that the position that he now wants to occupy either makes distributive decisions on morally arbitrary grounds or comes close to collapsing into equality of welfare. Nevertheless, he does highlight a possible source of injustice – harsh treatment of valuational judgment-based expensive taste – which the equality debate has thus far overlooked.[5] This points to the need for renewed focus on the relationship between principles of distributive justice and the content of responsibility. The paper starts in earnest in section II, which clarifies Cohen’s position on responsibility. Cohen is now clear that responsibility for possessing one’s taste is not itself important; it is rather responsibility for the expense of one’s taste that matters. Section III introduces Cohen’s new distinction between judgmental and brute taste. Here Cohen’s additional condition for compensation for a taste – that it is based upon valuational judgment – is discussed, and the most plausible interpretation of this condition is put forward. In sections IV and V the difficulty of finding cases in which Cohen may reasonably refuse compensation is demonstrated. This difficulty is due to the fact that, on the two available interpretations of Cohen’s account, the non- compensable expensive taste – chosen brute taste – cannot be formulated in such a way that it is morally distinguishable from (compensable) chosen judgmental taste. In section VI one special type of case in which Cohen might refuse to provide compensation is described, and a familiar kind of problem with such a refusal is identified. This puts Cohen’s view uncomfortably close to equality of welfare. Section VII notes that one might bite the bullet at this point, but only at the cost of facing Dworkin’s expensive tastes objection. It is argued in section VIII that a more promising strategy is to account for valuational judgment by construing it as an obstacle to full responsibility. We need not assume that valuational judgment is such an obstacle, but only that it might be. II. Responsibility and Expense In ‘Expensive Taste Rides Again’, Cohen states that he wishes to revise the following first sentence of the ‘flagship statement’ of his earlier essay: ‘I distinguish among expensive tastes according to whether or not their bearer can reasonably be held responsible for them’.[6] Cohen now says that this sentence (together with similar statements in ‘Currency’) is poorly formulated. A suitable reformulation would replace ‘them’ with ‘the fact that her tastes are expensive’. He adds the following: ‘It is, as I made it abundantly clear elsewhere in “Currency,” precisely that fact for which the question of responsibility is crucial’.[7] Without this clarification the first sentence of the flagship statement is at best ambiguous and at worst morally implausible. The original formulation may be read as suggesting that justice distinguishes between responsibility and lack of responsibility for the existence of tastes that now happen to be expensive. Persons are entitled to compensation if they have what Dworkin calls bad preference luck, ‘which is their bad luck in having the preferences that they do because these preferences are expensive’. [8] On this account, the bare fact that Steve chose to generate his taste (fact 5) disqualifies his claim for compensation. But suppose that Steve’s taste was not expensive when he cultivated his taste, and that he could not have predicted that it would become so in the future. He has been unlucky, not in having the taste that he does, but in that his taste has become expensive. He has an unchosen disadvantage (a taste for an expensive food) which would be overlooked by one concerned only with bad preference luck. This possibility suggests that responsibility for holding the taste does not matter in the absence of responsibility for its expensiveness. For this reason Cohen now makes it clear that it is responsibility for the expensiveness of one’s taste that matters.[9] People are entitled to compensation on this account where they have what Dworkin describes as bad price luck, ‘which is bad luck in the high cost of the preferences they have’.[10] This would be a relevant set of facts on such an account: 6. (a) At t a steak cost £2. (b) At t a hamburger cost £2. (c) Steve could not reasonably have been expected to anticipate the rise in price of steaks from t to the present. Cohen believes that it would be unfair to penalize Steve for the unforeseen economic changes that have replaced the favourable circumstances depicted in 6 with the unfavourable circumstances depicted in 3. No transfer would be due if Steve knew such changes were likely to occur but gambled that they would not, or hoped that they would, maybe because he views expense as itself desirable.[11] But this is not the case here. As it is, Steve suffers from a welfare deficit for which he is not relevantly responsible; such deficits are compensable on Cohen’s scheme, and the first element of the reformulation of his flagship statement reflects that. III. Brute Taste and Judgmental Taste The second sentence of Cohen’s original flagship statement reads thus: There are those [tastes] which he [an individual] could not have helped forming and/or could not now unform, and then there are those for which, by contrast, he can be held responsible, because he could have forestalled them and/or because he could now unlearn them.[12] Cohen believes that the failings of this second sentence are more substantial than those of the first. The problem is not merely that his position is poorly expressed: the position itself is at fault. While Cohen still believes that some chosen tastes are non-compensable, he now maintains that others do give rise to valid claims for compensation. The former category are brute tastes, ‘which do not embody judgments of valuation’.[13] Cohen offers the example of ‘my own liking for Diet Coke, which embodies no particular approval of it’.[14] Judgmental tastes, by contrast, ‘are informed by valuational judgment’. Cohen explains his change of stance as follows: I no longer think that the mere fact that people chose to develop and/or could now school themselves out of an expensive judgmental taste means that they should pick up the tab for it, and that is precisely because they did and do identify with it, and therefore cannot reasonably be expected to have not developed it or to rid themselves of it.[15] The moral idea driving the distributive proposal is that some persons are disadvantaged by their identification with certain tastes which turn out to be expensive, since that identification leads them to develop those tastes. Cohen observes that this takes his position even further away from Dworkin’s: So what Dworkin gives as a reason for withholding compensation – the subjects’ approving identification with their expensive tastes – is something that I regard as a reason for offering it, since, where identification is present, it is, standardly, the agents’ very bad luck that a preference with which they strongly identify happens to be expensive, and to expect them to forego or to restrict satisfaction of that preference (because it is expensive) is, therefore, to ask them to accept an alienation from what is deep in them.[16] Cohen’s move should not exactly be to turn Dworkin on his head. One may recall that Dworkin is interested in the presence or absence of identification with taste right now: hence, on his account, fact 4 in the hamburger-and-steak case mentioned at the outset justifies compensation for Steve (given fact 1). Here Cohen, like Dworkin, focuses on present (dis)identification: unfair disadvantage arises ‘where identification is present’. However, we have seen that elsewhere in the same paper he focuses on the history of identification as well: persons who did identify or who do identify with a taste ‘… cannot reasonably be expected to have not developed it or to rid themselves of it’. In other words, present (dis)identification matters for establishing whether it is reasonable to expect an expensive taste to be unlearned, while past (dis)identification matters for the key issue of establishing whether it is reasonable to hold someone accountable for their choice to develop a taste. So why should Cohen maintain that history of identification matters? An answer becomes clear when we consider the case of someone who initially had no identification with an expensive taste, but nevertheless developed it, and subsequently came to identify with it. Cohen has no reason for thinking that it would be unreasonable to ask such a person to bear the cost of their taste whether they can now be rid of it or not, for that cost has arisen directly from a choice that she was not predisposed to make. It is a history of identification that excuses an individual from bearing the cost of their expensive taste; present identification only tells us that an individual who has been so excused cannot be expected to dispose of their taste. IV. Chosen Expensive Brute Taste The discussion of the previous section indicates that the following would be an example of an additional relevant fact on Cohen’s scheme: 7. Prior to and at t, Steve disidentified with the taste he then had and identified with the taste he now has. This embodies my stipulation, given as an interpretation of Cohen, that the (dis)identification must not come after the possession of the taste – it must be what we might call pre(dis)identification. According to Cohen, the presence of this fact diminishes Steve’s liability for his present expensive taste. It shows that, even if Steve is responsible for the fact that the taste he holds is expensive, it would be unreasonable to make him bear its cost.[17] The implications of this revised account are no simple matter. To see them one must first try and establish what exactly a chosen expensive brute taste might look like. This is because they are (with one important exception to be mentioned in section VI) the only kind of expensive taste that Cohen now refuses to provide compensation for, and as such they appear to be the only source of disagreement between Cohen and equality of welfare. I will maintain that, from what little Cohen says about chosen brute taste, it is not clear that they can exist in a form that is differentiated from chosen judgmental taste by any morally significant characteristic. On one of the two available interpretations, Cohen’s account amounts to no more or less than equality of welfare,[18] insofar as it is concerned with welfare;[19] on the other, it amounts to something with (even) less justification. The feature that makes a taste brute in Cohen’s sense is that it ‘does not embody judgments of valuation’. For reasons that should be obvious from the earlier discussion, the kind of judgments of valuation that they do not embody are those that are temporally prior to the taste itself. Steve’s taste for steak does not go from being judgmental to being brute simply because he stops identifying with it some time after he has acquired it, for Cohen’s question of whether Steve could reasonably have been expected to turn down the opportunity to cultivate that taste, given the cost that would impose upon him, can obviously only apply before taste acquisition. The kind of taste we are looking for is one that is chosen and initially not supported by value judgments. How might the origins of Cohen’s own taste for Diet Coke meet both of these criteria? Were Jerry (as I will call this character) brought up to have this taste or hypnotized into having it, it would fail to meet the criterion of being chosen.[20] This would also be the case if his first sip had him hooked, and that sip resulted from, say, a bartending error or lack of alternative beverages. Contrariwise, if Jerry made a deliberate attempt to cultivate that taste it would fail to meet the non-judgmental criterion. That attempt must be based on some kind of value judgment, however sensible or fatuous it may be, for otherwise his behaviour is inexplicable. There are further possibilities that may appear to salvage chosen brute taste. First, suppose that Jerry makes no deliberate attempt to cultivate a taste for Diet Coke and has no preidentification with it, but that he tries it anyway. Perhaps he just wants to see what it tastes like. Or maybe he has decided to live his life as a ‘free spirit’, which right now means trying some Diet Coke for no particular reason. Either way, Jerry’s first sip results in taste formation. For the moment, let us grant that the taste that is formed in a case like this is brute. Although preidentification with taste is absent, other kinds of relevant value judgments are present. It seems to me that Jerry could only have not taken his sip by acting in a fashion that contradicted his value judgment that he ought to find out what Diet Coke tastes like, or that he ought to drink it just because that’s what he wants to do. Such contradiction is not directly related to taste. It is, however, obviously an incident of self-infliction of cost. Were it identical in every regard except the structure of its relation to taste, Cohen would say that it could not reasonably be expected of Jerry. If this kind of self-infliction is unreasonable where the judgment endorses the taste I do not see any reason why it is less unreasonable where the judgment endorses something else that can only be realized by performing an action that may or will result in taste formation. There is a slight causal difference but no moral difference: in both types of case we require individuals ‘to accept an alienation from what is deep in them’ if they are not to suffer a welfare deficit. On this natural reading of Cohen, he appears to base compensation on a morally arbitrary distinction. V. An Alternative Interpretation Suppose, then, that we take a more liberal reading. Here we describe a taste such as Jerry’s – one that is not preidentified with, but that could only have been avoided at the cost of violating otherwise independent value judgments – as an indirect judgmental taste. The resulting broad sense of judgmental taste, encompassing both indirect judgmental taste and direct judgmental taste (instances of which are preidentified with), is somewhat strained as a construal of Cohen, given that it includes ‘taste that does not track a judgment of the value of its object’.[21] But it – and the corresponding narrow construal of brute taste – sits far more comfortably with Cohen’s moral argument that it is unreasonable to expect persons to contradict their value judgments. In this case it looks like there is nothing morally arbitrary about compensating for chosen judgmental taste but not for chosen brute taste. But the revised position also radically limits the occurrence of chosen brute taste (that is, a chosen taste that could have been avoided at no self-inflicted cost). Usually a taste must be either unchosen or endorsed, somewhere along the line, by valuational judgment; in the odd case where it is both unendorsed and, in some sense, chosen, compensation is still appropriate. This is shown in two more kinds of cases. We might suppose that Jerry has no preidentification with Diet Coke but that he is a ‘diceman’ – one who makes his choices in a random fashion – and, when ordering his drinks, Diet Coke’s number comes up. Or we might suppose that, while Jerry’s action might look like it could be explained as either experimental or wilfully random, he is simply acting on impulse. In each case we again suppose that the first sip results in taste formation. When presented with cases like Diceman Jerry it is tempting to think that chosen narrow brute taste might ride in on the back of uncertainty about outcomes. The unexpected results of random but deliberate action may appear to be chosen without there having been any question of identification. It is therefore important to note that the actual likelihood of the taste arising from the action, and the individual’s perception of this likelihood, are quite irrelevant to the question of whether the taste is (indirectly) judgmental. All that matters is that the individual could not have performed otherwise without contradicting his own judgment. Even if the individual thought there was no possibility of taste formation, and it was in fact a million to one shot, that taste would be fully judgmental on account of the clear causal link between it, the action, and the endorsement of the action. Both this point and the more general point about the infrequency with which chosen narrow brute taste can occur is illustrated well in the Diceman Jerry case. Diceman Jerry’s taste is, I will allow, chosen and not itself preidentified with. But the endorsement of running one’s life on the roll of a dice is clearly a value judgment. That preidentification with randomizing is, in this instance, bound up with the taste. Even though the link between Diceman Jerry’s taste and his judgment is indirect and fortuitous, his taste is judgmental, for he could have declined the option of drinking Diet Coke only by alienating himself from his own judgment. This causal relationship is all that is required to show that the taste is not brute in the narrow sense. The final, impulsive case faces the initial worry that no taste formed in this way could be chosen. But I am willing to grant, for the sake of argument, that choice may be present here. This issue aside, impulsive action appears to present the best opportunity for non-compensable chosen brute taste to arise. Even so, Jerry’s behaviour is presumably explainable in one of two ways, and neither of these will do. Firstly, there may be a value judgment even further back in the causal chain than in the earlier two cases. This judgment may direct Jerry to act spontaneously and without regard to consequences. This case is not relevantly different from that of the self-defined free spirit. As before, the taste can be treated as non-compensable only if we demand that Jerry alienate himself from his judgment. On the alternative reading of Cohen, the taste is, again, not brute. Secondly, and more interestingly, Jerry may be following no value judgment whatsoever. He may by nature be an extremely impetuous person. It may have never crossed his mind that acting in a more considered way is a viable option. This is the one kind of case where the more liberal reading of Cohen can, like the more straightforward construal, identify chosen brute (i.e. unendorsed) taste. But the common denial of compensation in both instances is morally arbitrary. If it is unreasonable to expect someone to act against their valuational judgment, it is, if anything, more unreasonable to expect someone to act against their natural temperament, where no reasons for such behavioural modification have even been entertained. The complete absence of valuational judgment makes it inappropriate to punish (or reward) in this case as one might were a full moral agent involved. Even if we allow that there is choice, there is not responsibility.[22] VI. Costlessly Removable Expensive Taste The basic problem we have so far encountered is that one can generally choose to undergo a course of action that may inadvertently result in the acquisition of a taste only where one holds value judgments that recommend that course of action. A taste of this kind is an indirect judgmental taste. One can only avoid this conclusion by failing to acknowledge the cost involved in acting contrary to value judgments that are only indirectly linked to the acquisition of the taste (for example, ones recommending ‘taste testing’ or spontaneity). This is precisely what Cohen does on the more natural reading mentioned above. But I can see no moral justification for such a position. Indeed, it seems to run counter to the unreasonableness-of-expecting-self-imposed-costs justification that Cohen attempts to support his stance with. If, on the other hand, we work with the more liberally interpreted senses of brute and judgmental taste, chosen brute taste can only be held by the very impulsive. So interpreted, Cohen’s position is either, again, morally arbitrary (if we penalize purely impulsive behaviour) or virtually normatively identical with equality of welfare. In the latter case, it makes conceptual distinctions (between brute and judgmental taste, and between chosen and unchosen taste) that equality of welfare does not, but the only normatively relevant combination is all but impossible since non-purely impulsive chosen brute tastes are reclassified as indirect judgmental tastes. I say ‘all but impossible’ as there is one special case where it may appear that Cohen could still reasonably deny compensation for expensive taste. This is the case of costlessly removable expensive taste. Whether it was initially chosen or not, Cohen denies compensation for expensive tastes that can be removed at no expense – which is to say not only that the tastes are not presently identified with, but also that there are no ‘incidental’ costs that are unrelated to identification. Equality of welfare would even compensate in the case where an individual with such a taste obstinately refused to remove it. The absence of present identification marks these tastes as brute in the sense that was rejected in section III above. The absence of incidental costs is much the more demanding of the two requirements mentioned in the previous paragraph. Even where a taste is disidentified with, there will almost always be some financial expenditure, lost time or other inconvenience involved in schooling oneself out of it that results in a welfare loss. As costlessly removable expensive tastes are little more than a possibility, the space between the broad judgmental taste view and equality of welfare is barely a sliver. Yet the rarity of costlessly removable expensive tastes is not the most substantial bar to their serving as an effective prop for Cohen’s position. The refusal to compensate for these tastes is morally suspect for reasons that are very much parallel to those that applied to the natural reading of Cohen (where judgmental taste was narrowly construed). A deliberate failure to remove a taste, even where there is no cost involved, presumably itself involves valuational judgment. We might again ask why it is unreasonable to ask people to act against their value judgments in some cases, but not in others. Someone endorsing compensation for judgmental taste in either the narrow sense (admitting direct judgmental tastes only) or broad sense (admitting both direct and indirect judgmental tastes) faces an unpalatable choice. They could treat the value judgments that oppose costless removal of expensive tastes as the basis for compensation for those tastes (either because the judgments are necessarily tied up with judgmental taste or for independent reasons). Or they could refuse compensation, in which case their moral consistency is in doubt, given the original justification for compensation for expensive judgmental tastes. VII. Biting the Bullet The argument that expensive judgmental taste should not be penalized has some intuitive support. But consistent endorsement of what appears to be its key premise – that it is wrong to penalize persons for choices that are required by their valuational judgment – commits Cohen to more than he acknowledges. For the premise to be consistently endorsed, both the broad construal of judgmental taste and compensation for costlessly removable expensive tastes that are not removed is required. This treatment of tastes is identical to that of equality of welfare, even if the justification is different.[23] It appears that Cohen is still determined to put distance between his account and equality of welfare.[24] If the arguments presented above have helped to clarify the size of that distance and, in particular, the normative difficulties that are involved in retaining it, they would for those reasons alone have some significance. But what if Cohen, or someone endorsing similar premises, were to bite the bullet, and accept that something very close to equality of welfare is the correct account of egalitarian justice after all? This move faces the arguments against equality of welfare presented by Dworkin, which could be applied to the new position with minimal revision. The objection to social allowance for expensive tastes is of course foremost amongst these. One question, then, is whether Cohen’s focus on the (alleged) wrongness of penalizing acts required by valuational judgment suggests a new way around the objection. It seems quite clear that it does not. In Dworkin’s original presentation, ‘someone (Louis) sets out deliberately to cultivate some taste or ambition he does not now have’, for the specific purpose of ‘trying to make his life a better life in some way’.[25] So it seems reasonable to assume that, from the outset, Louis is characterized as acting on valuational judgment. But Dworkin thinks that of no help to his cause: This does not make his claim for extra resources any more appealing or less counter- intuitive … . On the contrary, the fact that he is acting so deliberately in his own interests seems to make his claim, if anything, less appealing than the claim of someone who tries an expensive experience on a whim, for the pleasure of the moment, and then finds that he is hooked.[26] Whether one agrees with Dworkin’s position or not is unlikely to be affected by Cohen’s change of emphasis. As Cohen acknowledges, Dworkin always accounted for valuational judgment – he just did not think it mattered. Takers of the bullet-biting strategy are squarely in Dworkin’s sights. VIII. Valuational Judgment as Responsibility-Diminishing It seems to me that there may be a less hazardous way of endorsing the premise that it is wrong for persons to be disadvantaged by valuational judgment-endorsed acts. On one prominent construal of equality of opportunity for welfare, which Cohen has consistently endorsed, a disadvantage such as an expensive taste is non-compensable only where the holder has genuinely chosen it – that is, where she is responsible for it in a deep, metaphysically valid way. The appropriate distribution is dependent on both which account of free will and determinism is correct and facts particular to each case.[27] The presence or absence of preidentification, or of other, less direct value judgments that have resulted in taste formation (such as those of Dicemen), are candidates for relevant particular facts. It might be thought that, although a history of valuational judgment is not grounds for assuming full non-responsibility, it may diminish an individual’s responsibility for developing their taste. As the strength of the value judgments – and the degrees of difficulty and pain involved in acting contrary to them – increase, so responsibility decreases. We might, then, hold that Cohen’s discomfort with the position he took in ‘Currency’ is based on a failure to appreciate the full range of considerations that responsibility may take into account. The complaint is really with certain prevalent conceptions of responsibility and related concepts, such as choice, which fail to recognize the significant role of valuational judgment. That a person is acting on valuational judgment is not typically taken to be a factor in favour of their case for compensation – Dworkin for one thinks quite the opposite. Once the argument shifts to this level there appears to be little problem with the position Cohen took in ‘Currency’. One gets the outcome of compensating for valuational judgment-based tastes by combining conventional equality of opportunity for welfare with a particular account of responsibility. Now it might be asked why this strategy should be preferred to the bullet-biting one. The answer is twofold. First is the way it accounts for the moral significance of variations in the ease with which preidentification and other relevant judgments can be contradicted. In some cases, an individual may only act against their valuational judgment at cost of considerable self- contradiction, but in others, while there is some cost involved, it is relatively minor. An agent who acts against their preidentification with an expensive taste for the theatre might consequently experience great inner-anguish or hardly anything at all – it just depends on the character of the preidentification. In light of this, the strategy suggested by the alternative reading of Cohen, under which full compensation is to be provided in each case of valuational judgment-based taste, is insufficiently discriminating. A possible retort here is this: why not just adopt a further interpretation – or really, by this point, reinterpretation – of Cohen, such that his position admits of degrees. The more promising version of this approach works with the alternative reading of Cohen’s position, avoiding the moral arbitrariness of the natural interpretation, but placing all tastes somewhere on a continuum of valuational judgment strength. On other words, we accept that all tastes are judgmental (broad sense), but hold that those tastes whose cultivation could only have been avoided at great personal cost give entitlements to higher levels of compensation than those whose cultivation could have been avoided at a smaller price. This is certainly not the position that Cohen takes, and not only because it relies on the alternative reading of Cohen. For Cohen also does not seem to consider the possibility of morally relevant variations in the strength of identification; indeed, his discussion of (direct) judgmental taste seems just to assume that there is strong identification. But it does at least retain Cohen’s identification of valuational judgment and non-responsibility as two distinct grounds for compensation. There is, however, an additional argument for a subsumption of the first of these grounds under the second which points to a weakness common to both the bullet-biting strategy and the ‘sliding scale’ view of broad judgmental tastes. The argument is that both positions give hostages to fortune, in the sense that (direct or indirect) taste-creating value judgments are specified as grounds for compensation, be it full or partial, regardless of which account of responsibility happens to be correct. I do not see any reason for thinking that, whatever expensive taste- excusing properties valuational judgment has, they are independent of responsibility. If we take the most straightforward case, that of direct judgmental taste, I submit that the appropriate response to a particular instance of preidentification is very likely to depend on questions such as these: What is the causal history of the preidentification? Was the agent aware of the costs of acting as suggested by the preidentification? How do other agents with similar levels of preidentification and information act? Which general account of free will and determinism is correct? On this account, can acts based on preidentification be genuinely voluntary, or not? Does this account, combined with an account of egalitarian justice, allow that some acts based on preidentification are compensable? If so, which acts, and how much compensation? These are questions of responsibility. I do not see how we can detach the moral significance of taste-creating value judgments from them, far less assume an overwhelming moral significance however we answer them. IX. Concluding Remarks The argument of this article suggests that egalitarians looking to occupy a position on expensive taste that falls between conventional equality of opportunity for welfare and equality of welfare cannot expect the brute taste/judgmental taste distinction to carry the justificatory load. Nevertheless, the distinction serves a valuable role, in highlighting the possibility that egalitarian justice, especially as depicted by Dworkin, treats the bearers of valuational judgment-based expensive tastes in unduly harsh fashion. I hope to have shown that the corrective to this state of affairs need not be at the level of distributive principle, as Cohen proposes, since the role of valuational judgment in taste formation might be accounted for by the account of responsibility that is utilized by distributive principles. This underlines the importance for political philosophy of the contents of what Susan Hurley calls the ‘black box of responsibility’.[28] The problems with the notion of compensating for valuational judgment-based acts reinforces Hurley’s general contention that different accounts of responsibility have differing implications for accounts of distributive justice. If, as is fairly likely, valuational judgment bears on responsibility, and takes much (if not all) of its moral significance from that connection, it is a mistake to decide on its weight in distributive decisions quite independently of responsibility considerations. Interestingly, this is one case where a ‘luck egalitarian’ – or more specifically, ‘thin luck egalitarian’[29] – position may actually be reinforced by filling in the details of responsibility, since the non-luck egalitarian addition (compensation for judgmental taste) to a luck egalitarian position (equality of opportunity for welfare) may thereby be shown to be otiose. It might be objected that the conventional-equality-of-opportunity-for-welfare- plus-valuational-judgment-sensitive-account-of-responsibility strategy defended in the preceding two sections gives a hostage to fortune which, while different from that of the Cohenian alternatives, is no less problematic. This is its commitment that valuational judgment is necessarily a condition – or even, on the strongest version of the view, the sole condition – for non- responsibility. The objection is sound. Even so, to me this seems like the most plausible way of ensuring compensation for expensive taste-creating value judgments, for reasons already mentioned. Since I am unconvinced that such compensation is demanded by justice (although it may well be) I would prefer to leave the question of exactly what responsibility is as open as possible. Of course, in doing this, I face the further objection that I am myself trying to keep the lid on the black box of responsibility. In reply, I would point out that there seems to be no structural deficiency or superfluity with the thin luck egalitarian position that, like special compensation for valuational judgment-based expensive taste, would be exposed as such once a full account of responsibility was provided. It is quite compatible with responsibility, whatever role the latter gives to valuational judgment. I am not so much keeping the lid on the black box of responsibility as refusing to prejudge its contents. The strategy I propose replies to the Dworkinian expensive taste objection in just the way that Cohen intended in ‘Currency’, arguing that, regardless of whether an expensive taste is presently identified with, compensation seems reasonable where the holder is not responsible for the expense of her taste. We allow for the possibility that a history of valuational judgment is responsibility-diminishing, but we do not insist that it is a responsibility-diminishing factor, irrespective of what else there is to know about responsibility, nor (as the bullet-biting and sliding scale Cohen-derived positions suggest) that it is a compensation-entitling factor, irrespective of whether it diminishes responsibility. This is the right tack to take in the absence of overwhelmingly strong arguments for the distributive significance of valuational judgment. ------------------------------------ * I would like to thank Richard Arneson, Hillel Steiner, Andrew Williams, and an anonymous referee for their helpful comments. Research for this article was supported by the Arts and Humanities Research Council and the Worldwide Universities Network. [1] Ronald Dworkin, ‘What is Equality? Part Two: Equality of Resources’, Philosophy and Public Affairs 10 (1981), pp. 283-345. Reprinted as chapters 1 and 2 of Ronald Dworkin, Sovereign Virtue: The Theory and Practice of Equality (Cambridge, MA: Harvard University Press, 2000). [2] G. A. Cohen, ‘On the Currency of Egalitarian Justice’, Ethics 99 (1989), pp. 906-44. See also Richard J. Arneson, ‘Equality and Equal Opportunity for Welfare’, Philosophical Studies 56 (1989), pp. 77-93. John Roemer, Theories of Distributive Justice (Cambridge, MA: Harvard University Press, 1996). [3] Cohen actually endorsed equal access to advantage, where advantage included both welfare and resources (‘Currency’, pp. 907, 916-21). I will set this difference aside, as Cohen allows (G. A. Cohen, ‘Expensive Taste Rides Again’, in Justine Burley (ed.), Dworkin and His Critics (Oxford: Blackwell, 2004), p. 4. (Cohen believes that ‘access’ and ‘opportunity’ are different concepts, but I can detect no difference between them on Arneson’s definition of the latter [see Arneson, ‘Equal opportunity for welfare’, pp. 85-6].) [4] Cohen, ‘Expensive Taste Rides Again’. [5] For an accessible account of the debate up to Cohen’s latest contribution see Jonathan Wolff, ‘Equality: The Recent History of an Idea’, Journal of Moral Philosophy 4 (2007), pp. 125-36. [6] Cohen, ‘Currency’, p. 923; cf. p. 920. [7] Cohen, ‘Expensive Taste Rides Again’, p. 7. [8] Ronald Dworkin, ‘Replies’, in Justine Burley (ed.), Dworkin and His Critics, p. 344. [9] This is also suggested on p. 927 of ‘Currency’. But note that Cohen introduces this thought as one of a number of additional comments on Dworkin’s position that come after his main argument against it. Despite Cohen’s protestations, it is far from ‘abundantly clear’ that his main concern was always with expensiveness of taste rather than taste itself, and the only textual support he offers is the passage cited above. [10] Dworkin, ‘Replies’, p. 344. [11] See Cohen, ‘Expensive Taste Rides Again’, pp. 12, 14. [12] Cohen, ‘Currency’, p. 923. [13] Cohen, ‘Expensive Taste Rides Again’, p. 7. [14] Cohen, ‘Expensive Taste Rides Again’, p. 7. [15] Cohen, ‘Expensive Taste Rides Again’, p. 7, original emphasis. See also Terry Price, ‘Egalitarian Justice, Luck, and the Costs of Chosen Ends’, American Philosophical Quarterly 36 (1999), pp. 267-78. [16] Cohen, ‘Expensive Taste Rides Again’, p. 7, original emphasis. A note makes it clear that by ‘standardly’ Cohen means only ‘barring the special case where people welcome the fact that their taste is expensive’ (‘Expensive Taste Rides Again’, p. 27 n. 13, original emphasis). [17] Cohen strongly hints that those who develop expensive tastes for snobbish reasons or who unsuccessfully gamble that they will be able to afford them should be denied compensation (‘Expensive Taste Rides Again’, pp. 12, 14, 20-1). But given that such tastes are patently based on valuational judgment, and that Cohen mentions no reasons for denying compensation in these cases of judgmental taste but not others, I find this puzzling, as does Dworkin (Dworkin, ‘Replies’, p. 347). [18] Dworkin said something similar about the position Cohen took in ‘Currency’: ‘that supposedly different ideal (equality of opportunity for welfare) turns out to be equality of welfare under a different name’ (Dworkin, Sovereign Virtue, p. 286). This conclusion is based upon the thought that all tastes are unchosen (and therefore compensable on Cohen’s scheme) as they are themselves the result of uncultivated second-order tastes; see Dworkin, Sovereign Virtue, ch. 7; ‘Replies’, pp. 346-7, 391 n. 22. My conclusion, by contrast, is based on the argument that, on one understanding of Cohen, some tastes can be chosen, but that all such tastes are nevertheless compensable for reasons of valuational judgment. [19] See note 3 above. [20] See Cohen, ‘Expensive Taste Rides Again’, p. 21. [21] Cohen, ‘Expensive Taste Rides Again, p. 8. [22] See sections VIII and XI below. [23] Cohen holds that, even if all expensive tastes were unchosen in the way Dworkin suggests (see note 18 above), and therefore compensable according to equality of opportunity for welfare, there would still remain a difference of principle between that account and equality of welfare; see ‘Expensive Taste Rides Again’, pp. 19-20. It is worth noting that the present situation is very different from this, since Cohen has given a reason for providing compensation for expensive tastes even if they are chosen. [24] Cohen, ‘Expensive Taste Rides Again’, pp. 19-21. [25] Dworkin, ‘What is Equality?’, pp. 229, 230. [26] Dworkin, ‘What is Equality?’, p. 230. [27] Cohen, ‘Expensive Taste Rides Again’, p. 22; Cohen, ‘Equality of What? On Welfare, Goods and Capabilities’, in Martha Nussbaum and Amartya Sen (eds.), The Quality of Life (Oxford: Oxford University Press, 1993), p. 28. See also Arneson, ‘Equal Opportunity for Welfare’, p. 86; Carl Knight, ‘The Metaphysical Case for Luck Egalitarianism’, Social Theory and Practice 32 (2006), pp. 173-89. [28] See S. L. Hurley, Justice, Luck, and Knowledge (Cambridge, MA: Harvard University Press, 2003). [29] ‘Thin luck’ is just the opposite correlate of responsibility; see Hurley, Justice, Luck, and Knowledge, ch. 4. citation_temp.pdf http://eprints.gla.ac.uk/49956/