Value Holism (draft of 3/18/09)
Richard Chappell

How does the value of a whole (person, society) relate to the value of its parts (timeslices, individuals)?  Utilitarians have traditionally treated the parts as axiologically fundamental, and held that we may simply sum the intrinsic values of the parts to obtain the intrinsic value of the whole.  But this presupposes what we may call 'value atomism', or the claim that the contributory value of a part depends only on its intrinsic (non-relational) features.  The value1 of a part is, in other words, taken to be independent of what else exists.  In this paper, I wish to challenge this 'independence assumption', and explore the prospects for an holistic axiology that takes extrinsic or relational properties into account.  We may thus take value holism to be the claim that the value of a part cannot be assessed in isolation; it depends on how the part stands in relation to the whole, and in particular on what other parts there are.  To speak even more generally: value holists treat the whole as axiologically fundamental, and hold that parts get their value in virtue of how they contribute to the 'shape' (so to speak) of the whole.  This rather abstract picture should become clearer when applied to the particular examples discussed below.

The first section discusses holism as it applies within a life.  I begin by noting some intuitive cases of extrinsic or relational properties affecting the value of our momentary experiences.  I then discuss some surprising evidence suggesting that additional moments of pain may be preferable to the immediate cessation of pain.  This leads to a closer examination of hedonic duration, or the sense of 'duration' that is relevant to hedonic assessment.  I offer a speculative reconceptualization of hedonic duration that strikes an appropriate balance between subjectivity and objectivity, and which reinforces the holist's idea that we cannot determine the value of a momentary state of affairs in isolation. 

The second section of this paper then applies value holism to the larger question of how individual lives contribute to the value of the world as a whole.  This discussion focuses on a central problem of population ethics known as 'the repugnant conclusion'.  I will suggest two distinctively 'holistic' solutions to the problem: the first taking the position that 'mere addition' - adding intrinsically good lives to a world - may make the world worse, and the second exploring the value of diversity.  In each case, I argue that the contributory value of a life cannot be assessed in isolation, but instead depends on what else there is.  I conclude by considering some objections to this position.

(I) Whole Lives and their Individual Moments

- Trends and Narrative Structure

We care about not only our net happiness but also its distribution.  Better to start miserable and die happy than the reverse, many of us would think.  This may partly be explicable through contrast effects: the order in which you eat brussel sprouts and icecream may influence how tasty each seems individually, i.e. changes in order might also lead to changes in hedonic quality.  Similarly, the memory of better days may render present drudgery all the worse; or memories of past difficulties may sweeten today's success.  But even if we correct for all of that, so that the momentary experiences are truly identical in hedonic quality, and differ only in order, we may still prefer an upward trend in its own right.  For in addition to the moment-to-moment experiences, many of us also care greatly about the overall "narrative structure" of our lives.  We prefer to live out a story that improves with time rather than declines.  (As an extreme case, consider the - perhaps apocryphal - tales of Japanese lovers who leap to their deaths mid-coitus, so as to end their lives in bliss.)

The significance of narrative structure has very important implications for how we think about longevity and life-extension.  Atomists assume that making a life n times longer also makes it n times better (assuming the additional moments are intrinsically as good as the earlier ones).  But this seems implausible.  A twenty-five year old should prefer dying at age 75 over an even gamble of either living to 125 or dying instantly.  This is not just a matter of risk aversion; rather, it seems that the latter option has a lower expected value than the former.  Living an extra hundred years (even assuming perfect health, etc.) simply isn't twice as good as living an extra fifty.  That's not to say that life-span always has diminishing marginal utility.  If I am fated to die in one year, this might be too short a time to achieve anything of great value.  Twenty additional years of life would plausibly be more than twenty times as valuable to me, if this would enable such important goods as raising a family.  I might reasonably risk instant death for a less than 1/20 chance of such life-extension.  It all depends on what I most want out of my life as a whole, and how long is necessary to achieve this.

So we cannot just assume that each additional century (say) adds as much value to a life as the first.  Even if the additional centuries are no worse in themselves, they may be less important overall.  Atomists cannot accommodate this vital insight.  Their axiological method is to define the value of a life as the sum of the values of its moments.  But we have seen that this approach is inadequate.  We should thus prefer the holist's axiological method of directly evaluating the life as a whole.  'Global' preference theories of wellbeing (such as Parfit's success theory) exemplify this holistic approach.2 Rather than assessing each moment in isolation, we can instead ask ourselves, 'How would I most like my life as a whole to go?'  Or: 'What is it that I most want in life?'  It is entirely possible that when we categorize our global preferences or 'life goals' into those that could be achieved in a hundred years, and those that could only be achieved with a second hundred, the former group will outweigh the latter.  This would justify our unwillingness to risk the former for a chance at the latter.  Even when atomists can't tell the difference, holism affords us the resources to recognize that some years contribute more to our lives than others.

Common sense thus allows that extrinsic features may influence the magnitude of the contributory value of some life experience.  We might further question whether even the valence of an experience (i.e. as good or bad) can be determined in isolation.3  An intrinsically painful experience might be considered good rather than bad if it is embedded within the right context, e.g. the strivings of a marathon runner to stretch her endurance to its limits, or a repentant wrongdoer enduring his just punishment.  (A false memory causing the same subsequent feelings of satisfaction would arguably not have the same value, at least if we value the challenge - or the penance - itself.)  When assessing their life as a whole, an agent may reasonably judge a moment of pain to have been non-instrumentally desirable, in light of its relations to other parts of their life, and the elegance of the overall pattern this gives rise to.  These examples help to bring out the prima facie plausibility of value holism, and may even convince some readers that they are implicitly committed to the view.  But I now want to focus on the more surprising phenomenon of 'duration neglect'.

- Preferring the Longer Pain

Kahneman et al. (2003) found that subjects sometimes prefer additional pain, if the appended moments are less painful than what has gone before.  After experiencing both the short and long episodes, and being offered a choice of which to repeat, most subjects choose the long episode.  More generally, the duration of an experience seems to have little effect on our retrospective evaluations.  Instead, these value judgments are largely determined by some combination of the 'peak' and 'end' moments.  Kahneman et al. consider these judgments erroneous on the grounds that they violate temporal monotonicity, an independence rule to the effect that "adding moments of pain to the end of an episode can only make the episode worse."4 But perhaps we should instead reject temporal monotonicity in favour of value holism. We should take subjects at their word when they tell us that the added moments of lesser pain made their overall experience better.  Or so I will argue.

To begin, let us note the default presumption that each individual is the best authority when it comes to determining the hedonic quality of their own phenomenal experiences.  To override this presumption, the atomist might offer two arguments.5  First, they could point out that the subjects themselves would presumably want the pain to end sooner if offered the choice during the episode.  After all, it is in the nature of painful/aversive experience that it be accompanied by a local preference for its own cessation.  So at each moment when we are experiencing pain, we wish it to stop.   True enough.  Yet when making an overall judgment from 'above the fray', so to speak, the subjects express a conflicting preference, and merely noting the conflict does not tell us how to resolve it.  As a general rule, we tend to privilege (reflective) global preferences over (momentary) local ones: such a hierarchy is, after all, essential for the exercise of self-control.6  So, again, some further reason is required to override the presumption in favour of the subjects' expressed global preference.

Alternatively, the atomist may grant the authority of (informed) global preferences, but suggest that subjects' retrospective assessments here are ill-informed.  The lesser pain, being more recently experienced, is more salient in their memory, so perhaps subjects misremember or fail to appreciate how bad the initial period of pain really was.  This is a difficult claim to assess.  There's something dialectically suspicious about it, insofar as the claim is motivated by our prior reluctance to credit the substantive assessment offered by the subjects, when the credibility of this assessment is precisely the question under dispute.  Independent evidence of error might be sought in the fact that most subjects claimed that the longer trial caused "less overall discomfort", which Kahneman et al. insist is "simply wrong".7 But for this to qualify as independent evidence of factual error, we must assume that subjects were interpreting 'overall discomfort' to mean 'aggregate momentary discomfort'.  This seems unlikely.  It's far more plausible to think that subjects were simply reiterating their holistic judgment that the longer trial was less unpleasant on the whole.  So these considerations leave us at a dialectical impasse.

- Subjective Time

So far, I have assumed that the third-personal perspective yields accurate insights into the aggregate momentary qualities of experience.  For example, I have assumed (with Kahneman et al.) that an episode of pain that has longer physical duration must thereby have longer experienced duration.  But this arguably presupposes a false picture of consciousness -- what Dennett calls 'the Cartesian Theatre', or a special place in the brain "where the order of arrival equals the order of 'presentation' in experience because what happens there is what you are conscious of."8 An important consequence of rejecting this picture of consciousness is that our subjective experience of time may not match up with the objective timeline.  We experience the content of conscious representations, not the vehicle doing the representing.  It is tempting to assume that the temporal representation in consciousness is somehow transparent -- that if our brains represent to us that A occurs before B, this must be because a vehicle representing A was followed, in objective time, by a vehicle representing B.  But of course there is no logical reason why this must be so.   As Dennett points out, we can represent time using a medium other than time itself.9  We can say "B occurs after A", and it represents the ordering (A,B) even though the sentence mentions them in the order (B,A). Similarly, you can subjectively experience the represented order of events (A,B), even if the brain regions doing the representing actually process the events in the opposite order.  And indeed several experiments have been conducted which demonstrate precisely this effect, at least on very short time scales.10

At this point, one may object that our understanding of phenomenal consciousness has become too subjective.  There is surely a difference between actually experiencing some temporally extended episode and merely believing this to be so.  But can we maintain this distinction without falling back on physical duration as our objective standard?  To bring out the problem, compare the following three 'experience machines':11

  • Machine A gives you pleasant experiences for 100 years.
  • Machine B (allegedly) gives you all the same experiences, feeling exactly the same from the inside, but packed into just a single physical day.
  • Machine C gives you one day of pleasure, and then simply implants in you the (presumably false) belief that it felt like it lasted for 100 years.

The challenge is to make sense of how Machine B could be anything other than a fanciful redescription of Machine C -- which we don't want to say really gives us 100 years of experienced pleasure.  The solution, I think, is found by reflecting on the relative paucity of the C-implanted representation.  It's one thing to write a story which says "100 years passed", and quite another to fill out the details for 100 years' worth of fictional events.  So, if we think of Machine A as taking a long time to impart rich informational content (lots of pleasure), and Machine C as taking a very short time to impart a very thin representation (very little pleasure), the question whether Machine B imparts a lot or only a little pleasure comes down to the richness of the representations it implants.  We can thus make the necessary distinctions without appeal to the dubious notion of a 'Cartesian Theater' in our minds where consciousness "occurs".

This suggests that a fairly radical reconceptualization of hedonic value may be in order.  The spirit of hedonism seems more consonant with a concern for subjectively experienced duration than mere physical time.  But this in turn is best analysed in terms of representational richness, or so I have proposed.  This may have significant practical implications -- let me note just two.  Firstly, evidence suggests that younger, inexperienced brains tend to lay down denser and richer memories, explaining why time seems to pass more quickly as we age.12 So if my above claims are correct, utilitarians should be disproportionately (relative to their merely physical duration) concerned to increase the hedonic quality of children's experiences -- possibly counterbalancing our prior preference for a life that improves with age.  Secondly, we may be led to the conclusion that repetitive experiences may be discounted, if merely repeating the same old information over and over again does nothing to enrich the representation.  (Suppose all you need is a ditto mark, or the cognitive equivalent of "times a million".)  For example, if I am groggily experiencing a long, painful operation,13 it's possible that after a while the moments will begin to blur together, to the point where it makes no subjective difference to me whether the operation lasts (say) two hours or four.  Of course, the mere retrospective inability to tell the two scenarios apart does not by itself imply that there was no hedonic difference.  But if the intervening momentary experiences are indeed qualitatively indiscernible in fact, and not broken up by any (perhaps later forgotten) distinguishing features, then this provides some reason to think that we should count the hedonic (dis)value as the same in either case.  On the other hand, providing a 'signpost' that dispels temporal blur may cause an experience to count for a lot more.

I should emphasize that most of the time, even intrinsically identical pains are embedded in discernibly different experiences (with different background thoughts running through our heads, etc.), and so count as recognizably distinct.  But we can at least imagine a case where extending the duration doesn't introduce sufficient qualitative differences.  After a while, many moments of hospitalized agony all blur together, and we may think the reason for this is precisely that there is truly nothing in the experiences to distinguish them. And so, on this account, they count for just one.

- Taking Stock

I began this section by bringing out our implicit commitment to value holism as it applies within an individual life.  There are many plausible cases in which it seems the value of a momentary experience cannot be assessed in isolation.  Then, by distinguishing physical and 'subjective' (or consciously represented) time, I suggested, in addition, a metaphysical basis for something that closely resembles holistic practice: for even insofar as the intrinsic qualities of experience are concerned, these may yet be underdetermined by a physical moment in isolation, instead depending on physically later (or earlier) events to precisify their phenomenal character.  Developing this idea further, I offered a speculative reconceptualization of hedonic evaluation, whereby the 'duration' dimension is replaced by an atemporal notion of 'representational richness'.  This provided a reason to think that duplicate experiences may count for only one, assuming that duplication may be represented cheaply.  In the following section, I will seek to establish some corresponding normative conclusions on the societal level, but without relying on such contentious metaphysical assumptions.

(II) Whole Populations and their Individual Lives

- The Repugnant Conclusion
Derek Parfit noted that total utilitarianism implies the 'Repugnant Conclusion' that for any finite flourishing population A, we can imagine some vastly larger population Z of lives barely worth living, which ends up counting as "better".14 This is partly because total utilitarianism is an atomistic theory according to which the contributory value of a life is simply its welfare value for the person living it -- a value that can be determined in isolation, simply by looking at that life in itself.  Since worthwhile lives presumably have some positive contributory value, atomism implies that astronomically more lives yield astronomically greater value.  Yet when we consider the imagined world Z as a whole, this conclusion no longer seems plausible.  Holistic judgment enables us to recognize the Repugnant Conclusion as repugnant; we may accordingly expect value holism to provide a fruitful starting point in our search for a solution.

It is not easy to avoid the Repugnant Conclusion, as Parfit's 'Mere Addition Paradox' demonstrates.15 'Mere addition' is when we add additional lives (all above the baseline of lives worth living) to a world, without affecting the prior inhabitants in any way.  Parfit claims that this process cannot make a world worse.  This seems prima facie plausible: after all, where's the harm?  How could it be bad to add intrinsically good lives, to no ill effect on anyone else?  This suggests the following principle:

(Mere Addition)  If the only difference between worlds A and A+ is that the latter contains additional lives above the baseline, then A+ is no worse than A.

Next, note that it can only improve a world to reduce inequality in such a way as also increases total welfare, while holding all else equal.  Call such a shift 'beneficial equality'.  Beneficial equality licenses the move from A+ to a world B where the worse-off group in A+ is benefited more than the well-off group is harmed by the shift.  If B is better than A+, which in turn is no worse than A, it follows - by transitivity - that B (a world of greater total but lesser average utility) is likewise at least as good as A.  We may iterate this process until we reach the repugnant world Z, with astronomic total utility but miniscule average utility.

These implications may lead us to examine the Mere Addition principle more closely.  Indeed, to a value holist, the justification for the principle will seem immediately suspect, for it assumes atomism from the start.  Recall that the justification appeals to the idea that the contributory disvalue of a life must be due to some badness in its intrinsic qualities: the life must be bad in itself, in a way that lives above the baseline are not.  But this is just to assume atomism.  To a holist there is nothing contradictory about the idea that adding an intrinsically good part may make the whole worse.  (Laughter and merriment are good in themselves, but not at a funeral.)  So while mere addition may sound harmless enough, the real test is to directly evaluate the two worlds in question.  And here is it entirely open to us to judge that A+ is indeed a worse world than A.  Why might we think this?  Because the addition of worse (though not bad) lives alters the shape of the world as a whole, and not for the better.  Whereas before we had a world full of flourishing, we now find mediocre lives in addition.  That's not to say that the mediocre lives are bad in themselves, or considered in isolation.  But given how the rest of the world is, their addition may be considered undesirable nonetheless.

None of this is to endorse anything so crude as the 'average utilitarian' principle that it's always bad to lower average welfare.  A world sparsely populated with only a hundred people, however well-off they might be, would plausibly be improved by adding more good lives, even if they are not as good as those already there.  For one thing, a world that's too sparsely populated can be expected to lack the full range of diversity that makes a (typical) larger population so desirable.  But note that diversity is a distinctively holistic value: whether a particular life or experience adds to the diversity of the whole depends on what else there is.

- Duplication and Holism
The Repugnant Conclusion brings to mind dystopian visions of a universe (Z) tiled with bland mediocrity -- the same old "muzak and potatoes",16 repeated over and over.  This is certainly a dreary scenario.  Muzak and potatoes were never all that great to begin with, but to simply add more of the same is arguably not to add any more value to the world whatsoever.  It would be different if we were to instead imagine a world Z* of very diffuse and diverse excellences, where each life contained only modest value when considered in isolation, but nonetheless offered a distinctive contribution to the world as a whole.  This no longer seems so repugnant at all; it may even be an open question whether, intuitively, it would be better to condense those diffuse glimpses of excellence into a smaller number of more consistently flourishing lives.  The holist may go either way on this question, so I will not attempt to settle it here.  What I want to highlight instead is the intuitive significance of diversity.  What seems most repugnant about the Repugnant Conclusion (as envisioned above) is not just that value is so diffuse, but that there doesn't really seem to be much value in the Z-world at all.  We're inclined to think that contributory value just can't be duplicated in this way.  Rather, to make the world as a whole more valuable, we must add lives of distinctive value.  So claims the 'diversity principle' that I now wish to consider.

(DP) Multiple evaluands count for less the less distinct they are.

DP has a number of significant implications.  We've already seen how it can explain the problem with (the most repugnant version of) the Repugnant Conclusion.  [Though it's worth noting that it suggests a new objection to the mere addition paradox: the 'beneficial equality' step may actually be for the worse if it reduces diversity.] For another example: if DP were false, then it would be of momentous importance whether ours was a world of Nietzschean 'eternal recurrence'.  So those of us skeptical about whether this would really matter so much must be relying on some DP-like holistic principle for discounting all those duplicate epochs.  For a more practical example: suppose it turns out that the experiences of most hens in factory farms are qualitatively identical (or nearly so).17 DP would then imply that the total disvalue here is less than we might at first expect, since the duplicates are subject to discounting.  This would be a very surprising, and somewhat disconcerting, result.

This may be turned into an objection: it seems absurd to think that it might make the world no worse were a billion people tortured instead of just one, for example.  But such intuitions gain their force from "ordinary" cases, where different people have different memories, etc., and which thus involve discernibly different experiential contexts.  (This background assumption of human cognitive diversity may be at least part of the explanation why this objection seems so much more gripping when applied to humans than hens.  The logic of atomism should apply equally to either case, by contrast.)  Special care is required to conceive of a situation where DP would actually apply – e.g. a dystopian 'farm' of duplicate brains-in-vats, programmed to have exactly the same series of experiences throughout their existence – but then the intuition is less clear.  To ensure that all variables are controlled, imagine a 'digital person' or conscious Artificial Intelligence, whose 'life' is constituted by the running of a computer program.  Would it matter how many duplicate copies of the program were run?  It isn't obvious.  Suppose that, as a digital person, you are about to undergo copying (digital 'fission'), but that not all of your future continuants will be identical.  If there were to be 95 copies of one programmed future, and just one copy each of five other, qualitatively distinct futures, would you rather improve the first program or the other five?  The latter preference would indicate a strong commitment to DP.

Clear intuitions are difficult to come by in these cases, so we may prefer to decide the issue on more general theoretical grounds.  The first section of this paper noted that we care about the shape of our life as a whole.  A stronger claim is that this is all that matters.  Mere multiplication – say where every event extends for twice as long – would then be dismissed as lacking in normative significance, so long as the broad contours of the life remained much the same.  The value of an extra year of life depends on what would be achieved with it, and whether it would contribute anything new or significant to the overall structure of the life.  If it is all much of a sameness with what has gone before, it may not have any significant impact on the quality of the life taken as a whole.  And the same may be true of the relation between individual lives and the value of the world as a whole.

This general holistic picture may cast some doubt on the importance of absolute quantities of mere duplicates.  But it doesn't immediately follow that all that matters is the number of distinct evaluands.  We may also care about their proportions, or the relative quantities of each duplicated kind.  For example, in the case of digital fission, perhaps I should care more about the future that occurs in 95% of instances, even if I shouldn't care whether the absolute number of instances is 100 or 1000.  This is another issue on which the holist could go either way; I won't attempt to settle the matter here.

- The Alleged Asymmetry
Some claim that there is an important asymmetry in our assessments of merely possible pleasures and pains (or benefits and harms more generally).  They claim that we have no reason to bring more good lives into existence, or to regret their absence, whereas we do have reason to prevent lives of suffering and to feel relief at their absence.18 We feel that it would be wrong to accept as a 'package deal' several good lives and one bad.  As Christopher Belshaw writes, "We can't justify starting this bad life by appeal to the good in other, separate, lives."19 But I propose that we can explain away this intuition by appeal to value holism.

According to value holism, the value of the above 'package deal' is not independent of what else exists.  It might be positive in some circumstances and negative in others, depending on how it impacts the overall 'shape' of the world.  As it happens, we have a high normal baseline: we assume that most lives in our society are pretty good. So, creating a good life is nothing exceptionally good, whereas creating a bad life is exceptionally bad. Given the more fundamental principle that exceptionally good or bad actions have greater moral significance (in virtue of their impact on the general shape or form of society), we find that a contingent asymmetry in social circumstances leads to the above moral asymmetry.

Note that things could have been different. If we imagine a dystopian world where the normal 'baseline' is much lower, i.e. where most lives are rather awful, then it seems to me that the moral asymmetry would be likewise inverted. Given the opportunity to bring about an exceptionally good life, people ought to do so. To prevent another typically bad one would be permissible -- good, even -- but not required. So, dystopians ought to embrace the package deal of good and bad lives.

On the view I've outlined, there is no fundamental moral asymmetry in the relative weighting of good and bad additional lives.  Any asymmetry here instead arises from the application of value holism to our particular circumstances.  We may be right to think that we shouldn't want six additional good lives at the cost of an additional bad one, in our current circumstances.  But this does not imply that more good lives couldn't - in other circumstances - outweigh bad ones so as to make a world better on net.

Some have proposed a more plausibly fundamental asymmetry: although additional good lives may make a world non-instrumentally worse (if they reduce the average welfare, say), additional bad lives cannot make a world non-instrumentally better (even if they are less bad than average).  I think it's open to a holist to reject this assumption.  But even if we don't go quite that far, Huemer proves that assigning any non-zero weight to average utility commits us to:

The Sadistic Conclusion:  In some circumstances, it would be better with respect to utility to add some unhappy people to the world (people with negative utility), rather than creating a larger number of happy people (people with positive utility).20 

This does seem counterintuitive, at least at first glance.  But further reflection reveals that it is not much of a move from the claim that adding mediocre lives can make a world worse.  For then we may expect that adding a great many mediocre lives could make a world much worse (transforming it from a predominantly flourishing world to a predominantly mediocre one).  If this is a harm at all, then it isn't surprising that it could outweigh the modest harm of adding a single moderately bad life.  We are tempted to draw a bright line between lives that are worth living and those that aren't, but the absolute difference in utility might be as small as you care to imagine (for arbitrarily small ε, compare the welfare values +ε/2 and -ε/2).  So we should not place as much weight on this difference as is relied upon in the above objection.

- Wrapping Up

While my first section explored the application of value holism within a life, this second section has explored how a holistic view might be applied to some of the central problems of population ethics.  Value holism naturally suggests two routes to avoiding the Repugnant Conclusion: the first by rejecting Mere Addition, and the second by rejecting any duplication that may sneak in during the 'beneficial equalizing' step.  (The latter leaves open a diversified version of the repugnant conclusion, but then it no longer seems so repugnant.)  The preceding sub-section addressed some objections that are based upon alleged asymmetries between harms and benefits.  I now want to wrap up by drawing out what I consider the two most pressing challenges that emerge from all this.

First, there is the question whether atomists can make a principled stand in defence of Mere Addition. Huemer suggests the following "almost irresistable" principle:21

Modal Pareto Principle: For any possible worlds x and y, if, from the standpoint of self-interest, x would rationally be preferred to y by every being who would exist in either x or y, then x is better than y with respect to utility.

As before, this principle is clearly atomistic: it considers each life in isolation ("from the standpoint of self-interest"), and leaves no room for holistic 'big picture' considerations.  But this can be defended by appeal to an independently appealing conception of ethics as fundamentally person-centered.  It is tempting to think that in order for something to be bad, it must be bad for someone.22  The underlying theoretical principle may be captured by something like the following:

(Individualism)  Individual persons are the ultimate bearers of value, or that for which we act. 

Now, individualists might accept holism within an individual life, since they accept that - morally speaking - the whole person is 'prior' to their temporal parts.  Value inheres in people first, and their momentary stages only derivatively.  But the individualist will reject my extension of holism to the interpersonal level.  They deny that there is any supra-personal collective entity (even 'the world' as a whole) that can serve as an independent bearer of value.  Whereas the holist talks seriously about making the world a better place, individualists will insist that this is merely shorthand for making all the individual people in the world better off.  (Compare Frankena's maxim that "Morality is made for man, not man for morality.")23

Such a view has significant traction in the western cultural tradition.  But I think even total utilitarians should ultimately reject it, for reasons that emerge from Parfit's work on the 'non-identity problem'.24 Suppose that we have an opportunity to bring about world Z, but choose to stick with world A instead.  Looking back, the total utilitarian will judge that we chose wrongly.   But who or what has been 'wronged' or harmed by our choice?  The only available answer, I want to suggest, is 'the world as a whole'.  After all, they can't very well insist that we should have brought about world Z for the sake of those merely possible people who will now never get to exist.  One cannot be harmed by non-existence, for there is nobody there to be harmed -- that's what it means to not exist.  This is not to deny that we may have reason to bring people into existence, or to lament our past failure to do so.  But it would be nonsensical to say that the non-existent people are themselves the source of the reason.  There are no such people.25  Rather, I suggest, the total utilitarian seeks to make the world better, and creating more good lives is a way to achieve this end.

As a general principle: instrumental reasons that truly exist can only stem from - and exert their force on behalf of - entities that (likewise) actually exist.  This forces us to reject individualism as a theoretical account of the source of reasons, if we do not think that actually existing individuals are the only source of reasons.  However, this places no practical limitations on what we may have reason to do, since among the things that exist is the world as a whole, and any putative reasons (e.g. concerning merely possible people) may be reinterpreted as speaking to us on its behalf.  For example, we may have reason to prefer a possible outcome with more happy people in it, if that would make for a better world.  We might even say that all those additional people are what 'gives us reason', in some sense, to prefer this imagined outcome over the actual situation.26  But this should not lead us into metaphysical confusion.  In particular, it is not to say that those non-existent people are the fundamental source of the reason, or that for which the reason exerts its normative force.  (Again, ex hypothesi there are no such people to provide such a basis.)  Rather, such loose talk merely serves to specify what features of the imagined scenario make it a desirable state for the world to be in.

It may be thought desirable, for the sake of the world as a whole, that it contain more happy people -- in which case we may reasonably lament that this isn't so.  We may regret that the world is worse than it could have been.  But we cannot sensibly think that non-existent individuals are worse-off than they might have been.  Though our world is worse, it is not worse for them (or, perhaps, anyone).  This means that defenders of the repugnant conclusion cannot appeal to individualism, and so their atomistic 'modal pareto principle' lacks independent support after all.

A second - more methodological - challenge may be suggested by the various unresolved questions I've raised about how to best flesh out the holistic position.  Rejecting atomism opens up a number of new variables, and the sheer range of options here may raise concerns about "curve-fitting".  Tailoring the view to accommodate our intuitions in particular cases may just seem less 'principled', somehow, than the atomist's bullet-biting resolve.  I think that can't be right quite as stated: the method of reflective equilibrium licenses moving "back and forth" between our judgments of particular cases and general principles, as we seek to bring them all into coherence.27 But the atomist might at least note that their theory has the advantage in terms of simplicity.  The big question is whether it is too simple in its neglect of the 'big picture' relations between lives.28 


References

 Benatar, D. (2006) Better Never to Have Been: The Harm of Coming into Existence. Oxford: OUP.
Dennett, D. (1991) Consciousness Explained. Boston: Little, Brown and Co.
Frankena, W. (1973) Ethics 2nd ed. Englewood Cliffs, NJ: Prentice Hall.
Huemer, M. (2008) 'In Defence of Repugnance' Mind 117 (468): 899-933.
Kahneman et al. (2003) 'When More Pain is Preferred to Less: Adding a Better End' Psychological Science 4(6): 401-5.
McNaughton, D. (1988) Moral Vision. New York: B. Blackwell.
Parfit, D. (1987) Reasons and Persons. Oxford: Clarendon Press.

notes



1 Hereafter when I speak of a part's "value", simpliciter, I mean its contributory value, i.e. how much it adds to the value of the world as a whole.  This should be distinguished from 'welfare value', or the value of a life for the person living it.

2 Parfit, appendix I.

3 Cf. McNaughton

4 Kahneman et al., p.401.

5 Versions of each can be found in Kahneman et al.

6 cite Frankfurt?

7 Kahneman et al., p.403.  I should note that some subjects made more straightforward factual errors, but that still leaves many others whose preference for the longer pain is yet to be 'debunked'.

8 Dennett, Consciousness Explained, p.107, original emphasis.

9 Dennett, pp.143-9

10 See Dennett, p.143, on the 'phi phenomenon', 'cutaneous rabbit', etc.

11 Suggested by Ben Bradley and Troy Jollimore at 'PEA Soup':
http://peasoup.typepad.com/peasoup/2008/02/objective-and-s.html#comment-101237434

12 http://www.livescience.com/health/071211-time-slow.html

13 Cf. the thought experiments in Parfit, chp 8.

14 Parfit, p.388.

15 Parfit, chp.19.

16 Parfit?

17 I owe this example to Michael Vassar.

18 See, e.g., Benatar, p.30; Parfit, p.391.

19 Belshaw, NDPR review of Benatar's book: http://ndpr.nd.edu/review.cfm?id=9983

20 Huemer, p.911.

21 Huemer, p.903.

22 Cf. Parfit, p.395.

23 Frankena, p.116. (Note that Frankena wasn't talking about this particular issue; but it's a suggestive turn of phrase all the same.)

24 Parfit, chp.16.

25 One may object that the people would exist if we acted rightly. So in that case, at least, the total utilitarian could point to these existing people as the reason why the act was right.  But this is insufficient, for we need an explanation that will also carry over to the case in which we fail to act rightly -- in which case there do not exist any such additional people to ground the claim that it was wrong of us not to bring them into existence. Hence the moral claim must stem from some other source.

26 That's slightly sloppy wording, however.  We arguably can't refer to the imagined 'additional people' individually, if they don't exist and so have no particular identities. To speak more carefully, what gives us reason here is the fact that the imagined outcome would contain x many additional people.  This makes it clearer that we're talking about the world, and not about any particular individuals in it.

27 http://plato.stanford.edu/entries/reflective-equilibrium/

28 Thanks to Nick Beckstead, Peter Singer, and Helen Yetter for helpful comments on an earlier draft of this paper.