2. Demandingness and the Deontic Asymmetry

We generally have some moral reason to make the world a better place, or bring about better states of affairs. So, to refute the Asymmetry, it will suffice to establish the following thesis:

An ‘awesome’ life, as I use the term here, is one that exhibits a very high quality of life, along whatever dimensions you take to be normatively relevant. So it might, for example, be a life that is subjectively happy, contains loving relationships, and exemplifies creative excellence and the achievement of worthwhile goals. Whatever you like.

Awesome Lives is, I think, intuitively highly plausible. When we think about what makes for a good state of affairs, the quality of life for the sentient beings contained therein is surely a (if not the) primary factor. A world full of awesome, flourishing lives is (intuitively) better than a world that lacks these good lives. Not everyone will agree with this, of course, but it is at least sufficiently plausible that anyone claiming to represent the broad consensus of ‘commonsense morality’ can hardly start from the assumption that Awesome Lives is false.

Given how plausible Awesome Lives seems in its own right, what motivation is there to reject it? The main concern, I think, is the fear that it will commit us to implausible procreative obligations.Footnote ¹ I certainly agree that it is not, in general, obligatory to procreate. But it’s just bad reasoning to conclude from this that there must be nothing good about bringing more (good) lives into existence. After all, the following is not a generally valid principle:

For example, it would be very odd for an opponent of stringent duties of beneficence to conclude that, because we aren’t obligated to donate all we can to help the global poor, it must be that helping the poor is not a good thing – that their welfare simply doesn’t matter. Yet that is precisely the reasoning of those who reject Awesome Lives simply on the basis of rejecting procreative obligations. If we think that we are not obligated to do certain things in order to benefit others, we should not conclude that those ‘others’ do not matter. A more promising strategy is to focus on the ‘certain things’ being requested of us, to see whether they might be too much to ask in a way that would explain the lack of obligation. Commonsense morality rejects the idea that we have to sacrifice everything to help the global poor, not because the global poor don’t matter, but because our sacrificing everything does matter, and would be too much for morality to demand of us.

And so it goes, I suggest, for the commonsense rejection of positive procreative duties. In our biological, psychological and cultural context, to require procreation would be massively intrusive and demanding. (Pregnancy alone is arguably more demanding than, say, giving up a kidney would be. And that’s before we even consider the life-changing implications of either raising a child or giving up one’s progeny for adoption. Or, for that matter, the idea that people have a moral prerogative, not easily overridden, over their genetic material.)Footnote ² If extreme financial duties of beneficence are too demanding to countenance, surely positive procreative duties are all the more so. So there’s no motivation here to deny Awesome Lives. Implausible procreative obligations are better denied on grounds of their excessive demandingness.Footnote ³

It’s not that possible people don’t matter, but just that actual people (who would have to bear and rear them) do. It’s worth noting that there seems no such problem of demandingness for the negative procreative duty to avoid bringing miserable lives into existence. No minimally decent person could possibly wish to have a miserable child, after all, so insisting that they avoid this bad outcome is no great burden.Footnote ⁴ Considerations of demandingness thus contribute to explaining the contingent truth of the Deontic Asymmetry: that it’s impermissible to (knowingly) bring a miserable life into existence, but permissible to refrain from bringing Awesome Lives into existence. The reasons that permit so refraining are specific to human biological procreation; they do not advert to bringing people into existence in general (which would include bringing unrelated people fully formed into a distant existence where they won’t impact upon the agent’s life at all).

Since the asymmetry in demandingness that underpins the Deontic Asymmetry is contingent, so too is the Deontic Asymmetry itself. If we imagine a scenario in which the act of ‘bringing into existence’ is very different in nature, such that a moral requirement here would not be intrusive or demanding in any way, then we should (if I’m right) find that our intuitive attachment to the Deontic Asymmetry likewise dissolves. So consider the following case.

It strikes me as intuitively clear that you have good reason, in this case, to refrain from plucking and eating the particular apple in question. This suffices to refute the Asymmetry – we can have moral reason to bring good lives into existence (or refrain from preventing their existence, which I take to amount to much the same thing in this context).Footnote ⁵ Moreover, it would plausibly be positively wrong to eat the apple in this case. The personal cost of refraining from eating a particular apple is so trivial, in comparison to the great value of Awesome Lives, that it would seem perverse, or morally indecent, to so act. So the Deontic Asymmetry does not hold of necessity, as a matter of moral principle. It merely happens to be generally true in our particular circumstances.

The key difference in real life, of course, is that procreation is not a trivial matter. But this just goes to show that it is these contingent complexities associated with child-bearing and rearing – rather than any fundamental asymmetry in the moral status of bringing miserable vs. awesome lives into existence – that is doing the real work here. Once this is recognized, I (and, I hope, the majority of readers) find no residual intuitive plausibility to the Asymmetry. Its prima facie plausibility is exhaustively captured by weaker related theses such as the contingent asymmetry in demandingness, and the Weak Asymmetry mentioned in the introduction.

3. Partiality towards the actual

Roberts (Reference Roberts2011, 772) writes: ‘[E]ven those theorists who do not find the Asymmetry particularly intuitive surely will find within themselves the deeper intuition that the existing Ann must in some way come before need-not-ever-exist-at-all Ben’. Suppose Ann’s quality of life is currently near zero. If you have the choice to either greatly help Ann, improving her quality of life by some amount X for several decades, or else bring into existence a new person ‘Ben’ with welfare X (and similar life expectancy of several decades), it seems clear that you ought to help Ann. Call this the Deeper Intuition.

There are many possible explanations of this intuitive datum. Roberts takes it to support a view she calls Variabilism, which in turn entails the Asymmetry. I will argue that there are other, better explanations of the Deeper Intuition that do not support the Asymmetry. Consider Roberts’ view:

On this view, losses due to non-existence, or benefits due to existence, simply don’t matter.

Variabilism secures some desired results. It explains the Deeper Intuition: bringing Ben into existence has no inherent moral value, so we do better to benefit the already-existing Ann instead. Bringing miserable people into existence does matter morally though, as in this case the ‘loss’ is located in a world where the agent exists, making it a morally bad outcome – just as it should be.

Variabilism also has some undesirable features, however. It may seem rather unprincipled or ad hoc – why don’t benefits due to existence get to count? Also, by entailing the Asymmetry, it inherits the flaw of being unable to accommodate our intuitions about the value of life in scenarios like A Distant Realm. It isn’t plausible that the benefits of bringing into existence should count for nothing at all. All that the Deeper Intuition supports is that they count for less than benefiting those who exist antecedently.

So it is worth exploring alternative explanations of the Deeper Intuition. One possibility is that average (and not just total) welfare matters, at least to some degree. This would explain why it’s better, all else equal, to increase total welfare without increasing the number of individuals who exist, and hence to benefit Ann rather than bringing Ben into existence. But this is not a sufficient explanation, because in a sufficiently large population the difference here could be negligible, whereas the importance of helping Ann does not seem to diminish with greater population. (If anything, the opposite seems true: bringing additional people into existence is better the fewer people there already are—cf. Hurka (Reference Hurka1983).)

A better explanation might appeal to the prioritarian idea that it’s especially important to benefit those who are ‘badly off’ (or to prevent harms that would leave one badly off), in the sense of having low or negative welfare (Part Reference Parfit1997). The non-existent have no (or undefined) welfare value, rather than low welfare, so there is no prioritarian reason to bring them into existence. But there is prioritarian reason to help Ann, as she has low welfare.

What if we imagine a variant of the case that involves a more privileged existing individual Annabelle? Suppose Annabelle is antecedently very well-off, so there is no (or negligible) prioritarian reason to favour her. Still, you have the choice to make her even more well off, or else to bring the new Ben into existence. Which should you do? In this case, I think the answer is less clear. This is some evidence that the prioritarian explanation of the original Ann and Ben case is the correct one (or at least a significant contributing factor, if not the whole story).

If you think we still ought to favour privileged-but-existing Annabelle over ‘need-not-ever-exist-at-all’ Ben, then I suspect your intuitions are instead tracking a principle of Partiality Towards the Antecedently Actual.Footnote ⁶ According to this principle, we should give greater weight to the interests of those who do or will actually exist independently of our current deliberations.Footnote ⁷

Something along these lines seems the most natural way to try to accommodate the Deeper Intuition. It’s the principle you get if you take the Deeper Intuition at face value: Ann seems to matter more because she exists independently; Ben, by contrast, need never exist at all, depending on what choice we make.

The principle may need refinement, however. Suppose we are faced with a choice between either preventing miserable Moe from coming into existence, or else preventing some lesser harm to already-existing Ann. Since Ann exists antecedently, Partiality Towards the Antecedently Actual would seem to tell us to count her interests for more. Suppose we do, and as a result let miserable Moe come to exist. Now Moe is actual, deserving of full moral consideration, and so in retrospect it looks like our previous discounting of his interests was regrettable. And predictably so.

Is that a problem? One might go either way on this. You could hold that in cases where our attachments change over time, we might reasonably come to later regret actions that were reasonable (indeed, right) at the time (Harman Reference Harman2009). I may rightly favour a friend, only to later befriend the person I previously slighted, and so come to regret the slight. Moreover, the intuitive wrongness of letting miserable Moe come into existence might be adequately explained on prioritarian grounds: If the lesser harm to Ann merely leaves her less well-off, but still generally happy, then we’ve much stronger prioritarian reasons to prevent Moe’s misery (by preventing his coming into existence). This could suffice to outweigh any (plausibly moderate) partiality towards already-existing Ann. If we suppose instead than Ann is antecedently badly off, and the ‘lesser harm’ would leave her in a position just as miserable as Moe’s, then it’s less clear that we do wrong by preventing this harm to Ann, and letting Moe come into existence. Some readers may thus be satisfied by the combination of prioritarianism and simple Partiality Towards the Antecedently Actual.

If you nonetheless think it would be wrong to favour Ann and allow Moe to come into a miserable existence, you could revise the partiality principle to preclude discounting interests in non-existence. We should then give greater weight to both the interests of those who do or will antecedently exist, and the interests of possible people in remaining non-existent, discounting only the interests of possible people in coming into existence.

Would this move be objectionably ad hoc? Perhaps not if you accept a general principle proscribing predictably regrettable actions (call this the Proscription of the Predictably Regrettable).Footnote ⁸ After all, it’s one thing to say that reasonable partial acts might become regrettable in light of unforeseen changes in one’s attachments. But what if the new attachment, and hence the regret, is entirely predictable? Can I reasonably favour a friend over someone who is not yet a friend but (I somehow know) will be in future? Can I provide some benefit to my child, at what I know will be a greater cost to my future (but not yet existent) second child? If not – if avoiding predictable regret is indeed normatively mandatory – then we have principled grounds for revising the Partiality principle as needed to ensure its conformity to this prior normative principle.

The simplest way to ensure such conformity is precisely to disallow the discounting of non-actual persons’ interests in non-existence. For if we do discount them and allow the miserable life to come into existence, then the person in question is no longer non-actual. The effect of discounting such interests undermines the very grounds one had for the discounting in the first place, and predictably so. The Proscription of the Predictably Regrettable thus requires us to instead give full weight to the importance of preventing miserable lives. We are forced to revise the principle of Partiality Towards the Antecedently Actual to only admit of partiality within these independently prescribed constraints: discounting possible people’s interests in existence, but not their interests in non-existence. This then implies the plausible Weak Asymmetry that we have less reason to bring good lives into existence (since, in deliberation, their interests should be partly discounted on grounds of lacking antecedent actuality) than to prevent bad lives (which must be weighted in full, on pain of violating the Proscription of the Predictably Regrettable).

This revision adds an extra element of complexity to the Partiality view. But it does not render it objectionably ad hoc in the way of Variabilism, because the revision was forced on us by broader considerations of principle, rather than just to accommodate our intuitions about particular cases.

Partiality views have other advantages over Variabilism. They avoid the implausibly strong Asymmetry, being compatible with our intuition (from cases like A Distant Realm) that we ought to forsake trivial benefits in order to allow large increases in value from people coming into existence. (As with any view involving partiality, it’s an interesting question just how much extra weight we can properly give to those to whom we’re partial, but I will not attempt to settle such details here.) And there are more subtle differences.

Consider that possible world w where Bob existed all along in place of actual Ann (with the same quality of life, and all else equal). What should our attitude be towards w, when comparing it with the actual world @? According to Variabilism, w and @ are normatively equivalent: each contains a person that isn’t in the other, but who doesn’t exist in the world at which the comparative personal ‘loss’ (of non-existence) takes place, so neither individual counts for purposes of this comparison. But I think it would be natural for those who share the Deeper Intuition to feel rather more attached to actual Ann than that, and so to consider w to be lamentable for its lack of her. (To help amplify the intuition, imagine that Ann is your actual child, and Bob is another child you might have had in her place.) Partiality Towards the Antecedently Actual accommodates this intuition.

Or, even more subtly, consider a couple of different world-pairs. Compare improving Ann’s welfare in @ by a certain amount, to yield an improved world @+, as opposed to improving Bob’s welfare in w by that same amount, to yield the improved world w+. If you think that you ought to more strongly prefer @+ over @, as compared to the strength of your preference for w+ over w, then again this would seem an advantage of Partiality views over Variabilism. (Since the benefited individual, be it Ann or Bob, still exists in the alternative world where they aren’t so benefited, Variabilism gives full and equal weight to both possible improvements.)

Variabilism might seem to do better in certain multiple-option cases.Footnote ⁹ Suppose that you must choose between either (i) creating no-one, (ii) creating Ann at +1 (low but positive) welfare and Ben at +9 (high) welfare or (iii) creating just Ann at +10 (high) welfare. Some might intuitively think that option (ii) is wrong, despite bringing about the same amount of total welfare as (iii) does, due to its being worse for Ann. Option (iii), by contrast, is not worse for anyone who exists in that situation. Variabilism thus easily secures the result that (ii) is a worse option than (iii). Partiality towards the Antecedently Actual does not immediately secure this result because the availability of option (i), where neither exists, ensures that neither Ann nor Ben qualify as antecedently actual here.

If one considers it important to proscribe option (ii), one need not resort to full-blown Variabilism. One may merely appeal again to considerations of regret and ratifiability. If you choose (ii), Ann’s existence gives you pro tanto reasons to regret not having chosen (iii) instead. Option (iii) creates no such grounds for regret. So if we accept a principle that directs us to favour ratifiable over regrettable options, all else equal, then that will suffice to rule out option (ii), as desired. There’s certainly no basis here for further claiming that the benefits of existence morally count for nothing.

We’ve seen that in order to explain the Deeper Intuition without entailing the Asymmetry, one can accept simple Partiality Towards the Antecedently Actual, the revised version constrained by the Proscription of the Predictably Regrettable, or just stick to plain old Prioritarianism. Either of these three options strikes me as superior to Variabilism, and their disjunction even more so. We thus find that the Deeper Intuition provides no support for the Asymmetry. The former is best explained without the latter.