Saturday, March 27, 2021

Stable Actualism and Asymmetries of Regret

Jack Spencer has a cool new paper, 'The Procreative Asymmetry and the Impossibility of Elusive Permission' (forthcoming in Phil Studies).  I found reading it to be really helpful for clarifying my thoughts on the procreative asymmetry.

Back in 'Rethinking the Asymmetry' (CJP, 2017), I argued for two main claims: (i) we have reason to bring good lives into existence, whereas "strong asymmetry" intuitions to the contrary can be explained away; and (ii) the intuition that we should prioritize existing lives is better accommodated by a form of modest partiality towards the (antecedently) actual than by Roberts' Variabilism (or any other strong-asymmetry-implying view).  To avoid incorrectly permitting miserable lives to be brought into existence, I argued, actualist partiality should be supplemented with a principle proscribing the predictably regrettable.

To illustrate (borrowing the evocative names from Jack's examples), suppose that Joy will be happy if created, and Misery will be miserable if created.  We can coherently discount Joy's interest in coming to exist, without this consequently generating new grounds for regret (though if we happen to bring her into existence, we may subsequently be extra-happy about this). By contrast, if we create Misery due to discounting her interest in non-existence, her new status as actual undermines the very basis for our prior discounting.  Our decision, in this case, is predictably regrettable, in a way that casts doubt on the coherence of the moral reasoning that led to it.  So we should instead regard interests (like Misery's) in avoiding bad existence to be not coherently discountable.  This places an important constraint on any kind of actualist partiality, or so I argued.

In his new paper, Jack nicely expands upon this sort of view.  According to his Stability principle, "If p makes it the case that an agent is permitted to choose a, then p would have obtained had the agent chosen a."  Discounting interests in non-existence would be self-undermining in just the way that this principle prohibits, which explains why it is a principled exclusion.  Acts cannot be, in Jack's terminology, "elusively permissible" -- permissible only if you don't actually choose them.

Now, Jack's Stable Actualism is built upon a narrow person-affecting foundation, which I think is a mistake.  But I don't see any principled barrier to weakening this to the (non-absolute) partiality that I prefer. (Jack characterizes his view as "[what] results from subtracting the possibility of elusive permission from strong actualism."  My view was, in effect, that we should subtract the possibility of elusive permission from modest actualist partiality.)

A second point of difference is that Jack's view appeals to (timeless) actuality rather than antecedent actuality.  This makes an important difference to the Joy or Lesserjoy case he discusses in section 9. (As you might have guessed, Lesserjoy is someone who, if created, will be merely moderately happy.) On Jack's view, given a choice between creating either Joy or Lesserjoy, "we're obligated to create whoever we create." (p. 34) 

While I agree that you should subsequently be glad of your decision, whichever person you create, I don't think that shows the choice was truly right. I instead agree with Liz Harman that subsequent attachments can make such gladness fitting without justifying the prior act (contra Jack's claim on p.36, there doesn't seem anything "odd" about this -- it's just how we should expect subsequently-formed attachments to affect things!).  All else equal, it would seem perverse -- irrational from a moral point of view -- to choose, or antecedently prefer, to create the less-happy life.

On p.37, Jack argues that those who accept the (strong) procreative asymmetry should reject the claim that it's wrong to choose Lesserjoy over Joy, since they think it's permissible to choose Nothing over Joy, and Lesserjoy over Nothing.  But, as noted above, I think we should reject the strong asymmetry. If creating Joy would truly be costless, then it would be straightforwardly wrong not to create her: the quality that her life would have provides some reason in favour of creation, and there is (by stipulation) no reason at all against it.

But these are minor disagreements.  Something I found especially helpful was the discussion in section 10 of cases -- like Misery or Equalmisery -- in which every option is predictably regrettable.  This poses a problem for principles like my proscription of the predictably regrettable. One could, as Jack notes, opt for the (dilemma-tolerant) "hardline" view that both options are impermissible. But I prefer Jack's "hierarchical" alternative, according to which agents should act so as to minimize regret (i.e., the distance between the value of their choice -- conditional on making it -- and that of its best alternative).

Anyway, I highly recommend that anyone interested in these issues check out the full paper!


  1. I have a few comments on this general topic. Feel free to delete if I'm going too far astray!

    1) It seems like a lot of analysis in this area "rests" on a distinction between people that one causes to exist and people who one merely "allows" to be created or even people who exist "independently" of you. I guess doing/allowing has been written about plenty elsewhere in ethics, but I think it's often a tough distinction to maintain, though an intuitive one in this context. This gets trickier when you talk about situations where someone could have created a new person but didn't or prevented a new person from being created. The word "situations" is misleading here since this is basically all the time for many individuals (and even more so if especially if you're government). I guess Parfit's book talked about this.

    2) Relating to point 1, there seem to be two types of uncertainty: non-causal and causal: You might be uncertain who will exist because you are uncertain about how things (more-or-less) beyond your control will play out. That's non-causal uncertainty. Causal uncertainty is where you are uncertain who will exist because you haven't decided what to do yet and your decision will determine who will exist. Hedden and Hare (2016) relates to causal uncertainty (though their examples are adoption and stuff, not procreation).

    3) Things get even trickier when we think about uncertainty. (This paragraph is a bit "out there".) Suppose we think of "actualism" as being concerned with both current people and actual future people but not with "merely-possible-not-actual" people (while possibly deemphasizing the causal/non-causal distinction). Then it seems that uncertainty about who will exist in the future should be treated similarly to uncertainty about who currently exists. But I think its highly plausible that people that we think might currently exist shouldn't be discounted because we are uncertain about their existence per se; that is, they shouldn't be discounted any more than usual expected value-style reasoning would recommend. For example, suppose we know that there are people trapped in at least some of a number of caves but we know neither their "identity", total number, nor the breakdown between different caves. I don't think we should discount these possible people because we are (in some sense) uncertain about their existence (again, any more than by usual probabilistic reasoning). But then, if actual (as opposed to merely possible) future people are on similar status with current people, how can we discount future people just because we are uncertain they will exist.

    4) I think there's a lot of interesting philosophy of language (sense and reference), logic (e.g. temporal and modal), and metaphysics (I guess Parfit did some) here. When I was looking at this, Nuel Belnap seemed impressive on logic relating to these issues. I hope more people write about these things.

  2. Richard,

    Looks like we agree in the main!

    One point on which we part ways is the connection between retrospective gladness and permission. If one should be glad to have created Lesserjoy instead of Joy if they have created Lesserjoy, then I think that it was permissible for them to have created Lesserjoy. 

    We might just have a verbal dispute. We can distinguish weak gladness, which does not entail preference, from strong gladness, which does. If I am choosing between $5 and $10, and choose $5, then weak gladness is fitting. After all, something good happened: I’m $5 richer! But strong gladness is not fitting, since the chosen option is disprefered to some unchosen option. 

    Weak gladness may be fitting after making an impermissible choice. But if strong gladness is fitting after an impermissible choice, that means the preference order of options changes: that the chosen option was dispreferred to some unchosen option at the time of choice but is so no longer. 

    I think that strong (moral) gladness is fitting if we have created Lesserjoy instead of Joy. Do you agree?

    If so, I do think the view is odd---unless it’s coupled with some view on which the future doesn’t yet exist. I can see why the preference order of options might differ from world to world depending on who exists at that world---I can see why the difference between existence and nonexistence might matter morally. But I can’t see why the preference order of options might differ from time to time---I can’t see why a difference in temporal location might matter morally. 

    The temporal view also leads to some awkward questions. If (because of special relativity) it is a relative matter whether Lesserjoy exists, is it a relative matter what the preference order of options is? If Lesserjoy “goes on” to be a backward time traveler, was the choice actually permissible at the time of choosing?

    1. Yeah, my thought was that our preferences should change in these cases -- that our partiality towards the actual should only kick in once Lesserjoy exists (or, better, once it is settled that, whatever choices people make, Lesserjoy will exist).

      It's not that temporal location matters. Rather, it's that agents must regard it as a deliberatively open question what choice they will make, and so -- prior to making the choice -- the agent must regard it as a deliberatively open question whether or not Lesserjoy will be actual.

      (Don't you get the intuition that it would be perverse to deliberately choose to bring about the worse life rather than the better life?)

      It could be that I'm more drawn to a "perspectival" understanding of normativity (treating the agent's available evidence and perspective, e.g. expected value facts, as what's relevant to determining permissibility) whereas you're assuming a more "objectivist" understanding (appealing to the actual facts, including about the future). That could make sense of our disagreement, as it would seem to make sense that bias towards choice-contingent future people would be "objectively" warranted even though rationally or "perspectivally" unwarranted.

      P.S. I don't think you should be weakly glad to have chosen* $5 rather than $10 (nor should anyone "weakly regret" choosing $10). Money is fungible, so the greater amount cancels (rather than merely outweighs) the reasons there would otherwise be for choosing the smaller amount. To get proper pro tanto (without overall) gladness or regret, you need a tradeoff between non-fungible goods. Like, getting 5 paintings instead of 10 superior sculptures, or something like that. (See 'Value Receptacles' for more detail.)

      *: You may, of course, be glad to have received the choice in the first place, and so found your subsequent position improving by at least $5. But I take this to be different from being glad, in any way, that you picked the lesser amount!

    2. The time-traveller case is fun, though. I guess Oracles, too, could similarly mess with our standard assumptions about what should be considered deliberatively open. Not sure what to think about all that! Thanks for bringing it to my attention.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)

Note: only a member of this blog may post a comment.