THE PROCREATION ASYMMETRY

(1)

THE PROCREATION

ASYMMETRY

The Existence-requirement Strategy and some

Concerns on Incompatibility

Jesper Söderstedt

Supervisor: Per Algander Examiner: Jan-Willem van der Rijt

Master Thesis (30 ECTS) Umeå University, Department of Historical, Philosophical and Religious Studies

(2)

Abstract

According to the procreation asymmetry there is no moral reason to create a new and foreseeably happy person just because this person will be happy, but there is however a moral reason against creating a new and foreseeably unhappy person just because this person will be unhappy. A common way to defend this conjunction of claims is by employing a so-called existence-requirement, according to which the happiness of a given person p in a world w depends on it being possible to understand p as an existing person in w. The aim of this paper is to consider whether this existence-requirement strategy is compatible with other intuitions and convictions held in normative moral theory and populations ethics. This aim will be achieved by considering whether the existence-requirement strategy is compatible with three plausible desiderata for a solution to the (normative) procreation asymmetry. Although some solutions to some potential incompatibilities are suggested, the thesis that will be argued for is that every instance of the existence-requirement strategy is incompatible with at least one of the relevant desiderata. Where the common denominating incompatibility for all instances of the existence-requirement strategy is to be found in an inability to be sufficiently action-guiding.

Abstrakt

(3)

(4)

Introduction

1

Consider the possibility of being able to foresee whether or not one’s child would be happy or unhappy before deciding to reproduce. In what sense would the happiness of one’s child have an impact on the decision to create a new child? Conversely, in what sense would the unhappiness of one’s child have an impact on the decision to reproduce? It seems fairly uncontroversial to claim that if one’s child would be foreseeably unhappy; this would constitute a reason against creating this child. This while Jan Narveson’s (1976, 73) famous claim that although there is a reason to make people happy, one ought to remain neutral about creating new happy people possess great initial intuitive appeal. Consequently, it similarly seems counterintuitive to claim that the fact that one’s child would be foreseeably happy is a moral reason or even morally obligates one to create that child.2

Despite the fact that both of these claims are intuitively appealing, certain theories and theoretical convictions highlights a tension between them. This in the sense that the justification of one of the claims seems to suggest the negation of the other. For example, one way to justify that one should not create an unhappy child is because it will negatively contribute to the total happiness in the world, a total happiness which one ought to maximize. This totalist utilitarian view is however incompatible with the claim that there is no reason to create a new happy child since this child would contribute to the total happiness in the world. It could thus be considered being a tension between the two relevant claims from this particular perspective. A similar way to illuminate this apparent tension is by means of the observation that agreeable and disagreeable experiences such as happiness and unhappiness are ontologically similar in a way that justifies them entering into parallel and symmetric moral principles (Rachels, 1998, 105). Why would goods such as happiness not constitute a reason to cause a person to exist and evils such as unhappiness constitute a reason against creating the person? As Jeff McMahan (2002, 300) notes, in the abstract it seems that it should be perfect symmetry here. Although, in virtue of the relevant claims it could be understood as being a tension to be found in the procreation asymmetry, it is nonetheless commonly understood as part of common-sense morality and it is

1_{I am very grateful for all the help that I have received in the process of writing this paper. In particular the}

advice and critique received from my supervisor Per Algander immensely helped me to improve the coherence and general quality of this paper.

2_{In a sense, the asymmetry that is mentioned here partly consists of yet another asymmetry. Namely that there is}

(6)

2

furthermore a widely held view in populations ethics.3 This combination of an apparent tension found within the procreation asymmetry and its intuitive appeal constitutes a very good reason for further theoretical consideration.

The relevant view was first discussed and endorsed by Narveson (1967; 1976) and later denoted by McMahan (1981, 100) as ‘the asymmetry’ (henceforth the procreation asymmetry). The procreation asymmetry can be understood as the intuition that there is a significant difference in terms of degree between the reasons one has to create a happy person and the reasons one has to create an unhappy person. Or even more radically that there is no reason to create a happy person and a reason not to create an unhappy person. The procreation asymmetry is notoriously hard to defend and this despite its very strong initial intuitive appeal (see for example McMahan, 1981, 2009; Singer, 1993; Persson, 2009).

In the literature concerning the procreation asymmetry a common way to provide a defence of this asymmetry is by means of an employment of an existence-requirement, a strategy that in this paper will be denoted as the existence-requirement strategy. What this strategy generally implies is an explanation of the moral significance of the happiness/unhappiness of a persons whose existence is dependent on the performance of the action under consideration in relation to a claim about this dependently existing person’s existence. For example, the relevant strategy will in some sense rely one a claim similar to the following: the fact that a person p exists in a particular world w makes the fact that p is happy in w morally significant.4 There are different ways of employing such a strategy and what distinguishes one employment of the strategy from the next generally comes down to two questions: (i) what does it mean to say that p exists in w? And partly depending on one’s answer to (i), (ii) is the moral significance of happiness or unhappiness dependent on p existing in the relevant way? A way to make the relevant strategy clearer is to determine its place within population ethics in general. The existence-requirement strategy could generally be understood as a strategy that in some sense depends on or at least is inspired by a person-affecting restriction of some kind. Very generally speaking, according to such a restriction the proper referent of moral claims concerning happiness or unhappiness (or sometimes both) are understood as being someone, i.e. a person that could be understood as, in some sense, existing.

3_{See, among others, (Narveson, 1967;1976; Roberts, 2011a;2011b; Cohen, 2019; Earl, 2017).}

4_{What will be meant by “morally significant” throughout the paper is basically that whatever have moral}

(7)

3

Populations ethics is a notoriously hard subject of inquiry and often contradictions and incompatibilities are found between the intuitions and convictions commonly held within populations ethics, normative moral theory, and philosophical theory in general (Parfit, 1984; Temkin, 2012). The aim of this paper is to consider in what sense the existence-requirement strategy is yet another instance of an attempt trying to accommodate for a particular intuition that ends up displaying such incompatibilities and contradictions. The way this aim will be achieved is by considering the compatibility of the existence-requirement strategy with three plausible desiderata for a solution of the (normative) procreation asymmetry. This will be done by considering whether three (or perhaps two depending on how one sees it)5 instances of defences relying on an employment of the existence-requirement strategy are compatible with the relevant desiderata. In particular the questions that will be asked to achieve the relevant aim is whether (i) the particular employment is incompatible with the intuition that there is not a reason against creating happy person? I.e. is it compatible with anti-antinatalism? (ii) is the particular employments of the existence-requirement strategy compatible with the intuition that there is a reason to create the happiest person possible when it is inevitable that one will create a new person and the happiness of the relevant person or persons is entirely transparent to the agent? And finally (iii) will the particular employment of the existence-requirement strategy be able to guide the actions of an ideal agent? The thesis that will be considered and ultimately defended is that the existence-requirement strategy, as it has been previously formulated, is incompatible with at least one of the three plausible desiderata, where the common denominating incompatibility of all defences employing the existence-requirement strategy is an inability to be sufficiently action-guiding.

The paper will be structured in the following way: In the first section different interpretations of the procreation asymmetry will be considered, and the definition and interpretation of the procreation asymmetry employed in this paper will be presented and briefly argued for. In the second section the three desiderata will be presented in detail and furthermore their intuitive appeal and general plausibility will be considered. The third section considers the compatibility of three (or two) ways of employing the existence-requirement strategy to defend the procreation asymmetry with the stated desiderata. First the compatibility of the harm-benefit defence and its two variations with the relevant desiderata will be assessed, and second the

(8)

4

compatibility of the conditional reasons defence with the same desiderata will be considered before concluding.

1. Defining the Procreation Asymmetry

When defining the procreation asymmetry, it is initially instructive to acknowledge a distinction between two ways of expressing the procreation asymmetry: the strong and the weak asymmetry (McMahan, 2009, 57). According to the weak asymmetry the happiness of a dependently existing person gives one a weaker reason to act than the unhappiness of a dependently existing person. The weak asymmetry can be spelled out by means of the following conjunction of claims:

W-unhappy: There is a relatively strong moral reason not to create an unhappy person. 6

W-happy: There is a relatively weak moral reason to create a happy person.

The weak asymmetry is prima facie plausible and is furthermore interesting in its own right, however what this paper will concern itself with is the arguably more controversial7 strong asymmetry (henceforth the procreation asymmetry if not stated otherwise) which can be formulated in the following way:

S-unhappy: There is a moral reason against creating an unhappy person.

S-happy: There is not a moral reason to create a happy person.

Although somewhat clear and limited in scope, further questions could be asked about how the procreation asymmetry will be understood. In what sense can other things than the happiness of the dependently existing person account for the conjunction of S-unhappy and S-happy? Furthermore, in what sense should one understand the asymmetry as saying something explicitly about the value of the world or what one has a moral reason to do? Such questions make it necessary to introduce two further distinctions between the fundamental and contingent procreation asymmetry and between the evaluative and normative procreation asymmetry. First, following Michael Tooley (1998) and Johann Frick (2014) the evaluative asymmetry will be understood as concerning itself with how the value of populations will be affected by the

6_{Although the term ‘person’ is employed, it is not meant to exclude beings that sometimes are considered as}

non-persons that arguably have a moral status such as non-human animals. For all intents and purposes, the term ‘person’ can be understood as much more including than it usually is, i.e. as a shorthand for ‘any individual with a moral status’.

7_{Some philosophers who deny the justifiability of the strong asymmetry are open to the weak asymmetry being}

(9)

5

creation of an additional happy person. The normative asymmetry concerns itself with whether or not the happiness of a happy dependently existing person will constitute a source of moral reasons. The main concern of this paper will be whether or not the normative asymmetry can be defended and consequently the evaluative asymmetry will only be relevant if it ought to be understood as a crucial underpinning of the normative asymmetry by a particular defence of the normative asymmetry. 8

Second, some philosophers (Hedberg, 2016; Yetter-Chappell, 2016) have argued that the appeal of the procreation asymmetry is merely contingent rather than fundamental, i.e. the appeal of the procreation asymmetry is contingent on circumstances other than the happiness of the dependently existing person in question. When considering the intuitive appeal of the procreation asymmetry, they argue that what is really at stake are other values and reasons than the unhappiness/happiness of a dependently existing person (such as bodily autonomy and the demandingness of morality). Consequently, what such philosophers argue is that under normal human procreative circumstances, the conjunction of S-happy and S-unhappy are defensible, but when these circumstances are stripped away, the conjunction loses its intuitive appeal. However, this seems initially rather implausible since the fundamental procreation asymmetry is deemed as intuitively appealing by many (the author of this paper included, but see also: Roberts, 2011a, 2011b; McMahan, 1981, 2009). Furthermore, the concerns of Trevor Hedberg (2016) and Richard Yetter-Chappell (2016) might provide a critique of one of the initial reasons for why one should defend the fundamental procreation asymmetry (although, arguably, implausibly so), i.e. its intuitive appeal. But they do not show that it cannot be coherently defended, something that arguably is of intrinsic theoretical interest.

One of the main reasons for why Hedberg (2016) and Yetter-Chappell (2016) can appeal to a contingent asymmetry is because they seem to primarily interpret the procreation asymmetry in terms of all-things-considered reasons. However, by posing this paper’s research problem in terms of pro tanto reasons, concerns of what one has all-thing-considered reason to do are bracketed. This furthermore neutralizes objections toward S-Happy appealing to the value of the continuation of humanity. If what is of concern is whether the dependently existing person’s

8_{How one perceives the plausibility of the normative asymmetry can have implications for how one perceives}

(10)

6

happiness is a source of pro tanto moral reasons, it is perfectly consistent to agree with S-Happy while also claiming that there is a reason to create a foreseeably happy person in virtue of the value of the continuation of humanity. Consequently, by formulating the procreation asymmetry in terms of pro tanto reasons, it will be an articulation of the fundamental rather than the contingent procreation asymmetry.

Before presenting the final version of the procreation asymmetry that will be considered in this paper, it is furthermore necessary to specify what is meant by “unhappy” and “happy”. If not stated otherwise, “unhappy” and “happy” will be understood in terms of how the prudentially good things compares with the prudentially bad things in a given person’s life. Consequently, a happy person is happy if the prudentially good things outweigh the prudentially bad things. While an unhappy person is unhappy in virtue of the fact that the prudentially bad things outweigh the prudentially good things in this person’s life.

Given the considerations above, the procreation asymmetry that will be discussed in this paper will be defined as follows:

Unhappy: The unhappiness of a foreseeably unhappy dependently existing person is a source of pro tanto unconditional moral reasons against creating the unhappy dependently existing person.

Happy: The happiness of a foreseeably happy dependently existing person is not a source of pro tanto unconditional moral reasons for or against creating the happy dependently existing person.

The formulation of the procreation asymmetry above, besides from being compatible with the claim that one still can have an all-things-considered reason to create the happy dependently existing person, is also compatible with the claim that it is morally permissible to create the person in Happy. The claim is that the happiness of the dependently existing person in Happy is not a source of pro tanto moral reasons, which means that it will not be a source of reasons for or against creating the person in question. All other things being equal, according to Happy, the creation of the happy dependently existing person is thus morally permissible but not morally required.

(11)

7

(Roberts, 2011b; McMahan, 2002, 2009). It is considered as either ‘probably impossible to dislodge’ (McMahan, 2002, 300) or as an important constraint on any adequate moral theory (Roberts, 2011b, 334) thus constituting a reason to further consider it.

2. Desiderata

2.1. Anti-Antinatalism

An intuitively appealing aspect of Happy is that it accounts for the claim that one ought to consider the creation of happy dependently existing people as morally permissible. This since, the happiness of a dependently existing person does not constitute a source of moral reasons for or against creating this person. It is however possible to perceive of a risk that any defence trying to prove that indeed the dependently existing person’s happiness is not a source of pro tanto moral reasons, will go too far and imply that one generally has a reason against procreating. In particular what is considered to be frequently occurring facts about the general person’s state of happiness or the act of procreation, is something that the defence in question will consider as a reason against creating the happy dependently existing person. It seems, for example, very likely that the happiness of the person p in Happy does not entirely consist of prudentially good things. It will almost always (if not always) be the case that in a happy person’s life there are some prudentially bad things that are outweighed by the prudentially good things. In virtue of this consideration it is important that the bad things in a happy person’s life are not considered sufficient reason to not create the happy dependently existing person. Briefly put, it is important that a defence of the procreation asymmetry does not imply antinatalism, and this is primarily important in virtue of two reasons.

First, it would seem that under normal empirical circumstances (i.e. that even a happy person will experience some prudentially bad things), a theory implying antinatalism would be incompatible with Happy. Since a crucial aspect of Happy is that the happiness of the happy dependently existing person is not a source of moral reasons, which means that it cannot be a reason for or against creating the relevant person. Thus, if one has the ambition to provide a defence that is applicable to circumstances that are so common they can almost be perceived as necessary, it is important that it does not imply antinatalism in the relevant circumstances.9

9_{Here it is possible to anticipate a worry that this desideratum would constitute a departure from the ambition to}

(12)

8

The second reason why antinatalism is considered as an undesired implication is the fact that it is counterintuitive. This by being (i) intrinsically counterintuitive, meaning that most would find the idea itself, i.e. that it is a reason against procreating even if the dependently existing person would be happy, highly counterintuitive. This while (ii) antinatalism have counterintuitive logical implications. One of these implications is that if everyone would take seriously that one, under normal circumstances, have a reason against procreating even when the person is happy slow extermination would follow. Although proponents of antinatalism does not necessarily view this as something that is morally problematic (Benatar, 2006, Ch. 6 esp. 182-200), it should be plausibly considered as something one should avoid in virtue of its sheer counterintuitiveness. One way to make this point is by arguing that the lives of individuals partly are granted meaning and value in relation to future generations (Scheffler, 2018), if so, it seems that there is at least a great value in the continuation of humanity for present people. Of course, this will not settle the debate on the value of future generations or what one has moral reason to do with regards to these future generations. However, what it does point to is that there is something that is, initially, highly counterintuitive with antinatalism. As noted in previous sections, one of the main reasons for why one defends the procreation asymmetry is because of its initial intuitive appeal (see also Roberts, 2011b, 772). If the implications of a defence are as unappealing as the procreation asymmetry is intuitively appealing, it will partly defeat the purpose of defending the procreation asymmetry. 1011 This is not to say that when weighing the relevant intuitions, it will not be the case that antinatalism will be considered an acceptable price to pay in order to defend the procreation asymmetry. But it seems nonetheless like a good initial desideratum to avoid counterintuitive implications such as antinatalism to avoid detracting from the intuitive appeal of the procreation asymmetry.

2.2. The Reproductive Maximizing Principle

The non-identity problem was coined by Derek Parfit (1984, Ch 16)12 and denotes the problem of explaining why it is wrong to create a not as happy person when one has the opportunity to create a happier but different person. A similar but nonetheless different case is the so-called same-person case, involving making the same, but previously non-existing person, happier. The

10_{Rivka Weinberg (2012, 28) for example argues that the implications of the axiological asymmetry (see section}

3.2.), i.e. antinatalism might be more counterintuitive than the axiological asymmetry is intuitively appealing. Arguably the same will hold for any defence of the procreation asymmetry that implies antinatalism, such as the axiological asymmetry.

11_{It is of course true that it will still be of intrinsic theoretical interest as noted in section 1.}

12_{Although the problem or at least the tension that constitutes the problem were identified somewhat earlier}

(13)

9

latter thus involves a decision of actualising two separate worlds containing the same person but with different levels of happiness in the different worlds, while the former involves a decision of actualising two worlds containing two different persons with two different levels of happiness. In this paper the same-person case and the non-identity case will serve as ways of proving the plausibility of the principle that will serve as one of the relevant desiderata that captures the intuitions relevant for these cases. The cases that will be considered are what will be denoted as the good-bad case and the good-better case, the former illustrated in figure 1 and the latter illustrated in figure 2. In the figures below both same-person cases and non-identity cases will be displayed in the same figure illustrating the respective version of the non-identity case and the same-person case.

Consider the good-bad case as illustrated in figure 1 below, where the agent ax can be

understood as occupying a world wx and from wx ax have the opportunity to actualise one of

two possible worlds w1 and w2. For some reason, ax will inevitably become pregnant but ax have

the opportunity to choose which person ax ends up creating or in the same-person case, how

happy the relevant person will be. In the non-identity version of the good-bad case if ax chooses

to actualise w1 ax will create the happy person p1 and if ax chooses to actualise w2 ax will create

the unhappy person p2. While in the same-person version of the good-bad case if ax actualises

w1 ax will create the happy p and if ax actualises w2 ax will create the unhappy p.

Figure 1 (The good-bad case)

w1 (non-identity: happy p1; same-person: happy p)

ax (at wx)

w2 (non-identity: unhappy p2; same-person: unhappy p)

It is plausible to assume that most have the intuition that ax, in the non-identity version of the

good-bad case, have a strong moral reason to actualise w1 rather than w2 and thus create the

happy p1 rather than the unhappy p2. The same result would arguably hold for the same-person

case version of the good-bad case. It might even be possible that the intuition that ax ought to

actualise w1 rather than w2 is even stronger in a same-person case than in a non-identity case.

Now consider the good-better case and figure 2 illustrating it, where the same conditions hold as does in the good-bad case, except that ax can actualise one of two, other, possible worlds w3

and w4. In the non-identity case version of the good-better case, if ax chooses to actualise w3 ax

(14)

10

person p4. While in the same-person case version of the good-better case if ax actualises w3 ax

will create p in a happy state and if ax actualises w4 ax will create p in an even happier state.

Figure 2 (The good-better case)

w3 (non-identity: happy p3; same-person: happy p)

ax (at wx)

w4 (non-identity: happier p4; same-person: happier p)

When considering the non-identity case version of the good-better case it is plausible to claim that ax have a moral reason to actualise w4 rather than w3 due to p4 being happier than p3.

Although this intuition is perhaps not as strong as the one pointing ax in the direction of w1

rather than w2, it is still sufficiently strong to be an intuition one should seriously consider.

Arguably, also here the intuition is even stronger when considering the good-better case as a same-person case, if one will inevitably create a person p, it seems that, all other things being equal, one ought to create the happiest possible version of p.13

A way to capture the intuitions described in relation to both the better case and the good-bad case (in both variations) is by means of a principle according to which, when one inevitably will create a person, there is a moral reason to create the happiest person possible, the reproductive maximizing principle (RMP):

RMP: When it is inevitably the case (as in the good-bad case and the good-better case) that an agent ax will create a person p14 and the happiness of p is fully transparent to ax, there is a

contrastive pro tanto moral reason for ax to create the happiest person that ax can possibly

create.15

13_{It is possible to note that the claim that is being made here in relation to the good-better case is weaker than}

Parfit’s (1984, 366-369) “no-difference view”, according to which the identity of the person in the good-better case does not make any difference and thus there is no difference between the non-identity case and the same-person case versions of the good-better case. Rather the claim being made here is that there could be a

difference, but the difference is a difference in degree. Exactly how this difference ought to be explained cannot be accounted for here, but it seems sufficient to conclude that it is nonetheless intuitively appealing. Even if one does not think so and instead thinks that the no-difference view is more plausible, it does not affect the

plausibility of the forthcoming reproductive maximizing principle, since it will hold regardless whether one thinks there is no difference or whether there is a difference but only in terms of degrees between the non-identity case and the same-person case.

14_{As alluded to by the discussion and inclusion of both non-identity cases and same-person cases ‘p’ should be}

interpreted in the de dicto rather than the de re sense, i.e. ‘p’ does not denote a particular person but is a place-holder for whatever person that ax ends up creating.

15_{Frick (2014, 51) suggests a similar principle although it does not include the stated epistemic condition about}

(15)

11

Julian Savulescu (2001, 415) suggest a principle similar to RMP, according to which one should, based on available information, create the person that is expected to have the best life possible. In this sense Savulescu’s suggestion involves a prima facie moral obligation for prospective parents to create the best possible person that they can create. An objection against Savulescu’s principle is that even when one knows about the genetic predispositions of a given person, it is nonetheless impossible (or at least near impossible) to tell whether that person will be happy or not. This since, as Michael Parker (2007, 281) argues, any notion of the good-life will display complexities that parents cannot, despite available technology, foresee as part of the dependently existing persons future life. 16 However, RMP is very limited in the sense that it only claims that when it is fully transparent what level of happiness this person will enjoy, is there a pro tanto reason to create the happiest person possible. Something that furthermore highlights a crucial difference between RMP and Savulescu’s principle, RMP is highly abstract and is not meant to fully capture the real-life decisions of prospective parents. It does not, for example, imply, as Savulescu (2001, 425) seem to argue his principle does, that prospective parents are morally required to conduct a “[…] selection for non-disease genes which significantly impacts well-being”. This kind of selection involves a level of uncertainty and complexity that implies that the transparency condition of RMP is violated. Only when the happiness of the relevant person is entirely transparent and thus free from any form of uncertainty does RMP apply.

It is furthermore important to note, that there is arguably an important distinction between the conception of moral reasons that is employed in RMP and moral obligations. RMP does not necessarily entail a ground for a moral obligation, this since the pro tanto moral reasons that holds for RMP are not necessarily of a requiring strength (although they can be). Here it is relevant to introduce a distinction regarding the strength of normative reasons proposed by Joshua Gert (2003, 2016), the distinction between requiring and justifying strength. Where a reason with justifying strength can make a previously impermissible option permissible, while a reason with requiring strength would make its corresponding action required. RMP is very limited in the sense that in order to fulfil it, all that is necessary is that the relevant theory could understand the fact that an action would entail creating the happiest possible person as something that could make an otherwise impermissible action permissible. Thus, in conclusion,

(16)

12

the fact that RMP is very limited while intuitively appealing, makes it a plausible desideratum that many would accept.

2.3. Action-guidance

Two general subjects of enquiry for moral theory is providing a theory of value and a theory of right conduct i.e. normative moral theory. As already indicated by the choice to consider the normative rather than evaluative procreation asymmetry, this paper primarily concern itself with the latter and this is what ‘moral theory’ henceforth will be understood as denoting i.e. moral theory will be understood as a short hand for normative moral theory. A common view is that it is crucial for a moral theory thus understood to possess the ability to guide our actions (see for example Vranas 2007).17 If such a theory prescribes that an agent ax have a moral reason

to do something, it is commonly considered a weakness of the relevant moral theory if this moral reason cannot guide the actions of ax. Such an action-guidance requirement can be

defined and understood in several different ways, but, following Erik Carlson (2002,72-75) it will generically be understood as meaning that a moral theory x can tell an agent ax what do to

assuming that ax is motivated to act in a way that is compatible with what x prescribes. This

action-guidance requirement will consequently be weaker than an action-guidance requirement prescribing that a moral theory need to be able to motivate agents. This choice is justified since if a moral theory cannot even prescribe actions to motivated agents, the fact that it cannot motivate these agents is of secondary concern.

In order to further weaken the relevant action-guidance requirement, not only should ax be

motivated to act in accordance with the prescriptions of x, it also need to be the case that ax

have the cognitive ability to process and understand what x prescribes. While furthermore ax

need to possess the ability, from the relevant empirical information, to infer what x prescribes, as long as these are in principal epistemically accessible to ax. What this means is that ax is

connotatively and cognitively ideal:

Action-guidance requirement: A moral theory should be able to guide cognitively and conatively ideal agents in their actions.

17_{Undoubtedly the most famous variation of this claim is Immanuel Kant’s “ought implies can”. It is however}

(17)

13

The reasoning behind this assumption is similar to the reasoning presented above. If the idealized agent will be unable to act in accordance with what a given moral theory prescribes, it is of secondary concern that it cannot guide a non-idealized agent.

The action-guidance requirement above only stipulates that a theory need to be able to guide the actions of an idealized agent ax. A theory can however fail in doing so in several ways, but

with regards to the purposes of this paper, there are two ways in particular that are of interest. First a moral theory x will fail to guide the idealized agent ax actions if it is impossible for ax to

comply with what x prescribes. Consequently, x will fail to guide the actions of ax if it fails to

satisfy the following satisfiability condition (SC):

Satisfiability condition (SC): For any idealized agent ax and any possible situation s, there is

an action φ such that if ax were to perform φ in s then ax would conform with the given moral

theory (Bykvist, 2007a, 116).18

One way a moral theory x would violate SC is if it would for example imply in a situation s that if an agent ax performs an action φ it is wrong for ax to perform φ. However, if ax does not

perform φ it is wrong for ax not to perform φ. What this means is that ax cannot comply with x

in s and if so, ax cannot use x in a moral deliberation about what to do in s, since no matter what

ax does in s, it is dictated by x that ax cannot use the relevant deliberation to act in compliance

with what x prescribes ax to do at s (Bykvist, 2007a, 117).

The second relevant way a moral theory x can fail to guide the actions of ax is if the prescriptions

of x are not epistemically accessible to ax at the time of the action, i.e. x fails to satisfy a

condition of epistemic accessibility (EA):

EA: The prescriptions of a moral theory x and the facts relevant for these prescriptions should in principle be epistemically accessible to any idealized agent ax in any situation s

where the prescriptions are relevant for the actions of ax.

18_{The satisfiability condition SC is not an uncontroversial requirement, this since it implies a rejection of moral}

(18)

14

As with SC there are several ways that EA can be violated, but one example is if in order to know what the prescriptions of x are, ax need to have knowledge about facts that necessarily

only are available to ax at a time that is subsequent to the time after ax have decided whether or

not to perform the relevant action. For example, a given prescription concern an action φ that ax can perform at a time t however, in order to know whether or not ax ought to perform φ x

requires ax to have knowledge about facts that are only available at a time t1 that is characterized

by being a time subsequent to ax decision whether or not to to perform φ.19 Another way that

EA can be violated is if it is impossible to discern for the idealized agent ax what x prescribes

in the moral deliberation of ax, this by the prescriptions being incoherent in virtue of, for

example, producing inconsistencies.

It might be possible to object to this requirement and argue that it seems too strict, or at least that, when considering certain moral judgements, action-guidance does not seem that crucial. However, it seems very plausible that a moral theory should be able to guide an idealized agent’s actions, indeed it is easy to agree with Bykvist’s (2007a) claim that:

[…] if a theory cannot even be used by agents who are ideal, then it seems to be devoid of all practical relevance, and one might wonder how it could be properly called a normative theory20_{– a theory that is supposed to tell you what to do (Bykvist, 2007a,}

112).

But perhaps a yet more convincing argument is that in order for moral concepts such as blameworthiness and praiseworthiness to make sense, an ideal agent’s ability to perform what morality prescribes is necessary, concepts that in turn have a conceptual connection with notions of rightness and wrongness (Österberg, 1998, 282). Furthermore, concepts such as rightness and wrongness can properly hold for entities that one, by means of voluntary actions, can choose to actualise or not to actualise . I.e. rightness and wrongness serve the function of evaluating events one is able to choose between, and provide guidance in that process, thus separating it from concepts such as goodness and badness (Smith, 1986, 342). In relation to this, it is furthermore important to note that the claim only is that normative notions such as rightness and wrongness have a connection with this action-guidance requirement. It is entirely possible that one’s notions of goodness and badness do not guide the ideal agents’ actions,

19_{It is however possible to reject that this kind of foreknowledge would be incompatible with action-guidance}

(see Carlson, 2002) although it seems plausible that it is and its plausibility will be assumed in this paper.

20_{Bykvist (2007a) uses the wider term “normative theory” that also captures theories of rationality. However,}

(19)

15

however if these are assumed to have some impact on notions of rightness and goodness they ought to be. In virtue of the relevant considerations, there is reason to consider the action-guidance requirement justified.

3. Defending the Procreation Asymmetry

Several attempts have been made to defend the procreation asymmetry, most of which seem to be a variation of the particular defence that is of interest to this paper, i.e. defences that in some sense relies on the strategy of stipulating an existence-requirement for the moral significance of happiness/unhappiness. 21

Exactly what is meant by the relevant strategy of defending the procreation asymmetry is perhaps not fully clear as of yet. But as already briefly explained in the introduction above, the relevant strategy often contains a claim similar to the following:

Generic existence-requirement (GER): The fact that p is happy/unhappy in w has moral significance iff p exists in w.

In addition the strategy could be understood as depending on or at the very least making use of a person-affecting restriction very generally speaking, i.e. the proper referent of moral claims and reasons are always someone, i.e. a person that could be understood as existing in some sense. The relevant characterization of the person-affecting restriction is of course very wide and purposely so. This since exactly how it is understood varies between particular employments of the existence-requirement strategy and there is furthermore not necessary to consider the relevant particularities of the different variations of the existence-requirement strategy in this section. It might however be useful to explain how they can vary. As already mentioned, how the relevant strategies vary depends on how they will answer the following two questions in relation to GER: (i) what is meant by p existing in w? Furthermore, (ii) does the moral significance of happiness or unhappiness (or both) depend on p existing in w?

One way to answer (i) and (ii) is to give the actualist answer, where the answer to (i) is that p exists in w if p is actual in w i.e. p does, will exist or have existed in w, while the answer to (ii) is that the moral significance of both happiness and unhappiness depends on p being actual in

21_{It is however important to note that not all do. Ingmar Persson (2009) for example suggests that the}

(20)

16

w.22 Narveson (1976, 73) suggests such an explanation of the underlying intuitions of Happy, i.e. that one ought to consider the creation of new happy people as morally insignificant (or neutral). Furthermore Parfit (1982, 150) suggests (but does not ultimately endorse) that actualism could solve the procreation asymmetry since the person in Unhappy could be considered actual, while the person in Happy cannot. Christine Korsgaard (2018, 91) similarly seem to propose an actualist explanation of the procreation asymmetry, where an atemporal understanding of moral status allows for an understanding of the person in Unhappy as actual and the person in Happy as non-actual. However, the actualist variation of the existence-requirement defence is ultimately unsuccessful, since it is not obvious why it is compatible with the asymmetry, while it also seems to violate the relevant desiderata. Roberts (2011b, 352-353) for example argues that actualism ability to account for the procreation asymmetry depends on whether or not the relevant procreative acts are performed. This furthermore does not only cause problems with compatibility with the procreation asymmetry but, as will be argued in section 3.4.1., will also be problematic in relation to the action-guidance requirement as stated in section 2.3.23_{Actualism furthermore, as noted by Frick (2014, 16), is unable to account for the}

correct result when applied to non-identity cases and thus will also be incompatible with RMP. In virtue of the considerations above, actualism is obviously incompatible with the relevant desiderata and thus is of little interest and will be excluded from further consideration. 24 Two more initially promising variations of the existence-requirement strategy are what will be denoted as the harm-benefit defence and the conditional reasons defence. As will be presented in this paper, the former comes in two further variations, where their uniting feature is an emphasis on a distinction between the moral significance of harms and benefits. While according to the latter, both happiness and unhappiness are understood as morally significant

22_{It is important to note that although it can be the case that sometimes defences of the procreation asymmetry}

relying on a particular notion of modal status can be understood as employing the existence-requirement strategy, this is not necessarily always the case. For example, variabilism consider all groups of people of equal importance and instead explains the betterness-relation between worlds in relation to an existence-requirement (more specifically, who exists in the relevant worlds) which in turn accounts for the procreation asymmetry.

23_{It is of course possible to argue for a weak actualism instead of this strong variation, where the moral}

significance of a person is understood in relation to whether or not the person would be actual if a given action φ were to be performed. However, Roberts (2011b, 353) similarly notes that this view would also be problematic since it would render permissible acts wrong depending on what action one is evaluating. Cohen (2019, 76-78) makes similar observation.

24_{It is noteworthy that Jake Earl (2017) proposes that an actualist defence of the procreation asymmetry can}

(21)

17

in relation to an existence-requirement, where a particular understanding of conditional reasons plays a crucial role to account for the procreation asymmetry.

3.1 The Harm-benefit Defence

The general distinguishing feature of the harm-benefit defence is the claim that there is a difference, in terms of moral significance, between one’s beneficiary actions and one’s harm-avoiding actions. One way to account for this difference is to argue that one’s reasons to provide benefits, in general, are weaker than one’s harm-avoiding reasons. Assuming that benefits and harms can be understood in terms of happiness, the claim could be interpreted as saying that the reason-giving force of happiness is, generally speaking, significantly weaker than the reason-giving force of unhappiness. One philosopher that considers and defend such a solution is Elizabeth Harman (2004), however, as McMahan (2009, 57-58) notes, if our reason to benefit are merely weaker than our reason not to harm, it will only lend support to the weak asymmetry. Consequently, since the focus of this paper is to consider a strategy used to defend a formulation of the strong asymmetry, Harman’s harm-benefits defence, and defences similar to it will not be considered.

It is however possible to defend the (strong) procreation asymmetry by claiming that there is a general asymmetry between one’s harm-avoiding reasons and one’s reasons to benefit. Where the claim ultimately comes down to that there are no reasons to benefit while there are harm-avoiding reasons.25 This would account for the strong asymmetry since the person in Happy arguably is benefitted by being brought into existence and there is no reason to benefit there is no moral reason to create the relevant person. However, since the person in Unhappy arguably is harmed by being created, there is a harm-avoiding reason not to create the relevant person. Although this suggestion accounts for the procreation asymmetry, to stipulate such a general asymmetry is rather implausible since arguably a reasonable moral theory need to be able to account for the fact that there is at least some moral significance in benefitting others.26 This while it also is a claim that obviously is incompatible with the reproductive maximizing principle RMP. If there is no reason to benefit, it is hard to account for the fact that one ought to create the happier person in either of the non-identity case or the same-person case version

25_{One such view is for example negative utilitarianism, where the only thing that is of moral significance is}

suffering. Thus, according to negative utilitarianism all one has moral reason to do is to minimize suffering. If so there is no reason to create a person that will be happy, and there is a reason to not create a person that will be unhappy. An early proponent of negative utilitarianism was Karl Popper (1952, Ch. 5. n. 6).

26_{It is plausible to assume that a lot of people tend to agree with the relevant claim, Roberts (2010, 90) for}

(22)

18

of the good-better case. A further concern is also the fact that it implies antinatalism since the pro tanto benefits lacks moral significance, and the inevitable pro tanto harms a person will experience have moral significance. Thus, there is only a reason against creating a new person and consequently there is reason to disregard the relevant kind of harm-benefit defence. A more promising attempt to defend the procreation asymmetry is by arguing that there is an asymmetry between the moral significance of bestowing existential benefits and existential harms (i.e. harms and benefits one suffers/are bestowed by being brought into existence) by means of an existence-requirement. In order to account for Unhappy this view makes the following claim: the fact that we create p in Unhappy rather than not is a morally significant existential harm for p. Similarly, creating the p in Happy amounts to bestowing the relevant p an existential benefit. However, it is claimed that this existential benefit lacks moral significance and thus Happy is accounted for.

A noteworthy and well-discussed general problem for defences similar to the harm-benefit defence is whether or not one can be benefitted or harmed by being brought into existence. The relevant controversy can be understood in relation to a common conception of harms/benefits, according to which for a person p to be harmed/benefitted by an action φ it is necessary that p is made worse/better off by φ, i.e. harms seems to be dictated by a counterfactually worse off condition. When for example considering the choice between creating the unhappy p and not creating p at all, it might be hard to understand how p actually is made worse off by being created and better off by not being created. I.e. it could be considered perplexing how someone whose existence is dependent on an action φ can be harmed by φ, this since arguably non-existence cannot be better or worse for p and hence there is no baseline with which it is possible to determine whether φ really made p worse/better off (see for example Parfit, 1984, 487-490, 489 esp.; McMahan, 2009, 55; Temkin, 2012, 420, esp. n25; Bykvist, 2007b). This can be solved by either granting the viability of comparisons with non-existence in different ways (see Holtug, 2001; Roberts, 2003; Johansson, 2010; Arrhenius & Rabinowicz 2010 for ways to defend this claim) or denying the comparative understanding of harm for an causal account of harm (see among others Shiffrin, 1999; Harman, 2004; Gardner, 2013).27 Although this problem is highly relevant for both the ability of the harm-benefit defence to account for the procreation asymmetry and its ability to account for RMP, it is a problem that have been well-discussed elsewhere (see for example McMahan, 2009, sec. 3.4) and will not be discussed here. Instead

(23)

19

it will be assumed that either of the cited solutions above can be accepted and thus the relevant problem will be bracketed in the forthcoming discussion.

3.2. An Existence-requirement for Morally Significant Benefits

A view that is suggested and later rejected by Peter Singer (1999, 103) is what is denoted as the prior existence view (PEV). According to PEV only the happiness and unhappiness of individuals existing independent of the performance of the action under consideration φ are morally significant for the appraisal of φ.28 Translated to the terminology of the harm-benefit defence, only the harms and benefits of people whom exist independently of the action under consideration are morally significant. An attractive feature of PEV is that it can account for Happy, this since the foreseeably happy person in Happy is a dependently existing person. Assuming that the happiness of the dependently existing p in Happy constitutes a benefit for p, the existential benefit p would be bestowed by being created lacks moral significance. To make this claim explicit, consider the following symmetric independent existence-requirement (SIER) capturing PEV:

SIER: An action φ harms/benefits a person p in a world w in a morally significant way only if the existence of p is independent of the performance of φ.

However, as argued by Singer (1999, 104; 2014, 368) and Roberts (2011b, 347) SIER cannot account for Unhappy. This since, the unhappiness and thus existential harm suffered by the person in Unhappy lacks moral significance because it is not the harm suffered by an independently existing person. What thus could be considered the underlying reason for why PEV cannot account for Unhappy is because it implies that both existential harms and existential benefits lacks moral significance.

A way to accommodate for both Happy and Unhappy is by claiming that existential benefits lacks moral significance, while existential harms are morally significant. A way to make this claim is to argue that the moral significance of benefits depends on them being bestowed upon an independently existing person, while the moral significance of harm is unconditional. I.e. by stipulating an asymmetric version of SIER (AER):

28_{Singer’s (1999, 103) own formulation is that only “[…] beings who already exist prior to the relevant decision,}

(24)

20

AER: An action φ benefits a person p in a world w in a morally significant way only if p:s existence is independent of the performance of φ, while φ harms p in w in a morally significant way irrespective of p:s existence being independent of the performance of φ. 2930

Something that initially could be considered to count in favour of AER is that it seems to capture aspects of other theories that are the main contributing reason to why these theories can account for the procreation asymmetry (or some variation of it). Consider, for example, David Benatar’s (2006, 39-40) axiological asymmetry, which can be expressed by means of the following matrix:

w (p exists) w* (p does not exist)

(1). p is harmed (bad) (3). p is not harmed (good) (2). p is benefitted (good) (4). p is not benefitted (neutral)

What (1)-(4) essentially amounts to is that the presence of harm31_{is unconditionally bad and}

the absence of harm is unconditionally good, while the presence and absence of benefits is good conditional on independent existence. The relevant axiological claims (1)-(4) could furthermore be plausibly considered the axiological underpinning of the following normative claims:

w (p exists) w* (p does not exist)

(1) p is harmed (morally significant) (3) p is not harmed (morally significant) (2) p is benefitted (morally significant) (4) p is not benefitted (morally insignificant)

Here it might be possible to question why existence in the particular case should be interpreted in terms of independent existence. However, if the axiological asymmetry wishes to account for the procreation asymmetry and especially Happy, it seems necessary. This since it enables one to rather than understanding w and w* as possible outcomes of one’s actions, as worlds

29_{For a discussion of a similar view see (Parfit, 1984, 393-6; Temkin, 2012, 416-35) It is however important to}

note that the view Temkin (2012) discusses, although similar, is significantly different since Temkin’s principle implies a comparative view of harm while it has an axiological rather than normative emphasis. Ben Bradley (2013) denotes the view he discusses as asymmetric necessitarianism. Bradley (2013) also suggests this view can account for the procreation asymmetry, however Bradley (2013) plausibly argues that it violates the independence of irrelevant alternatives.

30_{Arguably Benatar’s defence of this view (as the axiological asymmetry i.e.) could be considered objectionably}

circular. For an argument of this nature see for example (Magnusson, 2019).

31_{The originator of the axiological asymmetry, David Benatar (2006, 39-40), sometimes uses the term ‘pain’,}

(25)

21

from which and in which the actions are performed. In order to make this point, it is worthwhile to consider figure 3 below illustrating Happy, i.e. an agent ax is presented with the two-way

choice between making the foreseeably happy person p by actualising w5 or refraining from

creating p by actualising w6. The worlds are thus possible outcomes of the actions of ax that are

performed from yet another world wx:

Figure 3(Happy) w5 (p exists and is happy)

ax (at wx)

w6 (p does not exist)

In order to account for Happy as conceived of in this paper, a defence of the procreation asymmetry needs to consider it permissible for ax actualise either w5 or w6. However, if one

interprets “p exists” as “p will exist in one of the worlds which are under considerations to be actualised”, Happy cannot be accounted for. This since the benefit p is bestowed in w5 have

moral significance due to the fact that p indeed will exist in w5 and w5 is one of the worlds that

ax considers actualising. Consequently, there is a moral reason to actualise w5 and thus create

the happy p. However, if one instead interprets ‘p exists’ in terms of independent existence, ‘p exists’ means ‘p will exist independently of what ax do at wx ‘one will get the correct result

when considering figure 3. The fact that p will exist and is happy in w5 lacks moral significance

since p does not exist independently of what world ax opts to actualise from wx and thus there

is not a pro tanto reason to actualise w5 rather than w6.

(26)

22

action is an all-things-considered benefit.32 For further discussion (i) will be denoted as the permissibility interpretation, (ii) as the reason generating interpretation and (iii) as the risk interpretation.

In order to illustrate how the consent-requirement defence (in all three interpretations) can account for the procreation asymmetry, once again consider figure 3. The fact that p is created in w5 and is happy is considered a pure benefit for p, and when considering figure 4 below

illustrating the choice made in Unhappy, actualising w8 rather than w7 is considered a

harm-avoiding action:

Figure 4 (Unhappy) w7 (p exists and is unhappy)

ax (at wx)

w8 (p does not exist)

Arguably p cannot consent to either of the actions under consideration prior to their performance, this since p: s existence is dependent on the action under consideration. However, since bringing about w7 is a harm-avoiding action, consent is not necessary and thus the

consent-requirement defence (in all three interpretations) accounts for Unhappy. With regards to Happy how the consent-requirement defence accounts for it will differ between which one of the three interpretations one argues is the most plausible. According to the permissibility interpretation, the answer is supposedly rather straight forward. Since the fact that bringing about w5 is a pure

benefit bestowed on p, the action actualising w5 requires the consent of p. Something that p

cannot give the existence of p is dependent on the action under consideration. For the reason generating interpretation, the fact that p does not exist prior to the relevant action and thus cannot consent to the pure benefit that is bestowed on p means that there is no reason to actualise either of the relevant worlds and thus Happy is accounted for. Finally, with regards to the risk interpretation Happy is accounted for by means of a empirical assumption about happiness as understood in this paper, namely that when one creates a new happy person, there is a

32_{In the example that illustrates why being bestowed a pure benefit requires the consent of the beneficiary,}

(27)

23

substantial risk that they will suffer some pro tanto harms although they will, all-things-considered, be happy and thus benefitted. This risk in turn requires the consent of p, however, as already established, the consent of p is not possible since p:s existence is dependent on the action under consideration.

As might have become clear from the reasoning above, the reason why AER captures the consent-requirement defence is that the existence which is necessary for p: s consent is previous existence and thus also independent existence. Otherwise the consent-requirement defence will not be able to account for the procreation asymmetry. This point can be illustrated by considering whether the consent-requirement defence should employ a consent-requirement that relies on hypothetical or actual consent. If hypothetical consent is sufficient for an action to be considered consensual, it is sufficient that the action in question could be estimated to be reasonably understood as in the hypothetical interest of the person in question. But if so, it is hard to understand how the consent-requirement defence can account for Happy. Arguably, it can be considered to be reasonably in the interest of the dependently existing person to be created happy. This would thus mean that there is a moral reason to create the happy person, which would be in conflict with Happy.33_{If instead actual consent is considered necessary (in}

terms of justification) for bestowing someone a pure benefit, some form of decision, at the least,34 is required for it to be considered a valid consent. This means that the person ought to exist previous to the action under consideration in order to give her consent, meaning that the consent-requirement also is captured by AER and by doing so it also serves as an initial support for why one ought to accept AER.

3.2.1. AER and Antinatalism

A problem with both the axiological asymmetry and the consent-requirement defence is that they both imply antinatalism. According to the permissibility interpretation of the consent-requirement defence what is necessary for creating the p in Happy to be permissible is the actual consent of the person that is bestowed a pure benefit. Unfortunately, this interpretation will render the consent-requirement defence unable to account for Happy. This since Happy implies that both w5 and w6 are morally permissible, while the permissibility interpretation implies that

ax actualising w5 is impermissible since it is not possible to get the consent of p at wx. If so, it

33_{Indeed Shiffrin (1999, 133) expresses some concerns with regard to hypothetical consent and procreation. One}

of the points that Shiffrin makes is that there seem to be a subjective component with regards to available benefits and trade-offs that make hypothetical consent unsuitable. In general Shiffrin (1999, 131-133) seem to reject the hypothetical consent solution.

34_{On two common competing theories on the ontology of consent (subjectivism and performativism) consent}

(28)

24

means that it is impermissible for ax to create the happy p and thus the permissibility

interpretation implies antinatalism. Furthermore, the risk interpretation will arguably imply antinatalism when combined with the assumption made above that creating a new happy person will involve a risk of this person suffering a pro tanto harm. Given this empirical assumption it is not permissible to create a happy person p without the consent of p, however, as already explained, this consent is not possible to get and thus it is impermissible to create the relevant person. Lastly and in a somewhat similar way, the reason generating interpretation will imply antinatalism combined with the plausible empirical assumption that every happy life will contain some pro tanto harms. If pure benefits cannot generate any reasons to act, they will always be outweighed by the reasons against acting produced by the (almost) inevitably pro tanto harms in a given happy person’s life. Thus, there is a reason against creating even a happy person and this interpretation also implies antinatalism.35

In a similar way as the reason generating interpretation, the axiological asymmetry implies antinatalism (Benatar, 2006, esp. Ch. 2 and Ch. 3), this since arguably one of the features that makes the axiological asymmetry imply antinatalism is that it is primarily concerned with pro tanto harms and pro tanto benefits, where the moral significance of the former is dictated by an existence-requirement similar to AER. When ax considers whether or not to create p the pro

tanto benefits that p will foreseeably be bestowed by being created in virtue of, for example, the prudentially good things that p can be expected to experience, are morally insignificant since p is not an independently existing person. However, the axiological asymmetry claims that the pro tanto harms constituted by the prudentially bad things p will, almost inevitably, experience are morally significant irrespective of the modal status of p. If so, when considering the decision in figure 3, ax have a moral reason against actualising w5 and no reason against

actualising w6. Consequently, the axiological asymmetry implies that ax have a moral reason

against creating the happy person p and thus is incompatible with Happy given the relevant empirical assumption.

If neither the axiological asymmetry nor the consent-requirement defence can account for Happy (in virtue of implying antinatalism either in principle or given certain empirical assumptions), it would seem that there is little initial support to be gained by the considered theories for AER being compatible with anti-antinatalism. However, just because theories that

35_{Shiffrin (1999, 118) is somewhat unclear what she thinks about such implications and ambiguously notes that}

(29)

25

could be considered captured by AER implies antinatalism does not mean that AER necessarily need to imply antinatalism.

One way for AER to avoid antinatalism is to simply deny that p will experience things that could be considered prudentially bad and thus pro tanto harms. It is however quite hard to understand how this could be possible. Every life seems to involve some pain for example, pain that arguably is a paradigmatic example of a harm and furthermore a prudentially bad thing. In fact, by the two already briefly mentioned common understandings of harm it seems very unlikely that a life will not contain at least some pro tanto harms. First consider the possibility of not being harmed if ‘harm’ is defined by a counterfactually worse off condition (CWC):

CWC: A person p is harmed/benefitted by an event e iff e makes p worse/better off relative to what p would have been had e not occurred.36

A pro tanto harm according to CWC means that if there is one moment in the life of a person p where it is a set-back, p becomes less happy for example, p suffers a pro tanto harm. Consequently, in order for a person to avoid pro tanto harms in their life, they ought to be on a continuous path upwards without as much as a bump in the road. This, however, seems very unlikely to achieve. Consider for example the prospect of death, which assumingly is an inevitable part of every human life and furthermore assume that the termination thesis is true, i.e. when one dies, one also goes out of existence. At the least death seems to have a neutral value for the person in question, if so even if a person is on a continuous path upwards her whole life, the neutral value of death ought to be considered a worse state than p:s previous blissful existence and p ought to be considered harmed in the comparative sense. But perhaps there is a way to create a person that will not suffer any intrinsically bad states, if so, on a causal account of harm (CAH) it might be the case that the person will not suffer any pro tanto harms in her life:

CAH: An event e harms p if it causes a state of affairs s that is intrinsically bad for p.37

If this is possible it would mean that antinatalism is avoided and Happy is accounted for, however the relevant possibility is something that partly depends on how one conceives of

36_{There are several proponents of such an understanding of harm, see for example (Feinberg, 1984; Boonin,}

2008; Schwarz, 1978).

37_{A version of CAH is arguably espoused by Seana Shiffrin (2012, 383; 388) where p is in an intrinsically bad}

THE PROCREATION ASYMMETRY

THE PROCREATION

ASYMMETRY

The Existence-requirement Strategy and some

Concerns on Incompatibility

Table of contents

Introduction

1. Defining the Procreation Asymmetry

2. Desiderata

2.1. Anti-Antinatalism

3. Defending the Procreation Asymmetry