Peacefulness, nonviolence, and experientialist minimalism

This is part three of a series on minimalist axiologies (i.e. axiologies that essentially say “the less this, the better”).

Every part of this series builds on the previous parts, but can also be read independently.


This essay is my attempt to explore and respond to some concerns that people may have about minimalist views on intrinsic value. My own views have changed a lot over the years, and they might well continue to change in the future upon further reflection and discussion. Such discussions about values can understandably evoke strong feelings in most of us, yet it is my hope that we can nevertheless improve our views together, with the help of good-faith dialogue and a scout mindset.

1. Overview and scope

Some people may worry that minimalist axiologies would imply an affirmative answer to the following questions:

A. Would an empty world (i.e. a world without sentient beings) be axiologically perfect?

B. For any hypothetical world, would the best outcome always be realized by pressing a button that leads to its instant cessation?

C. Does minimalist consequentialism imply that it would be right to seek to turn our world into an empty world in practice, even by coercive means?

The scope of my responses is limited to purely welfarist consequentialism. That is, I assume that the value of outcomes and the rightness of acts depend solely on the outcome in terms of welfare, and not on any intrinsic disvalue assigned to things like acts, motives, or character traits independent of the consequences that they may have for welfare.

Additionally, the scope is limited to minimalist axiologies that are based on experientialist accounts of welfare (cf. van der Deijl, 2021). In other words, I assume that the welfare of any given being cannot be affected by things that do not enter their experience, and thus set aside views such as preference-based axiologies that imply extra-experientialism.[1] Relatedly, I further assume that the disvalue of any experiential state is wholly intrinsic to that experiential state, independent of the context or the rest of the life in which that state exists (cf. the “isolated view” that population axiology takes on individual lives, applied also to experiential states).

Regarding the theoretical value of hypothetical worlds (A), experientialist minimalist axiologies do imply that there is no world that could be better than an empty world; after all, an empty world would involve no trouble of any kind. Yet an equally ideal world, in theory, would be one in which all lives are perfectly untroubled. And because this second world is a much more widely shared ideal, it is misleading and needlessly divisive to talk of emptiness as the only ideal of perfection according to minimalist axiologies.

The other questions and potential worries are a lot less trivial. To respond to them in their proper context, the rest of this piece draws a sharp distinction between the hypothetical question (B) and the practical question (C). Sections 2 and 3 address these questions, respectively. The next two subsections provide an outline of the main points.

1.1 Overview of the hypothetical side

Section 2 is a six-part response to the hypothetical choice of cessation (B). To focus on the choices where minimalist axiologies imply that cessation is the best option, I assume that the ideal of an untroubled paradise is not an option. In other words, the response concerns only a restricted choice between the non-cessation and cessation of any hypothetical world that contains some amount of experiential bads (such as intense suffering).

The structure of the response is as follows.

  • 2.1 argues that we need to account for status quo bias and omission bias before we could hope to evaluate the choice between non-cessation and cessation from an impartial perspective.

  • 2.2 clarifies the choice by presenting a reversal test (to account for these biases) and highlights the need to be mindful of the radical assumption of ‘all else being equal’ often made in population ethics.

  • 2.3 responds to the choice (between the non-cessation and cessation of any hypothetical world that contains bads) from the perspective of experientialist and consequentialist minimalism.

  • 2.4 explores some implications of our possibly failing to maintain the assumption of ‘all else being equal’ when engaging in this thought experiment.

  • 2.5 zooms out to ask whether the principles provided by minimalist axiologies (to respond to non-cessation vs. cessation) are any less plausible than the principles provided by other consequentialist views (i.e. ‘offsetting views’). Furthermore, this section:

    • argues that in comparison with offsetting views, the cessation button does not constitute a unique objection against minimalist views in particular (if at all).

    • suggests that similar — and in many cases worse — hypothetical choices are implied by offsetting views, including the choice of ‘intense bliss with hellish cessation’.

    • notes that for some people, this may be a reason to reject all purely experientialist and consequentialist views; for others, it may be a reason to highlight the gap between consequentialist theory and practice.

  • 2.6 briefly highlights the gap between consequentialist theory and practice (before Section 3 does so at length). In particular, this section:

    • asks whether our practical anti-violence intuitions (strong and warranted as they are) might miss their mark in thought experiments that involve the cessation of causally isolated lives (be it ‘minimalist cessation’ or ‘intense bliss with hellish cessation’), and whether this might constitute an additional bias in such thought experiments.

    • looks at experientialist minimalist reasons to strongly oppose painless killing.

1.2 Overview of the practical side

Section 3 is a three-part response to the practical question (C). The hypothetical conclusions in Section 2 are based on assumptions that are completely unrealistic, so this section explores the altogether different considerations that are of key relevance once we drop those assumptions of ‘all else being equal’, cessation buttons, and the like.

The structure of the response is as follows.

  • 3.1 acknowledges that we can gain many benefits from orienting toward ideal states-of-affairs (i.e. ‘utopianism’). However, this section also argues that:

    • utopianism as a guiding principle has many pitfalls that can cause it to diverge from impartial consequentialism (and thereby do more harm than good).

    • impartial minimalists have no a priori reason to privilege an empty world as an ideal future.

    • instead of seeking possible paths to a single, ideal “endstate” of the world, sophisticated consequentialists ought to follow indirect proxy principles that tend to have the best consequences over all space and time in terms of carefully estimated expected value (and do so from a ‘marginal realist’ rather than a ‘broad idealist’ perspective).

      • this is a more impartial, realistic, and risk-aware outcome-orientation that does not privilege any subperiod in time, and nor does it privilege some uncertain prospect of realizing a hypothetical endstate at the risk of leading to an overall worse outcome.

  • 3.2 looks at the key practical considerations for assessing whether minimalist consequentialism, combined with sophisticated expected value thinking, would recommend or discourage efforts to create an empty world. Besides having no a priori reason to seek an empty world:

    • minimalists also have strong practical reasons to cooperate with other value systems and to seek mutual gains from compromise with them (not least because the prevention of worst-case outcomes is already common ground between multiple value systems).

    • minimalists (just like everyone else) should acknowledge that there is still considerable empirical uncertainty about how common life is in the reachable universe, and about whether an Earth-originating civilization would (over the long-term future) use astronomical resources in more or less suffering-conducive ways than would another civilization.

      • this uncertainty is a reason against prematurely concluding that the extinction of humanity would be desirable compared to human space colonization (from an impartial minimalist perspective).

      • and (as a practically relevant thought experiment), the more we would be both capable and goal-aligned enough to prevent all suffering on Earth, the more we might also be the kind of civilization that could play more positive roles by ensuring that the vast resources of the reachable universe would not become fuel for generating astronomically greater suffering.

    • minimalists would likely be wise — given the empirical uncertainty — to prioritize the widely shared (and robustly positive) goal of “improving the expected quality of future lives conditional on their existence”, such as by cooperating with other value systems to reduce the probability and severity of risks of astronomical suffering (s-risks) over the long-term future.

      • s-risk reduction will plausibly require large-scale cooperation between multiple value systems, whereas unilateral action toward an ‘empty world’ would likely increase s-risks in expectation.

  • 3.3 notes that the aim to minimize experiential bads is not by itself a directly suitable principle for guiding practical action. Rather:

    • minimalists (and consequentialists in general) need to deeply internalize and uphold more actionable principles — such as virtues and commonsense prohibitions — that indirectly tend to bring about the best consequences.

      • for minimalists, a practically optimal set of principles will likely include pragmatically absolute non-violence, non-aggression, and respect toward other sentient beings, not least because the erosion of these principles is a prime risk factor for s-risks.

2. The hypothetical side: Cessation versus creation of an imperfect world

2.1 Status quo bias and omission bias

Privileging non-intervention

The status quo bias is an irrational preference for the current state of affairs even when better options are available. Status quo bias can be probed with the reversal test.

For example, suppose Bob is against giving access to pain relief to those who lack it (even assuming no switching costs). Now, we may ask Bob whether he would be in favor of removing access to pain relief from the same patients assuming that the patients already had such access; if he would not, then status quo bias is probably influencing Bob’s judgment.

Status quo bias may often be partially explained by omission bias, i.e. the tendency to consider harmful inaction (omission) more acceptable than equally harmful action (commission).[2] Omission bias, in turn, may partially be explained by the fact that we often find it much easier to attribute harm to someone who took harmful action compared to someone who took no action, even if they both allowed equal harm to take place (in terms of the outcome). (The act-omission distinction is relevant on some nonconsequentialist views, but it is immaterial on pure consequentialism, which is assumed in this response.)

Regarding the cessation of a hypothetical world, clearly the choice of cessation would be a significant intervention to the status quo. To account for the influence of the biases above, we therefore need a reversal test where the choice that is equivalent to non-cessation entails a similar responsibility for the outcome as does the choice of cessation. In this case, the reversal test is to ask whether we would actively create a world that is identical to the one that the choice of non-cessation would allow to exist.

2.2 The reversal test: Creation at the moment of non-cessation

The reversal test is motivated by the consequentialist equivalence between action and inaction, provided that these involve identical outcomes. From a consequentialist perspective, the outcome of the choice of non-cessation is that a world does exist in place of an empty world, which is outcome-equivalent to the creation of such a world (from the moment of the choice onward) in place of an empty world. This equivalence is illustrated in the following diagram.

A diagram that shows the consequentialist equivalence between non-cessation and creation, and between cessation and non-creation.

A better framing for the initial question (B) is now the following choice between the creation and cessation of a hypothetical world W at some time T:

  1. ‘Creation’, which leads to the instant creation of world W (starting from time T) in place of an otherwise empty world.

  2. ‘Cessation’, which leads to the instant cessation of world W at time T.

In other words, does an overall better outcome result from the creation of world W at time T, or from the non-creation of world W at time T?

We, in this thought experiment, are simply an outside ‘chooser’. In other words, we do not live in world W, and neither are we in any way affected by our choice. (Section 2.4 will explore some implications of our possibly failing to maintain this assumption when engaging in this thought experiment.)

Additionally, in order to properly respect the standard assumption of ‘all else being equal’ in population ethics, we assume that the entire population of world W consists of ‘isolated Matrix-lives’ that are not affected by each other in any way.[3] This helps to ensure that our assessment of this hypothetical choice is not distorted by our practical intuitions, which in the real world might implicitly be tracking the overall value of individual lives not only in terms of their independent features but also in terms of their positive or negative roles for all other lives.

Put differently, we need to be careful not to accidentally overgeneralize our practical intuitions about the cessation or creation of an individual (real-world) life such that we misapply these intuitions to the hypothetical case of the cessation or creation of an entire world. This is because an individual (real-world) life practically always has effects beyond itself — which our intuitions may implicitly be tracking — whereas the set of all lives never does.[4]

2.3 Choosing a future with fewer problems

Now, to finally answer the reframed question of creation versus non-creation, we only need to specify

  1. what the future is actually like, in the creation scenario, for the isolated Matrix-lives of world W (from time T onward), and

  2. the axiological principle that we use to determine whether the overall better outcome (for the isolated Matrix-lives of world W) results from the choice of creation or non-creation (all else being equal).

Trivially, if the lives are perfectly untroubled, minimalist axiologies would not strictly favor the non-creation of such a world (nor strictly oppose its creation). After all, such a world is already perfect due to involving no experiential problem for anyone. So no one would feel harmed by its creation.

The more interesting case is the ‘near-perfect paradise’, where the isolated Matrix-lives are arbitrarily blissful, meaningful, and numerous, yet one of the lives is rendered ‘imperfect’ due to a near-negligible problem, such as (some) involuntary suffering, unmet need, or experienced preference frustration. All else being equal, would experientialist minimalist views favor the instant cessation (i.e. non-creation) rather than the creation of such an ’imperfect world’ at time T, before the problem occurs?

Hypothetically and strictly from a consequentialist point of view, yes they would. After all, regardless of how the experientialist minimalist views in question are formulated, they would always imply that the better future is the one that involves the overall lesser amount of involuntary suffering, momentarily unmet needs, experienced preference frustration, or the like. And regardless of the number of the isolated lives or the intensity of their bliss, a shared feature of minimalist views is that the suffering, need, or frustration of some beings cannot be positively counterbalanced or outweighed by the addition of subjectively perfect experience-moments elsewhere.[5] (Some problematic implications of contrasting offsetting views are explored shortly in Section 2.5.)

2.4 Minimalist creation for extrinsic reasons

Breaking the ‘all else being equal’ assumption

Minimalist axiologies imply that replacing an empty world with the future of world W cannot be an improvement within that world, because the empty world is perfectly fine to begin with (“no need, no problem”). Yet even if one holds this view, one might still intuitively feel that ‘creation’ would be the better choice in the case of the near-perfect paradise. This section explores some reasons why such an intuition might arise not necessarily due to our implicitly holding a non-minimalist axiology, but possibly due to our breaking the boundaries of the ‘all else being equal’ assumption when engaging in this thought experiment.

First, if we already feel more mental trouble contemplating the cessation or non-creation of a hypothetical world than what would be entailed by its existence, then this might in part explain why we intuitively feel that the price of cessation would be too high.[6] For example, we might implicitly feel as if even a hypothetical endorsement of cessation would have dispiriting implications for our own unmet needs for existential security, bliss, or meaning. Yet just because we might have such unmet needs, it does not follow that the creation of isolated beings with satisfied such needs would be an independently positive endeavor.

Second, our choice between creation and cessation may be affected by the awareness that even theoretical endorsements of emptiness over an imperfect world can seem like an ‘anti-life’ stance to others who care about which side we would take if we had to choose between conflicting ideals in practice (cf. “The side-taking hypothesis for moral judgment”, DeScioli, 2016). Yet such signaling-related factors should ideally not distort our theoretical evaluations in this thought experiment.

These considerations may be a reason to highlight the fact that the near-perfect paradise is as perfect as any world that could ever be practically realized according to both minimalist and other welfarist views (cf. Tomasik, 2013d). And as far as ideals are concerned, minimalists may thus have strong practical reasons to side less with the hypothetical ideal of emptiness or cessation, and more with the equal and more popular ideal of perfectly untroubled lives.[7]

2.5 Comparative theoretical implications of minimalist and offsetting views

This section contrasts the hypothetical cessation implications of minimalist views (cf. 2.3) with some hypothetical implications of other consequentialist views. For brevity, the other views will be called ‘offsetting views’ due to their assumption that independent bads can be ‘counterbalanced’, ‘outweighed’, or ‘offset’ by a sufficient amount of independent goods. Roughly, minimalist axiologies favor outcomes in which the notional sum of independent bads is minimized, whereas offsetting axiologies favor outcomes in which the notional sum of independent goods over independent bads is maximized. (Offsetting views include e.g. classical utilitarianism as well as ‘weak negative’ or ‘negative-leaning’ views.)

Trivially, minimalist views are not alone in sometimes favoring cessation in hypothetical scenarios, since offsetting views also favor cessation whenever the notional sum of future goods and bads is negative. So the question is not whether offsetting views imply cessation, but rather when they imply it. And the question is also what else they imply, because in comparing axiological views based on their hypothetical implications, it seems relevant to compare their (apparently) least plausible implications with each other.

Even before looking at the specific implications, one might defend an offsetting view by saying that it implies cessation only in the cases where the “sum” of goods over bads is negative (i.e. in the “correct” cases), and that minimalism would imply cessation in cases where this “sum” ought to be seen as positive. Yet the existence of this “sum” is a huge assumption to begin with, and one that can justify arbitrarily severe harms, including extreme suffering, for the creation of supposedly independent goods even if the absence of these supposed goods would cause no experiential problem in the first place.

For that reason, many people find the implications of offsetting views much worse compared to scenarios that involve no experiential bads, such as the hypothetical scenario of ‘minimalist cessation’, i.e. universal cessation or non-creation that prevents future problems. (Cf. these related surveys.)

Consider, for instance, that offsetting views have the following theoretical implications:[8]

  • ‘Intense Bliss with Hellish Cessation’. Assume a “minimalist paradise” that contains sentient beings experiencing perfectly untroubled contentment, but not pleasure. All else being equal, offsetting views imply that it is better to reject the infinite continuation of this paradise and to instead choose a sufficiently large paradise of intense bliss that ends in an arbitrarily large hellish cessation.[9]

An illustration of the choice of ‘Intense Bliss with Hellish Cessation’ (B), which is favored by offsetting views.[10] Minimalist views would favor the continuation of the minimalist paradise (A).
  • Creating Hell to Please the Blissful’ (expressed in hedonistic terms, but applicable to independent goods more broadly). Say we have a vast population experiencing nearly maximal bliss. All else being equal, offsetting views imply that it is a net benefit to add to this population a smaller population of maximally hellish lives provided that this fully maximizes the bliss of the (sufficiently) vast population of near-maximally blissful lives.

  • ‘More Suffering and a Greater “Sum” from Giving Space to an Alien Civilization’. Assume a world in which humans can either expand into space or give it to an alien civilization that would use the same resources to generate astronomically more suffering but also vastly more goods, such that the “offsetting sum” from the alien expansion would still exceed that from human expansion. Here, offsetting views imply that arbitrarily hellish human extinction is justified to prevent human expansion into space as long as the alien civilization generates sufficient goods, whereas minimalism implies that human expansion into space is the better outcome due to ours being the more peaceful and harm-free civilization overall (in this hypothetical).

Finally, we may draw an analogy between the case in which minimalist views would favor the non-creation of a near-perfect paradise (cf. 2.3) and the following cases in which all experientialist consequentialist views would have a similar implication. Namely, just like a single ‘near-negligible problem’ may lead minimalist views to prefer that world’s non-creation or cessation over its creation, so too would a single such tiny problem “tip the scales” according to all experientialist consequentialist views in some worlds.

Specifically, all such views would favor cessation for the following worlds that may commonsensically seem to contain a lot of value even though they are assumed to be “balanced” or “neutral” in terms of experientialist value:

  1. The Pinprick Argument’ (adapted).[11] Regardless of what else a world contains, if the “sum” of all the experiential goods and bads of its inhabitants is perfectly neutral, all experientialist consequentialist views would favor the instant cessation of the world over adding the tiniest of bads.

  2. ‘Extra-Experiential Fulfillment’.[12] Say we have a world with arbitrarily many lives full of knowledge, accomplishments, and relationships, yet where these lives experience none of the supposedly independent goods and bads (such as pleasure and pain). According to all experientialist consequentialist views, it would be better to painlessly end these lives than to add the tiniest of bads.

To commonsense intuitions, those may seem like absurd conclusions and thereby like compelling objections to all experientialist consequentialist views in general.[13] Yet one could defend such views in ways identical to the previous sections’ defense of minimalist views. That is, one could emphasize that we need to (1) account for our status quo and omission bias, (2) remember to carefully respect the radical assumption of ‘all else being equal’, and (3) count only that which has independent value.

Experientialist offsetting views already deny that accomplishments, relationships, or other things have value independent of their roles in relation to certain experiential features, such as the affectively valenced experiences of pleasure and pain. Similarly, experientialist minimalist views take the extra step of denying that attractively valenced experiences have value independent of their roles in relation to aversively valenced experiences. (Cf. footnote 49 in DiGiovanni, 2021a. Consider also tranquilism as a theory of need-based motivation, according to which the absence of pleasure is not a problem unless there is an unmet need for pleasure.)

For someone attracted to experientialism and consequentialism, is the existence of cessation implications a reason to reject minimalist views and to prefer offsetting views? Arguably not, since we have just illustrated that cessation implications exist for all experientialist consequentialist views. Moreover, the hypothetical implications of offsetting views include forms of cessation, as well as other implications, that are arguably worse than the implications of minimalist views (cf. ‘Intense Bliss with Hellish Cessation’ and ‘Creating Hell to Please the Blissful’).[14]

Are the cessation implications of experientialist consequentialist views a reason to reject all such views? One could argue that they are. Yet they may also be a reason to mind the gap between consequentialist theory and practice. After all, experientialist consequentialists of every kind already justify norms against killing and violence not directly at the level of their preferred axiology, but at the level of practical decision procedures (cf. Mayerfeld, 1999, pp. 120–125). (More on this below, and in a future post on multi-level minimalism.)

2.6 The gap between theory and practice

This section briefly highlights the gap between consequentialist theory and practice (before Section 3 does so at length). On the practical side, I argue that experientialist minimalists — like other consequentialists — indeed should follow strong prohibitions against killing and violence in general.

The next two subsections look at the following questions, respectively:

  1. whether our practical anti-violence intuitions (strong and warranted as they are) might miss their mark in thought experiments that involve the cessation of causally isolated lives (be it ‘minimalist cessation’ or ‘intense bliss with hellish cessation’), and whether this might constitute an additional bias in such thought experiments.

  2. whether there are experientialist minimalist reasons to strongly oppose painless killing.

Cessation and our practical anti-violence intuitions

Regarding the first question, one could emphasize that our practical intuitions are rightly sensitive to the negative effects, including indirect long-term effects, of violence, killing, and dying in the real world. At the same time, the hypothetical choice of minimalist cessation involves no subjectively felt harm, no negative secondary effects, no loss of positive roles, and no uncertainty about the outcome whatsoever. This is highly unrealistic, which might cause our practical intuitions to at least partially miss their mark in this thought experiment.

In other words, it makes sense that our intuitions might treat death as a great bad in itself, even if we may on reflection think that its (experiential) badness comes from these neighboring phenomena that are merely closely intertwined with it. And even if we abstract away these phenomena in theory, it could be that our intuitions on the badness of death are not easily moved by merely adding the magic words ‘instant’, ‘all else being equal’, or ‘neither are we in any way affected by our choice’ (cf. 2.2, 2.4).[15]

Together, the assumptions of experientialism and consequentialism imply that we ought to think of the hypothetical “post-cessation lives” just like we already think of any non-existing lives that could exist in the future (cf. 2.1, 2.2). So unless we would call ‘non-creation’ a form of ‘killing’, ‘destruction’, or ‘violation of autonomy’, then we ought to comparably avoid such connotations in our mental image of hypothetical ‘cessation’ as well (so as to avoid the misapplication of our practical intuitions in the fantasy world of ‘all else being equal’, wise as they are in the practical world).

To be sure, many of these points are applicable also to the offsetting implication of ‘intense bliss with hellish cessation’. After all, it also involves cessation. Yet in that case the cessation is preceded by an arbitrarily large hell (supposedly offset by the preceding bliss). So our harm-oriented intuitions may rightly raise concerns about the pre-cessation part of this offsetting implication, as it does involve experienced harm.

Minimalist reasons to strongly oppose painless killing

The discussion above raises the follow-up question: Without the concept of independent good, how can experientialist minimalist views oppose painless killing?

A common misconception of minimalist views relates to their non-use of the concept of independent good. Why else would people take great pains to protect life and to even create new life? Yet we need not jump to the conclusion that these pains could be worthwhile only for the sake of some independent good. After all, minimalist views are perfectly compatible with the concepts of positive roles and positive lives, even if only in a relational sense, which can explain why we may rationally take great pains to protect and promote a variety of things even beyond their immediate preventive benefits.

The concept of positive roles also extends to positive norms, such as a general respect for autonomy and nonviolence, provided that these have overall greater preventive benefits than alternative norms do. So minimalist views can also imply a strong opposition to killing and violence for the sake of upholding positive roles, lives, and norms.

Crucially, the upholding of overall positive norms will often justify the protection of overall negative lives (both on minimalist and offsetting views). Positive norms are not easily worth eroding for the sake of preventing marginally more experiential bads at the individual level.[16]

3. The practical side: Why we should not seek to create an empty world

The hypothetical cessation response provided by minimalist views in Section 2.3 was based on assumptions that were completely unrealistic. Thus, we need to separately consider the altogether different practical question (C): “Does minimalist consequentialism imply that it would be right to seek to turn our world into an empty world in practice, even by coercive means?”

Why would anyone think so? A common route to such a conclusion may be what my previous essay called a narrative misconception of consequentialism. Section 3.1 will consider how such a misconception might arise, how it does more harm than good, and what a more accurate starting point for addressing the practical question would be.

3.1 Against endstate-oriented utopianism

For minimalist views, any world that is free of problems would be ideal, and thus akin to a utopia. But should one be guided by the goal of creating an ideal world, or utopia, in practice?

We may, for some purposes, find it very useful to imagine a desired endstate of the world, such as a world free of painful experiences. When people imagine their ideal society, they start to see more flaws in the current one and become more willing to help close the gap between the two (Fernando et al., 2018). This is a form of mental contrasting, which is a likely mechanism underlying the motivating effect of utopian thinking on social engagement (Fernando et al., 2018). Additional positive roles of utopian thinking are mentioned in the following footnote.[17]

Yet utopianism can also be blind and dangerous, as exemplified by some of the greatest atrocities of the 20th century. Just like our need for closure may bring us to justify a system that we feel unable to change for the better (cf. Jost & Hunyady, 2003), it may also cause us to downplay or altogether ignore the possibility that we might accidentally make things worse when we think we are acting rightly (cf. Wiblin & Lempel, 2018).

The pitfalls outlined below are, in my view, some of the main reasons why people often draw fallacious practical conclusions based on consequentialist thought experiments, including when it comes to thought experiments about cessation in particular.

Pitfalls of utopianism

“It is the only way.”

Utopianism can take mental contrasting too far by focusing on a single image of how the future should go. By comparison, consequentialism can in principle imply multiple equally ideal futures. In particular, consequentialism based on minimalist axiologies would have an equal preference for all futures in which all lives are perfectly untroubled. To a priori privilege an empty world as the ideal future is therefore not an impartial form of minimalist consequentialism but rather some kind of partial or narrativized version of it.

“We must get there.”

Consequentialism is not about seeking some “endstate” of the world at a single point in the future (cf. the narrative misconception of consequentialism). The consequentialist understanding of “the end justifies the means” is that “the end” refers to the total consequences of our actions. It does not imply that reaching some ultimate destination would justify any means or risks necessary to get there.

“If only everyone followed along.”

Utopianism can be unrealistic due to requiring that everyone act in a certain way. Historically, all efforts to enforce utopian visions in a top-down manner appear to have failed. When trying to bring about the best consequences, it is useful to distinguish between actions that are optimal from a “marginal realist” versus a “broad idealist” perspective (cf. Vinding, 2022, Section 8.1). Marginal realism relates to how a relatively small group should ideally spend their limited resources in order to have the best impact on the margin. The answer to that question might, but will not necessarily, approximate the answer to the “broad idealist” question, which is the question of how an entire society should ideally act if there were clear goal alignment between all stakeholders.

For instance, it is reasonable to think that the whole world should ideally shut down all factory farms. Yet if a small group of people were to attempt a violent shutdown of factory farms, the effort would probably not only fail, but also antagonize this group and their ideology in the eyes of more powerful groups. At worst, taking unilateral coercive action can lead to an overall more ruthless world, which would likely be much worse than business-as-usual in the long term. A better strategy for small groups is arguably to seek broader and deeper support for their views without coercive means, by adopting a positively cooperative approach. (Such a marginal realist strategy for minimalist consequentialists is defended shortly in Sections 3.2 and 3.3.)

“Seeking perfection over absolute expected impact.”

Utopianism can sometimes involve a degree of completionism or perfectionism, as if the final step toward utopia were somehow especially significant. Yet impartial consequentialism would not value the “last step” toward an ideal state of affairs any higher than it would value any other “beneficial step” of equal magnitude. (This idea is explicitly captured in scalar utilitarianism, which understands rightness not in binary terms of “right” versus “wrong”, but as a matter of degree; cf. Tomasik, 2015.)

Related may also be the empirically documented effect called ‘proportion dominance’, which refers to people’s common preference to help a higher proportion of individuals even when the absolute impact remains the same. One study found that people preferred helping 225 out of 300 lives over helping 230 out of 920 lives, which indicates proportion dominance even at the expense of absolute impact. Yet most people, when they reflected on it, agreed that one actually ought to prefer the higher absolute impact (Bartels, 2006).

Proportion dominance is highly relevant in our practical context, since the aim of abolishing all suffering on Earth might intuitively override the aim of preventing suffering from spreading beyond Earth, even though the latter arguably involves far more suffering prevented in expectation. After all, when it comes to suffering on Earth vs. (the risk of) suffering beyond Earth, we are not just talking about 300 versus 900 lives, but about suffering on one planet versus suffering on billions of planets, suggesting that proportion dominance may represent a serious bias in our thinking about what to prioritize to best reduce future suffering.

A better alternative: Expected value thinking

Only naive consequentialism recommends that we look at just the direct and immediate effects of our actions. By contrast, sophisticated consequentialism recommends that we appropriately estimate all of the effects of our actions, including their indirect long-term effects. This requires that we estimate all the ways in which coercive or rule-breaking actions could accidentally cause more problems than what they prevent.

The standard way to approximate this impossible ideal is to compare our potential actions in terms of their ‘expected value’, which is a technical term for the sum of all their independently (dis)valuable consequences weighted by their probability (cf. Todd, 2021).[18]

When expected value is difficult to estimate directly (as it often is), we may instead focus on indirect measures that tend to serve as good proxies for what ultimately matters. A preliminary framework of proxies for reducing future suffering is presented in Vinding (2022, Chapter 9, “Identifying Plausible Proxies”). These proxies include greater levels of cooperation, increased and more impartial concern for suffering, and a greater capacity to achieve shared aims, including the reduction of suffering. Thus, an indirect way to estimate the (minimalist) value of different possible actions is to ask how much the actions seem likely to increase these proxy measures.

Sophisticated expected value thinking can, similar to utopianism, foster hope and collective action by highlighting ways in which things could go better (cf. mental contrasting). At the same time, it can guide us around the pitfalls of blind utopianism. Specifically, it offers us a more risk-aware way to consider the full range of possible outcomes — accounting also for risks of astronomical suffering — without fixating on a single, vivid image of what the future should look like.

3.2 Key considerations for estimating practically optimal aims

This section looks at the key practical considerations for assessing whether minimalist consequentialism, combined with sophisticated expected value thinking, would recommend or discourage efforts to create an empty world.

Cooperation and gains from compromise

From a marginal realist perspective, minimalists have strong practical reasons to stay on cooperative terms with others.

As noted in the NU FAQ, attempting to realize an empty world is a highly suboptimal expression of reducing suffering from the perspective of many other value systems. Yet, for minimalists, “the difference between ’no future’ (i.e. no Earth-originating intelligence expanding into space) and a decent future, where concern for suffering and thwarted preferences plays some role [...] is much smaller than the difference between a decent future and one that goes awfully wrong.”

Minimalists thus have reason to steer the future away from going ‘awfully wrong’ and toward going ‘decently’ — a goal that everyone can directionally agree with — and thereby even less reason to privilege an aim of ‘no future’.

Many of the main reasons to avoid needless conflict and antagonism are covered briefly in Vinding (2020b) as well as in Tomasik (2011, “Why we should remain cooperative”).[19] For instance, by accommodating each other’s wishes and avoiding costly fighting with each other, proponents of different value systems can achieve mutual gains from compromise (Tomasik, 2013b; Ord, 2015). This enables everyone to better steer the expected future in mutually desired directions.[20]

All else being equal, basically everyone can agree that the reduction of suffering is an important aim. Given this agreement, a reasonable starting point is to work toward this aim while standing on common ground between multiple value systems (Baumann, 2020b). In practice, this could imply a focus on improving the expected quality of future lives conditional on their existence, such as by reducing the probability and severity of worst-case outcomes (Tomasik, 2011; Baumann, 2017b).

I quote the following considerations essentially verbatim from Knutsson (2021, Section 4), with some relevant hyperlinks added, and connect them with some related points made by others.

(1) If merely all humans died, there would be room for more suffering wild animals (Tomasik, 2016), and humans would no longer be able to reduce wild-animal suffering, which we may do if we survive (Vinding, 2015).

Yet there are also scenarios in which wild-animal suffering would be multiplied by human space colonization (Tomasik, 2014b). Given such vastly higher stakes, possibly the best way to reduce wild-animal suffering in expectation is to argue against spreading it in the first place (Tomasik, 2013a). (A similar argument might also apply to factory farming.)

(2) Even if all sentient beings on Earth died, beings that suffer could still evolve again on Earth (Acton and Watkins, 1963, 96; J. J. C. Smart, 1989, 44). Also, if humans survive, we may reduce suffering in other parts of the universe (Pearce 1995, chapter 4, objection 32), or, at least, if we spread through space, it may result in less suffering than if other spacefaring civilizations do so instead (Tomasik, 2011).

For the sake of argument, imagine that our civilization had both the motivation and the know-how to prevent the possibility of life — and consequently all suffering — on our planet. From a minimalist perspective, would such “planetary euthanasia” be the practically optimal aim of such a powerful version of humanity?[21]

Arguably not, because that very same technical capacity and value alignment could (if stable) probably be put to better use if our civilization were to become a safeguard against extreme suffering taking place elsewhere in the reachable universe. Consequently, the closer we are to being the kind of civilization that actually could prevent all suffering right here on Earth via some high-tech “planetary euthanasia”, the less clear it becomes that we should adopt that as our final aim. After all, such a civilization could plausibly play more positive roles by ensuring that the vast resources of the reachable universe would not become fuel for generating astronomically greater suffering.

(3) Similarly, if all humans or all sentient beings on Earth were killed, a new spacefaring civilization may eventually develop on Earth, and if it were to colonize space, it is an open question whether it would result in more suffering than if humanity were to expand into space (Tomasik, 2013c).

More generally, a key factor to estimate in the big picture is the potential difference in suffering that results from one colonization wave versus another, including colonization waves that may stem from other civilizations (Vinding & Baumann, 2021). For instance, if one colonization wave would use its resources in more suffering-conducive ways than would another, then it is more desirable (all else being equal, from a minimalist perspective) that the second colonization wave acquires those resources instead.

Risks of astronomical suffering

A reason to take seriously the considerations of cooperation, compromise, and empirical uncertainty — cf. the previous sections — is that they may be of crucial relevance to risks of astronomical suffering (Tomasik, 2011), which are also known as suffering risks or s-risks for short (Baumann, 2017b). Roughly, s-risks can be understood as “events that would bring about suffering on an astronomical scale, vastly exceeding all suffering that has existed on Earth so far” (Baumann, 2017a).[22]

A likely optimal aim for minimalists in practice is to reduce s-risks (cf. Baumann, 2020a). This is because s-risks are not extremely unlikely, are bigger than present-day suffering in expectation, are neglected, and reducing s-risks seems reasonably tractable (Baumann, 2017b).

Additionally, many of the most promising interventions for s-risk reduction seem to be robustly beneficial for reasons besides the independent weight of such risks. After all, we can also better reduce near-term suffering, as well as other risks, by developing greater levels of cooperation, impartial moral concern, and capacities to achieve shared aims between multiple value systems (cf. the positive proxies mentioned in Section 3.1).

Strong reasons to prioritize safer and more widely shared aims

In sum, minimalists have no a priori reason to aim at ‘no future’ over some equally ideal hypothetical future (cf. 3.1). And on a closer empirical and marginal realist analysis, it appears that adopting such an aim is unlikely to be optimal in practice. To the contrary, doing so would likely increase s-risks, whose reduction is the strongest candidate for a practically optimal aim for minimalists.

In practice, minimalists will benefit (quote):

… by cooperating and compromising with other value systems in trying to make the future safer in regard to (agreed-upon) worst-case scenarios … It would be a tragedy if altruistically-concerned people split up into opposing factions due to them having different definitions of “doing what is good”, while greed and bad incentives lead the non-altruistically-inclined people in the world to win the race. Instead, those who share at least some significant concern for the reduction of suffering should join together. (Cf. the NU FAQ.)

3.3 A safeguard against worst-case outcomes: Pragmatically absolute nonviolence

Quote:

I’m as near as one comes to [being] a pacifist as is possible without being a pacifist. [Yes], there are exceptional circumstances in which violence may be unavoidable; we all know that life is messy. But other things being equal, I think the sanctity of life is a very good utilitarian principle because it promotes respect for other sentient beings. (David Pearce.)

Anti-harm ideas have inspired uniquely non-violent practices for literally millennia. Many Jains and Buddhists aim to follow the principle of Ahimsa: ‘never hurt another sentient being by word or deed’. Yet impartial minimalism is not about minimizing our personal “hurt-footprint”. Instead, it recommends that we aim to minimize overall hurt for all sentient beings, regardless of the act-omission distinction (cf. 2.1).

Do minimalist views in practice diverge from (absolute) Ahimsa any more than do offsetting views? All experientialist consequentialist views recommend that we cause the lesser hurt when it is the only way to prevent a greater hurt (other things being equal). However, minimalist consequentialist views are unique in saying that such situations are the only ones in which we could ever be justified in hurting others. (By comparison, offsetting views in theory provide more ways to justify hurting others — whether by act or by omission — such as in order to create more happiness; cf. 2.5.)[23]

When we zoom out from the personal “hurt-footprint” perspective, we may see sophisticated minimalism as being based on a principle that is similar in spirit to that of Ahimsa, yet which is expressed in fully impartial terms: “The less sentient beings hurt, the better, regardless of the source.”

At the same time, the aim to minimize suffering (or other experiential problems) is not by itself a directly suitable principle for guiding practical action (Mayerfeld, 1999, pp. 120–125). In practice, this abstract aim calls us to follow more actionable principles — such as virtues and commonsense prohibitions — that indirectly tend to bring about the best consequences. And in order to prevent our corruptible parts from opportunistically breaking such principles in self-serving ways, we likely need to internalize these principles deeply into who we are.

(The road to hell is paved with good intentions that allow us to convince ourselves that we are in exceptional circumstances that warrant discarding the anti-harm precepts of commonsense morality. We rarely are.)

Finally, the sanctity of life (in roughly the way it is understood in Ahimsa) is plausibly a key principle to uphold as a safeguard against worst-case outcomes. This is because it promotes positive norms against aggression and in favor of peace and cooperation. (In many cases, nonviolent movements have also been more effective at achieving their altruistic aims; cf. Chenoweth & Stephan, 2011.)

Overall, an optimal safeguard against risks of astronomical suffering may be to promote respect for all sentient beings in the form of unambiguously compassionate principles, such as something like impartial Ahimsa or pragmatic non-violence and non-aggression. Without such principles, the risks seem worse. With them, we have more hope.

Acknowledgments

I am grateful for helpful comments by Riikka Ajantaival, Tobias Baumann, Timothy Chan, Eleos Arete Citrini, Anthony DiGiovanni, Simon Knutsson, Winston Oswald-Drummond, David Pearce, Michael St. Jules, Emma Tulanova, Magnus Vinding, and anonymous commenters.

Commenting does not imply endorsement of any of my claims.

I also wish to thank the anonymous authors of the NU FAQ, who made many key points about minimalist axiologies accessible in a concise and readable way already in 2015.

References

Acton, H. B., & Watkins, J. W. (1963). Symposium: negative utilitarianism. Proceedings of the Aristotelian Society, Supplementary Volumes, 37, 83–114. DOI.

Ajantaival, T. (2021a). Positive roles of life and experience in suffering-focused ethics. Ungated; EA Forum.

Ajantaival, T. (2021b). Minimalist axiologies and positive lives. Ungated; EA Forum.

Althaus, D. & Gloor, L. (2016). Reducing Risks of Astronomical Suffering: A Neglected Priority. Ungated.

Badaan, V. et al. (2020). Imagining better societies: A social psychological framework for the study of utopian thinking and collective action. Social and Personality Psychology Compass, 14(4), e12525. DOI.

Bartels, D. M. (2006). Proportion dominance: The generality and variability of favoring relative savings over absolute savings. Organizational Behavior and Human Decision Processes, 100(1), 76–95. Ungated.

Baumann, T. (2017a). S-risk FAQ. Ungated.

Baumann, T. (2017b). S-risks: An introduction. Ungated.

Baumann, T. (2019). Risk factors for s-risks. Ungated.

Baumann, T. (2020a). Arguments for and against a focus on s-risks. Ungated.

Baumann, T. (2020b). Common ground for longtermists. Ungated; EA Forum.

Baumann, T. (2021). How can we reduce s-risks? Ungated.

Chenoweth, E. & Stephan, M. J. (2011). Why civil resistance works: The strategic logic of nonviolent conflict. Columbia University Press.

Danaher, J. (2018). Popper’s Critique of Utopianism and Defence of Negative Utilitarianism. Ungated.

DeScioli, P. (2016). The side-taking hypothesis for moral judgment. Current Opinion in Psychology, 7, 23–27. Ungated.

DiGiovanni, A. (2021a). A longtermist critique of “The expected value of extinction risk reduction is positive”. EA Forum.

DiGiovanni, A. (2021b). Tranquilism Respects Individual Desires. Ungated.

Fernando, J. W. et al. (2018). Functions of utopia: How utopian thinking motivates societal engagement. Personality and Social Psychology Bulletin, 44(5), 779–792. Ungated draft; DOI.

Gloor, L. (2017). Tranquilism. Ungated.

Gloor, L. (2018). Cause prioritization for downside-focused value systems. Ungated; EA Forum.

Jost, J. & Hunyady, O. (2003). The psychology of system justification and the palliative function of ideology. European Review of Social Psychology, 13(1), 111–153. Ungated.

Knutsson, S. (2016). How Could an Empty World Be Better than a Populated One? Ungated.

Knutsson, S. (2021). The world destruction argument. Inquiry, 64(10), 1004–1023. Ungated; EPUB.

Mayerfeld, J. (1999). Suffering and Moral Responsibility. Oxford University Press.

Ord, T. (2015). Moral trade. Ethics, 126(1), 118–138. Ungated.

Pearce, D. (1995). The Hedonistic Imperative. Ungated.

Popper, K. (1945). The Open Society and Its Enemies. Routledge. Wikiquote.

Smart, J. J. C. (1989). Negative utilitarianism. In Freedom and Rationality (pp. 35–46). Springer. DOI.

Todd, B. (2021). Expected value: how can we make a difference when we’re uncertain what’s true? Ungated.

Tomasik, B. (2011). Risks of Astronomical Future Suffering. Ungated.

Tomasik, B. (2013a). Applied Welfare Biology and Why Wild-Animal Advocates Should Focus on Not Spreading Nature. Ungated.

Tomasik, B. (2013b). Gains from Trade through Compromise. Ungated.

Tomasik, B. (2013c). How Would Catastrophic Risks Affect Prospects for Compromise? Ungated.

Tomasik, B. (2013d). Omelas and Space Colonization. Ungated.

Tomasik, B. (2014a). Reasons to Be Nice to Other Value Systems. Ungated.

Tomasik, B. (2014b). Will Space Colonization Multiply Wild-Animal Suffering? Ungated.

Tomasik, B. (2015). Is Utilitarianism Too Demanding? Ungated.

Tomasik, B. (2016). Strategic Considerations for Moral Antinatalists. Ungated.

Vinding, M. (2015). Anti-Natalism and the Future of Suffering: Why Negative Utilitarians Should not Aim for Extinction. Ungated.

Vinding, M. (2020a). Suffering-Focused Ethics: Defense and Implications. Ratio Ethica. Ungated.

Vinding, M. (2020b). Why altruists should be cooperative. Ungated.

Vinding, M. (2022). Reasoned Politics. Ratio Ethica. Ungated.

Vinding, M. & Baumann, T. (2021). S-risk impact distribution is double-tailed. Ungated.

van der Deijl, W. (2021). The sentience argument for experientialism about welfare. Philosophical Studies, 178(1), 187–208. Ungated.

Wiblin, R. & Lempel, H. (2018). Ways people trying to do good accidentally make things worse, and how to avoid them. Ungated.

Notes

  1. ^

    Regarding minimalist views formulated in extra-experientialist or nonconsequentialist terms, it is worth noting that such views would have additional ways to respond to the questions above. For instance, such views might say that the outcome resulting from instant cessation could be suboptimal due to its involving other independently disvalued things, such as preference frustration, premature death, or rights violations (i.e. extra-experiential disvalue). Such views might also hold that the rightness of acts, like that of pressing a button or of engaging in supposedly justified coercion, is dependent also on nonconsequentialist factors, such as some properties of the actor or the action itself, and thereby deny that the outcome is the full picture even after extra-experiential disvalue is accounted for.

  2. ^

    As an example of omission bias, say we have two identical populations, P1 and P2, living on identical deserted islands, yet only P1 is afflicted by an infectious disease. Say also that we could immunize P1 against this disease at no cost to us, and that the population-specific outcome of our non-immunization of P1 is basically the same as if we were to actively introduce the disease to P2. Given these assumptions, if we still tend to consider the non-immunization of P1 more acceptable than the active introduction of the disease to P2, this suggests that we have a bias in favor of non-intervention.

  3. ^

    To stress test our intuitions about the independent value of isolated pleasure, we may consider a thought experiment in which we choose between a) transforming Earth into a minimalist paradise in which everyone is perfectly untroubled, or b) ending this world and replacing it with a sufficient number of happy Matrix-lives enjoying maximal bliss. (H/​T David Pearce and his virtually identical utilitronium shockwave thought experiment.)

  4. ^

    A different question, considered on the practical side, is whether Earth-originating life could play positive roles beyond Earth, such as by reducing the suffering caused by an alien civilization.

  5. ^

    A seemingly common view on personal identity (among experientialist consequentialists) is empty individualism or “constant replacement”. Empty individualism simply counts each “experience-moment” as a separate being, which casts serious doubt on how exactly interpersonal value aggregation and compensation are supposed to work (cf. DiGiovanni, 2021b).

    For ease of speaking, the current essay will talk of “beings”, “lives”, and “experience-moments” interchangeably. And the reader may indeed usefully interchange “experience-moments” with “beings” and “lives” in the thought experiments presented here, to see whether doing so makes a difference. After all, the combined assumptions of ‘all else being equal’, experientialism, and consequentialism arguably imply that this substitution should not make a difference. So this could be a way to notice the influence of our practical intuitions, which may track things like positive roles even when such factors are supposed to be theoretically ruled out.

  6. ^

    H/​T David Pearce.

  7. ^

    After all, a shared goal between minimalists and all proponents of welfarist views is to increase the quality of all lives (other things being equal) given their existence. By comparison, the ideal of an empty world is not nearly as universal. For more on practical reasons to prioritize common goals between different value systems, see Section 3.2.

  8. ^

    Thanks to Magnus Vinding for suggesting all three thought experiments.

  9. ^

    This thought experiment was also partly inspired by Knutsson, 2021, Section 3.

  10. ^

    In particular, if we assume “equal intensities” for the supposedly independent goods and bads in the diagram, then classical utilitarianism favors the choice of ‘Intense Bliss with Hellish Cessation’ (B) at the proportional scale that is shown in the diagram. For “negative-leaning” offsetting views to favor B, the large paradise needs to extend considerably more than the hell does.

  11. ^

    H/​T David Pearce.

  12. ^

    H/​T Magnus Vinding.

  13. ^

    On the other hand, non-experientialist offsetting views may imply that it is a net benefit to create a world with arbitrarily many suffering lives, no happiness, and a sufficiently large amount of (e.g.) complex knowledge and accomplishments.

  14. ^

    For additional theoretical implications of offsetting views, see e.g. Vinding (2020a, Chapter 3).

  15. ^

    The previous points are similar to how total symmetric utilitarians often argue that our intuitions about the ‘Repugnant Conclusion’ are unreliable.

  16. ^

    See also Section 3.3 on the practical side.

  17. ^

    A theoretical model by Badaan et al. (2020) also predicts that utopian thinking increases hope, which may foster collective action in more constructive ways than does anger at injustice. By providing a salient standard of comparison for the status quo, utopian thinking can make a desired future more cognitively accessible and emotionally motivating (Badaan et al., 2020).

    Without hopeful visions, we may quite strongly and universally feel a need to justify the existing system so as to reduce the cognitive dissonance of wanting to improve the system yet feeling unable to do so. As proposed by Jost & Hunyady (2003), “system-justifying ideologies serve a palliative function in that they reduce anxiety, guilt, dissonance, discomfort, and uncertainty”. By highlighting ways in which things could be better, utopianism enables us to reduce such feelings through constructive action, in turn replacing the need to palliatively rationalize or tolerify the existing system.

  18. ^

    An advantage of expected value thinking is its commitment to an impartial sensitivity to risk: Expected value encourages calculated risk-taking when it makes overall sense, yet discourages it when the risk is too high. By contrast, endstate-utopianism seems prone to ignore key parts of the full risk landscape, especially when the risks are not emotionally salient or easy to think about.

  19. ^

    For a brief and accessible book chapter on the importance of cooperation for reducing suffering, see Vinding (2020a, pp. 205–213).

  20. ^

    Of course, there are also limits to how nice one should be to other value systems, which depends on factors such as whether they are also nice to yours (cf. Tomasik, 2014a).

  21. ^

    Thanks to Nil for the question.

  22. ^

    Note that sometimes s-risks are defined in a different way. For example, Althaus & Gloor (2016) use the following definition: “Suffering risks are risks of events that bring about suffering in cosmically significant amounts. By ‘significant’, we mean significant relative to expected [action-relevant] future suffering.”

  23. ^

    It has also been argued that offsetting views are more likely, in practice, to lead to totalitarianism and to the use of coercive means than is a focus on reducing suffering. Specifically, Karl Popper’s essay “Utopia and Violence” defends such a claim in relation to views that justify present misery so as to eventually realize a supposedly compensatory ideal state of affairs (see also Popper, 1945, ch. 24; Danaher, 2018).