Five Tests for What Makes a Life Worth Living

1 Introduction

What makes a life worth living? This is not the same question as what makes a life good for the one who lives it.¹ A theory that answers the latter question is a theory of well-being (i.e., “welfare” or “prudential value”). The two questions are clearly related and they are often conflated. But most likely worth is not strictly a matter of welfare, since one can live a life of great hardship and suffering that might nevertheless be worth living. Prima facie compelling examples abound: The proverbial soldier in a foxhole who throws himself on a grenade to save his comrades does not enhance his welfare. Far from it. But he does improve the worth of his life.

We should distinguish between (1) a life worth starting and (2) a life worth continuing. The phrase “a life worth living” is ambiguous.² It might mean either. Here, I am concerned with whether life is worth starting, not whether it is worth continuing. Clearly, throwing yourself on a grenade is not a great way to make your life worth continuing. The best way to understand the difference is to see it at as one of duration. A life worth continuing is one where the period ahead is worth living.³ A life worth starting is one where the life as a whole is worth living. When I refer to a life worth living, I have in mind the life as a whole.

Worth and well-being are easily confused, as we might say that a life is “worth living for” someone. In such a case, we are interested in what’s in it for the individual. It appears that this usage strictly tracks the notion of welfare or prudential value. Surely welfare is part of what makes a life worth living, but it is not the entire story. Here, I am interested in the worth of a life, not merely the worth of the life for the one who lives it.

Just as some things that do not promote our self-interest are nevertheless worth doing, some lives low in welfare appear to be worth living.⁴ When we wonder whether some activity is worth doing, our only thought isn’t “What’s in it for me?” Intuitively, the same should hold for lives. Conversely, some lives high in welfare are not worth living. Most plausibly, a supremely happy Hitler does not live a life worth living. It would be highly counter-intuitive to suggest otherwise. More needs to be said, but these considerations suggest that worth and welfare are distinct.

Similarly, one might live a meaningful life that is relatively low in welfare. Indeed, sometimes we rationally hold the meaning of our life in greater esteem than our own good. Achilles’ choice illustrates the point. Achilles knew that if he entered the battle against the Trojans, he would die soon, but that his name would live on for ages; alternatively, if he stayed out of the fight, his name would be forgotten, but his life would be long and happy. He chose a short meaningful life over a long life high in individual welfare. He was not merely motivated by a desire for fame, but for achievement and, more significantly, revenge for Hector’s slaying of Patroclus. By entering the battle, Achilles did not enhance his welfare. But, plausibly, he did improve the worth of his life by making it more meaningful.

Conversely, it seems that a life that significantly advances horrendous evil is not worth living. Hitler, Pol Pot, and Stalin lived worthless lives. Killing millions of people is hideously evil. And, most plausibly, moral repugnance is sufficient to sap a life of positive worth. But it is not the only thing to do so. Intense pain can suffice. By any account, lives spent in persistent, incapacitating agony are not worth living.

As these examples show, there are clear cases of lives not worth living. And there are not so clear cases. Worth comes in degrees. Lives entirely consumed by meaningless activities, such as counting blades of grass, collecting rubber bands, or making handwritten copies of War and Peace, appear to be less worth living than those spent in pursuit of valuable ends, but it is not so clear that they are not worth living.⁵ Worth appears to be distinct from, or over and above, welfare and meaning.⁶

…

I will assume that these distinctions are plausible. Clearly, claims of this significance require extensive support. But I do not intend to defend them here. That would require at least another paper. Nor will I attempt to develop a theory of worth or explore the policy implications any test for worth.⁷ It is not my goal to develop a test by which someone could effectively evaluate their life. Nor am I offering a handbook for evaluating lives. Instead, I undertake a preliminary task—to find a way to track the general extension of the concept.⁸

To that end, I identify four distinct tests for what makes a life worth living: (1) The Suicide Test (Camus), (2) The Recurrence Test (Schopenhauer and Nietzsche), (3) The Extra Life Test (Cicero and Hume), and (4) The Preferring Not to Have Been Test (Williams and the Book of Job).⁹ I argue that all four fail in their most plausible formulations.¹⁰

In response to Saul Smilansky’s objections, I tentatively defend a fifth—The Pre-Existence Test for what makes a life worth living: (5) A life worth living (LWL) is one that a benevolent caretaker with foreknowledge would allow. A life worth avoiding (LWA) is one that a benevolent caretaker would disallow. I consider three objections and note that there appears to be an indeterminate middle category of lives worth neither (LWN).

I argue that the pre-existence test usefully tracks the general extension of the concept of what makes a life worth living. But it does not tell us what makes a life worth living. For that we need a theory of worth. Further, in some cases we will require a theory of worth to say whether or not a life passes the pre-existence test. As I will argue, this is a limitation on any non-trivially true, objective test for what makes a life worth living. But this limitation is not a liability for the present project. I intend the test to aid the development of a theory of worth, not to serve as analyses of the concept. The test helps bring the concept into focus. But it is a ladder that we can kick away when we get to the point where the concept is in clear view. At that point we can develop a theory of worth.

The tests I consider are not as thick as full-blown theories of what makes a life worth living; nor are they as thin as mere definitions of the concept. A test is something in between: a useful test provides an epistemic indicator of when worth obtains. A test does not provide a metaphysical account of why some lives are worth living. Again, for that we need a theory of worth.

This paper has two central goals. First, I aim to provide a careful consideration of a suite of historically precendented and intrinsically interesting attempts to provide tests for what makes a life worth living. Second, through the evaluation of the candidate tests, I hope to identify some useful indicators of worth and to eliminate those that merely mislead. Hence, this project is both propaedeutic to developing a theory of worth and of value in itself.

2 Four Failed Tests

2.1 Test 1: The Suicide Test

As every sophomore knows, and every philosophy professor is tired of hearing, Camus claims that “There is but one truly serious philosophical problem, and that is suicide. Judging whether life is or is not worth living amounts to answering the fundamental question of philosophy.”¹¹ Apparently Camus thinks that there is a connection between whether a life is worth living and whether suicide is appropriate.¹² He says only that: “One kills oneself because life is not worth living, that is certainly a truth—yet an unfruitful one because it is a truism.”¹³ Unfortunately he does not make the connection clear. Accordingly, I will put aside Camus’s project and abstract two general versions of a suicide test: a subjective version and an objective version.

For reasons of historical accuracy, it pays to note that Camus would likely object to the objective version, as he purports to make no value judgments: “value judgments are discarded here in favor of factual judgments.”¹⁴ But this raises a question of the coherence of his project. The Myth of Sisyphus asks whether life is worth living given its absurdity. The problem is that if absurdity is a form of nihilism, then it is hard to see how it would exempt the value of worth or, alternatively, the value of his favored “revolt.”¹⁵ Calling a value judgment a “factual judgment” does not make it any less normative.¹⁶ This confusion gives us an additional reason to dismiss Camus and just look at the tests.

First consider an unrefined subjective version:

• Subjective Suicide Test (SST): Life L is worth living for person P iff P does not desire to end L.

SST is highly implausible as it stands, but its problems are instructive. It faces two fatal difficulties. First, and most obvious, we can be mistaken about whether our lives are worth living in ways that the defender of SST cannot explain. For instance, after breaking up with her latest boyfriend of two weeks, a depressed adolescent might think that her life is not worth living. Sadly, she might want to commit suicide. But this does not mean that her life is not worth living. She is mistaken. Hence, the desire to commit suicide is not sufficient for a life to be not worth living.

I will forgo consideration of an idealized subjective test. On any plausible idealization, even though she might know that she would be happy later, she might not care now. Perhaps this is irrational. If we add a qualification that P rationally desire suicide, the test suffers from the problems I will raise for the objective version below.

Hume exposes a second problem for SST. In Book X of the Dialogues Concerning Natural Religion, Philo argues for a thoroughgoing pessimism about the human condition. He briefly considers an objection to his dismal assessment: If life were so bad, why don’t we find more people committing suicide? Surely we are not all hopelessly deceived into thinking that our lives are worth living. Philo replies: The reason we don’t find people killing themselves in droves is because most everyone is afraid of death. We fear that death might bring something even worse than life: “We are terrified, not bribed into continual existence.”¹⁷ Or, as Hamlet puts it:

Who would fardels bear,

To grunt and sweat under a weary life,

But that the dread of something after death,

The undiscovered country from whose bourn

No traveler returns, puzzles the will

And makes us rather bear those ills we have

Than fly to others that we know not of?

The conscience does make cowards <of us all>¹⁸

Although I think that there are good reasons to reject Philo’s pessimism, his reply reveals a problem with SST: One’s life might not be worth living even though one does not desire to commit suicide. Hence the desire to commit suicide is not necessary for a life to be not worth living. Since the desire to commit suicide is neither necessary nor sufficient, SST is false.

Similar problems plague the more viable, objective version of the suicide test. According to an objective test, the worth of one’s life is not determined by the attitude that one takes toward that life; instead, the worth of the life determines the appropriate attitude. Rather than holding that a life worth living is merely one where the person does not want to commit suicide, the objective version holds that a life worth living is one where the person should not commit suicide. Conversely, a life worth avoiding is one where the person should commit suicide. (I will leave the nature of the “should” ambiguous for now; it helps expose a bigger problem.)

• Objective Suicide Test (OST): Life L is worth living for person P iff P should not end L.

To avoid problems that might arise from mistaken self-assessments, the most plausible version of the objective test must hold that P is well informed about L, about whether it will be filled with happiness or misery, whether it has good or bad effects on the world, etc. There is no need to spell out the idealization in more detail. Although OST is far more plausible than SST, it fails for the same reason Philo and Hamlet give us to reject SST.

Here we are primarily concerned with the evaluation of mortal lives, but I think we should also consider the fate of the tests if there is an afterlife. The literature on the topic from Cicero to Hume considers the possibility of an afterlife. I will follow suit.

For the time being, assume that an afterlife would be a separate life, that it is not part of L. The problem for OST is that P might live an utterly miserable, worthless life, but the afterlife for P might be even worse. Most plausibly, it would be better to prolong a torturous life than to hasten one’s trip to hell. Conversely, one might live a life worth living, but the afterlife might afford much greater pleasures and other goods. In such a case it might be rational to commit suicide. Hence, whether one should commit suicide is neither necessary nor sufficient for whether a life is worth living. Accordingly, OST does not track the general extension of the concept of worth.

Alternatively, if the afterlife should be considered part of the life of P, as I suggest later, suicide might be an appropriate way to hasten one’s trip to paradise, assuming that suicide does not preclude passage through the Pearly Gates. In such a case, one’s life (considered inclusive of the afterlife) would be worth living even though one should commit suicide.

We can construct a version of the test that gets around the problems arising from better and worse afterlives:

• Exempted Objective Suicide Test (EOST): Life L is worth living for person P iff P should not end L, assuming that the afterlife for P wouldn’t be worse than L.

Of course, it will be difficult to spell out “worse” without making the test circular, but there are more serious problems. EOST suffers from a similar defect to OST. Just as it might be better to quicken one’s transition to an afterlife, it also might be best to commit suicide if there would be tremendous prudential benefits to others, say, if self-immolation would result in the downfall of an oppressive regime. If human lives do not have infinite value, then it is possible that suicide could sometimes result in much greater value for the world. The problem is that one’s life might indeed be worth living even though one should commit suicide. Perhaps such a suicide might make one’s life even more worth living. Either way, EOST is wrong. (The suicide test simply cannot get off the ground if human life has infinite value. So, any plausible version of the test must handle this possibility.) The test requires yet another adjustment.

• Double Exempted Objective Suicide Test (DEOST): Life L is worth living for person P iff P should not end L, assuming that the afterlife for P wouldn’t be worse than L and the benefit to others from ending L does not outweigh the value of L.

DEOST is better, but it faces an inverse moral problem. A life that is not worth living might nevertheless be wrong to end, say, if one’s children would suffer greatly as a result.¹⁹ To get around this difficulty, we need yet a third exemption:

• Triple Exempted Objective Suicide Test (TEOST): Life L is worth living for person P iff P should not end L, assuming the afterlife for P wouldn’t be worse than L, the benefit to others from ending L does not outweigh the value of L, and it wouldn’t be wrong for P to end L.

At this point, one might worry that TEOST does not give us an informative test for what makes a life worth living. As noted above, the “worse” in the first exemption appears to make the definition circular. Even more exemptions are on the horizon. If the “should” is an all-things-considered should, the test tells us nothing. It merely amounts to a statement that one should commit suicide if one’s life is not worth living.

Putting aside concerns about the informativeness of this convoluted test, there is a fatal problem. It’s a matter of timing. At what point in the life should P end L? The life ahead of P might be miserable, but this does not mean that P’s life on the whole is not worth living. Conversely, the past may include so much pain and suffering that P’s life on the whole is not worth living, nevertheless the period ahead might be good. The life might be worth continuing, but not worth living on the whole.

These timing considerations show that the distinction I drew in the introduction between (1) a life worth starting and (2) a life worth continuing is important. Remember, here we are concerned with whether life is worth starting; we are concerned with the life as a whole, not whether it is merely worth continuing. The only point in time at which TEOST could be used to assess the life as a whole would be at the beginning, but this is no longer a suicide test. No, it is an infanticide test. Infants cannot kill themselves. I will consider a similar suggestion later, but for now, we can put aside the suicide test. No matter how many exemptions we tack on, it will fail to track the extension of the concept of what makes a life worth living.

2.2 Test 2: The Recurrence Test

We find the seed of a more plausible test in the writings of Schopenhauer and Nietzsche. Both make use of the Pythagorean notion of eternal recurrence. Regardless of their intentions, the trope can serve as a test for what makes a life worth living.

Although Schopenhauer thinks that no human lives are worth living, he acknowledges that some people are more optimistic than he purports to be. The most optimistic embrace life; they are able to will their eternal recurrence. But Schopenhauer thinks that this requires a degree of blindness to the facts of life:

A man who had assimilated firmly into his way of thinking the truths so far advanced, but at the same time had not come to know, through his own experience or through a deeper insight, that constant suffering is essential to all life; who found satisfaction in life and took perfect delight in it; who found desire, in spite of calm deliberation, that the course of his life as he had hitherto experienced it should be of endless duration or of constant recurrence; and whose courage to face life was so great that, in return for life’s pleasures, he would willingly and gladly put up with all the hardships and miseries to which it is subject, such a man would stand “with firm strong bones one the well-grounded, enduring earth,” and would have nothing to fear.²⁰

Later in World as Will and Representation, Schopenhauer claims that those not so blind to the realities of existence would be unwilling to live again:

No man, if he be sincere and at the same time in possession of his faculties, will ever wish to go through it again. Rather than this, he will much prefer to choose complete non-existence. The essential purport of the world-famous monologue in Hamlet is, in condensed form, that our state is so wretched that complete non-existence would be decidedly preferable to it.²¹

Unlike Schopenhauer, Nietzsche’s familiar use of the trope appears to be prescriptive. He suggests that we should live in such a way that we could will our eternal recurrence.

What, if some day or night a demon were to steal after you into your loneliest loneliness and say to you: “This life as you now live it and have lived it, you will have to live once more and innumerable times more; and there will be nothing new in it, but every pain and every joy and every thought and sigh and everything unutterably small or great in your life will have to return to you, all in the same succession and sequence—even this spider and this moonlight between the trees, and even this moment and I myself. The eternal hourglass of existence is turned upside down again and again, and you with it, speck of dust!”²²

Nietzsche’s probe easily lends itself to a recurrence test for a life worth living: A life worth living is one that we should choose to live again. We can construct strong and weak versions of the recurrence test.²³ Here are two candidates:

• Weak Recurrence Test (WRT): Life L is worth living for person P iff P should will to live an intrinsically qualitatively identical life over again.²⁴

• Strong Recurrence Test (SRT): Life L is worth living for person P iff P should will to live an intrinsically qualitatively identical life over and over again eternally.

WRT appears to set a much lower bar for a life worth living than does SRT. But one might doubt that the tests are extensionally inequivalent. According to WRT, if life L1 is worth living, then one should will to live it again, to live L2. L1 and L2 would be largely identical. They would be intrinsically qualitatively identical. And they would have the same impact on the world. They would differ only in that L2 came after L1. Accordingly, it would be odd to say that L1 is worth living but that L2 is not. This sets up a regress: If L2 is worth living, according to WRT, one should be willing to live it again, to live L3. And so on. If there is no principled way to stop the regress, any life that passes WRT must also pass SRT. If a life is worth living once according to WRT, it would have to be worth living over and over again eternally.

But there is a way to stop the regress. It shows that WRT and SRT are extensionally inequivalent. The source of the difference is also the source of the problem for both tests. The central problem with WRT is that one can reasonably decide not to live a life worth living over again. This is because we have good reason not to repeat ourselves.²⁵ Just as we do not want to endlessly row in place, we might not want to repeat our entire lives, much less do so indefinitely.²⁶ Herein lays the difference between L1 and L2.

Intuitively, each time one relives a life, some of what makes it worth living is diminished. It is not entirely clear what accounts for the disvalue of this peculiar kind of stagnation, but there is at least one good candidate: the achievement value is diminished. Although achievement value appears to be largely a matter of the relative difficulty of a task and the value produced, intuitively, the historical originality of an achievement is also relevant.²⁷ Consider a serial amnesiac who repeatedly writes a great poem. Most plausibly the accomplishment is less valuable the 100^th time he composed the poem than it was the first time. Similarly, even though it is a wonderful achievement to invent calculus, it is not as wonderful if someone beats you to it.

The difference in value between repeated lives only needs to be tiny for WRT and SRT to be inequivalent and for both to fail as tests for what makes a life worth living. Once again, consider L1 and L2: If L1 is just barely worth living, L2 might lack the value needed to cross the threshold. In this scenario, L1 would be worth living, but P should not will to live it over again, since L2 would not be worth living. The greater the disvalue of repetition, the more significant the problem for WRT and SRT. Accordingly, both tests fail to track the general extension of the concept of a life worth living.

There is an even more troubling problem for the recurrence test. The problem is revealed by simply asking for one more detail: Would we remember our past lives? Some think that this is the best way to conceive of the test.²⁸ Regardless of Nietzsche’s intention, this is a horrible mistake. It would not be the same life; it would not be qualitatively identical, if one remembered one’s previous existence. Memory of former lives would radically alter one’s experiences. Knowing everything that will happen before it happens would be maddening, especially if one could do nothing to alter any outcome. Imagine experiencing a life-long episode of déjà vu. This would be far worse than Phil’s (Bill Murray) fate in Groundhog Day (Ramis, 1993). I would not wish that on anyone. All lives fail this test.

Alternatively, as Ivan Soll notes, if the future person lacks memory of her former lives, it is far from clear that she has any prudential reason to care about the future person.²⁹ Whether a life exactly like the one I am now living will repeat seems entirely irrelevant. I couldn’t care less. I could will that almost any life recur for all infinity. Either way, the test fails.

2.3 Test 3: The Extra Life Test

Cicero imagines that on his deathbed the great Cato, who lived a life of learning and accomplishment, said that if he were given the option of living yet another life that he would decline:

Assuredly no one will easily draw me back, or boil me up again, as if I were a Pelias. Nay, if some god should give me leave to return to infancy from my old age, to weep once more in my cradle, I should vehemently protest; for, truly, after I have run my race I have no wish to be recalled, as it were, from the goal to the starting place. For what advantage has life—or, rather, what trouble does it not have.³⁰

In The Dialogues Concerning Natural Religion, Hume has Philo refer to Cicero’s remark as a source of support for his avowed pessimism.³¹ If, as Cicero imagines, Cato, who lived one of the best possible human lives, would not choose to live another life, then human life in general must not be worth living. The only way to draw this pessimistic conclusion from Cato’s remarks is to assume that someone who lives a life worth living would choose to live yet another life if given the chance.

Here we find a third test for what makes a life worth living:

• Extra Life Test (ELT): Life L is worth living for person P iff P would choose to live another life, L′, after L.

This test suffers from the principal problems of the suicide test and more. Philo fails to mention that Cato’s remarks appear in the context of a discussion of immortality. Just prior, Cato professes a belief in the immortality of the soul. He refuses to live another life because he eagerly awaits the afterlife, not because his life was not worth living. We could adjust ELT to get around this problem, as we did with the suicide test, but there is no need. ELT faces more serious difficulties.

Unlike the suicide test, ELT faces pressing problems of personal identity. Echoing Soll’s worries about the coherence of the recurrence test, Bernard Williams notes in “The Makropulos Case” that the difference between birth and rebirth is dubious at best.³² Lucretius states the worry concisely:

For if the mind has undergone a transformation so vast

Enough to cancel out all the remembrance of things past,

That is a state approaching little short of death, I’d say.

Therefore you must admit that what it was has passed away,

And that which it is presently has been created fresh.³³

It’s not clear that on any plausible theory of personal identity we should say that P could live two different lives, that P would be “reborn” into L′. Tiresias was not reborn as a woman; he was changed into a woman (and back again). ELT proposes something much different. P is not faced with the option to continue living with different sex organs; no, P is asked if she wants to live yet another life. The problem is that it is not clear how it could be P who lives life L′ and not another person, P′.

We can put this worry aside. Even if we assume that it would be P living life L′, it is conceivable that L could be worth living, but P might, nevertheless, not be willing to live L′. Although Cato’s refusal certainly expresses a degree of world-weariness, his reluctance to live again does not give us a plausible test for what makes a life worth living. He might simply have had enough. In fact, he says as much: “Old age is the final scene, as it were, in life’s drama, from which we ought to escape when it grows wearisome and, certainly, when we have had our fill.”³⁴ In a subjective version, ELT merely tracks one’s zest for life. Cicero’s Cato was world-weary. This is the conclusion we should draw from his deathbed remarks, not that his life was not worth living.

There is another, related problem for ELT. Walter Kaufmann suggests that the best kind of life would be one where the person lives in such a way that more would be less: “The life I want is a life I could not endure in eternity. It is a life of love and intensity, suffering and creation, that makes life worth while and death welcome. There is no other life I should prefer. Neither should I like not to die.”³⁵ He quotes from a poem by Holderlin that expresses this sentiment:

A single summer grant me, great powers, and

A single autumn for fully ripened song

That, sated with the sweetness of my

Playing, my heart may more willingly die.

… Once I

Lived like the gods, and more is not needed.³⁶

Although Kaufmann’s remarks are directed at immortality, the sentiment is applicable to ELT. Having lived an intensely engaged life, it is conceivable that one might not want to live yet another life. To refuse another life one need not be world-weary, merely world-sated.

As for an objective version of ELT, I have no idea how to make such a test credible. I am at a loss for any reason that someone should want to live another life. Perhaps one might argue that by declining we would miss out on the goods of another life. But, once again, it is not clear to me how a different life could be an object of hope, even if identity holds. If I were faced with the options of death or permanent irrevocable amnesia, I’m not sure that I could be bothered to flip a coin. The same holds for another life.

2.4 Test 4: The Preferring Not to Have Been Born Test

In “Resenting One’s Own Existence,” Williams develops a fourth test for what makes a life worth living: “I see no way of denying that one who resents his own existence prefers that he should not have existed; and no way of interpreting that preference except in terms of thinking that one’s life is not worth living.”³⁷ Williams captures the thought behind Job’s wish.³⁸ After Job loses his children and his wealth, he curses the day he was born:

Let the day perish in which I was born, and the night that said, ‘A man-child is conceived.’ Let that day be darkness! May God above not seek it, or light shine on it, Let gloom and deep darkness claim it, Let clouds settle upon it; let the blackness of the day terrify it. […] Why did I not die at birth, come forth from the womb and expire?³⁹

In Job’s lament and Williams’s suggestion, we find a fourth test for what makes a life worth living:

• Preferring Not to Have Been Born Test (PNT): Life L is worth living for person P iff P does not prefer not to have been born.

This is the best test we have considered so far. PNT avoids the temporal problem facing the suicide test; it addresses the value of the entire life, not just the time remaining. It avoids the disvalue of repetition facing the recurrence test. And it avoids both the problem of personal identity and the problem of lack of zest facing the extra life test. But PNT is not without difficulties.

Saul Smilansky raises several problems with Williams’s subjective, first-person test.⁴⁰ The central problem is this: One might think that one’s life is worth living, but nevertheless think that it would have been better not to have been born, or vice versa. One might think this for a variety of reasons. For instance, severe self-loathing might make one wish to have never been born, but one might nevertheless live a life worth living. Alternatively, an intense episode of pain in the past might make one’s life not worth starting, but not make it unworthy of continuing.

Although it appears troublesome, on further consideration the episode of pain objection is not a problem for PNT. Smilansky offers an example of a cancer survivor who finds her life worth living, but thinks that an intense bout of suffering in the past makes it such that it would have been better if she were never born. Her life is worth living, but she thinks it would have been better never to have been. Hence, the objection concludes, PNT is false. In response, we should note that this objection conflates a distinction I drew earlier, a distinction between: (1) a life worth starting and (2) a life worth continuing. The cancer survivor’s life is worth continuing, since the period ahead is worth living; nevertheless, the life might not be worth starting. There is no problem for PNT here.

Smilansky’s other objection is more serious. He asks us to imagine a self-loathing, but otherwise happy child molester who wishes that he were never born, but whose life is worth living. The child molester objection reveals a fatal problem for Williams’s test, but not for the reasons Smilansky suggests. Since Smilansky restricts his discussion to lives that are subjectively worth living, he fails to see that lives might be objectively worthless in absence of a preference to never have been. On some subjective standard, it might be that such a life is worth living in spite of self-loathing. But this is not the sense of “worth” that I am pursuing.

Just as we might ask if an activity is objectively worth performing, we can ask the same of lives. Most important, one can be wrong about whether one’s life is worth living. Hitler might not have wished that he had never been born. In the moments before he committed suicide, although upset that the final solution would not be completed, he might have been pleased that he killed so many Jews. He might have thought that although his life could have been better (had he managed to exterminate all the Jews), it was nevertheless worth living. He might have thought this. But he would have been wrong.⁴¹ By any plausible account, a life that furthers hideous evil is not worth living. Since orchestrating the murder of six million Jews and six million other ‘undesirables’ — Poles, Gypsies, and homosexuals—is not an activity worth doing, it is not worth doing well. And a life devoted to activities that are profoundly not worth doing is not a life worth living. It might be a supremely happy life. That’s not at issue. Nevertheless, a supremely happy Hitler does not live a life worth living. That much seems clear. As a subjective test, PNT cannot account for this fact. Hence, it fails as a test for what makes a life worth living.

Nor can PNT account for simple mistakes, such as the distraught adolescent who wishes that she had never been born because her boyfriend of two weeks broke up with her. In this regard, PNT is no better than the first subjective test that we considered—SST, the subjective suicide test. It appears that any plausible test for what makes a life worth living must be an objective test.

I turn now to develop a test that avoids the problems facing the four we have so far considered.

3 Test 5: The Pre-Existence Test

I propose a test close to an objective version of PNT. Whether a life is worth living is not a matter of whether the person living the life wishes she had never been born, but whether she should have been born. Here is a compelling way to think about the test: A life worth living is a life that a benevolent caretaker, given a synoptic preview, would allow someone to live rather than to never have been.

• Pre-Existence Test (PET): Life L is worth living for person P iff a benevolent caretaker with foreknowledge of the facts about L would allow P to live L.⁴²

Imagine that the caretaker is a sympathetic judge who wants everyone to live a life they should not resent given all the facts, that is, their benevolence extends beyond concerns about welfare.⁴³ I specify that the caretaker is benevolent to preclude scenarios where the person is to be used merely as a means to some greater cosmic end without regard to his own good. Benevolence precludes this.

PET allows that the caretaker might allow P to live a miserable life for the sake of a valuable afterlife or for the purposes of soul-making. Rather than build in an afterlife exemption as I did for the suicide test, for the sake of argument, I will assume that an afterlife is part of the same life of P. This is a personal-identity first notion of a life, rather than a biological notion. I will not defend a formal analysis of a life here. But, roughly, the life of P is the collection of facts about P’s experiences, actions, and effect on the world. Either way, it seems perfectly reasonable to think that whether or not there is an afterlife would affect the worth of our biological lives. If so, the benevolent caretaker should take this into account.

As stated, PET captures the thought behind another version of Job’s complaint:

Why did you bring me forth from the womb? Would that I had died before any eye had seen me, and were as though I had not been, carried from the womb to the grave.⁴⁴

PET sets a fairly low bar for a life worth living. If a life passes the test, it is worth living. But the test does not tell us what makes lives worth living. Nor does it tell us what makes some lives more worth living than others. Instead, it is merely a plausible way to track the general extension of the concept. The test captures the idea that a life worth living is a choiceworthy life.⁴⁵

PET avoids the problems and absorbs the lessons learned from the suicide test. When considering the suicide test, I noted a timing issue. The life as a whole must be evaluated from the beginning. But that precludes suicide; it only allows for infanticide. PET has the virtues of an infanticide test while avoiding concerns about the morality of infanticide and abortion. Morally, PET is akin to a counterfactual contraceptive administered by a well-informed guardian.

In addition, the pre-existence test avoids the problems facing the recurrence test. The recurrence test asks whether one would choose to live one’s life over again, not whether one should allow it to start.⁴⁶ The two questions are not equivalent, since, as we saw, one can coherently decide not to live a life over again that one should choose to start. PET concerns but a single life.

Unlike the preferring not to have been test, PET is not based on subjective preferences; instead, it is a test for whether a life is objectively worth living.⁴⁷ But this does not mean that the pre-existence test excludes the subjective experience or the felt satisfaction of the one living the life. Most plausibly, these are factored into the overall welfare value. And certainly the welfare value of a life is relevant to whether it is worth living. Once again, the question is not whether given your preferences would you choose to be born, but rather, would a benevolent caretaker who knows all the facts of your future life allow you to be born.

We might call this the “pre-crib test.” Fred Feldman proposes a “crib test” for isolating the concept of welfare: Imagine looking down at your infant child in his crib.⁴⁸ Think of all the things that you want for the child: close friends, a good education, an interesting career, and the like. These are things that contribute to the child’s welfare. The crib test is designed to hone in on welfare considerations. But it does not do so cleanly. An insane parent might have strange desires for her child. As Ben Bradley imagines, she might want him to become the first person to consume an entire helicopter by grinding it up and eating it bit by bit.⁴⁹ Or a religious fanatic might want her child to become a martyr for the faith. But these desires are not for the welfare of the child.

The problems for the crib test are not confined to the desires of insane and fanatical parents. Most people, apart from gangsters and other psychopaths, prefer that their children be decent people, even if virtue does not always make them better off prudentially. Hence, it appears that the crib test does a better job at tracking the sane, moral parent’s conception of what makes a life all-things-considered good than it does at tracking what is merely good for their child.

As a pre-crib test, PET is designed to explicitly incorporate broader considerations, such as concerns about meaning. Many think that meaning is best evaluated from a synoptic first-person perspective—the deathbed.⁵⁰ Imagine lying on your deathbed thinking about all the things you wish you had done. Such thoughts likely concern meaning. For instance, you might wish that you had taken a rare chance to do something big, rather than chosen the safe, comfortable option. The deathbed test is a useful tool for tracking concerns about meaning, but it, too, suffers from over-inclusiveness. In the moments before death, just as one might regret a dearth of accomplishments, one might also regret not having had enough fun. Perhaps, once again, the deathbed test tracks what makes a life all-things-considered good better than it tracks concerns about meaning alone.

I suspect that the crib test and the deathbed test both help us identity what we think makes a life worth living. But neither is wholly adequate to the task. The pre-existence test fares better. It is designed to incorporate concerns about both welfare and significance. In addition, the pre-existence test is superior to the crib test and the deathbed test, since it avoids the troublesome relativity of both.⁵¹ A selfish parent wants the wrong things for her child. And an evil person on his deathbed might regret not causing enough pointless suffering. Plato reveals a similar problem in the “Myth of Er,” where he imagines souls deciding on the kinds of lives they wish to lead in their next incarnation. Just as some of Plato’s souls make poor choices in future lives, we can safely imagine that they might also poorly assess their former lives.⁵² In contrast, the ideal evaluator in the pre-crib test suffers no such defects.

4 Objections

As a method of identifying lives worth living, the pre-existence test is not without problems. In this section, I consider two difficult problems for PET in particular and a third problem that limits the usefulness of any plausible objective test.

4.1 The Sacrificial Problem

A problem for the pre-existence test comes from lives that involve a great deal of self-sacrifice. One might be willing to live such a life, but not obligated to do so. It might be supererogatory. But we can ask: What should the caretaker decide? It appears that by allowing one to be born, the caretaker would thereby make the sacrifice obligatory. That would be unfair. So, we might think that a benevolent caretaker should prevent such lives. The problem is that such lives are often worth living. Hence, the pre-existence test gives the wrong answer.

In response, note that the caretaker does not make any life obligatory. The caretaker is aware of the choices we make. Certainly, the moral worth of a choice to go beyond the call of duty enhances the value of a life. Accordingly, the caretaker would factor this into her evaluation. Hence, she will not prevent all forms of supererogatory self-sacrifice. However, I see no reason to think that all supererogatory actions are worth the price. The caretaker knows the value. Unless supererogation has infinite value, it is not guaranteed to make a life worth living. Once again, there is no problem for the pre-existence test here.

Perhaps the issue is more compelling if we consider a related kind of life. There are surely lives that produce a great amount of good at the cost of great hardship. We can imagine lives much like that of Victor Frankl’s medical ape, who is repeatedly poked with long needles to extract a spinal fluid required for the production of a life-saving serum.⁵³ The ape cannot see the good that results. He cannot hope to understand the point of his suffering. Similarly, we might not always be able to see the good that results from our lives. The question is this: Should the caretaker allow one to live such a life? This is a complicated problem, but it is not a problem for the pre-existence test. Instead, it is a problem for the caretaker, or more precisely the criteria by which the caretaker’s decision should be made. Since it is out of scope to develop a theory of worth here, I will simply note that although such lives might be meaningful, it is not always the case that they are worth living.⁵⁴

This shows that in hard cases we will not be able to specify the extension of the concept without a theory of worth. But this does not jeopardize the usefulness of the pre-existence test for aiding theory development. It merely shows why we need to develop a theory.

4.2 The Borderline Problem

A second problem for the pre-existence test comes from borderline cases. There are clear cases of lives worth living. Most plausibly, these are high in various objective goods, such as friendship, romantic love, knowledge, pleasure, meaning, and moral worth. Informed of all the facts of such a life, a benevolent caretaker would surely allow someone to live such a life. Conversely, some lives are clearly not worth living. As noted earlier, a life that significantly advances horrendous evil is not worth living. Likewise, lives spent in persistent, incapacitating agony are not worth living. As these examples indicate, there are clear cases of lives not worth living.

The problem for the pre-existence test is that there are not-so-clear cases. Worth comes in degrees. Lives entirely consumed by meaningless activities, such as collecting rubber bands or making handwritten copies of War and Peace, appear to be less worth living than those spent in pursuit of valuable ends.⁵⁵ They appear to be wasted.⁵⁶ But are they not worth living? It is hard to say in cases where the welfare level is at least modest. In addition to the problems that one encounters when trying to weigh various kinds of goods and bads, there appears to be a grey zone. Yeates proposes that between the thresholds of lives worth living (LWL) and lives worth avoiding (LWA), there are lives worth nothing (LWN).⁵⁷ I am not sure that this is the best term for the category, since these lives might have some value. But his suggestion that there is a middle category is plausible. To avoid the misleading connotations of Yeates’s label, we can refer to these as “lives worth neither” (LWN).

Assuming that there is a middle category, the question is this: Should a benevolent caretaker prevent one from living a LWN? It is not clear. But the test forces a decision. If the caretaker does not consent, then we are certainly not in life worth living territory. However, we also might not be faced with a full fledged life worth avoiding. But then why should the caretaker prevent it? The problem is that the test is binary, but threshold concepts are not. This problem is not insurmountable. A plausible solution is to say that LWN’s are simply those where the balance of reasons favors neither.

There is no need to explore this issue further here. Since the pre-existence test is not an analysis of the concept of a life worth living, the objection does not cast doubt on its usefulness in helping us track the general extension of the term. For present purposes, what matters is what makes a life worth living. The key question is this: What should ground the decision of the benevolent caretaker? To answer that question, we need a theory of what makes a life worth living.

4.3 The Circularity Problem

PET is an objective test for what makes a life worth living. Unlike a subjective test, it does not hold that the worth of a life is constitutively dependent on one’s reactions to one’s life. All subjective tests face grave difficulties accounting for mistaken self-assessments. Again, before committing suicide Hitler might have thought he lived a life worth living, but he would have been wrong. Conversely, Abraham Lincoln might have thought that his life was not worth living, but, most plausibly, he would have been wrong. Subjective tests make similar mistakes. Any plausible test for what makes a life worth living must avoid this problem. The only viable way to account for these kinds of mistakes is to incorporate an objective evaluative standard. PET does just that. Rather than appeal to the reactions of the one living the life, PET relies on an ideal observer—a benevolent caretaker with foreknowledge of the facts of that life. Any plausible test will have a similar structure. At minimum, the evaluator will require full information of the facts of the life. But this raises a serious problem.

To see the problem, consider the fate of the test if instead of a caretaker, we substitute God. Put aside concerns about the problem of evil. There are more serious difficulties for the test if one substitutes God or any other omniscient being. An omniscient being would know all the facts about L, including the evaluative facts. This would make the test vapid and circular: Life L for person P is worth living iff a being with knowledge of whether the life is worth living would allow P to live L. That is not the least bit informative.

This problem has important implications. All objective tests that base the worth of a life on evaluative facts suffer from this issue. Any plausible test will be an objective test, but any objective test will be threatened with circularity. I see no clear solution to the problem.

We cannot preclude evaluative facts altogether, since the moral worth of a life is certainly relevant to whether it is worth living. To avoid vicious circularity, we could think of the caretaker as determining the worth based on the other facts of the life, not based on knowledge of the particular evaluative fact of whether the life is worth living. But this is an unsatisfying suggestion. As the previous objections show, in order to explain how the caretaker would determine whether a life is worth living in absence of knowledge of that evaluative fact, we would have to build a theory of worth into the description of the caretaker. But that defeats the purpose of the test. Why not simply appeal to the theory and discard the test? In fact, this is precisely what I think we should do, eventually. I intend the test merely as an aid to theory development. The circularity problem is the principal reason we need a theory of worth.

5 Conclusion

In this paper, I identify four distinct tests for what makes a life worth living: (1) The Suicide Test (Camus), (2) The Recurrence Test (Schopenhauer and Nietzsche), (3) The Extra Life Test (Cicero and Hume), and (4) The Preferring Not to Have Been Test (Williams and the Book of Job). I try to develop plausible versions of each before exposing their flaws.

(1) The suicide test fails in both subjective (SLT) and objective (OLT) versions. The subjective version cannot account for mistaken self-assessments. The objective version faces a timing problem: When is suicide appropriate? The only plausible version of the OLT would be an infanticide test, but this requires abandoning the central concept. Babies cannot kill themselves. (2) The recurrence test fails in both weak (WRT) and strong (SRT) versions. One might live a life worth living, but nevertheless not want to repeat oneself. Repetition of this sort amounts to grand stagnation. A repeated life is most plausibly less good. If so, a life could be worth living, but its repetition might not be. (3) The extralife test (ELT) fails for similar reasons. One might live a life worth living, but not be willing to live another. One might be world-weary like Cato, or simply world-sated, as Kaufmann would prefer. (4) The preferring not to have been born test (PNT) suffers from a serious problem: Hitler might not have wished that he had never been born, but his life was nevertheless not worth living. This problem does not face an objective third-person assessment.

I defend the heuristic value of a fifth test. It avoids the problems facing the other four. My proposal is a revised version of Bernard Williams’s claim that a good test of whether one’s life is worth living is whether one prefers not to have been born. The Pre-Existence Test holds that a life worth living is one that a benevolent caretaker with foreknowledge would allow one to be born into. A life worth avoiding is one that a benevolent caretaker would disallow. I consider two objections and note that there appears to be an indeterminate middle category of lives worth neither. I argue that the pre-existence test usefully test tracks the general extension of the concept of what makes a life worth living. But it is not a theory of worth. I also consider a third problem, the circularity objection. It shows that any plausible test will risk circularity if the evaluator has knowledge of evaluative facts, or the test will require a theory of worth to be viable. Hence, any plausible test has limited heuristic value.

My principal task here is preliminary to developing a theory of worth. I pursue both negative and positive goals. The negative goal is to exclude those tests which will do little but interfere with the development of a theory. Overall, this is likely the paper’s most important contribution. The positive goal is to identify a test that can serve as an aid for theory development. I hope that the pre-existence test supplies a mechanism by which one can evaluate candidate theories of what makes a life worth living. But as with other objective tests, it will sometimes be impossible to say whether a life passes the test without having a theory of worth.⁵⁸ No informative test can serve as an analysis of the concept.