59 Comments

Why focus on "worth living"? Seeing that majority of future people will be EMs and will not be able to physically die, is there really a lower bound to their utility?

Expand full comment

 Total measure can't be increased or decreased, only rationed (and maybe not even that, but here's hoping futility theories are wrong)

Expand full comment

Rationally choosing what food you eat, how much exercise you take, whether to cross the road, etc., all require comparing the utilty of being alive with the utility of being dead.

If as I'm suggesting, the utility of being alive and dead are incommensurable (for want of the zero point's being defined), then it's not rational to compute the food you eat, etc. in the manner you (and apparently the utilitarians) believe we do. What we must calculate--to avoid adding apples and oranges--is the relative utility-disutility of the decision to die, as determined by affects such as guilt and (most of all) fear. I may think I don't want to die, but I really don't want to choose death. It's not necessary to conceive of the "utility" of death to answer this question. Utility, after all, is an abstraction from our choosing between different states of being.

Expand full comment

Utilitarians should try to maximise:

Sum_( i = 1 to N) (U_i - U_min)

where N = number of people, U_i = utility of ith individual and U_min is the minimum utility to make existence worthwhile. This is neither max average nor max total and avoids the problems of both.

Similar to in finance corporations should maximise (profit - cost of capital).

Expand full comment

but only if one concedes that it makes any sense to compare the utilities of being dead to being alive. But this abstract choice is never actually presented in real life

Actually it's presented pretty much continuously. Rationally choosing what food you eat, how much exercise you take, whether to cross the road, etc., all require comparing the utilty of being alive with the utility of being dead.

Of course humans are not ideal rational agents, and don't have a well defined utility function, which makes the point moot.

Expand full comment

You mean as opposed to behaviorial utilities--Kahneman's distinction? I suppose I'm talking about behavioral utilities, since they're seemingly the relevant sort when you try to use suicide as a zero point.

This line of thinking about suicide not being an absolute zero suggests some persons would be "better off dead"--but only if one concedes that it makes any sense to compare the utilities of being dead to being alive. But this abstract choice is never actually presented in real life, so there's no reason to presume the concept of utility applies.

Expand full comment

Are you talking about preference utility?

Expand full comment

Notice, also, that if you conceive a life worth living as one substantially exceeding rather than barely exceeding the subsistence level, RH's EM dystopia then involves a trillion lives not worth living.

Expand full comment

Suicide doesn't provide a zero point for utility because the very process of ending one's life or even deciding to end it incurs massive disutilities, that being the real reason many people don't commit suicide.

Expand full comment

EY is prone to think that clarifying a term resolves serious questions. Granting your presuppositions for argument's sake, the problem remains: a sufficiently large population will always outweigh a much smaller one, regardless of there being a much higher level of welfare in smaller one. When EY's only reason for believing utilitarianism is true is that he finds it corresponds to our "moral intuition," the shock to the intuition should be decisive here--except that utilitarianism is based on a moralistic faith (http://tinyurl.com/cxjqxo9 )

Expand full comment

Total utilitarianism has also absurd implications.I don't think there is any way to salvage act utilitarianism.

Maybe you could make the case for rule utilitarianism, but you would still have to solve the hard problem of interpersonal utility comparison and find some reasonable way to constrain the specificity of the rules (if you allow for arbitrarily specific rules then you reduce to act utilitarianism).

Expand full comment

I'm surprised to read that you are an average utilitarian, since this theory has clearly absurd implications. As I write in this Felicifia post:

For instance, consider a world A in which people experience nothing but agonizing pain. Consider next a different world B which contains all the people in A, plus arbitrarily many people all experiencing pain only slightly less intense. Since the average pain in B is less than the average pain in A, average utilitarianism implies that B is better than A. This is clearly absurd, since B differs from A only in containing a surplus of agony.

Expand full comment

I'm on unfamiliar ground here, so perhaps I'm off target. Nodoubt I will be informed or ignored appropriately.

Since this post relates only population to utils I'll focuson any plan we might have to determine future population, and I'll ignore someof the other factors that might alter future utils, even though they may beindirectly related to future utils - e.g. undesirable climate change caused bynear future population growth that harms more distant future populations.

Two possible future worlds X-world and Y-world, withpopulations X > Y, with a difference D = X - Y, and with utility Uy > Ux(based on inverse population/utils, which seems to be assumed here).

If Y is the outcome there will be D of X that never got toexist. As such the state of D, their lost lesser utils, could be ignored. Whyworry about people who will never exist, from our perspective, or do notexists, from a Y-world perspective?

So, if planning future populations, planning for X insteadof Y is planning a state where utlility is less than it could be. Could membersof population X look back and think badly of us for not maximising 'their'utility (really the utils of Y that would have been born had Y-world been thecase). We might think that provided individuals of X could not determinewhether they would have been alive, as members of Y, or never conceived, asmembers of D, they have no way of knowing whether they would have benefitedfrom our plan for Y-world or not.

But, now thinking of yourself as some member of X at timeTx, if you learned from planning and controlled population data that you wouldhave been a member of D and so would not now (at Tx) exist, why should itbother you at all? You should note that had you not been born you would not bealive to regret your lack of birth. It would be irrational for a member of D ofX in X-world to lament his possible non-birth and not to lament more thereduced utils of being in X-world.

In other words members of X-world could lament their beingin X-world. They could regret not being in Y-world with Uy or not being born toregret anything. This knowledge of members of X would subtract further fromtheir Ux. On the other hand members of Y-world considering their possiblemembership of X-world would be happier for their greater Uy, and so add futherto their Uy.

The above might imply that the best future world consists ofreducing X in order to increase Ux, possibly reducing X to 1 for some maximumUx. Or maybe it implies the best state is X = 0: zero suffering but infinite,if unused, utils - i.e. extinction of intelligent self-aware species is thebest possible outcome for them, if utils is inversely related to population.

But I don't see any reason to suspect that there is acontinuous inverse relationship between X and Ux across all possible X. Theremay be some point where humans fail to flourish under small populations (e.g.lack of human resources required to maintain a world that can maintain Ux).There may well be some optimum Y and Uy: X > Y > Z and Ux < Uy > Uz.

If this latter distribution of utils across populations isthe case, and given that there is no reason to worry about the non-existence ofpeople who do not exist, it would seem obvious to aim for the lowest populationY that does not decrease Uy further.

That leaves the problem of figuring out what Y and Uy are.

Expand full comment

I also suspect another reason why we find the repugnant conclusion repugnant is a confusion about what it means for a life to be 'barely worth living.'

I suspect people imagine this as a life which is just barely preferable to suicide. However, pragmatically speaking our drive to survive, our need to condemn the murder of unhappy people, our need to account for an uncertain future even when emotionally we are convinced things won't get better and our selfish desire for our loved ones not to commit suicide all conspire to drive our sense of what constitutes a life barely worth living down to a condition of mild (but not horrific) suffering.

Expand full comment

Robert Wiblin is Australian. 

Expand full comment

I reject the existence of individuals in general. The appearance of individuals is merely an artifact of our history. Many species that were well-adapted to an aquatic environment that subsequently move out of it might well use 'integument' and 'immune system' strategies that generally require and ensure a same-reproductive-fate for all genes within the integument or immune-system domain - so individualism is likely a reasonable and reasonably common form of life.

However, life could certainly exist without individuals.

Furthermore, I reject summation of utilities of different individuals. Utility functions are valuable in making decisions for yourself, and even in building semi-autonomous entities. But summing decision-theoretic utilities of different individuals is a violation of units.

Behaving well, social welfare, is just not that simple; path-dependence is frequent. For example, in deciding whether to enfranchise someone, to decide that they are a citizen or a person, you might reasonably use soley the preferences of the people who are already enfranchised. But in deciding whether to disenfranchise someone, that person's preferences do matter, as they already have "the vote".

Expand full comment