Designing an Ideal Utilitarian Welfare Scale

Confidence: 8 | Importance: 8 | Novelty: 9
Post #: 9 | draft | Topics: ethics

BLUF: I formalize my conception of an ideal utilitarian welfare scale here and state why I think total, non-negative, act, hedonic utilitarianism is right.

Epistemic Status: This is in very philosophical territory, so I am still very uncertain this is true. Future qualia research could definitely upend this. However, I believe this scale is currently the simplest explanation that fits the facts and wins the support of Occam’s razor.


The door is basically sealed for utilitarianism as the best formal ethical decision-making structure. For example, Nobel Prize-winning economist John Harsanyi wrote, “…Bayesian rationality postulates, together with the hardly controversial Pareto optimality requirement, entail utilitarian ethics as a matter of mathematical necessity.” Cost-benefit analysis is clearly the golden road to optimal ethical decision-making, at least for the most rational people when they have the time and are answering the most important questions, like those of public policy.

Mathematically, much of the structure of utilitarianism has been specified, but we are still short, in my opinion, on bridging the gap between the math of utilitarianism and physical and subjective reality. Here, I will outline my conception of the ideal utilitarian welfare scale and explain the flavor of utilitarianism which I think is true.

Grounding Ethics in Physics

The seemingly human universal experience of qualia–our subjective mental states which range from being astonishingly desirable to extraordinarily undesirable–as well as logical reductionism, yields that our best guess of ethically-relevant phenomena should probably be the simplest discrete units of qualia, whatever they may be. In other words, since our joy and suffering are probably grounded in fundamental physics operations (which include substrate-independent computations), even if these operations are currently not well-understood, we should still probably value them. Our world is probably not going to be worthy of being deemed as overwhelmingly good until we drastically increase the instances of these well-being fundamental physics operations and decrease the instances of these net-negative fundamental physics operations. In order to guide our actions in this ethical space, I believe we need a map/scale.

Designing the Scale

We can theoretically define the least desirable, but still net-desirable, qualia-producing fundamental physics operation as a +1 utility on our scale. All the other assigned utilities of all other qualia-producing fundamental physics operations will be relative to this +1 utility.

Suppose this +1 fundamental physics operation is not the only kind of physics operation which produces net-desirable qualia. Suppose other types of physics operations produce different, more desirable moments of qualia. The next number could be a 2, a 7.5, a 554, a 452243.7653, or any other positive real number. It would be based off how long one would want to feel the first state to make up for feeling a moment of the latter state. The same goes for additional unique, net-desirable, qualia-inducing fundamental physics operations.

The most desirable of the net-undesirable qualia-inducing fundamental physics operations gets a negative score equivalent to the least amount of positive utility which would be required to make up for it. This could be a -1, a -337.2, or any negative real number. In other words, 100 moments at the +1 level would, by definition, cancel out a moment of qualia at the -100 level. Further, a single moment at the -100 level plus 102 moments at the +1 level would, by definition, be preferable to 1 moment at the +1 level.

Importantly, there is no reason to think a negative or a positive infinity must exist here. First, I am not aware that physicists have established that any infinities actually exist in reality outside of math, so I am not sure why this would be the first. Second, many people are willing to experience truly undesirable states because the experience of the endorphins afterward is worth it. If no infinities exist regarding the desirability of states of qualia, all states of qualia are somehow commensurable with one another.

Determining the Utility of an Event

I’ll define an event as all the physics operations which happen in a given n-dimensional space over a period of time T. Theoretically, within the confines of a person, there is a net-utility over a period of time T. The confluence of negative and positive utilities in a person’s brain and body can be simplified to single utility number over a time T. Importantly, this is a separate measure than that of a person’s instrumental utility.

Total Utilitarianism, not Average

According to total utilitarianism, it is the net score at the end of the day that matters. To say otherwise requires eliminating any below-average hedonic states, even if those states are positive and are by definition, worth experiencing, just in order to raise the average quality of hedonic experiences in the universe. Average utilitarianism is clearly deeply flawed.

Non-Negative versus Negative

There are many different types of negative utilitarianism, but the common thread among them is they give greater priority to suffering rather than well-being. The weakest versions of negative utilitarianism, where positive qualia still matters, are not categorically different than what my scale permits. If one values positive utility but thinks the negative qualia available to us are much worse than the positive qualia, the first negative operation, for instance, can be a -100 or a -1000. In my opinion, that’s not negative utilitarianism.

For the strongest version of negative utilitarianism where one believes positive utility doesn’t matter or that some negative states are infinitely bad, my scale simply says that loses the support of Occam’s razor. Again, I think one needs a lot of evidence to assert that an infinity exists in nature, and further, empirical evidence shows that people are willing to undergo suffering for better hedonic states. That being said, I know our evolution-derived biological preferences are weak evidence–I’m just saying that what little evidence exists is probably in favor of non-negative utilitarianism.

Act, not Rule

I write about the deep flaws with deontology here. I’m not saying that people shouldn’t rely on heuristics in everyday life, and that we ought to punish people for failing to achieve ideal outcomes when they stuck to promising heuristics and things didn’t work out, but there is no free lunch regarding the capturing of utility. The points are in capturing and realizing utility, not in sticking to rules, especially those not backed by formal proofs (please show my a rule that beats out cost-benefit analysis because I don’t know any that do).

Hedonic, not Preference

Preferences are high-order abstractions which correspond to utility estimates of various outcomes. They are useful guides, but some people clearly have suboptimal utility functions and preferences. I am a proponent of reducing theories and truths to their simplest forms. Thus, satisfying preferences is not what matters most in the grand scheme of things. Hedonic utilitarianism is much more fundamental and, I believe, envelops all that is morally desirable about preference utilitarianism.


This is still far from concrete application, and much more qualia research needs to be done, but as it stands, this utilitarian welfare scale has some important implications. First, it supports the long-term value thesis, which says that most of the remaining value to be captured is probably in the long-term future. Its practical implications generally involve reducing existential risks. Second, we may be able to capture a lot more value through making hedonium which is a hypothetical structure designed to produce high-utility qualia states as efficiently as possible. It may be unintuitive how this is intrinsically good, but I believe this may be adequately explained by our evolution-derived preferences and the psychological distance we feel towards pure, thoughtless ecstasy. Of course, our moral uncertainty can also allow us to pursue other goods like human flourishing and mathematical research. Lastly, I believe this conception of utilitarianism is significant evidence against negative utilitarianism and its concomitant world-destruction arguments. It may be a feature of our universe that the expected value of future Earth-originating life is net-negative, but we do not know this yet, so there is no good reason to pursue extinction. That being said, we must be careful to avoid being unduly optimistic, and at least some of our global priorities researchers should be considering failure scenarios which could lead to things like far-future astronomical suffering.