Effective altruists use terms and notions in utilitarianism very inconsistently. Lots of effective altruists come from classical liberal arts education at universities like Oxford. A lot of them are philosophically uneducated folks from the internet. That’s fine, as much of academia has stagnated as is, and so intellectual development is as likely to come from the blogosphere as anywhere else. The rationality community is notorious for using idiosyncratic terminology to describe concepts which have existed in “real”, i.e., academic philosophy for a long time. But there are all kinds of effective altruists who do the same thing. And the inconsistent connotation of different kinds of utilitarianism is an example that makes conversations about utilitarianism become confused quickly.
“Utilitarianism” seems to be used to mean any of the following by effective altruists:
optimizing for those conditions which maximize the satisfaction of one’s full moral intuitions.
- Optimizing for one’s full moral intuitions transformed into an expected value function in the sense of the hedonic calculus of classical utilitarianism.
- Optimizing for one’s full moral intuitions transformed into an utility function in the sense of Von Neumann-Morgenstern utility theorem from decision theory.
- Maximizing the happiness and or/minimizing the suffering of moral patients in a generic sense, typically amenable to moral patients’ preferences.
- Maximizing the pleasure and/or minimizing the pain of moral patients, including down to a neurobiological level of utility assessment, e.g, wireheading, virtually any
creature being capable of experiencing pleasure and pain, etc.
- Consequentialist morals which lead to classical utilitarianism, as opposed to negative or “positive” utilitarianism.
- Any form of consequentialism excluding egoism.
I put “positive” in scare quotes because in the history of utilitarianism over the last 200 years, “positive utilitarianism” doesn’t appear to have been used as a term to describe any common variant of utilitarianism. But I keep seeing it used among effective altruists. “Negative utilitarianism” as a term and a variant of utilitarianism is an intellectual development of the late 20th century appears still to be a more niche position outside of effective altruism.
“Negative utilitarianism” seems to be used in context to mean any of the following by effective altruists:
1. Optimizing for those conditions which minimize the probability of maximally and intuitively bad outcomes, as opposed to those conditions which maximize the probability of maximally and intuitively good outcomes.
2. Utilitarianism with utility defined as giving more proportional weight to reducing suffering than increasing happiness.
3. Utilitarianism with utility defined as giving moral weight only to reducing suffering and zero to increasing happiness.
4, 5. Respectively the same as (2) and (3) but with “happiness” switched out with “pleasure” and “suffering” switched out with “pain”.
6. Utility defined as eliminating the possibility, or minimizing the expected probability of, any suffering, up to and including eliminating or minimizing the expected probability of any life-forms or experiencing beings.
I’m not as familiar with the term “positive utilitarianism”, so I haven’t observed all these instances of different meaning for it used. However, they can be derived by inverting the meaning of a corresponding definition of “negative utilitarianism”.
1. Optimizing for those conditions which maximize the probability of maximally and intuitively good outcomes, as opposed to those conditions which minimize the probability of maximally and intuitively bad outcomes.
2. Utilitarianism with utility defined as giving proportionally more moral weight to increasing happiness relative to decreasing suffering.
3. Utilitarianism with utility defined as giving moral weight only to increasing happiness and zero to decreasing suffering.
4, 5. The same definitions as (2) and (3) respectively, but with “happiness” switched out for “pleasure”, and “suffering” switched out for pain.
6. Classical utilitarianism.
The biggest problem for effective altruists is thinking through these definitions can lead one to the conclusion self-identified utilitarians of the opposing kind of having goals which seem intuitively catastrophic. There is a definition of “negative utilitarianism” where one can model an agent as having the primary objective of eliminating life. There is a definition of “positive utilitarianism” where one can model an agent as willing to go to any extreme to maximize happiness without regard to suffering, including permitting arbitrary amounts of torture-level suffering for indefinite periods of time.
If some effective altruists suspect each other of such motives because each describes themselves as a certain kind of utilitarian, that’s a fast way for things to become political and everyone getting mindkilled. My estimation is most effective altruists aren’t so formally utilitarian in any sense that they would pick upon a definition of utility/utilitarianism that would permit an outcome of either the total elimination of all life and experience, or arbitrary amounts of torture-level suffering. This appears to be a minor problem in effective altruism which could in theory but isn’t in practice leading to internal movement conflict rather than cooperation and coordination.
This is only one example of how effective altruists using different definitions of utilitarianism can lead to confusion. I think a lot of disagreements among effective altruists over utilitarianism, including those effective altruists who aren’t utilitarians, stems from everyone using different definitions of utilitarianism without being aware of that. Laying out the structure of the problem above describes it. This is the first step to solving the problem. I don’t know what one would do next though.