A Longer Introducton to Utilitarianism
Epistemological Foundation
In order to even consider questions of morality, one must first hold some basic assumptions about the way the universe works. Below are the epistemological assumptions on which utilitarianism depends.
Assumption. It is intelligible to think and talk about concepts, such as "existence."
The purpose of this admittedly vague assumption is to overcome nihilist criticisms of philosophy in general. Nihilists deny basic ideas about intelligibility that we take for granted, and this axiom assumes away some of those challenges. Since there are so many ways in which a nihilist could deny the intelligibility of philosophy, I am unable to give a more precise statement of this assumption. But below is an example of the type of challenge that this axiom is intended to dismiss.
Example. Suppose I am conversing with a nihilist, and I make this statement: "Okay, so you deny lots of things. You deny any underlying structure to reality. You deny the meaningfulness of logic. But surely, you must admit that at least the following statement is true: You are thinking. How could that not be the case?"
To this the nihilist says, "What do you mean by 'true'? That statement is unintelligible. Indeed, even the words that I'm articulating right now are unintelligible."
In a vain attempt at refutation, I retort, "Your statement itself is a contradiction. You assert that words are unintelligible, but if words are unintelligible, then the words with which you made that statement are unintelligible. In order for your assertion to be true, it must be false!"
In reply, the nihilist merely sighs and says, "I refuse to acknowledge that anything you say is meaningful. Just stop trying." Indeed, there is no way to convince the nihilist to acknowledge intelligibility.
Like a spacecraft that has passed through a black hole's event horizon, the nihilist cannot be pulled back away from her position by conventional argument--because, of course, such conventional argument is unintelligible.
The purpose of this assumption is to move the discussion beyond the above dispute. If the assumption is accepted, then we can steer clear of the black hole and proceed ahead to other matters.
Further Assumptions: The law of causation holds. Basic logical rules of inference like modus ponens and modus tollens hold. Induction holds (that is, the past can help predict the future). And finally, assume everything else that I've neglected to mention.
The purpose of outlining these assumptions is to make clear exactly what I am taking for granted. I have tried to do that here, but there are undoubtedly other assumptions that I am taking for granted because I am so accustomed to thinking within a framework in which they hold. If readers notice any important assumptions that I have left out, please let me know.
Ethical Foundation
Definition. A state of the universe is simply a reference the way things are in a universe.
Example. In one state of the universe, person A has lost his keys. In another state of the universe, A has not lost his keys, and everything else is the same. Thus, different "states of the universe" just refer to different ways the universe might be.
Definition. A preference is any explicit or implicit emotional desire on the part of an organism to exist in one state of the universe rather than another.
Example. Presumably, person A has a preference for existing in the state of the universe in which he has not lost his keys.
Definition. An explicit preference is a desire that an organism thinks about abstractly as being a desire. This means that the organism speaks, writes, or internally verbalizes what it wants.
Examples of explicit preferences.
Definition. An implicit preference is a desire that an organism feels but does not think about as abstractly being a desire.
- A person feels hungry and thinks, "I wish I had some food!"
- An adult sees a violent cartoon show and says, "It's wrong that people make these types of shows for kids to watch."
Examples of explicit preferences.
- A person has a stomachache and endures the pain. Even though the person doesn't have an abstract thought that "I wish this pain would stop," she nonetheless dislikes the experience. Thus, she has a preference for it to stop even though she doesn't put that preference into words.
- A battery-cage laying hen feels hungry, agitated, and miserable.
- A baby mouse looks forward to the return of its mother to the burrow.
Definition. Sentience is the ability to have explicit and/or implicit preferences.
Assumption. Some organisms are sentient.
How do we know whether a given organism is sentient? We don't--at least not with certainty. But we do have various subjective probabilities of sentience of different organisms.
Examples.
Assumption. Preference strength is cardinalizable. That is, different preferences may be held with different degrees of intensity. The strength of a preference has a cardinal value (e.g., 2 or 14).
- I would assign a subjective probability of approximately 1 to the proposition that these organisms are sentient: people, monkeys, cows, chickens, and bluejays.
- I would assign a probability of approximately 0 to the sentience of the following: amoebas, bacteria, trees, flowers, rocks, water molecules, and air molecules.
In reality, we don't know what an organism's preferences are to any precise degree, so we can't assign numbers like this in practice. I do so here and elsewhere merely for the purpose of clarity. This does not mean, however, that preference strength does not have some value in principle, even though we can never measure it exactly. In any event, we can at least get a good sense of relative magnitudes of preferences, and that's generally sufficient for a utilitarian analysis.
Assumption. Intersubjective and intertemporal comparisons of preference strength are possible.
Suppose person A wants a blueberry muffin, person B wants a chocolate-chip cookie, and person C wants orange juice. This axiom says that its possible to describe the preferences of A, B, and C in a common unit of preference strength. Suppose A has a preference strength of 5, B has a preference strength of 7, and C has a preference strength of 5. Then A and C have preferences of the same strength and B has a stronger preference.
In "Utilitarianism and Interpersonal Comparison: Some Implications of a Materialist Solution to the World Knot," Social Choice and Welfare, 8 (1992): 1-15, Yew-Kwang Ng argues that it may be possible in principle for a single person to experience the mental states of at least one other person. Since our left and right brains can operate somewhat independently yet also combine into a single mind, might it not be possible to similarly connect two different minds? Even if this isn't even conceivable, we can still say that utility comparisons ought to be possible in principle, because two different minds have some values for preference magnitudes, and those values can be compared.
Definition. Unconsciousness is the state in which an organism has no awareness, no sensations, and no preferences. Think of unconsciousness this way: time stops, everything in your life stands still, and you feel (or, well, don't feel) the sensations of nonexistence.
Definition. Utility is a quantity representing the strength of fulfilled or abridged preferences. The unit of utility is a util/unit time. To make comparison of utilities easier, I define preferences to be relative to unconsciousness. Thus, an organisms experiences positive utility when it's in a state that it prefers over unconsciousness. An organism experiences negative utility when it's in a state where it would prefer to be unconscious. The magnitude of the utility the organism experiences is proportional to the strength of the organism's preference.
I should emphasize that choosing unconsciousness as a reference point is arbitrary. I could just as easily have chosen utility to be defined relative to the state of stubbing one's toe. All that really matters for utilitarianism are changes in aggregated utility that moral agents can bring about, and choosing a reference state simply makes it easier to determine what those changes in aggregated utility will be. There is no such thing as "absolute utility"--since "fulfillment of a preference" only makes sense relative to some other way things might have been--but I think the definition that I used comes pretty close to any intuitive sense of absolute utility that people might hold.
Examples.
- A cow wants to find a patch of fresh grass. It would prefer the state of finding the grass over the state of unconsciousness by a magnitude of 5. Hence, finding grass will furnish 5 utils/minute.
- A turkey in a factory farm had its beak trimmed two days ago, and it is now enduring chronic pain. Moreover, it feels cramped, stressed, and sick due to overcrowding. If it could imagine the abstract state of unconsciousness and compare that to its present condition, it would prefer unconsciousness with a strength of 150. Since the turkey is not in its preferred state, its preference is abridged, and it feels -150 utils/minute.
Definition. Aggregated utility is the expected-value sum total of the utility in the universe over all of time, where utility is counted equally regardless of which organism happens to experience it or when it happens to be experienced. The unit is the \textit{util}.
Assumption. Aggregated utility is finite. This assumption requires that
- No organism ever experiences an infinite amount of utility at a given time,
- Only a finite number of organisms experience a nonzero amount of utility at a given time, and
- The sum of all utility experienced at a given time is never nonzero for an infinite duration.
The first condition is true so long as there's no such thing as an infinitely strong preference. The second and third present more serious difficulties that will not be addressed here (see "Infinite Ethics" by Nick Bostrom).
Definition. Utilitarianism is an ethical doctrine whose goal is to maximize aggregated utility.
Definition. A utilitarian is a moral agent who adheres to utilitarianism. She aims, through her life's work, to effect the greatest possible increase in aggregated utility.
Utilitarianism as Happiness Maximization
At its most basic level, utilitarianism is about preference fulfillment. But in this section, I pursue an alternate, and perhaps more intuitive, formulation: utilitarianism as maximization of happiness, satisfaction, and welfare. What I am trying to do here is show that the definition of utility that I gave above coincides closely with the way people usually understand these other terms.
Good Examples.
There are some situations in which utility does not correspond perfectly with happiness--at least not according to the way that happiness is generally understood. In these cases, I think utilitarianism gives the correct moral evaluation of the situation.
Not-So-Good Examples.
This last example seems troubling, if not viscerally revolting. "How can you even think of inflicting pain on another creature?" screams one's intuition. "That's horrible!" Yet, I do actually think that the utilitarian position in this case is the right one. It's hard to wrap one's mind around exactly how it would feel to be the frog. We--by virtue of being evolutionarily selected organisms--have such an ingrained reaction against pain that it's impossible for us to imagine wanting more. But this is a limitation of our capacity for empathy, not a genuine reason against torture. We oughtn't deny the frog its desire simply because we can't conceptualize its internal state of mind. Perhaps the best thing that the hypothetical humans could do would be to compare the torture in their minds with something that they could imagine desiring, such as extreme pleasure. That way, they could feel better about doing what really would be the right thing.