I want to present you with three hypothetical scenarios. Read each one and decide your answer before moving on to the next.
Scenario #1: You and 99 other random strangers are each given a panel with two buttons on it – one orange and one purple. You are all told the rules (everyone gets the same information): You must each press one of the buttons. If 51 or more of you press the orange button, then everyone in the group of 100 people lives. If not (i.e. if a majority press the purple button), then only people who pressed the purple button live, and everyone who pressed the orange button dies. Everyone presses their button choice at the same time. What button do you press? Option A: Press the orange button. Option B: Press the purple button.
Scenario #2: You and 99 other random strangers are in a condemned office building (the reason isn’t important) that has caught on fire. Again, everyone has the same information: The fire suppression system has already been deactivated, but if it was turned back on, it would put out the fire and everyone would be saved. The fire suppression system is on the roof, and while it requires no special skill, it does require 51 people to activate it. If fewer than 51 people go to the roof, they’ll die in the fire. The other option is to simply leave the building. The exits are clear and nothing is preventing you from doing so; there’s plenty of room & time to exit safely. Because of the smoke, you can’t see what other people choose until you’re either on the roof or outside, and it will be too late to change your mind. Which do you choose? Option A: Go to the roof. Option B: Go outside.
Scenario #3: You and 99 other random strangers are at a party when it’s discovered that the punch that has just been put out has been accidentally poisoned, though no one has drank any yet. Everyone has the same information again: If at least 51 people split the punch, then the dose of poison won’t be sufficient to be lethal and no one will die. But if 50 or fewer people drink the punch, the dose will be high enough to kill them. You can also not drink the punch at all. Do you drink the punch or decline? Option A: Drink the punch. Option B: Politely decline.
Did you answer each one? Great, thank you. Now, one more question for you:
If you didn’t answer the same way for all three questions, why not?
All three questions are mechanically identical. In all three scenarios, there are two options: Option A saves everyone if a majority chooses it but kills you if a majority chooses the other option. Option B saves you (and doesn’t contribute to the majority necessary for Option A, obviously), but there’s nothing preventing everyone from choosing it. The rational choice in every scenario is Option B, which not only introduces no risk to you personally, it also doesn’t inflict any harm or risk on anyone else, since everyone else can also choose Option B. So the rational choice for you personally is also the optimal choice for the group, assuming rational actors.
Yet, as I’ve been asking these questions to a lot of people lately, I’ve noticed that a huge number of people pick Option A for the first question. Most people pick Option B for the second question, though there are a few outliers. And no one has picked Option A for the third question so far. This absolutely fascinates me, so I’ve been diving deep into respondents’ reasons behind what caused them to diverge in their answers. Here are the broad categories of reasons why some folks answered A for #1, but B for #2/#3:
- Inserting assumptions into scenario 1. Several people said they simply assumed that part of the scenario involved some number of people needing to pick the orange button, or there being a limited number of purple button pushes available, etc. In other words, they assumed that the only way for everyone to live was the orange option, and they didn’t put together (or accept?) that everyone could simply push the purple button and be fine. It felt either “too easy” (because of the unconscious assumption that suffering is zero-sum; if someone avoids harm, it must be because they’re harming others) or they simply made the morally intuitive jump that since the orange button is framed as the “selfless” choice, it must also be the correct one. Or at least, they didn’t want to make the “selfish” choice and didn’t consider the effects further than that intuition.
- Mentally “swapping” the stated scenario for an allegorical one. Lots of people seemed to defend their answer of Option A in scenario one by speaking as if the scenario was an explicit metaphor for another one. People used it as an allegory for voting/democracy, or simply for acting selfishly at the expense of their broader community, and then defended the more community-minded choice in that context. (One person even hilariously commented, “What if everyone thought like you did?” to a person defending Option B, which is very funny because the literal answer is, “Then everyone would be fine.”) So some people seemed to be assuming that the scenario was being presented disingenuously and was trying to trap them into advocating for a philosophical position they otherwise wouldn’t hold. Notably, many people maintained this stance even when it was explained that that wasn’t the case.
- Believing the scenarios aren’t actually mechanically identical. This was less common, but some people suggested that there actually was a difference between the three scenarios. Namely, that in #1 you make an “active” choice to push the purple button, while in #2 the choice is less active since it’s “just leaving,” and in #3 it’s completely passive since you just don’t drink the punch. In reality, those are still identical, as they involve a decision on your part, even if the physical manifestation of that decision appears more “active.” But it’s true that psychologically people feel differently about them – I asked numerous people if their answer would change if scenario one has been “push an orange button or don’t push the orange button” rather than a choice between pressing two buttons, and many said yes, that felt different and would make it easier for them to choose not to push the orange button! (I didn’t think to ask, but I wonder what the responses would be like if the choice was between the orange button, and a button labeled ‘I Choose to Abstain.’)
If you’re familiar with philosophy, you may realize that these questions follow the structure of the Trolley Problem – present a mechanically identical scenario three different ways with different social cues and then examine why people answer differently. In case you’re unfamiliar, the full trolley problem isn’t just “Do you pull a lever diverting a runaway trolley so it kills only one person instead of five?” The full thought experiment asks that question, and then two more: “A runaway trolley is on a collision course with five people. There’s no lever, but there’s a very large man who is big enough to stop the trolley if pushed in front of it, and you’re close enough to push him. He’ll die if you do, but the trolley will stop. Do you?” And then the last question: “A doctor has five patients all about to die from injuries to five different organs, with no replacements available. A homeless but healthy man wanders into the hospital to warm up. Is it okay for the doctor to kill the homeless man and take his organs in order to save the five patients?”
Mechanically, those questions are identical: “Five people are about to die if you do nothing, and their deaths aren’t your fault. You can save them by actively causing the death of a different person. Should you?” But because the questions are framed differently with different social cues, people’s answers to them vary greatly. It’s fascinating to explore not only the reasons people will say “Yes, I’ll pull the lever,” but “No, I won’t murder a homeless man and harvest his organs,” but also to explore their explanations for why those scenarios are different.
Both the trolley problem and my button questions are exploring the same overall concept: We make what we believe to be rational judgements while under the influence of a great many irrational things. Without even realizing we’re doing it, we bring in our own biases, project our own beliefs, make decisions based on gut instincts trained by a lifetime of social desirability cues, and view the world from our own stance as if it were universal.
(A note on that last one: People who have a tendency to believe in win/win scenarios overall seemed more likely to recognize that in Scenario #1, everyone could simply press the purple button and be fine. People had a harder time seeing that solution if their general mindset was one where all “wins” must be someone else’s “loss.”)
It is incredibly worthwhile to examine why people make the decisions they do – including ourselves. It’s often uncomfortable to do so, especially when we realize the flaws and inconsistencies in our own thinking. But diagnosis is the only way to improve, and so if that discomfort allows new clarity to bloom, I’m happy I caused it.