The Stupid Button Thing
If you spend your time as poorly as I do, you’ll have encountered the “red vs blue button” hypothetical scenario.
The “discourse” around this question resembles “the dress” or, for the sophisticates, Newcomb’s paradox. Red-buttoners perceive red-buttoning as obviously correct, blue-buttoning as irrational, almost suicidal behavior. “Dog, who do you think you’re saving? They can just press red.” On the reverse, team blue heroically accepts the risk to their own life for the good of the many, even those no-good, narcissistic Reds. The battle is all the more fierce for being entirely hypothetical.
Overthinking entirely hypothetical scenarios was 80% of my graduate education, so let’s do a rant on this and maybe we’ll put a stop to the dastardly red-pressers before they can come for us and our blue way of life.
Part One. Why Darwin Had to Get On That Stupid Ship (Yes, Really, We’re Starting Here)
If you’ve read On the Origin of Species (which I haven’t), you’ll see that the case for evolution by natural selection can be put quite simply. Here, for example, is the summary to chapter four.
If during the long course of ages and under varying conditions of life, organic beings vary at all in the several parts of their organisation, and I think this cannot be disputed; if there be, owing to the high geometrical powers of increase of each species, at some age, season, or year, a severe struggle for life, and this certainly cannot be disputed; then, considering the infinite complexity of the relations of all organic beings to each other and to their conditions of existence, causing an infinite diversity in structure, constitution, and habits, to be advantageous to them, I think it would be a most extraordinary fact if no variation ever had occurred useful to each being’s own welfare, in the same way as so many variations have occurred useful to man. But if variations useful to any organic being do occur, assuredly individuals thus characterised will have the best chance of being preserved in the struggle for life; and from the strong principle of inheritance they will tend to produce offspring similarly characterised. This principle of preservation, I have called, for the sake of brevity, Natural Selection.
TL;DR: babies are similar to their parents; traits that helped the parents survive and have babies in turn help those babies have babies, becoming more common in the process. Throw in a bit of mutation into the mix and you’ve got yourself some species going, baby.
The logic here is straightforward on its surface and, if you were so inclined, you could create a mathematical model of evolution by natural selection. (Here are some receipts from a boffin so inclined.) So our starting question here is, if evolution by natural selection is a necessary consequence of certain starting assumptions (traits are heritable and have positive or negative influences on the reproduction chances of organisms bearing them) and those assumptions “cannot be disputed,” then why didn’t Darwin stay home, dash off a little pamphlet laying out this argument and call it a life’s work?
This is not by any means an especially difficult question, but it is not entirely an idle one. The mathematical logic of Darwinian evolution is load-bearing. If Darwin had merely presented his observations about finches and fossils and some speculative generalizations, the book would hardly have made a splash. Rather, the arguments give the theory a sense of inexorable truth: the examples are illustrative but can almost seem besides the point. (You might show a child that 2 + 3 = 5 by counting out two piles of stones, but it would be misleading to describe that act of counting as evidence for a mathematical necessity.)
Now, while Darwin might have wasted his time on barnacles, the voyages were time better spent. The mathematical abstraction of Darwinian evolution is too structural to tell us much about the world. To illustrate this, consider that a proponent of Intelligent Design can accept the formalism of evolution by natural selection. Of course they would quibble with the word “natural,” but whether the mutation of traits happens by some naturalistic random mechanism or divine intervention, whether God’s hands “tip the scales” to favor certain traits beyond what Nature herself would prefer, these differences need not even show up in a statistical model of trait distribution across a population.
This point becomes clearer when we consider a failed cousin of Darwinian evolution: memetics. The idea, courtesy of Dawkins, is just as organisms can be conceptualized as bundles of inheritable traits that reproduce or die, cultural expressions can be similarly imagined. A music genre, say, might be the equivalent of a species. When a band hears a song and decides to make a cover, this is the equivalent of reproduction.
Memetics has had some cultural currency, giving us the word “meme,” but attempts to make a real scientific theory out of this description have come to very little. Now, half a century since Dawkins floated the idea, after some modest attempt at working out the details, memetics is essentially dead as a research program.1
The difference, in short, is that while there is considerable similarity between the basic model of Darwinian evolution and that of memetics, there is much more fuzziness in the application of the model. While of course advancements in genetics, the discovery of DNA, have brought empirical rigor to Darwin’s notion of heritability, even before these came to light the idea of a “heritable unit” in folk biology is crisper than any notion we have of a cultural unit. (And no gene-equivalent has ever been identified.) So genetics is a science where memetics remains a metaphor.
The lesson here is that mathematical formalism is cheap, but empirical grounding of a formalism is rare to come by.
Part Two. Now That’s a Theory. A Game Theory
Let’s continue our whirlwind tour of dead Englishmen by talking about Hobbes for a bit. Hobbes, in his off time from being a cartoon tiger, was history’s first atheist neo-monarchist debate bro. In his Leviathan he argued, roughly, that the pre-political, default state of society is an anomic nightmare, a “war of all against all.” No one has enough stuff. Everyone wants everyone else’s stuff. Violence is a pretty good way of getting stuff. So people use violence and any political arrangements struggle to survive for long given that backstabbing your friends and taking their stuff is a pretty good way to get stuff.
The way out of this terror, for Hobbes, is overweening, dictatorial power. If there’s some guy, the sovereign, who has enough thugs at his disposal, then people will be too scared of him to try and take hist stuff. And if he uses that power to lay down some rules across the board for how people under his power should act, then people will actually fall in line with those rules. Both because they’re scared of what the tyrant will do to them if they break them but also because a world of even crappy rules is a much better world than this war of all against all. You would have loved some rules in the old chaos but you couldn’t trust other people to follow them; now that the guy with a bunch of armed thugs say’s he’ll behead people who don’t follow rules, you now have enough reason to expect other people to cooperate and so can cooperate yourself.
One recent interpretive trend is to read Hobbes’ argument as game-theoretic. Often, specifically, as a version of the prisoner’s dilemma.2 A quick introduction/refresher. The prisoner’s dilemma is a “game” (situation might be a fairer term) played between two accused and guilty prisoners and conspirators. The criminals are held in separate cells, with no way to communicate. The police don’t have enough evidence to convict, so the pair will walk away if both stay quiet. To get their conviction, the police offer a bribe to each if they rat on the other: you walk out with $5,000 and the other takes the full penalty of 2 years of jail. But they both know that if they both snitch, neither will be walking out and they’ll get hit with a slightly smaller penalty of 1 year.
Don’t think too hard about the details of the scenario. The basic idea of this situation is that from each prisoner’s perspective, there’s one of two possible cases.
The other person snitches on me. The best thing for me to do in that case is to snitch back, because that way I’m not stuck with the full time myself. (Assume the criminals don’t care about honor or reputation or anything. This is their last gig and they just want to minimize jail time.)
The other person doesn’t snitch on me. The best thing for me to do is to snitch on them, that way I get some cash. (Again, they don’t care about decency or honor or being retaliated against, but they do like cash.)
The way we say this in the biz is that snitching dominates honesty. Whatever anyone else does, snitching is going to get you a better result than honesty. The thought for Hobbes is that the state of nature has a similar incentive structure. If two people cooperate, that’s neutral to positive for each. If one cooperates and the other is malevolent, the bad guy steals the good guy’s stuff and is much better off. If they’re both bad guys, maybe they slap each other around a bit and come out of it similarly scuffed. Since nice guys finish last in this setup, you really have to play it safe and be an asshole. The sovereign, meanwhile, by hitting assholes with swords, adds just enough cost on being an asshole that the calculus flips and you’re better off being nice.
Objections to this argument come from one of two places.
There are way better (more effective, more humane) forms of social organization than dictatorship for getting us out of this anarchic “state of nature.”
The “state of nature” as Hobbes describes it does not really exist.
I’m here interested in this second line of objection. We can start with the basic observation that I had to put a couple of long parentheticals into my description of the prisoner’s dilemma in order for the textbook, “intro to game theory” conclusion to follow. If you took two actual people and tried to recreate the scenario (hint: don’t do this), in all likelihood it doesn’t go anything like the textbook description.3
And this for a range of reasons: the precondition that the pair’s decisions are isolated, independent of one another, is hard to meet (in most real situations, we get some information about what other people are attempting as they are attempting it that we can respond to in the moment); jail time and cash are not the only incentives by any means that matter (“honor among thieves”); and the strategic consequences are not actually going to be neatly self contained (maybe at some point the guy you snitched on gets out and beats you up, maybe his friends get word and beat you up).
There are tools to try and model bits of this real-world messiness. For instance, “iterated games” give you some insights on how strategies can take into account future interactions. For instance, in iterated versions of the prisoner’s dilemma, the best strategy is some version of “tit for tat:” play nice at first, but take revenge on double crosses. But it is absurd to imagine that the incredibly complex strategic space of a human community, with its web of relations, its social pressures, long-term interactions, the necessity of some sort of cooperation in the face of the dangers of the world, could be fairly reduced to a 2x2 pay-off matrix in a way that could license any clean inference about dominant strategies.
There are with game theory some of the same empirical problems as beset memetics. To be fair to game theorists, their formalism is rich and highly flexible. There is a bit more of an idea of how one would apply the formalism (decisions and incentives are things we have intuitions about). But game theory is a highly flexible formalism and a flexible theory is one that too easily survives contact with reality. Take rational choice theory which, because you can have whatever bullshit preferences you want, puts essentially zero constraints on what choice counts as rational. Or Bayesian epistemology, which, because you can have any old fucking priors, counts as rational almost any inference soever.
Part Three. Finally, It’s Button Time
Quick reminder. Press the red button: you live for sure. Press the blue button: you die unless 50% of all button pressers press the blue button.
Quick reminder. Press the blue button and we might all come out of this alive. Press the red button and only the stone-cold psychopaths survive.
Around the late Obama years / first Trump presidency, psychology found itself in crisis. For decades, researchers racked up many apparently groundbreaking apparent insights into the human mind. PhDs turned into tenured faculty. Citation indices bloomed. All was good and right in the world. Then for a lark some boffins tried rerunning an old, much celebrated experiment. They didn’t get the original result. They tried again, some other people tried some other old, important papers. Not across the board, not all the time, but disturbingly often the results came back negative.
This was the replication crisis and it threw into question much that you would read in a psychology textbook. In some cases the replication failures were ethically sinister: the original data had been tampered with, the results were fraudulent. Certain careers were ruined. More frequently, the explanation came down to sloppy statistics or the ethical hinterlands of motivated reasoning (p-hacking). The implicated parts of the discipline went through rounds after round of self-reproach and proposals for reform.
Among the diagnoses that make the most sense to me is Tal Yarkoni’s.4 Start with a stylized example of a psychological experiment. Say you want to investigate the relationship between disgust and political orientation. You’ve read Leon Kass’s The Wisdom of Repugnance and you want to do a science on the bitch. So you get 30 undergrads. You put fifteen of them in a clean room and give them a questionnaire about their political beliefs. You put fifteen of them in a room where you’ve stashed some rotten eggs and give them the same questionnaire. Crunch some numbers. The sulfur-addled are Trump voters. Boom, publication.
Now, if you have any instinct for this sort of thing, you’re probably already thinking of problems with this imagined study. For one, I didn’t tell you about the effect size or the p-value. (Folks, the best effect size, the best p you’ve ever seen.) But also, 30 (?) undergrads (?!). Come talk to me when you have an actual sample size beyond the nearest-to-hand WEIRDos.
Yarkoni’s objection goes a step further. Why do we care about sample size in the first place? Well, people are different. If you tried this test with a couple people, the result might just be a weird fact about those two people. Maybe you just got lucky and put the Clinton voter in the clean room and the despicable in the dirty room. People are a random factor, so you need enough of them to cut through the noise.
Great, but by that token there are more random factors to control for. I mean, couldn’t this just be a weird fact about eggs? Do other disgusting stimuli have the same effects. And once you start down this train of thought, you can just keep going imagining factors that might be relevant to the outcome. Maybe in one room the questionnaire was delivered by a woman and in the other by a man and the difference is explained by subconscious misogyny. Maybe the dimensions of the room matter. Maybe the time of day. Maybe…
Psychology budgets are not large enough, could not feasibly be large enough to collect the samples needed to control for all of these factors. With this in mind, the replicability crisis is not surprising. The studies are systematically underpowered, more likely to be telling you a fun fact about the sample the researcher had to collect than anything about the research questions the experiments were supposed to answer. Crucially, this mismatch arises with the imprecise terms of the questions. A chemist investigating the nature of, pretending to remember anything about chemistry here, covalent bonds, may be asking questions precise enough about a limited enough domain that just the right observations collected with sufficient care will actually answer her questions. Not so with the mind.
Part 4. I FUCKING LOVE BUTTONS.
If I had a button that would poof Donald Trump out of existence, I would press it. I don’t think this is an unusual thought for liberals. We are, of course, repelled by the idea of political violence, at least for the chaos we presume it would unleash. Perhaps we might hesitate at the thought that in the power vacuum some more eldritch horror may slither its tentacles around the nation. But it’s a risk I’d take on. I’m not vindictive. I don’t need the button to hurt. Even if the button sent Donald Trump’s soul to everlasting paradise, I’d press it. “For the good of the country.”
Of course, this is pointless grandstanding. There is no such button. It does not matter whether I or anyone would press it. And yet, pressing buttons is about the most I can imagine for our politics.
In 2015, I pickled my brain on Facebook. LA, the worst city on the planet, went spongiform and unreal at the periphery of an endless feed of the fluffiest forms of science journalism heretofore invented and “life updates” from high school acquaintances I was more or less happy to see relegated to a digital fuzz. The feed consumed and then metastasized the spectacle of the Democratic establishment cheating Sanders out of the nomination and then own-goaling into the election of Donald Trump, like a cobra swallowing a live rabbit. The machine said be angry. The machine said be hopeless. I was angry. I was hopeless. In the background, our understanding of human nature fizzled in breathless article titles parodied into even-more breathless BuzzFeed headlines that all failed to replicate. At some point the Cambridge Analytica scandal broke—the void I shouted my hatred of Donald Trump into had quietly gone and elected him. The most popular method of suicide is to pipe the car exhaust back into the interior.
“If you push the button,” Mr. Steward told him, “somewhere in the world someone you don’t know will die. In return for which you will receive a payment of $50,000.”
In Richard Matheson’s short story “Buttons, Buttons,” a mysterious figure presents a couple with the above proposition. The husband rejects the offer on principle, but the wife (spoilers for a fifty year old short story) slowly rationalizes pressing the button. They sure could use $50k, and besides, who is to say they’d actually kill someone. Probably this is some weird psychological experiment. So, she pushes the button. Except, whoopsie daisy, the monkey’s paw curls and her husband is the stranger who died. “‘My dear lady,’ Mr. Steward said. ‘Do you really think you knew your husband?’”
Matheson is here riffing on a thought experiment from a 19th century apologetics, François-René de Chateaubriand’s The Genius of Christianity. De Chateaubriand’s assumption here is that one wouldn’t press the button. Or at least one would naturally feel the whole weight of conscience arrayed against the prospect. And it is the existence of the conscience that shows the immortality of the soul and the existence of God. What’s the argument? “It would be an insult to the understanding of our readers, did we attempt to show how he immortality of the soul and the existence of God are proved by that inward voice called conscience.”5
To be fair, de Chateaubriand says enough elsewhere that we can loosely reconstruct the argument he might have given.
Humans are fundamentally self-interested: other motives are derivative on our self interest.
Humans are frequently motivated to do good to others in ways that cannot benefit us in our earthly existence.
Therefore, there must be an afterlife in which we are rewarded for the good we do for others.
It would be an insult to the understanding of my readers to point out the ways in which this argument is shit. What interests me more is the context of this argument. There is a parallel to be drawn between Hobbes and de Chateaubriand, even though Hobbes was notoriously a skeptic (of sorts) and de Chateaubriand a Catholic apologist. But both wrote in the aftermath of a revolution—the English Civil War for Hobbes and the French Revolution for de Chateaubriand—and urged a conservative reform to “the old ways”—absolutist monarchism for Hobbes and Christian suprematism for de Chateaubriand. And in the structure of the conservative politics, we can say Hobbes’ sovereign is more or less an earthly version of de Chateaubriand’s God: both ground morality by enforcing a regime of rewards and punishment that aligns prosocial and self-interested behavior.
Part 5. Hot, Fast, Button-Pressing Gameplay
A few pages after de Chateaubriand (this is still an essay about the red and blue buttons) presents his argument the fact that he wouldn’t murder a stranger for money to the immortality of the soul and the existence of God, comes an even stranger, racist argument for the immortality of the soul.6
Pseudoscientific (quaint) assumption: In general, nature is more vigorous towards the equator. Elephants be big.
Pseudoscientific (extremely racist) assumption: Humans are more vigorous away from the equator. White Europeans be the best.
So there is something in human nature that is contrary to natural law.
That thing is the supernatural, immortal soul.
There comes a point with apologetics of this caliber where the arguments are so piss poor that the question changes from “How did the author make such blunders in reasoning?” to “Is the author even reasoning at all?” And if the author is not really reasoning, then what they are doing instead?
I have nursed a suspicion that much apparent argumentation is not sincere but rather a sort of epistemic game. I mean “game” in Bernard Suit’s sense of an activity aimed at a goal but in which the actors deliberately constrain themselves to inefficient means.7 It would be much easier to get a soccer ball into the goal if you held the ball in your hands and ran it in, but you wouldn’t be playing soccer if you did so. So in an epistemic game the goal is expressive (assert some conclusion) but the player has voluntarily taken on the burden of first putting on a display that has similar verbal tics to an argument.
This way of thinking about what the apologist is doing we can draw out by contrast to two alternatives. The first is motivated reasoning. The speaker is genuinely trying to make an argument for their conclusion, but it’s a hasty post-hoc argument. They’re sincere, they are aiming to produce rational conviction, they’re just not doing a good job at it. The second is the conner. Their argument sucks, they know this, but they hope the audience is going to be too stupid to catch them, that by charisma or rhetorical tricks they’ll charm or stupefy the audience into belief.
The player of games is not really aiming at conviction, in the same way that it’s not quite right to say that the aim of a game of soccer is to put the ball into the net over there more times than the other team puts the ball into the net right here. The aim of the soccer player is to win the game by scoring the most goals, and the rules of soccer are part of that aim (they define what it is to score the most goals). This aim is artificial because the game is artificial, it arises from some other more foundational aims combined with a context. (I want to have fun with friends. I want to be a famous athlete.) In the same way there’s an inner ludic goal of saying a predefined conclusion after a sequence of apparently argumentative sentences and an outer non-ludic goal.
That outer goal I think is often one of social identification and value projection. Take conspiracy theorizing. If you’ve ever dipped a toe into conspiracy theories, you’ll observe that the quality of evidence, the quality of argumentation is abysmal. Conspiracy theories are as a rule bad theories. Conspiracy theorists seem unbothered by this lack of rigor. More interestingly, conspiracy theorists tend to be highly promiscuous with their conspiratorial beliefs. The Venn diagram of flat earthers and Q-Anon believers is a small circle inside a larger circle.
This makes complete sense if argumentation is not the point. Indeed, specific conspiratorial beliefs tend to be superficial, easily abandoned. What is core is rather a group identity: we are the people who are figuring out what’s really going on beneath the lies and evil of those in power. To be able to hold this identity requires the simulation of investigation in the same way that being a soccer player requires using your feet to manipulate the ball. Actual investigation carries too much risk of disappointment, of piercing the veil of the conspiracy.
Carry on pretend thinking for long enough and you become unable to identify real thinking. It will bother you less that your own position is fundamentally a vibe in argument’s clothing if you think this or worse is true of the alternatives. Thus to the conspiracy thinker the mainstream is fake, bought-off: our premises may be shoddy, our reasoning fallacious, but at least our conclusion is correct, unlike those bozos over there.
The year I lived in LA, for the only time in my adult life, I went to church most weekends. A progressive, Episcopalian church to be sure, more a coffee circle than a sect. Still, even though our convictions were more than a little halfhearted, the services had ritual power, produced the vibe of belief though not belief itself, a certain flavor of fellow feeling. Our Father who art in Heaven… And the sermons too, felt nice, were nice. It’s good to be nice.
There our pastor was, laying out how this or that chapter of Luke meant we should be treating immigrants like our neighbors which was to treat them as ourselves and though no ectoplasmic iota of my almost certainly mortal soul believed in the Father, the Ghost, or the Holy Spirit, Hell yeah, Linda, fucking tell ‘em. The point was not how we reached the conclusion, or even the conclusion itself. The point is it’s nice to be agreed with. It’s nice to hate the outgroup. It’s nice not to have to think about why your life is so nice.
Social media is the most efficient mechanism ever delivered for the frictionless delivery of such social affirmation. Indeed, at that point Facebook was almost finished with reoptimizing itself from being a public notice board for sharing pictures of your kid’s graduation to a smooth, sealed chamber offering the frightfully wonderful stimulation of complete oxygen deprivation. I Fucking Love Science for the easy simulacrum of being informed. Blueposting for the simulacrum of political participation. We are the only good citizens of this most blighted nation.
Coda. The Actual Thing With the Button
The button is the McGuffin of this essay. I’ve avoided talking about it because there’s little interesting to say. You should press the blue button. I don’t think this is a Newcomb’s Paradox situation where a consistent disagreement reveals something profound about the philosophical landscape. Some people are just brain damaged in such a way that saying they would press the red button feels attractive, maybe contrarian attractive, maybe enlightened by my own ego attractive, maybe “I watched the Matrix 8 too many times” attractive. It’s not really a big deal. It’s essentially impossible not to be brain damaged by the current state of things.
Chvaja, Radim. "Why did memetics fail? Comparative case study." Perspectives on Science 28, no. 4 (2020): 542-570.
“The development of a memetic framework was followed by many logical, theoretical, and empirical objections… but there are also more objective indicators of memetics failure. First, currently only two well-known scholars could be identified as active memeticists…. Also, Edmonds (2005) shows that between 1986 and 2004, only 41 articles used the word “memetics” indexed in the database Web of Science. Second, memeticists themselves have realized that memetics is no longer considered a serious scientific theory.” (544)
E.g., Boulting, Noel. “Ought Hobbes’s Natural Condition of Mankind Be Represented As A Prisoner’s Dilemma?.” Hobbes Studies 18, no. 1 (2005): 27-49.
If you’re interested in the good version of the project of deriving norms through game theory, there’s Bicchieri, Cristina. The grammar of society: The nature and dynamics of social norms. Cambridge University Press, 2005.
Yarkoni, Tal. "The generalizability crisis." Behavioral and Brain Sciences 45 (2022): e1.
de Chateaubriand, François-René, and Charles J. White. Genius of Christianity, Or, The Spirit and Beauty of the Christian Religion. Murphy, 1884. 190.
ibid. 194f.
Suits, Bernard. “What is a Game?.” Philosophy of science 34, no. 2 (1967): 148-156.
I'm very happy to read a new essay from you, it's been a while! This was well written, reasoned and, of course, witty as usual. Great read! :)