Interesting things I read or thought about in February 2026
Curling cheating, assisted dying, why debates over 'poverty' are often meaningless, weird religious art, trying to understand Nigerian infanticide, witches, speech, poetry, QFT
At the Fitzwilliam I wrote about Lafcadio Hearn, the half-Irish half-Greek American journalist who became Japan’s pre-eminent interpreter to the Western world.
Apart from Hearn, over the last month I have been thinking about:
Curling ‘cheating’, why football is shit, and why assisted dying would be good
‘Poverty’ is an arbitrary line, but ‘subsistence farming’ is a real category
Curling ‘cheating’, why football is shit, and why assisted dying would be good
So. The curling ‘cheating’ scandal. (If you’ve not been paying attention whatsoever, here’s a BBC article.) You probably already have opinions, but here are mine.
If you read the reporting and opinion pieces about this carefully, you’ll notice that curlers are generally much more nuanced about it than non-curlers, or even entirely nonplussed (relevant Reddit thread). I’m not a curler, I have very little context for the sport (so take what I say with many grains of salt), but I’ve tried to research this as best I can.
As far as I can figure it out, most people agree that Marc Kennedy’s ‘double touch’ infraction didn’t help him and wasn’t done maliciously; it was just a ‘tic’ or force-of-habit. But it was explicitly against the letter of the rules.
The rule against double-touching was written quite broadly (no double-touching at all past the hog-line) in order to make it easier to apply. If some qualification had been introduced—say, the rule was that you couldn’t double-touch ‘in a manner that provides advantage’—then you would have endless arguments over what counted or not; better to have a blanket ban.
But also, if someone did an obviously-honourable accidental violation that led to no advantage, other curlers always had the option to just … ignore it. Traditionally, curlers wouldn’t have been surveilling their opponents as aggressively as they were this time around (this is why people are talking about the ‘spirit of curling’ being violated), and there wouldn’t have been high-def slow-mo TV cameras to prove that a violation had happened. So an extremely strict law with no exceptions could work!
But now curling is more surveilled, and the rules are being enforced more vigourously. A rule that worked well in the past now creates vicious disputes (in a sport with a heretofore unblemished reputation for honourable conduct).
More importantly, you can’t just wish away an increased level of surveillance. Curling doesn’t officially employ any surveillance technology, as far as the rules are concerned the TV cameras might as well just not be there, the officials aren’t allowed to use them to adjudicate disputes; and yet the controversy happened anyway. The TV cameras are there, people will watch slow-mo replays, and that will influence the discussion around the sport—and the honour or dishonour that is accorded its participants. Since sportspeople are mostly in it for the honour, this really matters.
The moral of the story is that rules can only make sense relative to a fixed level of surveillance and enforcement. When the level of surveillance and enforcement changes, the rules need to change.
It is often difficult to know how to change the rules to capture the spirit of the game in the context of the moment: it usually requires extremely deep knowledge of the sport, which I don’t have in the case of curling, so I can’t tell curlers what to do. But regularly changing the rules to reflect a changing meta is just what healthy sports do. F1 is probably the most famous example (outside of more esoteric things like esports and card games), because it has to change its rules so often in the face of technological innovation. But it’s far from alone. To cite the example I care about the most, artistic gymnastics’ 2006 overhaul of the Code of Points was the best thing to happen to the sport for at least half a century.1
In unhealthy sports, by contrast, participants and fans walk calmly and slowly into high-surveillance quicksand, then decide that it’s the fault of biased referees / stupid officials / the other team that they’re sinking. Football is the perfect example. As TV cameras became higher-tech and achieved more detailed coverage of the whole pitch, with slow-mo and zoom, participants and spectators regularly used ‘replays’ to unofficially relitigate decisions made by referees who didn’t have access to the tech. Obviously this was undermining trust, and so the sport’s governing bodies gave the green light for referees to use higher-tech solutions like the infamous VAR.
The fact that I could write ‘the infamous VAR’, though, shows exactly what happened: fans got mad that VAR was resulting in perfect and pedantic enforcement of rules which (up till then) had been enforced with wiggle room and grace and generosity.2 This is not unreasonable in itself, but it demonstrates the dilemma. Either the refs are allowed to use advanced tech, and everyone hates the fact that the rules are being perfectly enforced; or the refs aren’t allowed to use advanced tech, and everyone yells at them for being so bad at enforcing the rules (or at least, fans of team x yell at them when their decisions hurt team x, ∀x).
Fundamentally, the current rules of football are based on the assumption of imperfect enforcement; now that enforcement has gotten more perfect, they are badly adapted. The solution is to change the rules.
Of course, football is too group-think-y and corrupt a sport for sufficiently radical changes to be made. It’s ‘the beautiful game’ (apparently), and the possibility that it’s badly-adapted and the rules need a radical overhaul is as far outside the football fan Overton window as open-the-borders-and-abolish-the-state-pension libertarianism is outside the political Overton window. (Plus, a significant minority of football fans are mostly in it for tribalist yelling, and the current set-up gives them more opportunities for yelling.)
Luckily you and I, being reasonable and well-adjusted people, can simply decide that football is not worth paying attention to. But sadly, less group-think-y and corrupt sports also struggle to make the leap; rule changes are rare and hard to reason about, and it’s easier to hold the rules fixed when you think about how to respond to infractions and controversy. As such, in response to the current controversy, some curlers are calling for VAR, to make decisions ‘dead certain’. If my above analysis is correct, this would solve nothing and just make the sport more pedantic and pointless. Nonetheless, it’s true that figuring out the right solution will demand creativity and careful analysis. I hope that curlers, who are generally an extraordinarily intelligent bunch, will manage to apply their minds to come out of this with an improved sport. But it’s not guaranteed they will.
And if we can’t guarantee it in curling, where everyone involved is intelligent, how can we guarantee it in public policy? (lol sorry.) The same dilemmas described above for sport apply: much of the law has been designed on the assumption of imperfect surveillance and enforcement. This paper makes this point in the context of AI: well-designed and extremely liberal laws can become actively dystopian if we assume perfect enforcement. What you need are ‘enforcement-native’ policies.
A great example is assisted dying. A common misconception (actively promoted by pro-life ideologues on one side of the debate) is that the ban on assisted dying reflected Christian pro-life ideas: there was a widespread moral horror and revulsion at the very idea of assisted dying. This is simply not true. Rather, it’s a bit like the curling double-touching rule. A blanket ban allowed for easier conviction in cases of murder or manslaughter, without having to adjudicate impossible questions around whether or not the deceased wanted to die. But also, in the early twentieth century, patients (or rather, well-off patients) tended to have family doctors who they had a personal relationship with, and who were minimally surveilled by the state. But if the doctor has talked to the patient and (to ensure nobody will go to the police) their family, and then decides to slip them a little extra morphine, the state would be none the wiser. And the state was happy to be none the wiser.
Nowadays, however, most sick and dying people spend a lot of their time in hospitals or hospices or care homes, which are highly-bureaucratised public spaces with immense levels of surveillance, usually either run by the state or at least heavily regulated by the state. Doctors still from time to time give lethal doses to patients who want to die—don’t trust anyone who tells you this doesn’t happen—but it’s becoming rarer because drugs are logged and tracked, and doctors rarely have long-term personal relationship with patients and their families: patients instead cycle through a suite of different specialists, assisted by untold numbers of nurses and PAs, all overseen by one of the vastest bureaucratic apparatuses in the world.
In the UK, the response to increased surveillance has been artificially reduced enforcement. In England and Wales, the Crown publicly says that it won’t prosecute people who help family members die, if prosecutors are satisfied that the dead person wanted it. In Scotland the status quo is very unclear, but a lot of people assume that the Crown would act in basically the same way. It should be clear that there are constitutional worries when bureaucrats and lawyers refuse to enforce the very clear law passed by parliament. There are also policy worries: in the first place, it shifts the responsibility to help patients die away from doctors (who know what they’re doing) to family members (who don’t); it also means that the only time anyone outside the family asks ‘does this person really have a settled and rational wish to die?’ is after they have already died, which is, um, not ideal.
What we need to do is leverage the new surveillance environment rather than wishing it away. Thanks to the new environment, what I above called ‘impossible questions around whether or not the deceased wanted to die’ are no longer impossible. Dying people previously would have come into contact with the state once or twice at the end of their life; now they are virtually never out of contact with the state, and this gives the state access to previously-inaccessible information. In principle it can now be established to within a high degree of reliability whether or not a terminally ill patient has a settled, rational wish to die (sooner than they otherwise would).3 Given the extraordinarily high degree of public consensus that such patients should be allowed assistance to end their lives, the right policy follows directly: if a dying patient expresses a wish to die, a bureaucratic information-collection procedure should begin, and if it is established that this wish is settled and the patient has capacity, then a doctor ought to help them end their life.
From this perspective, the policies of ‘assisted dying on demand’ and ‘assisted dying is completely banned’ aren’t opposed, they’re actually alike: neither is native to bureaucratic surveillance, and so neither takes advantage of the possibilities for information-collection. By contrast, live proposals in the UK to legalise assisted dying do take this seriously. It seems somewhat unlikely that these will actually pass in the near future in England and Wales, thanks to constitutional chicanery from a bunch of pro-lifers. Things in Scotland are less legible from the outside, but there’s also constitutional chicanery going on. In the Republic of Ireland, the policy—previously endorsed in principle by the Dáil—seems unlikely to proceed under a minority government. It’s the kind of policy that struggles to become law, because pro-life opposition is concentrated, but support for the policy is diffuse and the benefits are stochastic: voters don’t know today whether or not they’ll need it when they’re at the end of their lives. But in places where it has been implemented it works extremely well, precisely because it doesn’t wish away the realities of increased enforcement.
‘Poverty’ is an arbitrary line, but ‘subsistence farming’ is a real category
Tibor Rutar writes that ‘Extreme poverty before capitalism wasn’t as prevalent as you think’. He cites some pretty interesting economic analysis, but I’m not sure he should change your mind on the vibes.4
The thing about ‘poverty’ is that it’s not a well-defined concept. We all think we know what it is, but when it comes to actually saying what we mean, it kind of vanishes. The real definition is either that you say someone is in poverty if they are living on less than the equivalent of £x a day (‘absolute poverty’), or you say that they’re in poverty if they are living on less than y% of the average wage in their society (‘relative poverty’). But regardless of which definition you choose, the values of x and y aren’t forced on you by the situation: they’re free parameters you can set however you want, for whatever particular purpose you have.
And indeed, different people set different ‘poverty lines’ for different purposes. In particular, Rutar picks up on Robert C. Allen’s work that tries to apply a ‘basic needs poverty line’ to societies where we lack good economic data. Allen does some quite clever analysis to ensure that he can get accurate estimates for ‘basic needs poverty’ in societies without modern-style economic data; it’s quite clever. But fundamentally, Allen’s conclusion—that most people in preindustrial societies were not in poverty—depends crucially on his definition of ‘basic needs’; his definition is actually pretty reasonable, but it should also be pretty obvious that this is kind of subjective.
This is all to say that ‘poverty’ is not what philosophers call a ‘natural kind’, a concept that picks out objective similarity in the world. There’s no ‘natural’ place to draw the poverty line.
But there is a different term that does reflect real objective phenomena about how humans lived before the Industrial Revolution. That term is ‘subsistence agriculture’: farming that doesn’t create products to sell on the market, but which just about allows you to provide for yourself and your family. (Maybe subsistence farmers have some surplus, but this is traded for other necessities like clothes, or taken as rent or taxation.) Subsistence agriculture is a real, objective phenomenon in human social organisation, one that was widespread across the entire globe from the Neolithic until a couple of hundred years ago.
If you translate Allen’s argument out of the language of poverty, what he’s really saying is that most subsistence farmers, in most places, most of the time, actually did manage to achieve subsistence and weren’t starving to death or facing major nutrient deficiencies. (This makes intuitive sense: if most subsistence farmers in a society weren’t meeting their most fundamental needs, probably ‘Malthusian’ forces would kick in.)5 But whether or not it’s reasonable to say that someone is ‘out of poverty’ because they’re above this threshold, even though they are still subsistence farming: that is a matter of vibes, not science.
A few years ago I watched this series of videos, where a YouTuber tries to make his own bread literally from scratch: buying the seeds, growing the grain, threshing, milling, and baking. There’s a lot of ‘trad’ content creators, trying to cash in on nostalgia for an imagined golden age, who do something like this, but usually they defeat the point of the whole exercise by using modern industrial agricultural equipment that could only have been built in non-‘trad’ societies. But this guy took the exercise more seriously than that, and used only equipment he had lying around anyway.
Anyway, his bread was shit. It came out low-gluten, bad-tasting, poorly-risen—and it was full of rocks. When he threshed out the grain, it picked up a bunch of little rocks and sand from off the ground outside, and he couldn’t really manage to separate them back out. So he asks the obvious question: how did ancient people get the grit out? He then cuts to the expert who he’s been interviewing to help guide him through this process (Catherine Zabinski), who laughs and says:
That moment stuck deep in my head, and has never left. People in many subsistence-farming societies got the vast majority of their calories in the form of bread with little rocks in it. People eating this bread maybe had their ‘basic needs’ satisfied according to Allen’s definition, but: have you ever bitten down on something hard in your food that shouldn’t have been there? Imagine that being every single meal. You could try to be more careful about chewing, but you can’t avoid chewing altogether. Not all subsistence farmers relied on wheat bread, of course, and rocky bread was certainly not universal before 1800; but hopefully this gives you a taster of the bad vibes of subsistence agriculture.
It wasn’t just the subsistence farmers themselves, either, it was the wider societies built around subsistence agriculture. I have been reading Cyril Aldred’s book about Akhenaten, the Egyptian pharaoh, and this same point appears there in a passage about Akhenaten’s predecessor Amenhotep III. Amenhotep III was plausibly the richest man in the world at the time. He was in the same dynasty as Tutankhamun, which was plausibly the high-point of ancient Egyptian wealth and power, as seen in the immense quantities of gold and precious stones and luxuries found buried in Tutankhamun’s tomb—but Amenhotep III’s tomb would have been much larger and even more luxurious. And that’s only the wealth he was buried with: how much more wealth he would have had during his life!
And yet. Because of Egyptian mummification practices, we can examine his body:
he must have suffered miserably from dental disorders in his later years… [His] teeth had been worn down to the pulp by a lifetime of masticating the staple of Egyptian diet, bread made from flour with a high percentage of grit… [Aldred, Akhenaten, p. 106, emphasis mine.]
Think about the pain that Amenhotep III must have felt every day with teeth that bad. Think about the number of times he must have bitten down on a hard bit of grit, in order that his teeth were literally worn down to pulp. And keep in mind that, had there been any better alternatives in Egyptian society, he of all people would have been able to take them.
People in my circles like to tell a story about Agatha Christie looking back on her early-twentieth-century middle-class upbringing: ‘she never thought she would ever be wealthy enough to own a car—nor so poor that she wouldn’t have servants.’6 This is a really striking line that really gets at how society changed in the last hundred years. But the change over the last 4,000 is even more striking. Amenhotep III probably couldn’t have imagined being so poor as to not afford a massive tomb, carved into rock, filled with gold and jewels and intricate artworks; nor so rich as to be able to afford bread without rocks in it. (And he was lucky: just as most people in Christie’s society could afford neither cars nor servants, most people in Amenhotep’s society ate rocky bread and didn’t even get the gold to make up for it.)
So those are the facts. The vast majority of people in all societies between the Neolithic and the Industrial Revolution lived as subsistence farmers; and in subsistence-farming societies, people—even those who were not themselves subsistence farmers—suffered serious afflictions. There’s no fact of the matter as to if those afflictions count as poverty, it comes down to a purely subjective choice of how to define the word. But whether you use the word ‘poverty’ or not, subsistence agriculture fucking sucks.
Weird religious art
Speaking of Akhenaten: as per the famous XKCD, maybe you’re one of today’s lucky ten thousand who get to learn about him.
So, the aforementioned Amenhotep III was pharaoh, and then in the 1350s B. C. E. he died and was succeeded by Amenhotep IV—all well and good and normal for ancient Egypt. But then, a few years into his reign, Amenhotep IV suddenly decided that Egyptians were no longer allowed to worship their traditional gods. Instead, they were to worship the ‘Aten’. Unlike the traditional dudes-with-animal-heads images of Egyptian gods, the Aten was portrayed as a sun with hands at the end of its rays?
The temples of the Aten looked radically different to the entire previous architectural tradition of Egyptian temples: they were open air, directly below the sweltering sun. Amenhotep IV also changed his name (which contained the name of the old god Amun / Amen) to Akhenaten. He also moved his court to a brand new city built in the middle of fucking nowhere.
Also, at exactly the same time, the entire artistic style of ancient Egypt—which had stayed extremely conservative and consistent for millennia—changed radically. Like, really radically. Especially in art from the early period of Akhenaten’s reign, it’s extremely freaky and the people almost look like aliens:
Akhenaten himself is always portrayed in this incredibly androgynous way, with wide hips and even sometimes boobs. And his face looks bizarre, elongated, with these massive lips.
By the end of his reign, things begin to move in a more realist direction, but they still look nothing like traditional Egyptian art.
And then, suddenly, very shortly after Akhenaten dies, it all stops again. People go back to worshipping the old gods in the old temples; Egyptian art goes back to the old style; the new city in the desert was abandoned. All of this happened, and then un-happened, in a space of about twenty to twenty-five years.
Before long the Egyptian government decided to try to erase all evidence of Akhenaten’s existence from history. And they basically entirely succeeded. We didn’t rediscover him until the nineteenth century, and we still don’t really know for sure why any of this happened or what the art means, despite many theories.
Art is so often about context: understanding where and when an artist was working, what tradition they were working in, what events they were responding to, what their beliefs were. (For example: if you are going to walk around any gallery of historical European art, you either need to already have a good understanding of Christian theology and scripture, or you will have to be willing to do a lot of Googling in order to pick up information about Christianity.) But we just don’t have any context for Amarna art. All we really know is that this happened, we have no idea why. The few texts that remain from the period only raise more questions, and don’t answer anything: in one inscription by an Amarna sculptor, he says he is ‘an apprentice whom His Majesty taught’, which could suggest the art style was designed by Akhenaten himself? or could just be an obsequious display of loyalty. Or something else.
This wouldn’t necessarily be all that interesting except that the art itself is so surprising—weird as hell.
And yet, it’s also got a very affecting intimacy. There are a lot of family scenes in Amarna art, of the king and queen acting like normal parents with their kids, or their kids playing with each other, or stuff like that—none of which were subjects of Egyptian art. It’s this freaky mix of the alien and the deeply human.
Akhenaten’s successors (the bastards) have probably succeeded in making sure that we will never know what happened. But the art means that we’ll never stop trying to find out.
Generalising from this example. Religious art is often the weirdest and hardest for a new audience to understand, precisely because you need to actually understand the religious beliefs and practices that lie behind it. But there is something really compelling in its weirdness and difficulty: the question ‘what does this mean?’ keeps appearing in your mind.
I spent a lot of February looking at Mandaean art. The Mandaeans are probably the only surviving descendants of the ancient religious trend often called ‘Gnosticism’ (although they claim to in fact be following the authentic religion of John the Baptist); they mostly lived in Iraq before 2003, but now they’re scattered in a diaspora.
The only proper study I’m aware of of Mandaean art is by an academic in Australia, Sandra van Rompaey, who did her PhD thesis on the topic only a couple of years ago. But it’s not open-access because it’s been published as a (stupidly expensive) book, and the book isn’t available through the … usual channels. So I was reduced to just browsing through images I could find online with no context.
But these images are extremely cool and strange.
Infanticide in Nigeria, and finding loose ends
Rotating back around to the topic of how awful subsistence farming is: I read, and was deeply affected by, this piece in the Guardian about infanticide among subsistence farming communities near Abuja in Nigeria.
The basic story told in the article is that these communities—which are usually connected to the outside world only by unreliable roads that become impassable, and which are thus relatively isolated from outside culture and law enforcement—have superstitions around certain babies being ‘evil’. The superstitions are not universally held, but elders and witch-doctors take the lead in encouraging the killing of these children, often against the wishes of their families.
There’s a lot of tantalising and tragic detail in the piece. And it all chimes with a large anthropological literature about historical infanticide practices in Nigeria, that I started reading. The Guardian article and the anthropological literature both emphasise that infanticide was/is commonly carried out against twins, and secondarily against children with disabilities or albinism, and children of mothers who died during birth. Twins were taken to be ‘evil’ or ‘demonic’ in some way in many communities across Nigeria.
As I read this, I began to feel like I had a pretty good handle on the phenomenon. The attitude that twins should be killed was historically widespread in many parts of Nigeria; it was attacked and successfully repressed thanks to cultural pressure, especially from missionaries (the person most often mentioned is Aberdonian missionary Mary Slessor); but it persists in a few extremely poor and somewhat-isolated pockets of the country. This paper is an economics-y overview that analyses a lot of data from a bunch of different African countries, not just Nigeria; this one is a good example of the anthropological literature, focussing especially Igbo practices; this one discusses various explanations for why twins were killed, and while I’m not entirely happy with its analysis (they ‘ignore the parents’ emotional costs of having the twins killed’!!!), it’s helpful for orienting yourself in the debate. I want to stress that existing analyses really focus on the killing of twins in the first instance, and secondarily of disabled or albino children.
But the Guardian piece had a brief mention of polling carried out by the charity ActionAid in the subsistence farming communities that still carry out infanticide, asking people about their attitudes to the practice. Not liking to leave a loose thread dangling, I dug the report out. And the results were… completely out of line with the existing literature?
Twin infanticide is supported by about 6.4% of those surveyed, albeit men are twice as likely to support it as women. But the killing of the next child born after the twins is more than twice as popular—women are 40% more likely to support this type of infanticide than men are to support twin infanticide!
This practice, the killing of the next baby born after twins, is—as far as I can tell—not discussed anywhere in the extant English-language literature. (Please please please correct me if you know of an example, I would really like to learn more about this.) The Guardian article, which drew on ActionAid’s work, didn’t mention it at all, despite many mentions of twin and albino infanticide. Yes this is by far the most-supported form of infanticide in these communities. This is a really noticeable ‘loose thread’ that nobody else has picked up.
It raises some obvious questions that I have no answers to. In the first instance, why the child born after twins? The most obvious theory is that there’s some kind of ‘contamination’ logic, the evil of the twins rubs off on the next kid too—but this makes no sense given that people who don’t support twin-killing do support killing the next one.
There’s also a practical question. Obviously if a woman gets pregnant she doesn’t know if she’s going to have twins, or if her baby will be born with a disability. I can see how twins could therefore be interpreted as an unforeseen ‘disaster’. But if you’ve just had twins, by definition you know that your next baby will be a baby-born-after-twins. So do parents avoid having children after they’ve had twins to avoid bringing ‘evil’ into the world? If so, do they use some sort of contraceptive method (which one?) or do they just abstain altogether from sex? Or do they just continue on having sex, knowing that their next child will definitely be born with a target on their back? After all, large families are often a necessity in subsistence farming communities, since you need labour for the harvest. Knowing that they need to have more children, do women go through post-twin pregnancies with the knowledge that the child might immediately be murdered, just so they can then have more children after that? and do more than one in ten of them, and around one in six of their husbands, support this?
There’s only one bit of the existing literature than I can see maybe helping us understand baby-born-after-twins infanticide, but it involves some irresponsible speculation and guesswork. This paper by T. J. H. Chappel discusses a really interesting phenomenon among Yoruba people, who used to practice twin infanticide: their historic belief that twins were evil got reversed into a belief that twins are especially auspicious and should be celebrated. Yoruba twins often get special names, with one twin being called ‘Taiwo’ and the other called ‘Kehinde’, and there are all sorts of traditional beliefs about how important twins are and how they bring benefits to the family. There’s even a practice whereby it is considered bad for the family if one of the twins dies, and so—if that happens by misfortune—the parents will bring a special wooden sculpture into the house to ‘stand for’ the dead twin, and pay it appropriate homage. You can kind of see how this idea might emerge, by radical negation, out of the contortions required to reject a widespread cultural belief in infanticide. An analogy might be the special role that firstborn children have in some books of the Old Testament, in light of the very strong evidence we have that at one point Yahweh-worship involved the sacrifice of firstborn children.
Interestingly, Yoruba people today think the next baby born after twins is also special. They also get a special name, ‘Idowu’, and there are also cultural traditions about them; but these seem, to this outsider at least, to be weirdly contradictory. On the one hand, this baby is supposed to help the family return to normalcy; on the other hand, they will be tricksterish and troublesome, and are associated with the trickster god Eshu. It is not apparent to me from the outside how this combination of beliefs makes sense, which is itself interesting and potentially productive.
Chappel highlights that this third child, Idowu, is ‘the “key child” in the twin-birth complex’: ‘unless an Idowu is born after twins the consequences for the parents may be serious. The twins will be unhappy, for Idowu is the “Eshu of twins” and all twins must have their Eshu. If no Idowu is forthcoming the twins will take revenge on their parents; the mother may fall seriously ill or become mad.’
Chappel speculates about the origins of these ideas. One suggestion he makes is that the idea of Idowu came out of the ban on infanticide, as a projection of cultural uncertainty about twins onto an individual ‘trickster’. But another—and this is the one that I want to pick up on—is that Yoruba people perhaps believed that twins resulted from the actions of spirits, perhaps spirits having a supernaturalised version of sexual relations with the mother:7 ‘the woman [is] the “victim” of Eshu’s caprice in the field of sexual relations’. Chappel doesn’t quite say this explicitly, but I think that the idea is: twin births are caused by an Eshu-like spirit, which then becomes incarnate in the next child born after the twins. There’s an old 1928 anthropological paper that does make this claim explicitly; it’s suitably racist (apparently in West African people ‘the negro strain is … as nearly as possible unmixed’), but even if its value as evidence is some tiny ε, still ε > 0.
Chappel doesn’t discuss the killing of these children-born-after-twins, only the killing of twins, but I think his ideas point in a very clear direction. In a culture where twin births were anti-auspicious (as was historically the case among the Yoruba, and other groups in Nigeria) then that might well give rise to twin infanticide. But if there’s a further belief that this terrible event was the fault of a tricksterish or even evil spirit, and this spirit will become incarnate in the next child born after the twins, then that definitely seems like it would give rise to child-born-after-twins infanticide.
The communities around Abuja that still practice infanticide are not Yoruba. But, given that Yoruba people were historically not alone in practicing infanticide, maybe we can use inferences about historical Yoruba practices to help understand practices among other groups? Maybe? This is obviously extremely speculative and inexpert and clumsy, I really don’t know what I’m talking about here. But it’s the best I can do.
Why, then, does everyone talk primarily about twin infanticide, if (as I am speculating) child-born-after-twins infanticide may well have been even more prominent? Presumably this is in part because beliefs about twins still seem to be ‘primary’ in some sense. But also, maybe because it was easier for Christian missionaries to notice twin infanticide. It can be very tricky to tell the difference between a society that practices infanticide and a society with high rates of infant mortality, unless you literally witness the killing. This goes double when the method of killing is poisoning, carried out by witch-doctors who won’t cooperate with missionaries, as described in the Guardian article. But twins are inherently high-salience, and it’s much easier to notice if they’re dying at a higher rate than the infant population as a whole, especially if there’s a pattern (e.g., only the older of the two twins is killed, as happened in some communities). So Christian missionaries and (later) governing elites figured out that twin infanticide was happening, and then also infanticide against other high-salience groups of children like disabled and albino children, but they missed the more common practice, because ‘child born after twins’ was not a salient category for them.
The fact that Yoruba beliefs about the child born after twins seem less consistent and less positive than Yoruba beliefs about twins reflects two things: (a) it was harder to switcheroo the Idowu from negative to positive, because they were previously thought to be so negative that the cognitive distance was greater; (b) missionaries and the wider colonial culture, focussed so much on preventing twin infanticide, didn’t encourage thinking of Idowu as auspicious as much as they encouraged thinking of twins as auspicious.
And in turn, in the small number of communities where the missionaries had little influence and infanticide still happens, it is still most-supported when carried out against the child born after twins.
This is all extremely speculative and I have almost no evidence for this and one of my sources is a century-old anthropology paper that talks about ‘the negro strain’. But as far as I can tell, I am literally the first person writing in English to ask these questions about baby-born-after-twins infanticide and propose a theory. This is not good because I know so little about Nigeria and have never even been to the country. I cannot begin to imagine the depths of wrongness I might be plumbing here.
I would love to hear more about this topic from someone who actually knows what they’re talking about! So I tried asking ActionAid about their polling. I obviously didn’t offer them my pet theory to comment on, I just said I was a potential donor who wanted to understand their work, and was confused by the divergence between the reporting (focussed on twin infanticide) and the polling results (which showed significantly higher support for baby-born-after-twins infanticide).
I started out by reaching out to ActionAid’s UK branch, because the reporting was in a British newspaper and I’m in Britain. They seemed completely unable to comprehend what I was asking about. The first person I spoke to told me to speak to the press team, but then the press team were confused that I wasn’t a journalist. More than that, they didn’t seem familiar with the reporting about their charity’s work that had appeared in a major British newspaper that week. Nonetheless, they said they’d ask ActionAid’s Nigerian branch about my inquiry. I didn’t quite trust them to represent my questions accurately, so I sent my own email to ActonAid Nigeria; I have heard nothing at all since from either the British or Nigerian branches, and it has been several weeks.
It really seems that the charity just has no model whatsoever for what to do when a potential donor starts asking questions about their reports. But then, why bother making the reports public at all, if not to make donors aware of the work they’re doing? I remember hearing one of the founders of GiveWell somewhere say that the thing that shocked them the most when they were starting out was how few charities were able to answer really basic questions relevant to their effectiveness, and in particular how few of them had any capacity to answer these questions when asked by donors rather than major institutional funders. Still, even though I’d heard this before, it still kind of shocked me to come face-to-face with it. Presumably the whole point of doing this research, publishing it, and speaking to a Guardian journalist about it was because they thought it would convince donors; so why are they unable to do the work of trying to convince donors?
Witches and witch-doctors
The ideas described in the Guardian piece—there are ‘evil’ members of the community who will cause harm by magical or supernatural means, and this requires the community to take steps up to and including killing the ‘evil’ person—rhyme very heavily with Ronald Hutton’s book The Witch, which I read in February.
One of Hutton’s concerns is with whether or not it’s reasonable to use terms from European culture, like ‘witch’, to describe cultures radically different from our own. For comparison, Hutton discusses the term ‘shaman’, which refers to a specific social and religious role in the context of indigenous Siberian culture. Divorced from that context, it has just come to mean ‘somebody who does some kind of magic or has some kind of mystical experience or something like that’ in most writing, and this is pointlessly vague.
But Hutton thinks ‘witch’ is different. He defines the term very precisely, in terms of five conditions; he then shows that belief in witches so defined is widespread across the entire world: people fear witches on every continent,8 in hunter-gatherer societies, agricultural societies, and even industrial societies. The concept of ‘witch’, unlike that of ‘shaman’, is pretty much a cultural universal.
I say ‘pretty much’: there are obviously exceptions. Ancient Egyptians didn’t believe in witches; pre-modern Siberians didn’t believe in them either; modern North Africans sometimes say they believe in witches but in practice they don’t appear too worried; Ireland and the ‘Celtic fringe’ in Britain seemed pretty reluctant to hunt witches even at the peak of the European witch-hunts. And of course European culture is now free of witch-belief.
But given the really precise similarities in witch-belief across cultures, and the fact that there are only a tiny handful of exceptions, I think Hutton takes the right approach: we should treat fear of witches as a universal, and then analyse the exceptions as exceptions, thinking about them as departures from an overwhelmingly-common norm. There’s a lesson there for understanding cultural universals more generally. It is, of course, extremely depressing that literal witch-hunting is a human cultural universal.
Tragically, Hutton notes that some societies believe that witches are usually children, and act accordingly. I actually wrote to him to ask if he knew any more about Nigerian witch-belief, to see if maybe he knew whether the idea of ‘evil children’ discussed above fell into his tightly-defined category of ‘witch’; he responded politely but said he didn’t know enough to help me. But certainly, in the book he emphasises that witch-hunting is a big craze across sub-saharan Africa today.
He blames colonial-era Christian missionaries for this, in a really interesting way. To be sure, missionaries usually told people that witches didn’t exist; but this claim is straightforwardly inconsistent with Christian scripture, and African Christians who picked up the Bible would have found an unambiguous instruction to kill witches. Meanwhile, the missionaries also told natives that other entities they had previously believed in, like ancestor spirits or land spirits etc., didn’t exist; and on this point Christianity was entirely consistent. There are other categories of malevolent entities in Christianity, like demons or fallen angels or ‘ha-satan’; but these had little overlap with pre-Christian beliefs in Africa, whereas witches were universal.
Historically, sub-saharan Africans would have sometimes blamed misfortunes on witches, but also often on land spirits or ancestor spirits or trickster gods or etc. Yet newly-Christianised people could only find scriptural justification for one of these, and so they ended up blaming it all on witches. The white missionaries probably would have preferred them to blame it on Satan, or just to accept the hardship as a test of faith; but this kind of thing wasn’t familiar, whereas witch-hunting was. So although witch-belief definitely predated colonial missions (and definitely still exists among Islamic or otherwise non-Christian cultures in sub-saharan Africa), Christian missionaries accidentally intensified it, and made the fear of witches a much bigger part of many African cultures.
There’s a if-you-don’t-laugh-you’ll-cry story about the 1996 Ralushai Commission report from South Africa. This Commission was set up to try to tackle a huge uptick in witch-hunting in Limpopo after the end of apartheid. Those appointed to the commission came from the most Westernised and elite segment of the province’s native population (think: judges and academics); but witch-belief was so widespread in Limpopo that even these people believed in witches! And so their recommendation to the government was: extra-judicial witch-hunting is inefficient at accurately identifying witches, and it can create public disorder, so witchcraft should be officially criminalised and witch-hunting moved into the courts.
This recommendation was not accepted, thank God—but not necessarily for the reasons you’d expect. South Africa, like most societies with witch-belief (including early modern Europe), has a corresponding social role of the witch-doctor: a helpful magic user who will identify witches and protect you against them, perhaps for a fee.9 So the obvious standard of evidence to use for witch-trials would require testimony from witch-doctors. But in order for witch-doctors to serve as expert witnesses, there needed to be a regulatory regime that could distinguish a ‘real’ witch-doctor from a ‘fake’ one. And the problem of setting up occupational licensing for witch-doctors was so thorny and difficult that even the proponents of a witch-hunting law eventually mostly abandoned the idea.
There was also the problem of how to de facto exempt white South Africans, who (after all) don’t believe in witches, while somehow respecting the post-apartheid taboo against racialised legislation. Lest I get racists in my comments who take the lesson of this story to be ‘black people are savage witch-hunters and white people are civilised and rational’, I’ll mention that Hutton briefly but very persuasively argues that the 1980s ‘Satanic panic’ in the US was a literal witch-hunt, not just a metaphorical one: the figure of the ‘ritual Satanic sexual abuser’ meets all the criteria Hutton uses to define ‘witch’. The people involved in the Satanic panic didn’t use the word ‘witch’, of course (in post–The Crucible America that word had way too many connotations), but that was a mere rebrand. You even saw a new class of witch-doctors, the so-called ‘recovered memory therapists’, emerging in tandem with the witch-hunt; and unlike black South Africans, white Americans didn’t realise that letting witch-doctors act as expert witnesses in court might cause a problem until after the problem had already emerged.
Speech is scary!
You might well know the story about megafauna extinction. In the period between ~50000 B. C. E and ~9000 B. C. E., a bunch of animals went extinct: in particular, many big and evolutionary successful animals like wooly mammoths or sabre-tooth tigers died, and other heavy animals like hippos were made extinct everywhere outside Africa. Indeed, distance from Africa is a pretty good proxy for how well megafauna fared: African megafauna largely (though not entirely) survived; European and Asian megafauna gradually died away; American and Austronesian megafauna were wiped out extremely quickly.
What happened? The most effective animal predators in history emerged in Africa, and then spread outwards. Upon arriving at a new location, these predator apes rationally focussed their hunting energy on the biggest animals, which have the best ratio of calories expended in the hunt vs calories received in meat.
Now, the whole time that these apes were evolving, African megafauna were exposed to them; this meant they were able to develop protective strategies before the apes became so smart that they achieved complete evolutionary dominance. Asian and European megafauna had some exposure to smart primates and so had some protection, but ultimately they couldn’t match these extremely efficient hunters. And in the Americas and Austronesia, megafauna species stood no chance and these new apes eliminated them in a frighteningly fast space of time.
I knew this story, but I never looked into what exactly African animals did right to escape destruction at the hands of these new apes, I just accepted that they figured something out. It turns out that there are lots of different useful strategies. But one of the most obvious-in-hindsight but also smartest adaptations is that African animals are terrified of language.
The scientists behind this study hid a bunch of video cameras and speakers in a South African national park, and then played a variety of sounds: lions (another African apex predator), dogs, gunshots, birds (as a control), and finally speech. Lions, guns, and dogs freaked out the animals in the national park much more than birds, as you’d expect. But the animals freaked out even more at the simple sound of conversation. You can watch some videos if you follow the link, the reaction is crazy.
The same effect has been demonstrated outside of Africa: many of the same scientists worked on this study, which shows similar results for badgers in Britain. The evolutionary rationale is obvious: if badgers hadn’t developed this fear, they’d not still be around; and they were lucky enough to have a period of time to evolve it while humans in Britain were focussed on hunting megafauna. The same evolutionary effect still exists today: this study found that when opossums live closer to humans, they become more scared of human noises rather than habituating to them.10
After all, speech is the tool by which the most effective predator in history actually achieves effectiveness. If Homo species couldn’t coordinate and plan hunts involving a large number of hunters, they wouldn’t be much different from other apes, but speech takes them to the next level. In a pretty objective sense, then, speech is utterly terrifying.
I think this is absolutely fascinating. I’ve only ever had one opportunity to observe the development of an African species ‘up close’: the juvenile ape who I’ve been keeping in my house for the past few months. And this specimen loves language! Indeed, exposure to speech can actually reduce the level of distress the specimen exhibits, the complete opposite of the behaviour of other African animals.
Obviously, the evolutionary incentives for my little ape point towards love of language. The evolutionary incentives for most other animals point in literally the opposite direction. I wonder how things are for the intermediate case: other great ape species? On the one hand, gorillas and chimps were in Africa long before the genus Homo emerged, and had time to adapt to our hunting abilities, so maybe they developed the same strategies as other African mammals. On the other hand, they seem likely to have mental modules that (in a different evolutionary lineage) developed into speech and language, ‘proto-speech’ modules (though extremely ‘proto’). Maybe that cognitive similarity to the new speaking apes meant they didn’t learn to fear speech? This is all pure speculation, I don’t think there’s any studies akin to the ones described above that included great apes, but I’d love to know.
Poetry-as-repetition, novels as novelty
Anyway: yeah, my little guy really likes hearing speech or singing. Nothing unusual there. It’s best to actually improv fake conversations, which is a huge amount of fun, but when he’s distressed and I’m exhausted, that’s a bit of a tall order. You need something ready-to-order to say or sing to him.
The default cultural solution is nursery rhymes. But these are terrible for the purpose: they’re far too short! Seriously, are you telling me that an extremely distressed and loudly crying baby can possibly calm all the way down before you are done singing the four lines of ‘Baa, baa, black sheep’? (There actually seems to be a serious failure of Burkean cultural evolution here.) You can repeat nursery rhymes over and over, of course, but that gets boring real fast.
Do you know what doesn’t get boring fast when you repeat it over and over, though? ‘On First Looking Into Chapman’s Homer’.That’s been my default: I know it by heart which means I can go through it without much cognitive effort, and it’s got strong rhythm and rhymes so even if I trip up briefly I can usually find my way through to the end. And, of course, it’s immensely satisfying when you’ve actually got him to sleep and you put him down just as you say ‘Silent, upon a peak in Darien.’ So any time Ernie needs comforting, I sit him on my knee and off we go to the realms of gold.
I also experimented a bit with ‘When I Consider How My Light is Spent’, after a shopping trip where we saw a brand called ‘Milton’ and I started chatting away with Ernie about the man.11 I eventually decided that it was a bit too Jesus-y for a long-term refrain, but it was great for a period of time. We did ‘Among School Children’ for a bit too, though it’s a little long so it’s easier to lose your place in it. (I always remember the general structure, it starts with the Maud Gonne bits and then goes onto spooky occult stuff about creativity; but when you’re running on fumes at two in the morning it’s easy to lose track of the specifics and skip stanza 3.) I’ve tried a few other poems: my general recommendation is that rhymes are a plus, and metre is absolutely necessary—free verse becomes impossible to remember during a night wake.
Repeating these poems over and over and over again has not put me off them. Quite the opposite, I’ve begun to appreciate them even more. It might sound stupid, but it was only by repeating it so much that I noticed how specifically Reformed ‘When I Consider’ is, rather than just generically Christian. (All that stuff about the sovereignty of God.)
I’ve heard from many parents that once kids are old enough to speak, they want you to read the same books to them over and over and over again. Lots of parents find this strange or even frustrating, but I increasingly see where the kids are coming from. Literature that has strong rhythmic patterns and momentum (like most books for young children) is ideal for memorisation and repeated recitation, and when the literature is actually good this is super enjoyable. Parents tend to have more critical acumen than their kids, so they’re frustrated that the books their kids want to read aren’t all that good. But this is different from being frustrated by repetition itself, even if it’s easy to confuse the two. My experience over the last little while has shown that there definitely are works that you can hear repeated endlessly and still enjoy, or even come to appreciate more.
It wouldn’t be completely crazy to try and define poetry as ‘literature that you can repeat over and over again and still like’. There’s a reason that the main non-poetic form in literature is literally called the novel!
In January I read Possession by A. S. Byatt, and it reminded me of the degree to which many people in the nineteenth century spent their time reading poetry out loud at each other, often poetry they already knew well and were just revisiting. Even in a literate age where quiet novel-reading was definitely a real practice (as per that famous scene in Pride and Prejudice), hearing poetry read out loud was the way to go. And in many premodern societies, literature was a small number of poems recited out loud. (Ion could only recite Homer.)
Probably our desire for novelty, which the novel is definitionally well-suited to provide, is a historical aberration. Kids, who have yet to be socialised into the dominant literary culture, nonetheless have a natural inbuilt desire for literature: but it’s specifically a desire for rhythmic literature, repeated over and over and over and over. Without abandoning the novel or anything crazy like that,12 maybe we should all try to reinhabit that childish mindset from time to time, and regain the childish desires, at least to some degree.
Is quantum field theory modern-day phlogiston?
Quantum field theory has resulted in the most accurate predictions in the entire history of science, it is clearly getting something right. But having spent time in February trying to make my knowledge of QFT a bit more rigourous, I am increasingly unsure exactly what that something is.
Philosophers of science have a number of ‘classic’ case studies that they always refer to because they’re so neat and tidy and have a clear lesson. Probably this is bad and they should do more rigourous and large-scale studies of many historical episodes, rather than always relying on the same small-n not-randomly-sampled dataset, but hey. One of the most important is phlogiston theory.
This was an early modern attempt to explain what happens in combustion: the phlogiston theorists said that combustible substances contained an invisible fluid, ‘phlogiston’, and that combustion was caused by the emission of this fluid into the air. This is almost there, but gets things backwards. It’s not that phlogiston that was in the combustible material flows out into the air, but rather that oxygen in the air combines with the material. Phlogiston was kind of ‘negative oxygen’.
Being ‘almost there’ was actually good enough for phlogiston theorists to make really major scientific advances. They figured out that rusting was the same kind of thing as burning,13 they figured out that plants must absorb phlogiston (= give off oxygen), and one scientist even managed to ‘dephlogisticate’ air (= isolate oxygen). But it wasn’t quite good enough. In particular, phlogiston theorists couldn’t quite explain why emitting phlogiston sometimes seemed to cause substances to gain weight, and other times to lose weight. Contra a common myth, it’s not that the phlogiston theorists falsely insisted that substances must lose weight when they emit phlogiston, in spite of the data. Instead, their response to the weight-change data was ‘eh, idk’. After all, when things burn, almost everything about them changes: their colour, their texture, their hardness, their solidity, their reactivity with other elements, their nutritiousness to humans, etc. Why not their mass? The question of how exactly reactions affect mass was just outside the scope of phlogiston theory.
Phlogiston theory genuinely reflected something true and important about the structure of reality. But it also got some important things wrong. But the true bits and the false bits never came into conflict with each other, because phlogiston theorists had essentially ruled the difficult topics (especially mass) ‘out of scope’. This was worse than the theory just being wrong: it made it harder for chemists to figure out how it was wrong. Progress required someone like Antoine Lavoisier to think rigourously about exactly the topic (mass) that phlogiston theorists were teaching students to shrug their shoulders at. This is part of what Lavoisier meant when he said that acceptance of phlogiston in chemistry ‘retarded its progress considerably by introducing a bad style of philosophising’.
Anyway. Probably philosophers of physics aren’t willing to say this publicly, because the physicists would probably find it a bit rude, but I get the sense that many of them think that quantum field theory is a bit like a modern-day phlogiston theory.14 I confess I haven’t seen anything like survey data on this, but I have been reading quite a lot of the literature. The consensus position seems to be a bit ‘live and let live’, but if I’m right in getting the subterranean sense that some philosophers are trying to minimise confrontation with angry physicists, maybe the real majority position is much more hostile to quantum field theory. Certainly, philosophers who are willing to forthrightly defend the mainstream approach to QFT, as it is taught to physics students literally everywhere, are apparently exceptionally rare: David Wallace is the only one I know.
And as I learn more about QFT, I more and more get the sense that maybe the philosophers are onto something. Suppose you want to build a quantum theory of fields. Here are some problems you will run into:
The standard way to develop a quantum theory of some system is that you start by describing the system you want to figure out in purely ‘classical’ terms; and then you follow a rigourous procedure to ‘quantise’ your description, systematically applying a set of correspondences to replace classical objects with quantum ones. In the case of classical theories with only finitely many degrees of freedom, this works beautifully well: there are some interesting technicalities, but in essence Stone–von Neumann guarantees that there is only one way to ‘quantise’ any given classical system.15 But if you have a theory with infinitely many degrees of freedom—that is to say, a field theory—Stone–von Neumann no longer applies. There are multiple different incompatible ways to quantise any given classical field theory. So how do you pick? Physics texts tend to circumlocute at this point, but Wallace is admirably direct:
Stepping back from that problem. Even if you could quantise a classical system, you’d need a classical system to start with. How exactly do physicists know what classical system is most analogous to the quantum system they’re studying? Technically: how do they know what the Lagrangian is supposed to be? Again, in the case of ordinary quantum mechanics with no fields, this is easy enough—‘charged particle in a potential’ is a system that classical mechanics handles easily. But QFT deals with crazy stuff like quarks and gluons that we have no classical intuitions for. So physicists have to answer the question ‘which Lagrangian?’ differently; and for … reasons, they answer it with ‘whichever Lagrangian is invariant under the action of my favourite Lie group’. For example, the ‘standard model’—the current frontier of fundamental physics—has SU(3) × SU(2) × U(1) as its ‘gauge group’, and this determines the Lagrangian and thus the entire downstream structure of the theory. How do you pick which, out of the whole uncountably-infinite universe of Lie groups, is the right gauge group? Partly the answer is ‘experimental data’, but this only gets you so far. The actual answer, as far as I can tell, is ‘it would be easiest if it was a direct product of some of the classical groups, so let’s say that it is’.16
OK, let’s assume that your favourite Lie group really is baked into the fundamental structure of reality.17 Still, that doesn’t get you a unique Lagrangian! There are infinitely many possible Lagrangians that are invariant under the action of any given Lie group. Which one do you pick? In practice the answer has been ‘the simplest one, I’m not arsed to deal with anything more complicated’.18
Even assuming all of these problems away, even assuming that God himself came down from heaven and handed you the right Lie group and the right Lagrangian and the right quantisation and said ‘bro this is it, don’t worry how I came up with it, just trust me I’m literally God’, you still have to take the theory and do the calculations. In ordinary QM, this isn’t a problem, we teach undergrads how to do this all the time. But in QFT, if you try to calculate the answers… the integrals diverge lol. Physicists solve this by just
ignoring the problemchoosing not to consider physical effects on scales shorter than a certain ‘cutoff’ (usually ~10-10 m). If this method is applied then you can get numbers out of the theory. Notably this does not mean that QFT assumes fields do not vary on arbitrarily small scales: such ‘lattice theories’ exist, and they are not in general equivalent to QFT, even in the continuous limit. QFT rather assumes that the field is a perfect continuum, smoothly varying from point to point, all the way through the process of theory construction, right up until you have to do the calculations … at which point it declares ‘actually it’s discrete lmaooooooooooo’ and runs away laughing.In response to all of this, you might say (with David Deutsch) that this whole ‘quantisation’ business is nonsense from the get-go. Reality is fundamentally quantum, classical physics is just an approximation in the limit ℏ → 0, and so classical theories should be built from quantum theories—not the other way around. It should hardly be surprising that trying to invert this relationship results in problems. In ordinary quantum mechanics, this is no problem whatsoever: beginning with the brilliant work of von Neumann and continuing down to the present, the theory has become so mathematically well-understood that we can eschew quantisation entirely, and base quantum mechanics solely on a series of well-understood and minimal axioms. And this approach gives us a lot of insight into why, even though it’s philosophically improper, nonetheless quantisation ‘works’. But trying to apply the same ground-up approach to quantum field theory has overwhelmingly resulted in failure. Proving axiomatically that there is a consistent and non-trivial quantum field theory in 3D space is literally one of the million-dollar unsolved Millennium problems. And there are very strong reasons to expect that the field theories that are actually used by physicists simply cannot be placed on the kind of axiomatic footing that ordinary QM can be.
(4) is the most famous problem. But I think (1)–(3) are just as serious. At every stage where the question arises, ‘why would reality work like this rather than that?’, physicists have been taught to answer: ‘idk, but this is much easier mathematically so it’s what I’m going with’. This isn’t Occam’s razor, the resulting theories are not necessarily simpler in any objective sense, merely more convenient. As Wallace says, physicists are just assuming that Nature will be ‘kind’ to them. I think many philosophers are (rightly) inclined to say: ‘Why on earth would I assume nature will be kind? Nature is not your friend.’
In turn, an uncharitable—but not, for that reason, unreasonable—reaction to (5) would be to say ‘probably QFT is literally not mathematically consistent’. And if you try to apply a mathematically inconsistent theory, you can get literally any results you want.
Physicists insist that QFT ‘works’, it gets the right answers to the questions that physicists ask, and not because in theory it could output any answer: there’s something to the methods that physicists use that matches the structure of reality. But phlogiston theory ‘worked’: it really did get results, it really did contribute to scientific progress. And there really was something about it that matched the structure of reality.
But in order to ‘work’, phlogiston theory had to exclude certain questions from its scope. And this is the same as QFT. The problems, or even inconsistencies, in QFT can definitely be ‘tamed’: there are procedures physicists follow to prevent the seemingly-contradictory aspects of the theory from encountering each other, and therefore avoid making the theory completely unusable. But following these procedures necessarily comes at the cost of ruling certain questions out-of-scope.
This comes through very clearly in Wallace’s work: as the most prominent QFT defener19 in contemporary philosophy, I think he’s pretty well-placed to act as a stand-in for the best arguments in favour of QFT. He makes an analogy with the study of crystals. Crystals are often modelled as continuous media, despite the fact that we know crystals have atoms! But this is OK because we have mathematical assurances that the large-scale effects of atom-level interactions can be almost entirely accounted for with just a small handful of parameters. So rather than trying to model these interactions directly, we just estimate these parameters empirically. The continuum model will still sometimes spit out absurdities in cases where the fact that crystals aren’t continua are relevant, and we will have to pragmatically switch back to a discrete model in these cases, but that’s ok: we’re not looking for a fundamental theory, we’re just looking for a ‘good enough’ model of crystal behaviour, and that model can be messy and a bit inconsistent.
Wallace argues that we have very similar assurances that, in the context of quantum field theory, the interaction effects at scales smaller than ~10-10 m can be almost entirely accounted for using a small number of parameters. So long as we have good empirical estimates of these parameters, then he thinks it’s OK for QFT to do the discrete-continuous switcheroo described above: it’s like mostly working with a continuous model, and then pragmatically departing from it when you get into difficulties.
You will note that this presupposes that QFT cannot be a fundamental theory, and there must be some other theory describing interactions at scales smaller than ~10-10 m. Wallace is admirably explicit about this: QFT only makes sense because, whichever fundamental theory is correct, its cumulative divergence from quantum theory on scales >> 10-10 m can be expressed in terms of a small number of parameters; and so QFT practitioners don’t need to worry about where the values of these parameters come from, they can just take them as ‘given’ in the data. The tools physicists have adopted to make QFT self-consistent only work if you rule questions about fundamental interactions ‘out of bounds’. If you try to model these fundamental interactions within QFT, the whole theory falls apart; but, says Wallace, that’s fine, QFT is just trying to be a ‘good enough’ theory of a non-fundamental phenomenon.
The issue, unfortunately, is that QFT is the most fundamental theory we have. There is definitely a more fundamental theory out there, but we don’t actually know what it is. And the only methodology that has ever worked to develop better and more fundamental physical theories is to start from our current best theories, look at their anomalies and weirdnesses, and poke at them until patterns start emerging. Maxwell’s theory of electromagnetism is not a fundamental theory, but by treating the theory as fundamental—pushing it ‘to its limit’ and aggressively examining the seeming inconsistencies about light propagation / interchange between electric and magnetic fields / the nature of the aether / etc.—we got special relativity.
So I think it’s striking that the most prominent contemporary philosophical defender of QFT is the one who says that QFT ‘seals off’ large areas of questioning—that the very structure of the theory means you cannot use the framework of the theory itself to ask questions about what happens at arbitrarily small scales, or where exactly the gauge symmetries come from, or why the ‘most obvious’ quantisation choices tend to work even though they’re mathematically arbitrary, etc.
Lee Smolin, and others after him, have argued that fundamental physics has gotten in a bad way because it has de facto begun to use mathematical consistency as the main constraint on theorising; but this is barely a constraint at all, the landscape of possible mathematically consistent theories is literally infinite, and so search through this landscape has probability zero of returning the true fundamental theory. I have always felt there was something to this complaint. And if I take Wallace seriously, I think I can see the outlines of an explanation of why Smolin might be right: it’s not about scientific culture, it’s actually baked into the structure of QFT itself. If you can’t poke at the weirdnesses and anomalies in your existing theories, then you won’t really have any framework for making progress: all you can do is speculate, and the only possible constraint on the speculation is self-consistency.20 And over the years, physicists have adopted procedures to ‘wall off’ all possible weirdnesses and anomalies in QFT, to the point that if you try to ‘poke’ at them the whole theory falls apart and you end up with nonsensical divergent integrals and inequivalent quantisations of the same system etc. Wallace is right that walling off the troubling phenomena is ‘good enough’ if you want a pretty accurate model of a non-fundamental phenomenon; but if the most fundamental theory you have looks like this, you’re gonna be stuck.
I’ll quote Lavoisier again: perhaps the field has ‘retarded its progress considerably by introducing a bad style of philosophising’.
Sometimes, the magic of coordination and common knowledge allows sports to avoid explicit rule-changes. Racewalking is a prime example, where everyone knows that most competitors are breaking the rules regularly, and because everyone knows, nobody cares about the slow-mo footage that proves it (pace that one episode of Malcolm in the Middle). But this is essentially a rule change—just an unwritten rule change. While de iure the rule is still ‘you can’t allow both feet to leave the ground at the same time’, de facto it’s ‘you can do it so long as it’s not so blatant that it can be clearly seen with the naked eye’.
This isn’t how the fans themselves describe it, of course: ask the average football fan down the pub and they’ll say that VAR results in objectively bad decisions. It should be obvious to any neutral observer that this literally isn’t possible: refs are getting more and better information from VAR. What’s happening here is that fans aren’t particularly well-acquainted with the precise letter of the rules, and their expectations for how the rules should be enforced have been calibrated by watching actual refereeing decisions on the pitch—which often were ‘bad’ decisions if interpreted strictly to the letter of the rules, made only because the ref didn’t have enough info / had to make a decision quickly. Now that refs have more info and more time to reflect, they make decisions that are more in line with the letter of the rules, but this is out of step with the expectations of fans who’ve never sat down to read the Laws of the Game.
Oh sure, high reliability isn’t certainty; you can keep adding nines of reliability without ever reaching one; no amount of bureaucratic information-collection can solve the problem of other minds. But nobody in any other context demands that the state solve the problem of other minds before implementing an otherwise-good policy. I therefore take criticisms based on ‘won’t the state make mistakes?’ to be bad-faith special pleading from people whose real objection to assisted dying is either (a) nobody should be allowed an assisted death even if we knew with 100% certainty that they rationally wanted it, or (b) it is literally impossible to rationally wish to die, wishing to die is per se irrational.
I’m not even sure if he wants to? He agrees that life in preindustrial societies sucked, he just doesn’t think it counts as ‘poverty’. I argue below that this is a fake distinction.
Two qualifications to this are necessary, one pedantic, one substantive. The pedantic one is that Malthus’ own theory didn’t actually work like this, he didn’t think of ‘checks’ to population as operating naturally and smoothly, he was much more catastrophist and in particular providentialist (he thought God intervened to keep things running smoothly). So the term ‘Malthusian’ in this context is a misnomer. But I’m definitely fighting a losing battle by insisting on this, ‘Malthusian’ is far too embedded in the language.
More substantively, these ‘Malthusian’ forces operate over much longer timescales than people often assume: a population obviously can’t survive widespread famine, but it can persist for many generations with everyone getting somewhat too few calories, or having some degree of nutrient deficiency, or etc. Eventually these factors will show up in birth rates, especially if the society faces competitive pressures, but this can take a while—and other forces can act in the meantime. In his article Allen argues that British colonialism was such a force in India, which caused many people’s living standards to fall below his ‘basic needs’ threshold for many generations even after the British left. My parenthetical is just meant to be an intuition pump.
This story has mutated through the internet’s standard process of Chinese whispers, and people often present the passage I have placed in inverted commas (with ‘she’ replaced with ‘I’) as though it were a quote from Christie herself. In fact it is a paraphrase of a line in Christie’s biography, where she writes the following:
In describing my life I am struck by the way it sounds as though I and everybody else were extremely rich. Nowadays you certainly would have to be rich to do the same things, but in point of fact nearly all my friends came from homes of moderate income. Most of their parents did not have a carriage or horses, they certainly had not yet acquired the new automobile or motorcar. For that you did have to be rich. [Christie, Autobiography, part 4 ch. 2.]
To be sure, Christie’s ‘description of her life’ earlier in the autobiography certainly did mention having servants, so it’s a fair paraphrase and nobody’s being misled. I just like to take the opportunity to correct misquotations.
The actual person who wrote the popular paraphrase is unclear. It first appears on the internet on Brad DeLong’s old blog; DeLong said it was in his email inbox, but not who actually sent it to him.
The idea that this belief is relevant to infanticide practices is explicitly criticised in this iconoclastic paper, which (more generally) is sceptical of the very idea of twin infanticide: the authors argue that twins are more commonly killed because twins happen to be more likely to fall into a category of children that their society kills anyway (disabled, low birth weight, etc.); they’re not killed because they are twins. I think a weak version of this position can be made compatible with the ActionAid data, but only a very weak version: clearly there’s something special about twin births in the minds of people in these infanticide-practicing communities.
Except maybe Antarctica, obviously. Though who knows what the researchers down there have nightmares about?
In a classic case of what you might call ‘politically correct terminology shifting’, anthropologists today deprecate the term ‘witch-doctor’, but they are happy with terms like ‘traditional religious leader’ (used in the Guardian piece about infanticide) or ‘service magicians’ (Hutton’s term) that apply to exactly the same people as ‘witch-doctor’; the conceptual structure remains unchanged. Compare American progressives swapping out ‘homeless’ for ‘unhoused’ without making any other changes. ‘Witch-doctor’, it is worth noting, is not an inherently judgmental term: indeed, it involves a comparison with a respected and high-status job (that of a doctor). Of course, it comes to sound judgmental if the speaker also believes that the ‘disease’ (viz. being bewitched) is fake. But in that case, we should be judgmental! People who offer cures for nonexistent diseases are bad, that’s fraudulent and harmful! I know that anthropologists have to suspend these kinds of judgments for the purposes of fieldwork; but if anthropologists can successfully suspend their disbelief in witches, then I see no reason why they can’t also suspend the judgmentalness associated with the term ‘witch-doctor’.
The study also looked at skunks and bobcats, and there wasn’t significant evidence that these species became more scared of humans in close proximity. But importantly, they definitely didn’t habituate and become less scared of humans in proximity.
Babies are great in large part because you have a captive audience and can chat total shite to them. ‘Aye, so Ernie, Milton was this fella back in olden times, and he fucking hated bishops. Aye, bishops, the lads with the big hats! Anyway, he hated bishops so much he made friends with Oliver Cromwell. But we forgive him because he was also very good at poems. Melvyn will tell us all about him when we get home.’ And indeed, when we got home I stuck on the In Our Time on Milton.
After all, many historical aberrations are good. The lesson from Hutton (above) is that ‘not believing in witches’ is a historical aberration.
Arguably John Mayow got there first, but his theory grouped together rusting, burning, and a whole bunch of other phenomena that don’t actually have anything to do with each other; being charitable, he had a genuine insight and got carried away; being uncharitable, he got this one connection right mostly by luck. The first person who definitively connected rusting and burning in the ‘right sort of way’ was Johann Becher, the first phlogiston theorist (who albeit used the name ‘terra pinguis’ rather than ‘phlogiston’ but it was the same thing).
More precisely: quantisation is unique up to isomorphism. The tricky technicalities come down to the fact that the relevant isomorphisms don’t compose nicely, and so quantisation isn’t a functor (in the category-theoretic sense) from classical systems to quantum systems.
This might seem unfair, but I really don’t think it is. Usually it is insisted that the Lie group has to be (semi-)simple and compact, which narrows you down to the classical groups plus a few exceptions. There are indeed a bunch of weird mathematical artefacts that arise if the Lie group isn’t compact. But I’ve nowhere seen it rigourously proven that these artefacts fall into the ‘literally inconsistent with experience’ bucket rather than the ‘it would be mathematically awkward to figure out how to contain these’ bucket. (I haven’t read it but this paper looks relevant.) I think very few physicists have actually thought this through from first principles because it’s just easiest if you only have to deal with the classical groups, and ‘it has to be compact’ is a useful thing to say in order to get the awkward philosophers off your back.
I mean, actually, given what physicists usually say about gauge symmetries, it seems that they don’t think that the gauge groups are fundamentally baked into the structure of reality, they actually represent exactly those bits of the theory that don’t correspond to the structure of reality. But yet the groups play an essential role in theory construction. For some reason. This is philosophically extremely hard to understand: you may find this SEP article just as frustrating as I did.
Sometimes physicists instead claim ‘actually it doesn’t matter for getting empirical results’, because the so-called ‘non-renormalisable’ terms that differ between different possible Lagrangians only matter at small scales, below the ‘cutoff’ scales used for calculations (discussed below). I suspect this answer is literally incoherent, as discussed below. But even if it’s correct, it doesn’t change the fact that our fundamental theory of physics is underdetermined.
I’ve been corrected on this spelling before, so let me just say: sic erat scriptum.
Well, consistency with the data is also an important constraint. But if Wallace is right, then in practice this doesn’t matter in the QFT case: we’d get the same approximation whatever the true fundamental theory is, so long as it produced the right parameter values; and so our only information is contained in a small handful of parameters.
Erm if it looks like an alien, and it builds alien pyramids, then it's probably an...