Insights

In 2006, a judge in Worcester, Massachusetts was asked to rule on one of the oldest questions in food categorisation: is a burrito a sandwich? The case arose from a dispute between a shopping centre tenant and a Panera Bread franchise. Panera had a lease agreement that prevented the mall from renting to another sandwich shop. The mall then signed a lease with a Qdoba Mexican restaurant. Panera argued that Qdoba sold sandwiches — burritos being, in their view, a form of sandwich. The judge disagreed. A sandwich, he ruled, requires two separate pieces of bread. A burrito, wrapped in a single tortilla, does not qualify.

The ruling was both legally significant and philosophically unsatisfying — because the question of what constitutes a sandwich has never been simple, and the debate is far older than American tort law.

THE EARL WHO INVENTED NOTHING

The word "sandwich" derives from John Montagu, the 4th Earl of Sandwich, who according to popular legend requested meat tucked between two slices of bread in 1762 so he could eat without leaving the gambling table. The story — first recorded by Pierre-Jean Grosley in his 1772 travelogue of London — is almost certainly embellished. Bread had been used as an edible vessel for other foods across many cultures for thousands of years before Montagu was born. Ancient Roman soldiers ate flatbreads topped with olive oil and cheese. Jewish Passover tradition includes the Hillel sandwich, a practice dating to the first century BC in which bitter herbs and lamb were placed between pieces of matzo.

What Montagu actually did was lend his name and aristocratic cachet to a format that already existed — and in doing so, unwittingly launched a definitional debate that courts are still being asked to adjudicate 260 years later. Because once you have a named category, you have the problem of what belongs in it.

THE TAXONOMY PROBLEM

Food taxonomies are notoriously contested. The USDA classifies a hot dog as a sandwich. The New York State Department of Taxation considers a burrito a sandwich for tax purposes. The British Sandwich Association — yes, this is a real organisation, founded in 1990 — defines a sandwich as "any combination of bread, rolls, or other bakery products filled with either savory or sweet fillings." Under this definition, a wrap is a sandwich. A burger is a sandwich. A hot dog, if served in a bun, is a sandwich.

SplitTheWorld asked the hot dog question directly: "A hot dog is a sandwich." The split was 34% in favour and 66% against — a decisive majority rejecting the sandwich classification, despite the fact that most formal definitions would include it.

This result illustrates something fundamental about how people categorise food: cultural identity overrides formal definition. A hot dog is not experienced as a sandwich by most people who eat them, regardless of what the USDA says. The cultural context — the ballpark, the condiment ritual, the specific bun that cradles rather than sandwiches — creates a categorical identity that sits outside the formal taxonomy. Definitions are made by committees. Categories are made by culture. They are not always the same thing.

WHY THE DEBATE WILL NEVER END

The sandwich debate persists because it is not really about sandwiches. It is about the human need to categorise, and the equally human resistance to having that categorisation imposed from outside. When a court or a food standards agency defines what a sandwich is, they are asserting authority over a domain — what we eat, what we call it, how we think about it — that most people consider deeply personal. The pushback is instinctive.

Every generation rediscovers the same debate in a new form. In the 1990s it was whether a hot dog counted. In the 2000s it was the burrito ruling. In the 2010s it was whether a calzone was a sandwich (it is not — it is a folded pizza, a separate and equally contentious category). The internet has ensured that each of these debates now plays out at global scale, with millions of participants and no possibility of a final verdict that everyone accepts. Which, in a way, is exactly as it should be.

In 1917, Sigmund Freud introduced a concept he called the "narcissism of small differences" — the observation that communities which are close to each other in geography, culture, or belief tend to harbour the most intense mutual hostility. Neighbouring towns. Adjacent religious sects. Political parties separated by narrow policy differences. The closer two groups are, the more fiercely they tend to defend the distinctions that separate them.

Freud was writing about ethnic and national tensions, but the principle scales down with remarkable fidelity to the pineapple pizza debate. The reason people argue so passionately about whether fruit belongs on pizza is not that pizza is important. It is that the question touches something that feels identity-defining — taste, culture, culinary values — in a domain low-stakes enough to argue about without real social risk.

THE SOCIAL FUNCTION OF TRIVIAL ARGUMENTS

Anthropologists and social psychologists have long noted that humans use trivial disagreements as a form of social bonding. Arguing about whether Die Hard is a Christmas movie, whether the toilet paper goes over or under, or whether you should ever recline your airplane seat serves a social function that has nothing to do with the actual outcome of the debate. These arguments are a form of play — a way of testing and establishing group membership, demonstrating wit, and experiencing the pleasure of strong opinion in a context where nobody gets genuinely hurt.

Research published in social psychology literature on this phenomenon suggests that low-stakes debates activate many of the same cognitive and emotional mechanisms as high-stakes ones — the sense of being right, the frustration of being challenged, the satisfaction of a well-made argument — while carrying none of the real-world consequences. They are, in essence, practice arguments. Safe spaces for the exercise of conviction.

This is why the questions that generate the most engagement on SplitTheWorld are rarely the philosophical or political ones. Questions about cold pizza, toilet paper orientation, or whether a diagonally cut sandwich tastes better than a straight-cut one get shared and debated far more widely than questions about nuclear energy or the nature of free will. The lower the stakes, the more freely people engage.

THE HISTORY OF TRIVIAL DEBATE AS SOCIAL RITUAL

The use of trivial debate as social ritual is ancient. Ancient Greek symposia — the drinking gatherings of the educated elite — regularly featured structured arguments about absurd hypothetical questions as an intellectual warm-up for more serious philosophical discussion. Medieval scholars engaged in formal debates about questions that strike us as comically pointless today: how many angels can dance on the head of a pin being the most famous, though it is worth noting that this question was not actually as trivial as it sounds — it was a proxy debate about the nature of physical substance and spiritual existence.

The Victorian era produced a remarkable flowering of trivial debate culture, particularly in the form of parlour games that involved arguing for predetermined positions regardless of personal belief. These games — direct ancestors of modern debate competitions — trained participants to construct arguments for any position, building the intellectual flexibility that serious discourse requires.

The internet has democratised this tradition. What was once the preserve of educated parlours and university common rooms is now conducted by millions of people simultaneously, in comment sections and group chats, with the added twist of real-time vote counts that tell you how many people are on your side. SplitTheWorld is part of this tradition — a structured arena for the ancient human pleasure of having a strong opinion and finding out who agrees.

WHEN TRIVIAL ARGUMENTS REVEAL NON-TRIVIAL THINGS

The most interesting thing about low-stakes debates is what they sometimes accidentally reveal. A question about whether you should always finish a book you are not enjoying is, on the surface, trivial. But the split on that question — and the reasoning people give for their answers — reveals a great deal about how people think about commitment, self-discipline, sunk cost reasoning, and the value of their own time. The question is a vehicle for something deeper.

Psychologists who study self-concept use exactly this kind of indirect questioning. You learn more about a person's relationship with authority by asking how they feel about toilet paper orientation than by asking them directly about their views on rules and compliance. The trivial question disarms the self-monitoring that more obviously serious questions trigger, and lets something genuine come through.

This is, in a quiet way, what SplitTheWorld is actually measuring — not just preferences, but the underlying values and experiences that shape them. The splits are data. But what they are data about is more interesting than the questions suggest.

The first known opinion poll in American history was conducted in 1824, when the Harrisburg Pennsylvanian newspaper sent reporters into the crowd at a political rally to ask people who they intended to vote for in the presidential election. Andrew Jackson won the straw vote. He also won the popular vote in the actual election, though he lost the presidency to John Quincy Adams through the Electoral College. The poll was rough, unscientific, and based on a convenience sample of whoever happened to be at the rally. But it established a template for measuring public opinion that has been refined and contested ever since.

THE LITERARY DIGEST DISASTER

For over a century, the dominant method of opinion polling was the straw vote — large-scale, self-selected surveys that accumulated responses from whoever was willing to reply. The Literary Digest magazine perfected this approach in the 1920s and 1930s, mailing millions of postcard ballots to households across America and tabulating the responses. In 1932, it correctly predicted Franklin Roosevelt's landslide victory over Herbert Hoover. In 1936, it predicted that Alf Landon would defeat Roosevelt in a decisive victory — and was spectacularly, career-endingly wrong. Roosevelt won in one of the largest landslides in American electoral history.

The Digest had surveyed ten million people. The problem was not sample size. The problem was selection bias. The Digest drew its mailing list from telephone directories and automobile registrations — a methodology that had worked reasonably well in previous elections but, in the depths of the Great Depression, systematically oversampled wealthier Americans who were more likely to own phones and cars. Poorer Americans, who voted overwhelmingly for Roosevelt, were largely absent from the survey.

The Digest folded within months of the debacle. But the lesson it taught — that a biased sample is worse than a small sample — shaped every serious opinion research methodology that followed.

THE GALLUP REVOLUTION

George Gallup had been conducting polls for advertising research throughout the early 1930s, and had developed a rigorous sampling methodology based on demographic quotas rather than raw volume. In 1936, the same year the Digest failed so spectacularly, Gallup correctly predicted Roosevelt's victory — using a sample of just 50,000 people, a tiny fraction of the Digest's ten million. More importantly, he had published a prediction of how the Digest itself would get it wrong, weeks before the election results came in.

The Gallup poll established the modern template for scientific opinion research: a carefully constructed representative sample, standardised question wording to minimise leading language, and statistical analysis to calculate margins of error. These principles have been refined across nine decades of polling practice, but the fundamentals Gallup established remain the foundation of serious opinion research today.

THE DEWEY DEFEATS TRUMAN MOMENT

Even rigorous polling gets it spectacularly wrong sometimes. The 1948 US presidential election produced the most famous polling failure in history. Every major polling organisation — Gallup, Roper, and Crossley — predicted a comfortable victory for Republican Thomas Dewey over incumbent Harry Truman. The Chicago Tribune was so confident in the predictions that it printed its famous "Dewey Defeats Truman" headline before the results were in. Truman won by over two million popular votes.

The failure had two causes. First, the major pollsters had stopped collecting data several weeks before the election, missing a late shift in voter sentiment. Second, the "undecided" voters in their samples — who the pollsters had distributed proportionally between the candidates — broke heavily for Truman on election day. Both failures were methodological: insufficient attention to late-breaking changes, and poor handling of uncertain respondents.

The lesson was that a snapshot of opinion at a single moment in time is only valid at that moment. Public opinion moves. The polls that track it need to move with it — which is why SplitTheWorld's running vote count, updated in real time throughout a question's active week, gives a more honest picture of opinion dynamics than any single-point survey could.

ONLINE POLLING AND ITS LIMITS

The internet created a new class of opinion data that is cheaper, faster, and more plentiful than anything Gallup could have imagined — and simultaneously more vulnerable to the selection biases that destroyed the Literary Digest. Online polls draw from whoever happens to visit a website, click a link, or follow a social media account. These samples are self-selected, non-representative, and easily manipulated by coordinated groups who share poll links to push a particular result.

SplitTheWorld is not a scientific poll and does not claim to be. It is a self-selected sample of people who found their way to the site — a group that skews toward whatever demographics share the site most actively. The numbers are real, but they are not a representative sample of global opinion. They are a record of how SplitTheWorld's specific audience answered — which is interesting and sometimes surprising, but should be understood for what it is.

What SplitTheWorld does differently from most online polls is the answer-first mechanic. By preventing respondents from seeing results before they vote, it eliminates the bandwagon bias that affects nearly every online poll that shows live results. The numbers you see after you vote reflect genuine, unanchored responses — which makes them meaningfully more honest than a poll that tells you 73% picked option A before you make your choice.

In the early nineteenth century, it was not unusual for industrial workers in Britain and America to work sixteen hours a day, six days a week. Children as young as seven worked in cotton mills and coal mines. The concept of a "weekend" did not exist. The concept of a "work-life balance" would have been completely incomprehensible to the average factory worker of 1820 — not because they lacked imagination, but because the economic and social structures of the time simply did not allow for it.

The two centuries since have seen one of the most consequential and bitterly contested shifts in the history of human labour: the gradual, uneven, still-incomplete reduction of the working week. It is a story with no clean ending — as the SplitTheWorld result showing 81% support for a four-day work week makes clear. The debate is not settled. It is simply at a different stage.

ROBERT OWEN AND THE TEN-HOUR MOVEMENT

The first serious organised resistance to industrial working hours came from Welsh textile manufacturer and social reformer Robert Owen. In 1817, Owen coined the slogan "Eight hours labour, eight hours recreation, eight hours rest" — a division of the day that would have seemed utopian to most working people of the era. Owen had already implemented shorter working hours at his New Lanark cotton mills in Scotland and documented the results: productivity held steady, worker health improved, and the community surrounding the mill was visibly more prosperous and stable than comparable industrial towns.

Owen's experiment was influential but not immediately imitated. The dominant view among industrialists and many economists of the era was that shorter working hours would inevitably mean lower output, higher costs, and competitive disadvantage. This argument — essentially unchanged in its structure — remains the primary objection to the four-day work week today, more than two centuries after Owen first challenged it.

HENRY FORD'S COUNTERINTUITIVE DISCOVERY

The single most important experiment in the history of working hours was conducted not by a reformer or a politician, but by one of the most commercially minded industrialists in American history. In 1926, Henry Ford reduced the working week at Ford Motor Company from six days to five, and capped the working day at eight hours. The move was widely predicted to reduce output and increase costs. The opposite happened.

Ford's own analysis, published in the following years, concluded that worker productivity per hour increased enough to compensate for the reduced hours — and that the additional free time workers now had encouraged them to spend money on consumer goods, including automobiles, stimulating the broader economy. Ford framed the five-day week not as a concession to workers but as a sound business decision. His competitors, initially sceptical, gradually adopted the same schedule as the evidence accumulated.

The five-day, forty-hour working week became the legal standard in the United States under the Fair Labor Standards Act of 1938. Most other industrialised nations followed over the subsequent decades. The standard that now feels permanent and natural was, within living memory, a radical innovation bitterly opposed by the majority of employers.

THE FOUR-DAY WEEK: FORD'S ARGUMENT, APPLIED AGAIN

The contemporary four-day work week movement is, in structural terms, almost identical to the arguments Ford made in 1926. Productivity per hour increases when workers are less fatigued and more motivated. Reduced hours attract and retain better talent. The economic activity generated by workers who have more free time offsets the apparent loss of one working day. The employers who resist the change cite concerns that, historically, have not been borne out.

The evidence from recent trials is consistent with this analysis. Iceland's large-scale trial of a 35-to-36-hour working week, conducted between 2015 and 2019 and involving around 2,500 workers — roughly one percent of the country's entire working population — found that productivity either held steady or improved in the vast majority of workplaces. Worker wellbeing improved significantly. The trial was so successful that the majority of Iceland's workforce subsequently moved to shorter working hours through renegotiated contracts and agreements.

SplitTheWorld's 81% majority in favour of the four-day week reflects an intuition that is well-supported by the available evidence. The 19% who voted against are not necessarily wrong — there are industries and roles where compressed hours create genuine operational difficulties. But the historical precedent is clear: every previous reduction in working hours was resisted on the same grounds, and those grounds turned out to be less solid than claimed. The question is not whether the four-day week will eventually become standard. It is how long it will take.

In the 1990s, geneticist Linda Bartoshuk at Yale University identified a group of people she called "supertasters" — individuals with a significantly higher density of taste receptor cells on their tongues, who experience flavour with a measurably greater intensity than average. Supertasters find bitter compounds, in particular, overwhelmingly unpleasant. Broccoli tastes vile to them. Black coffee is almost undrinkable. Alcohol is harshly bitter rather than pleasantly warming. They are not being precious or difficult. Their physiology is genuinely different.

Bartoshuk estimated that around 25% of the population are supertasters, 50% are medium tasters, and 25% are non-tasters who experience relatively muted flavour sensations. This single biological variable goes a long way toward explaining why food debates are so intractable: in the most literal possible sense, different people are not tasting the same thing.

THE CILANTRO GENE

The most well-documented genetic taste division in the popular imagination is the cilantro debate. A substantial minority of the population — estimates range from 4% to 14%, with the variation partly explained by ethnic background — experience cilantro not as herbal and bright but as soapy and deeply unpleasant. This is not a learned preference or a cultural bias. It is the result of genetic variants in olfactory receptor genes, particularly OR6A2, which causes those who carry it to perceive the aldehydes in cilantro as chemically similar to the compounds found in soap.

SplitTheWorld put the question to a vote: "Cilantro tastes like soap." The result was 39% in favour and 61% against — which aligns reasonably well with the scientific literature on genetic cilantro aversion, given that SplitTheWorld's self-selected audience is not a representative population sample.

The cilantro case is important because it illustrates something that applies across all food debates: strong opinions about taste are not arbitrary or irrational. They are rooted in genuine sensory experience. When someone insists that cilantro is disgusting, they are not being difficult. They are accurately reporting what they taste. The person who finds cilantro delicious is equally accurate. They are having a different experience of the same ingredient.

THE HISTORY OF FOOD DISGUST

Psychologist Paul Rozin, who has spent decades studying the emotion of disgust, argues that food disgust is one of the most powerful and culturally variable emotional responses humans experience. What triggers disgust is not primarily a function of objective properties of food — it is a function of what a particular culture has defined as appropriate to eat, combined with individual learned associations built up over a lifetime.

The foods that people in one culture find repulsive are frequently delicacies in another. Fermented shark in Iceland. Balut — partially developed duck eggs — in the Philippines. Haggis in Scotland. Insects in much of Southeast Asia, Latin America, and sub-Saharan Africa. None of these foods are objectively disgusting. They are culturally categorised as disgusting by people raised in contexts where other foods were normalised.

This cultural variability in food disgust has a long history. Anthropologists studying food taboos across cultures have found that disgust responses cluster around a small number of universal categories — bodily waste, decay, certain animals considered unclean within specific religious traditions — but that within those broad categories, enormous variation exists. The specific items that trigger disgust, and the intensity of that response, are almost entirely learned.

Which means that virtually every food debate — pineapple on pizza, cold pizza versus hot pizza, whether ketchup belongs in the fridge — is, at its root, a debate between people whose food associations were shaped by different experiences in different environments. Neither side is wrong. Both sides are reporting their genuine sensory and emotional response to the question. The disagreement is real. It just cannot be settled by argument alone, because the disagreement is not about logic. It is about biology, memory, and culture.

In 399 BC, the citizens of Athens voted by a majority to put Socrates to death. The charges were impiety and corrupting the youth of the city. The vote was 280 in favour of execution, 220 against. By the democratic standards of Athens — which had invented the concept of democratic governance — the will of the majority had been expressed and carried out. By virtually every subsequent ethical and philosophical standard, a grave injustice had been committed.

The death of Socrates is the founding case study for one of the most enduring debates in political philosophy: when, if ever, can the majority be wrong? And what protections, if any, should exist for minority positions?

MILL AND THE PROTECTION OF DISSENT

John Stuart Mill, writing in his 1859 essay "On Liberty", articulated what remains the most influential philosophical defence of minority opinion. Mill's argument was not that majorities are usually wrong — he accepted that they are often right. His argument was that even a majority that is correct derives enormous benefit from the existence of a dissenting minority, because challenge and opposition force the majority to understand and articulate why they believe what they believe. A truth held without challenge becomes, in Mill's phrase, "a dead dogma" — something people repeat without understanding.

Mill went further: in cases where the minority is right and the majority wrong, the suppression of minority opinion does not merely harm the minority. It harms the majority too, by depriving them of a correction they need. The majority that silences dissent is, in a meaningful sense, acting against its own long-term interests.

This argument has obvious application to trivial debates as well as serious ones. The person in the 21% minority who prefers toilet paper under the roll is not simply wrong by definition. They may be responding to a genuine functional consideration — household pets, small children, a specific bathroom configuration — that the majority has not encountered. The minority position contains information. Dismissing it entirely means losing that information.

THE HISTORY OF MINORITY POSITIONS BECOMING MAJORITY POSITIONS

The history of ideas is substantially the history of minority positions eventually becoming majority positions. The view that the Earth orbits the Sun rather than vice versa was, for most of human history, held by a tiny minority of astronomers against an overwhelming consensus. The view that women should have equal political rights was a fringe position in most democracies less than 150 years ago. The view that smoking causes cancer was, in the 1950s, a minority position — and one that faced organised, well-funded opposition from the tobacco industry, which understood better than most how powerful it is to be on the majority side of a contested question.

In each of these cases, the minority was not merely inconvenient to the majority. It was right. And the mechanism by which the minority position eventually became the majority position was not force — it was the accumulation of evidence and the gradual persuasion of individuals, one by one, until the balance tipped.

SplitTheWorld captures snapshots of opinion at a specific moment. The splits it records are historically contingent — they reflect what a particular audience believes in the mid-2020s, on questions that would have generated very different splits in other eras. Some of the current majority positions will, in a generation or two, look as obviously wrong as supporting women's suffrage looked radical to a Victorian majority. Others will look just as right as they do today. The hard part is knowing which is which — and the only honest answer is that you usually cannot tell in real time.

The first known Western etiquette manual was written in ancient Egypt, around 2400 BC, by a court official named Ptahhotep. His "Maxims of Ptahhotep" included advice on table manners, how to behave in the presence of social superiors, and the correct way to conduct oneself during a meal. Some of his guidance would not look out of place in a contemporary etiquette column. Do not speak when you should be silent. Do not reach for food before your host. Show respect to those who know more than you do.

The fact that someone felt the need to write down rules for social behaviour in 2400 BC tells you something important: people have been arguing about manners for as long as there have been manners. The debates that SplitTheWorld captures — about reclining airplane seats, checking your phone during meals, speaking loudly on speakerphone in public — are the latest iteration of a conversation that has been running for more than four thousand years.

THE INVENTION OF THE TABLE FORK AND THE SCANDAL IT CAUSED

Etiquette debates are not trivial. They are, at their core, debates about power, class, and social inclusion — about who belongs and who does not, about what behaviour signals education and breeding, about where the line falls between acceptable and unacceptable. The introduction of the table fork into European dining in the eleventh century is a useful illustration. When the Byzantine princess Theodora Doukaina married the Doge of Venice in 1005 AD and used a small golden fork to eat her food rather than her fingers, she was subjected to widespread mockery and religious condemnation. A Venetian cardinal declared that God would be displeased by such excessive refinement — that fingers were God's gift for eating, and to reject them in favour of a metallic instrument was an act of pride bordering on impiety.

The fork won eventually. It spread across European courts over the following centuries, carried partly by the social prestige of those who used it and partly by the genuine practical advantages it offered. By the seventeenth century, using a fork at table was a marker of education and refinement. By the nineteenth century, not using one was a marker of ignorance. An instrument that had once been considered offensively pretentious had become definitionally proper.

THE AIRPLANE SEAT DEBATE IN HISTORICAL CONTEXT

SplitTheWorld asked whether reclining an airplane seat is inconsiderate. The split was 58% in favour of calling it selfish and 42% defending the right to recline — which is to say, a majority consider recliners to be in violation of a social norm, while a sizeable minority believe they are simply using a function they paid for.

This debate has the structure of most etiquette debates throughout history: a tension between individual entitlement and collective obligation. The 58% who consider reclining selfish are applying a norm based on collective welfare — the person behind you loses space they were implicitly promised. The 42% who defend reclining are applying a norm based on individual right — you paid for a seat with a reclining function; using it is not a violation of any rule.

Neither position is obviously correct. Both reflect genuine and defensible values. The debate will not be resolved by argument — it will only be resolved if airline seats are redesigned, or if a clear social consensus emerges that is strong enough to make reclining feel genuinely antisocial rather than merely inconsiderate to some. The fork took six centuries to move from scandal to standard. Airplane seat etiquette is a newer debate, and the norms are still in motion.

WHY ETIQUETTE DEBATES ARE ALWAYS PARTLY ABOUT CLASS

Etiquette historian Cas Wouters, who has spent decades analysing the evolution of manners codes in Western Europe and America, argues that etiquette debates are never purely about behaviour. They are always, at some level, about social positioning — about who has the cultural authority to define what is correct, and who is implicitly judged as deficient by those definitions.

The rules that govern how one should behave in shared spaces — airplanes, restaurants, public transport — are disproportionately written by and for people who use those spaces most. Frequent fliers developed the etiquette of air travel. Frequent diners at restaurants shaped expectations around tipping. The norms feel universal and obvious to those who created them. They feel arbitrary and sometimes actively hostile to those who encounter them from the outside.

This is not an argument against etiquette. It is an argument for understanding where etiquette comes from, and why people whose experiences differ from the norm-setters sometimes push back against the norms in ways that seem, from the inside, inexplicably rude.

In the eighteenth century, the coffee house was the primary venue for public debate in Britain. A penny bought entry and as much coffee as you could drink. What you got in return was access to newspapers, pamphlets, and the conversation of strangers from every class and profession. The great coffee houses of London — Will's, Lloyd's, the Chapter, the Grecian — were places where merchants discussed trade, writers shared manuscripts, scientists argued about natural philosophy, and politicians aired views they would not dare express in Parliament.

Historians credit the coffee house with a significant role in the intellectual ferment that produced the Enlightenment — the period from the late seventeenth century through the eighteenth century during which rational inquiry, individual liberty, and scientific method became dominant intellectual values across Europe. The coffee house made public debate physically accessible in a way it had never been before, drawing people from different backgrounds into the same room, forcing the exchange of ideas across social barriers that had previously kept those ideas separate.

The internet is, in many ways, the coffee house scaled by a factor of a billion.

FROM USENET TO SOCIAL MEDIA

The earliest large-scale internet debate forums were the Usenet newsgroups, established in 1980 as a distributed discussion system accessible to anyone with a university computer connection. By the late 1980s, Usenet hosted thousands of discussion groups covering every conceivable topic — science, politics, religion, hobbies, popular culture — and had developed its own recognisable culture of argument. It was on Usenet in 1990 that lawyer Mike Godwin observed what became known as Godwin's Law: "As an online discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches one." The tendency of internet debates to escalate toward extremity had been noted within a decade of the medium's existence.

The structural features of internet debate that Godwin observed — escalation, hyperbole, the collapse of nuanced positions into binary opposites — are not accidental. They are consequences of the medium. Written text, produced asynchronously and read without tonal or facial cues, strips away much of the social information that moderates face-to-face argument. Anonymity reduces social accountability. Algorithmic amplification rewards outrage over nuance because outrage generates engagement. The internet has not changed what humans argue about. It has changed how they argue about it, and who gets to hear them.

THE FILTER BUBBLE AND THE FALSE CONSENSUS

The most consequential structural feature of modern internet debate is the filter bubble — the tendency of algorithmic content curation to show people more of what they already agree with, and less of what challenges them. The concept was developed by activist Eli Pariser in his 2011 book of the same name, though the phenomenon had been identified by social scientists before it was named.

Filter bubbles make the false consensus effect — already a powerful feature of human psychology — dramatically worse. Not only do people naturally assume that their opinions are more widely shared than they are; modern social media actively reinforces this assumption by populating their feeds with people who think like them and filtering out evidence that other views exist and are held by thoughtful, reasonable people.

SplitTheWorld was designed, in part, as an antidote to this dynamic. By collecting anonymous votes across an audience that arrives from many different directions, and displaying the real split rather than the filtered version that any individual's social network would show them, it offers something that is genuinely rare in contemporary online media: an unedited view of how diverse the actual distribution of opinion is. The filter bubble says you are normal and your opponents are fringe. SplitTheWorld shows you the actual numbers.

THE PERMANENCE OF DISAGREEMENT

It would be comforting to believe that more information, better connectivity, and broader access to debate would eventually resolve the fundamental disagreements that have characterised human society throughout history. The evidence suggests otherwise. Four thousand years of etiquette debates have not produced consensus on table manners. Two hundred years of polling have not resolved which policies governments should pursue. The internet has made more information available to more people than any technology in history — and political, cultural, and social polarisation in most democracies has increased, not decreased, over the period of its dominance.

This is not an argument for despair. It is an argument for appropriate humility about what debate can achieve. Arguments can change minds. Evidence can shift opinion. Exposure to different perspectives can soften certainty and build empathy. But the fundamental diversity of human experience — different biology, different culture, different history, different values — ensures that complete consensus will always be a horizon rather than a destination.

SplitTheWorld is not trying to resolve any debate. It is trying to show honestly where people stand. The splits it captures — sometimes surprisingly lopsided, sometimes almost perfectly balanced — are a snapshot of a conversation that has been running since Ptahhotep wrote his etiquette manual in 2400 BC and will still be running long after the current questions have been answered and forgotten. We are a species that debates. The medium changes. The impulse does not.