Your thumb gets there before you do.
You're at a corner in a city, waiting to cross. There's a metal button on the pole with a worn patch in the middle where forty thousand other thumbs have been. You press it. Then, and this is the part worth admitting, you press it again. Slightly harder. As though the first press was a polite request and the second is the follow-up email.
The button isn't connected to anything.
I mean that specifically, and somebody went and counted. On February 27, 2004, the New York Times ran a piece by Michael Luo under the headline "For Exercise in New York Futility, Push Button." The sentence everyone has been quoting off it ever since is this: more than 2,500 of the 3,250 walk buttons that still existed in the city functioned essentially as mechanical placebos.
Around 750 were live. The rest had been cut loose as computer-controlled signal timing spread through the system, because once a central computer is coordinating a whole avenue's worth of lights the pedestrian phase is baked into the cycle and there's nothing left for your thumb to ask for. The housings stayed up. Removing them cost more than leaving them there looking official.
The update nobody repeats, because it ruins the joke: the buttons going in now do work. They're Accessible Pedestrian Signals, at 3,247 intersections as of the last day of 2024, and they deliver a locating tone, a spoken message and a vibrating arrow under your hand, for people who can't see the white figure light up. They don't shorten the wait. So the modern button is wired to something. The something isn't what you thought you were asking for.
You'll press it twice anyway. You can know every word of the above and you'll still stand at that pole and press, and then the light will change on its own schedule, and you'll step off the curb with a faint sense of having handled it.
Anyway. That impulse, the one your hand has and you don't, used to run an empire.
Reading birds was a job with a title, a uniform and a payroll
The thing modern people assume about omens is that they live at the edges of a society. Grandmothers, hedgerows, folk stuff, the kind of knowledge passed along sideways because no institution will touch it.
In the Roman Republic it was the institution.
There was a formal college of priests, the collegium augurum, and membership was one of the senior offices open to a Roman aristocrat. Livy says there were four members before the lex Ogulnia of 300 BCE opened the priesthoods to plebeians. Then nine. Sulla raised it to fifteen, and by Caesar's day it was sixteen.
You held it for life. Not "generally for life" or "in practice for life." Plutarch says a Roman priest convicted of a crime and exiled could be stripped of his priesthood, except an augur, who kept it however grave the conviction. One reason Plutarch offers is that augur names a skill, and you can't vote a man into no longer having a skill.
How you got in was a political football for fifty years. The old way was cooptation, meaning the existing members chose. Then in 104 BCE a tribune put the choice in front of seventeen of the thirty-five tribes, chosen by lot. Sulla repealed that. Labienus restored it in 63 and Mark Antony abolished it again in 44, which is four changes of procedure inside sixty years over who watches the birds, and tells you what the job was worth to the people fighting over it.
A templum was a piece of marked-out sky
This one reorganised my head, so I'm putting it early.
Varro says templum gets used three ways: from nature, meaning the sky; from augury, meaning a marked-out field of observation; and by resemblance, meaning a building. The building is the last of the three and the latest. In the augural sense, a templum is a region a man has defined out loud, in a fixed verbal formula, and a sign that happens outside the line he drew isn't a sign.
Varro preserves the formula. The augur speaks it sitting down, naming a tree over there as the boundary on the left and another tree as the boundary on the right, and the three words for what he's doing between them are conregione conspicione cortumione, roughly the line of the boundary, the line of sight, and the act of fixing the thing in the mind. So the result is an observation window. It's made by a man announcing in advance where he's going to look. Rome had permanent versions, called auguracula, roofless and bounded by stones.
The best illustration is a story Cicero tells about Attus Navius, who promised the god the largest grape in his vineyard if he got his lost pig back. He got the pig back. So he stood in the middle of the vineyard, divided it into four, and watched. The birds ruled out three. He took the surviving quarter, divided that into regions, watched again, and found his grape.
Read that as a search algorithm and it works perfectly. He's running a rule that narrows a field until one answer is left standing, and no information has to arrive from anywhere for it to do that.
Seeing a sign and being allowed to say so were different powers
Most popular accounts get this backwards. The augurs didn't usually take the state auspices. The magistrates did. Consuls, praetors and censors held what Gellius calls the auspicia maxima. The right came bundled with the office. So the man about to do the thing was the man watching for permission to do it.
Cicero, who was in the college and therefore knew, states the split in one line in the Second Philippic: nos enim nuntiationem solum habemus, consules et reliqui magistratus etiam spectionem. We augurs have only the announcement. The magistrates have the watching as well.
Two distinct powers, then. Spectio is the magistrate's right to go and deliberately look for a sign before acting. Nuntiatio is the augur's right to announce that a sign has occurred. And obnuntiatio is the announcement of an adverse one, which suspends business.
That last word is the whole apparatus. An announcement of bad signs stopped public business. Not "discouraged." Stopped.
Cicero's own draft law code in the De Legibus puts it in legal language. Whatever an augur declares unjust, unholy, faulty or ill-omened is void and undone, and whoever fails to obey suffers the capital penalty. A page later he asks what could be a greater power than dissolving assemblies convened by the highest authorities in the state, then answers himself with the sentence I keep coming back to. What is weightier, he says, than an undertaking broken off because one augur said alio die, another day.
Two words, and everybody goes home.
The chickens were kept in a cage and kept hungry
Festus lists five categories of sign the public augurs watched: from the sky, from birds, from the feeding of chickens, from four-footed animals, and from alarming events. On campaign, the one that got used was the chickens.
An official called the pullarius, whose title just means the chicken guy, travelled with the army and kept them.
(Which was a job. Somebody in the Roman army got up in the morning knowing that the timing of the day's engagement, and therefore the reputations of men enormously above him, and therefore in a fairly direct sense the afternoon's casualty figures, rested on whether the birds in his particular crate happened to be peckish, and that he was the one who had to keep them fed and watered and alive across a mountain range in a wooden box, and then walk over in front of everybody and tell a consul exactly what they'd done about it.)
Anyway. The procedure was to offer food and watch. The favourable sign, the tripudium solistimum, was food falling out of the beak and striking the ground. Cicero walks the etymology through in the De Divinatione: terripavium, striking the earth, becoming terripudium, becoming tripudium.
Then he takes the whole thing apart with a straight face.
He asks what could possibly be divine about a sign this forced and squeezed out. The college's own old decree said any bird at all could give a tripudium, which would only be an auspice if the bird were free. Then the line that does the damage: as things stand now, the bird is shut in a cage and half dead with hunger, and if it goes at a lump of mash and something drops out of its mouth, do you call that an auspice, and do you imagine this is how Romulus used to take the auspices?
Starve the bird, and the food falls, and the sign comes back favourable. More or less every time, which was the point.
The sceptic isn't the only one who noticed. Over in Book 1, where Cicero's brother Quintus is arguing for divination, Quintus concedes it himself. Our magistrates use forced auspices, he says, because a fragment has to fall out of the mouth of a hungry chicken. So the defence counsel concedes the point and carries on defending, which should look familiar to anybody who has sat in a meeting.
Drepana, 249 BCE, and what the sources actually say
The famous one. Publius Claudius Pulcher, commanding a fleet off Sicily, is told the sacred chickens won't eat, has them thrown into the sea with a remark about letting them drink instead, sails anyway, and loses catastrophically.
Great story. Also the kind that improves in transit. So here is how it's actually attested.
The earliest surviving source doesn't mention the chickens at all. Polybius, writing about a century after the battle, gives four chapters to Drepana with no omen in them and nothing religious anywhere in the account. His explanation is material throughout: Rome had lost most of its trained crews at Lilybaeum and had shipped over ten thousand hastily enlisted sailors, Carthaginian ships were better built and better rowed, and Claudius ordered his fleet to back out of the harbour, so ships fouled each other and snapped oars doing it, after which he pinned his own line against the shore and lost ninety-three of them. Polybius says he was tried, heavily fined, and narrowly escaped with his life.
The chickens surface two centuries later, in Cicero's De Natura Deorum, in a passage whose entire job is arguing that the gods punish contempt for religion. That's the earliest surviving version of the joke, and it's in indirect speech: ut biberent, quoniam esse nollent, so that they might drink, since they were unwilling to eat.
After that the tradition scatters. Livy's summary of the lost book has the chickens drowned and no quip, Florus has the chickens and the wrong first name for the man, and Eutropius has no chickens at all. Suetonius reports the quip as a saying rather than quoting it, and Valerius Maximus has the crispest direct-speech version of all, which reaches us through a late-antique abridgement rather than the manuscripts.
The detail I find most telling is that Cicero drops the anecdote himself when he writes technically. In the De Divinatione, where he's on augural law rather than piety, the same episode comes up twice and the chickens are gone both times. It's just vitium, a flaw. They sailed under bad auspices and lost the fleets.
So a real defeat with mundane causes acquired a religious punchline somewhere in the two hundred years afterward, and the punchline is never quoted in the same words twice. That's an exemplum, a story kept in circulation for the job it does, which is a different category of thing from a report.
The aftermath is stranger. Ordered by the senate to name a dictator, Claudius named his own messenger, a man of the lowest rank, who was promptly made to resign. And when the case against him came to trial, Valerius Maximus says a sudden downpour broke it up, and since it looked as though the gods were interrupting, nobody wanted to start a fresh one.
The rain got him off.
Bibulus stays home for a year
Two hundred years later the same machinery was being worked the way a filibuster is worked.
Marcus Calpurnius Bibulus was consul in 59 BCE alongside Julius Caesar. He lost every argument with him. So he stopped arguing. He shut himself in his house for the rest of the year and, as Suetonius puts it, did nothing but issue edicts of obnuntiatio, while Dio's version has him declaring the whole remainder of the year a sacred period on which assemblies could not lawfully meet, which amounts to one man alone in a private house attempting to switch off the legislature of the Roman Republic by writing notes.
Caesar carried his legislation anyway. Bibulus asked the senate to annul the agrarian law on the grounds that it had passed by force and against the auspices. The senate declined. Three years later two praetors moved an inquiry into the consular year, and the senate let that die too. Which leaves the position that on a strict reading of augural law, some of the most consequential legislation of the late Republic was flawed, was challenged twice by people with a real case, and stood.
Sometimes it did bite. In 163 BCE Tiberius Sempronius Gracchus presided over the consular elections. Later, reading the sacred books out in his province, he worked out that he'd pitched his observation tent on flawed ground, because he'd re-entered the city boundary for a senate meeting and crossed back out again without retaking the auspices. So he wrote to the college, and the augurs took it to the senate, and the senate told the consuls to resign, and they resigned, and since they had already taken office and gone out to their provinces, one of them had to be shipped home from Gaul and the other from Corsica to stop being consul.
Nobody made him write that letter. He worked out a technical fault against himself, at a distance, from a book, and reported it to men who could act on it.
Cicero was an augur and did not believe in it
He was co-opted in 53 or 52 BCE, replacing the younger Crassus, who'd been killed at Carrhae. Pompey and Hortensius nominated him, a detail he mentions in a way that makes clear he was pleased about it.
Then in 44 he wrote the De Divinatione, two books, the second finished after the Ides of March. Book 1 gives his brother Quintus the Stoic case in favour. Book 2 is Cicero taking it apart, and he says plainly at the start that he denies divination exists.
He's funny about it, too. He quotes a remark he attributes to Cato, who used to say he wondered that one haruspex could look at another without laughing. (A haruspex read entrails, a separate profession, so the joke is aimed at the other department. Which I think is the most Cicero fact available.)
And then he says the part that makes the whole institution legible. The augural law, he writes, is retained both for the opinion of the crowd and for the great advantages of the state. A few sections later he sides with one of his own colleagues against another, and says he thinks the augural law, though established in the first place on a belief in divination, was afterward kept up for reasons of state.
A serving priest, in a published book, saying the machinery is worth keeping for what it does rather than for what it detects.
Careful with that, though. Cicero says close to the opposite in the De Legibus, where the same dispute comes up and he sides with the other man and says he believes divination exists. Different book, a different stance, and nobody has reconciled the two. The sharpest modern work on augury, Lindsay Driediger-Murphy's, argues against reading the system as political theatre, on the grounds that augural results regularly cut against the interests of the people obtaining them, which isn't what a rigged procedure does.
What the specialists do agree on is narrower and more useful. Jerzy Linderski's long study treats augural law as a technical juristic discipline with its own experts and its own logic, in which the augurs' business is never to predict the future. It's establishing whether Jupiter approves a specific act, right now, before it starts, which is a different question from what will happen and a far more answerable one.
The living traditions, which are not a museum exhibit
I've spent this whole piece inside one dead republic, so let me be clear about what that does and doesn't establish. Reading signs is ordinary and current in a great many places, and plenty of people reading this belong to those traditions.
Ifá divination, from the Yoruba world and its diaspora in Cuba, Brazil and Trinidad, was inscribed on UNESCO's Representative List of the Intangible Cultural Heritage of Humanity in 2008. UNESCO's description makes a point that rhymes with everything above. Ifá doesn't rely on a person having oracular powers, but on a system of signs interpreted by a trained babalawo against a corpus of 256 chapters. The Tibetan mo, a dice system, gives one of thirty-six outcomes from two throws. The Chinese almanac codifies which days suit a wedding, a move, or opening a business.
None of that is a Roman bureaucracy and the comparison isn't tidy. What they share with the Roman material is structural: a defined procedure, a trained reader, a fixed set of answers, and a moment when the answer is in and the deliberating stops.
What the birds were for
Line up what the office did and the shape of it is unmistakable. It ran before elections, before battles, before the founding of a colony, and before a magistrate could take up the office he'd already been elected to. Varro describes colonies founded with a plough drawn round the boundary on a day taken by auspice, and says the resulting furrow is where the city auspices end. Observation happened after midnight, business after noon, and the sign had to occur inside a field of view somebody had announced in advance. The answer was binary: proceed, or alio die.
There's no bandwidth in that for a message. Jupiter can't use this system to tell you your marriage is in trouble, because the only two things it can carry are yes and not today.
What it can do is close a question at a moment when nobody arguing about it has any way of being right. Whether the fleet sails today. Whether a man already elected consul gets to take the office, or has to go home. No fact decides that, there are men on both sides with real interests, and the deliberation could run forever because nothing in it will ever run out. So you have a procedure instead, and it's public, and its rules were fixed long before anybody knew what would be at stake, and it's run by men who trained for years and hold office for life and can't be removed even by a criminal conviction, and when it returns an answer, the answer is over.
That's why senior officials could be openly sceptical while the institution carried on around them, and nobody experienced a contradiction. Cicero isn't a hypocrite in the De Divinatione. He's describing public equipment accurately. The birds were the part that couldn't be argued with, which is the only specification that ever mattered, and a starved chicken in a cage worked as an oracle because nobody could accuse anybody of having arranged it.
Which brings the thing back round to the pole on the corner. That button isn't wired to the light, and knowing it costs you nothing, because the light was going to change anyway and your thumb was only ever making the wait survivable. The Roman button was wired. Press it and a fleet sails or doesn't, an election stands or is voided, two consuls get on ships and come home. Nothing about the future came back down that wire, and something enormous went out along it.
So when you notice something and it lands with more weight than it should, the interesting question isn't whether the universe sent it. There's a large literature on why you noticed it, and it's good, and it isn't this. The question is what the noticing is being used to close: the argument you've been having with yourself since March, or the choice that ran out of facts a long time ago and has been sitting there ever since, waiting for somebody to be allowed to end it. Because an omen is a procedure for ending an argument that nobody was ever going to win on the facts.
