Forty-Six Percent of Emperors Were Overthrown. The Empire Lasted Anyway

Anthony Kaldellis puts the burden of proof on anyone who claims the Eastern Roman Empire was something other than Rome. Every source, he tells Lex Fridman, is clear: the state called itself the Roman Empire, its subjects were Roman citizens, and they called themselves Romans down to 1453 and beyond. Western Europeans spent a thousand years calling it an "Empire of the Greeks," then switched to "Byzantine Empire" for political reasons. Kaldellis describes the older habit as a kind of cognitive dissonance: people knew the facts and carried on as if they were not true. That naming scheme, he says, is now collapsing into a longer Roman history that runs from archaic antiquity to the fifteenth century.

The pair then sprint the 2,200 years from the legendary founding in 753 BC to the fall of Constantinople. Kings last until around 510 BC; a republic dominated by aristocracy, in cooperation with the Roman people, lasts about five centuries; civil wars around Julius Caesar and Augustus produce an imperial monarchy from roughly 27 BC that continues for a millennium and a half. Kaldellis divides that monarchy into two phases, one centered on Rome and one on Constantinople, and warns that the neat years in textbooks are labels on processes. 395 was one of many allocations of jurisdiction between sons, not a hard split; 476 is simply the last date in a decline that took decades.

Asked for holes in a whirlwind timeline, Kaldellis finds none that are simply wrong, but he would organize the story around three swift territorial disasters: the Arab conquests of the 630s, lost in a decade; the Seljuk takeover of Asia Minor in the 1070s, partly reversed; and the Fourth Crusade of 1204, when mostly French armies sacked Constantinople and dismembered what they could. The Romans regrouped after the last of those. For the rest of the history, he says, the state is consolidating, regrouping, and growing slowly in wealth and territory. Historians dance between discrete events with long consequences, such as Constantine's conversion, and conventional dates that freeze a much longer process.

Is there a through-line across 2,200 years? Kaldellis reaches for the Ship of Theseus: every plank can be replaced and the ship is still that ship because of the story told about it. Roman history, unlike Christian history or Greek cultural history, is the history of a specific political community, the Roman citizens. Christianity can live in any country; ancient Greeks were never united in one state, yet people still read Homer and Plato. At no point, he argues, did people inside the Roman story feel a rupture that put them outside it. Changes were real and eventually total, but gradual enough that the narrative of the Roman people is what holds the whole thing together.

Kaldellis argues that the late Roman state projected a consistent *persona* to its subjects, broadcast through laws, praise, and answers to petitions. Anyone could petition local officials or even the emperor; the expectation was that the petition would be read and answered, even if not granted. The message was always the same: authorities are responsive and accountable, they exercise public power only for the benefit of subjects, they plan ahead, and they work themselves to exhaustion. Justinian was publicly the "sleepless emperor," a workaholic even his enemies conceded, hauling officials in at odd hours to talk about war, architecture, law, and theology. (A rival rumor held that his head floated around the palace on its own, a demonic version of the same insomnia.)

Propaganda, Fridman notes, implies a gap between rhetoric and action. Kaldellis's research on the East Roman state finds the emperors generally sincere: they did, overall, what they said they would do, with the usual human corruption and abuse. The surprise is that men with enormous formal power worried so much about making subjects relatively content. The structural reason is that there is no right to the throne. Roman emperors did not found the state; they emerged inside a republic. Augustus had to pretend for a long time that he was not a king, stacking republican offices the Senate had "given" him. The facade mattered, Kaldellis says, because ending one civil war does not prevent the next. If no one has a right to the throne, violence remains available to the next claimant.

In the thousand years Kaldellis studies, there are something like 120 civil wars, usually swift and never ideological. Both sides talk about the responsibilities of government in the same way; they are competing over who is the best man for the job. Palace coups and street fighting in the capital count too. The statistic he keeps returning to is stark: 46 percent of the emperors of Constantinople are overthrown through violence. Every emperor knows he is vulnerable. Arresting everyone you suspect, Caligula- or Nero-style, gets you killed. The rational path is to make people happy enough that they will not back a rebel.

Constantinople ran what Kaldellis calls a perpetual referendum. Elections give a modern government a few years of legitimacy even when it is unpopular; the capital had no such cushion. When the emperor appeared, crowds were expected to chant. A Hippodrome of perhaps 30,000 to 100,000 people roaring "Worthy" was good news; sullen silence or booing meant the grain supply, or something like it, had failed. In the 1190s, Alexius III announced a "German tax" to buy off a threatening German emperor and had to withdraw it on the spot when the Hippodrome erupted. Sometimes a wannabe marched down the central boulevard toward Hagia Sophia and watched his crowd melt away; sometimes the people dragged an emperor out and cut off his nose. Institutionally the emperor looked absolute. In practice, Kaldellis says, the ones who behaved that way did not end well.

In 212, the emperor nicknamed Caracalla, officially Marcus Aurelius Antoninus, issued the Constitutio Antoniniana, which extended Roman citizenship to almost all free inhabitants of the empire. Citizenship had already been spreading for a long time through block grants, generals, and even foreign cities allowed to elect people into Roman status. By the early third century, Kaldellis estimates, at most a third of free people were Roman citizens; everyone else held a local citizenship, Alexandrian or Athenian, and lived under local law. The edict was an explosion, not just another expansion.

Motives are guesswork. Caracalla said he wanted people to thank the gods as Roman citizens after he survived a coup; contemporaries also alleged a tax grab; jurists around him may have wanted a unified community. The consequences matter more. The rights and opportunities of Roman citizens, including a path toward the Senate, were now open to provincials, and within a generation the emperors and the most powerful officials were provincials. "They did it, and they meant it," Kaldellis says. He asks listeners to imagine the British Empire at its peak granting British citizenship to everyone in India and opening London's highest offices, even the throne, to them. Race and colonialism made that unthinkable in the nineteenth century. Rome did not have that barrier.

We do not even have the full Latin text. What survives is a torn Greek papyrus fragment that happens to preserve the key clause, plus a sudden explosion of people named Aurelius, Caracalla's family name, across the provinces. Kaldellis suspects a few jurists, in an age of figures such as Ulpian, were in the room, but there are no minutes. A century later, when Constantinople is founded on the Bosporus, Romanness is already in the East. Creating a center of Roman power there is not a leap. It is a relocation inside a community that the edict had already made imperial in the civic sense.

What follows the edict is the Crisis of the Third Century, a name Kaldellis thinks it deserves. Twenty-six emperors are murdered in fifty years, almost all of them generals murdered by their own men. Civil war is the norm; there is hyperinflation and economic crash; a plague from 249 to 262, possibly smallpox or Ebola, kills up to 5,000 people a day in Rome and wrecks the tax base and recruitment; the empire fractures into three pieces until Aurelian reunites them in five years. Fridman stresses the body count. Kaldellis agrees it is tremendous instability at the top, then asks what life looked like in Egypt, where papyrus documents suggest ordinary people were not living through a crisis at all.

The so-called Gallic Empire is the tell. It was not provincials tired of Rome and peeling away. It was provincial generals trying to take the whole throne and getting only as far as the land they held. The only state they knew how to set up was a Roman one. By this point, Kaldellis says, there are almost no alternatives in the political imagination. Diocletian, who takes power in 284, turns the problem into the solution: too many would-be emperors, because one man cannot cover every frontier, so he deputizes colleagues. They are Illyrians from a Yugoslavia-sized patch of the Balkans, the size of Scotland, that produced emperors for centuries because poor young men joined the army and rose.

The Tetrarchy was a collaborative military allocation, not a hard partition of the empire. Administration hummed along wherever the emperors happened to be. Diocletian builds bigger government, possibly a bigger army, and therefore more systematic taxation, census, and bureaucracy. Fridman asks whether this is a "deep state." Kaldellis has not found sinister bureaucrats undermining the emperor; the closest glimpse is Diocletian complaining that nobody ever tells him anything. What Diocletian did create, successfully enough, is the administrative model East Roman specialists study, used in its main lines for three centuries and in some elements far longer.

Until Diocletian, Italy was a tax haven, the land of the conquerors living off everyone else. The Balkan soldiers around him could not see why the richest part of the world paid nothing. They taxed Italy. From that point, Kaldellis says, all Romans had the same citizenship, the same law, and the same tax system. A universal census listed every taxable asset, which meant the state could begin to budget. Asked to put a finger on one factor that explains how the society worked, he does not hesitate: taxation is the heart of it.

Fridman raises the claim that Diocletian legally froze people in their jobs and on their land. Kaldellis treats this as tax enforcement, not a caste system. Soldiers' sons often succeeded their fathers the way apprentices do; shipping guilds passed assets and obligations down the line. The harder case is farmers bound to land for fiscal purposes: they could physically leave, but they remained responsible for the revenue those fields owed. Even so, he sees considerable class and geographic mobility. People moved; Constantinople would soon prove it. Complaints of sons trapped in the army against their will are not what the sources show, and the army had tremendous perks.

Aurelian had already reunited the empire territorially. Diocletian reduced rebellions by putting four large armies in the field, beat back the Persians, and created the stability that lets people reclaim land and expand production. Without that, Kaldellis notes, foreigners maraud through, steal cattle, and enslave farmers for Mesopotamian estates. Economic and demographic growth return when people believe that will not happen. The tax machinery is what made the rest of East Roman government possible, including the army that was always the single largest item in the budget.

Constantine was the son of one of Diocletian's colleagues, passed over by a system designed to be non-hereditary, and backed by armies that liked him. Civil wars among the successors dwindle from as many as six emperors to one; in 324 he becomes sole ruler after defeating Licinius, whose headquarters included the old city of Byzantium on the Bosporus. Constantine founds Constantinople, New Rome, as a copy-and-paste of Rome in the East. Emperors had become itinerant frontier commanders, and a third-century historian had already said that Rome is wherever the emperor is. Other cities had been called New Rome. Constantinople was the last and most successful.

Geography is the point. Rome sits safely in the middle of the Mediterranean and too far from the frontiers these military emperors actually had to walk. Constantinople sits halfway between the Danube and the Euphrates, and almost every emperor who could choose East or West chose the East. It is also the hinge between the Black Sea and the Mediterranean, Europe and Asia. Civil wars had a habit of breaking at the Bosporus; the new city functions as a clamp holding the Balkans, Asia Minor, and Syria together. Constantine and his successors recruit perhaps 2,500 to 3,000 senators from the eastern elites, men who have to invest in the city to sit in its senate. After that, the empire does not break there again. The fault line moves to the Adriatic.

On a podcast ranking of East Roman emperors, Kaldellis put Constantine first, and not because he would want a beer with him. Constantine was ambitious, ruthless, and murderous toward the in-laws and alliances that had helped him up. He also executed his capable, popular son Crispus and then his wife, and offered no justification at all. When emperors have even a thin story, they blast it through every channel; silence, Kaldellis infers, means the act was unjustifiable. Contemporary writers praise Crispus while he lives and then simply stop mentioning him. The ranking is about consequence: founding Constantinople and setting the empire on a path to Christianity are world-history decisions almost no other emperor makes. A second criterion, used for the rest of the list, is whether a ruler did the job so that subjects lived better.

Fridman asks whether Christianity triumphed over Rome or Rome captured Christianity. Kaldellis leans toward the second. Christianity tells a victory story, which modern historians often repeat: Zeus and Apollo go down, animal sacrifice ends. Look the other way, he says, and the Roman imperial state is still there. Christianity neither defeated it nor tamed it; it became part of it. Islam, by contrast, arrived without a preexisting empire to receive it and had to build a state of its own. Christians in Constantine's day may have been 10 percent of the empire, and that figure might be high: mostly urban, not especially elite, in a world that was still rural. They were not a constituency large enough to explain a cynical conversion.

Kaldellis has never read a convincing account of the political gain. Bishops are useful social influencers who will not seize the throne, and some of them quickly become more trouble than they are worth. That leaves a personal explanation: Constantine believed, and may have thought the Christian God had helped him in battle. The Milvian Bridge vision is reported decades later, including in a dinner conversation with the emperor, which Kaldellis will not take at face value. Constantine had visions of Apollo too, and tended to have one whenever he wanted to rebrand. He remained a pragmatist, naming Christ to Christians and "good religion" in general laws, while a colossal gilded statue of himself in the guise of Apollo, naked, with rays from the head, stood at the center of his new city.

Conversion of the society is slow. A majority Christian world is perhaps mid-fifth century; a solidly Christian one, with non-Christians down to 10 or 15 percent, is early sixth, a half millennium after Jesus and two centuries after Constantine. The tools are incentives and disincentives: funds for bishops to run charity, land given to the church, then bans on sacrifice, on tying ribbons to sacred trees, on high office and wills for non-Christians. Christianity spread, Kaldellis argues, because it grew many faces: intellectual theology for elites, public processions that could replace ancient festivals, baptism and a personal afterlife, language in the court, oaths to the Trinity in the army. Exclusive truth-claims also made it both a stronger glue and a stronger source of conflict than ancient religions had been. He will not decide which side weighed more, because both are true.

Asked whether a Roman would say he was Roman or Christian, Kaldellis answers both, then adds that a moment's thought would show they are not the same thing. Bulgarians and Rus were Christians without being Romans; the churches commemorated martyrs killed by Roman emperors almost every week. Some men in the sources barely bother with Christian things and just do politics; others go all in on Orthodoxy and monasticism and ignore the Roman machinery. A third option is Greek: the language and the literature, with people who disappear into Homer and Plato. It is, he says, the only society in history that is Roman and Christian and Greek at once, usually in mixed ratios, occasionally in purist form.

Constantinople grows from perhaps 25,000 people, the old city of Byzantium, to half a million in two centuries. That rate is impossible by urban reproduction alone. Ancient cities were death traps, losing perhaps 1 to 3 percent of their population a year to disease, so even a stable city had to import people. Kaldellis does a back-of-the-envelope count: a few thousand senators required to reside there, each with a household. In Rome a senatorial household might be a hundred people; even at a conservative thirty in the new city, senators alone bring the population near 100,000. Constantine also provides a grain dole for perhaps 200,000. The grain comes from Egypt. New Rome is, in his phrase, stealing food from the mouth of Old Rome, which then draws on North Africa and Sicily. The signal about where the state is investing is unmistakable.

The Western collapse, as the East experienced it, is not a mystery. Barbarian armies put increasing pressure on the fifth-century state. Fail to suppress one group, settle them, or lose a province, as with the Vandals in North Africa, and you lose that province's taxes. Without taxes you cannot pay armies; with fewer armies you cannot handle the next wave. The Goths cross the Danube in 376; Adrianople annihilates field armies and kills the emperor Valens; a four-year war ends in a truce that, for the first time, accepts a quasi-independent Gothic army inside imperial territory. Geography then does the East a favor. Without a fleet, Goths stuck in the Balkans cannot cross to the richer Asian provinces. They are pushed into the poorer western Balkans and eventually decide Naples looks nicer.

The same geography works in reverse: Persians or Arabs coming from the east usually cannot cross into the Balkan provinces. An impregnable Constantinople is a command bunker that can call reinforcements from whichever wing is quiet. Flamethrowers, later, help too. None of that, Kaldellis insists, is the deep reason the East lasts. The deep reason is that local communities choose to stick with the Roman state, refuse to side with invaders, and restore ties with Constantinople as soon as trouble passes. By the 430s, Constantinople is already codifying Latin law for the West. In the sixth century the East reconquers parts of the West and recodifies Roman law on a far larger scale. The law we call Roman, he notes, comes to us in East Roman versions. Rome itself has become a provincial ruin with wolves in the Forum.

East Roman government splits, broadly, into military and civilian, with the church plausible as a third. Paper strength of the army runs from perhaps 100,000 at a low to 250,000 at a high, large for a medieval society and comparing well with the later Caliphate. Most of the time it is defensive. Civilian administration exists first to find the money for that army, through census, tax bureaus, and courts of dispute with the fisc, plus ordinary law, diplomacy, and the huge palace and church endowments. There is no tidy tax code. The system runs on exemptions: a monastery, a village hit by barbarians, a well-connected courtier petitions, and emperors often say yes. Officials start granting exemptions on their own; emperors periodically cancel the lot and start over. By the eleventh century the paperwork is a maze. The coverage, though, is dense. Taxes may come three times a year, in coin, kind, labor, and recruits. Add church, calendar, and coinage, and no village is outside the net.

Kaldellis wants to write a chapter against the myth of isolated peasants who do not know the emperor's name. He finds no evidence for such communities. The closest cases are Christian hermits described as having gone so far into the mountains that even the census takers did not show up, a line that reveals how people defined isolation. Every arable plot has been listed for centuries. Eunuchs, meanwhile, are the emperor's household staff, the West Wing: bedchamber, banquets, wardrobe, the people who handle access and collect bribes and absorb hatred when the court disappoints. They are useful as a parallel track of power because many came from outside the empire or from slavery, with no local family networks, and could not found dynasties. Emperors sometimes put a wardrobe eunuch in charge of an army to shake up the general staff. Narses, Justinian's tiny elderly commander, defeated the Goths in Italy after they laughed at him. Some boys were enslaved and castrated by force; some Paphlagonian families sent a son as strategy.

"Byzantine Empire," Kaldellis says, is the wrong phrase twice. There is no Greek term in this period that matches our "empire." The words in use are *basileia*, monarchy, and *politeia*, a translation of Latin *res publica*: the common affairs of a people, regardless of regime. Cicero could imagine a *res publica* under a king. The emperor's job, they keep saying, is to serve that polity. Romania is the state's proper name. Kaldellis's term is monarchic republic: a Roman *res publica* governed monarchically, a combination you do not find in dynastic empires that did not grow out of an ancient republic. He also rejects "military dictatorship." Emperors control armies, and armies often pick emperors, but they almost never use those armies as tools of social control. Soldiers are recruited from the people they protect and, for much of this history, live among them. Asking an army to fire on civilians is, as modern cases from the Arab Spring illustrate, a weapon that can destroy the regime that uses it. The exception who did it is Justinian.

Justinian rules from 527 to 565, nephew of Justin I, both of them from poor agricultural communities in the western Balkans. Justin rose through the army, made a disastrous church decision that lands him on Kaldellis's worst-ten list, and positioned his nephew as successor. Of Justinian before about age 38 the sources are blank: we know where he came from, not what he was doing. He arrives with Roman law and Christian theology rather than a lifetime in camp. He marries Theodora, a former sex worker, which was illegal for a senator until he got his uncle to change the law; Kaldellis says he clearly loved her, and Justinian almost says so in a statute. He staffed his government with talented, sometimes disreputable people and ignored class. That eye for talent, plus the nerve to use people others looked down on, is part of why he was so consequential.

He appoints a committee, led by the jurist Tribonian, to condense Roman law into the Corpus Juris Civilis. Overwhelmingly, what the world calls Roman law is what that committee said it was. It still underpins civil-law countries; even where common law prevails, Roman ideas of property and ownership shaped later theories of sovereignty. Most of the content is property. Marriage is a legal status with rules of age, citizenship, and prohibited degrees. Justinian thought homosexual acts caused earthquakes and tried to punish them, ultimately with death, but Kaldellis adds that this was not a society with an inquisition or a sex police. Scandal that became public was a problem; private lives were generally left alone.

The Nika insurrection of 532 is the other half of the ledger. The capital rises against him; street fighting lasts about ten days; he brings in outside troops and orders a slaughter of more than 30,000 people. Much of Constantinople burns, including the church he later replaces with Hagia Sophia, one of the great buildings of the world and a monument whose origin is also his. He thought God had authorized him to retake the lost western provinces, and he was opportunistic enough to hit barbarian kingdoms at their weakest. Italy was ravaged for decades. He disbanded forces meant to protect Constantinople, sent them west, and left the home front thin. The Persians marched in and destroyed Antioch. "I don't think you get a pass for conquering North Africa by surrendering Syria," Kaldellis says. North Africa probably paid; Italy and southern Spain probably did not. Conquest, on his scorecard, is not enough for the title "the Great."

The plague of 541 is bubonic, Yersinia pestis, a pathogen now confirmed in the lab after a generation of guesswork, and it comes and goes for two centuries. Some scholars argue for 50 percent mortality. Kaldellis does not believe a figure that would have halted logistics. The Black Death stopped fourteenth-century European societies for years; Justinian is fighting on four or five fronts with no pause in the war narratives, and taxation and the courts continue. The tax base and recruitment would have shrunk, and there would have been dislocation, but not half the population. He still leaves successors an overextended state with too many fronts and too few troops in the homeland. Good and bad, Kaldellis says, are mixed in such equal measure that, like Buridan's ass, he cannot choose a bale of hay.

Heraclius surprises admirers by missing Kaldellis's best-emperors list. In Western medieval literature he is almost a first crusader, the restorer of Jerusalem and the True Cross; modern fans, Kaldellis thinks, continue that tradition without always knowing it. The ledger is a wash or worse. Rome and Persia go to war around 602. Khosrow II shifts from raiding to permanent conquest. The emperor Phocas is himself a usurper; Heraclius rebels from Carthage and fights a civil war from 608 to 610 that pulls Roman armies off the Persian front and ravages Egypt, previously untouched. He takes Constantinople, executes Phocas, and inherits a disaster worse than the one he rebelled against. After some fifteen years he wins by striking into the Caucasus and Mesopotamia, ignoring Persian armies in Asia Minor and even a 626 siege of his capital by Persians and Avars, then calling in Turkish allies from Central Asia. Khosrow is killed in a palace coup. Both empires are wrecks. The Arabs arrive a few years later.

Kaldellis does not find a deep mystery. Arab warfare first aims to unify Arabia as a Muslim community, then strikes out at two exhausted neighbors from a direction neither empire had ever treated as a real invasion route. Occasional raiders, yes; a conquering army out of Arabia, no. Defensive orientations are all wrong, manpower and resources are spent, and the outcome is almost inevitable. Egypt is the killer loss: without Egyptian grain, Constantinople's population falls sharply. The rest of the seventh century is spent stopping the bleeding. Only around the 660s and 670 do the Romans begin to win on land and hold the sea. Armies pull back into Asia Minor, fortify, and live off local systems. The state that emerges is smaller, highly militarized, and logistically different.

Greek fire is a state secret so well kept that scholars still debate the mechanism: a heated, pressurized compound sprayed through a nozzle as flaming napalm that burns on water, plus clay grenades better against wooden siege engines than against a single person. It is used against Arab fleets and, later, against Vikings in the 940s and 1040s. The second Arab siege of Constantinople, in the early eighth century, finds a competent usurper, Leo III, in charge, quite possibly because elites knew the attack was coming. Kaldellis sees little luck in the defense: walls, fleet, Greek fire, starvation of the besiegers in contained zones, and Bulgar allies who had only recently settled south of the Danube. Leo rules another twenty years and founds what is called the Isaurian dynasty, though the family was not from Isauria.

What follows is iconoclasm, the fight over whether sacred images should be venerated or destroyed, the reason Orthodox churches today are covered in icons, which, Kaldellis notes, is not how Constantinople's churches originally looked. He puts a hard asterisk on the drama. In his research, ordinary people showed no interest. A few emperors, bishops, abbots, and monks, perhaps a hundred or two hundred people, cared intensely, and they wrote the texts. He could not find the people of Constantinople rising to demand icons on or off. When emperors changed policy, almost everyone fell in line. The pro-icon side won, and their view filled the buildings, literally, down to the present. The intensity a reader feels is the intensity of the small group that left the paper trail.

Kaldellis wrote *Streams of Gold, Rivers of Blood* to see whether he could write narrative rather than analysis; its success led to the thousand-page *The New Roman Empire*. The book covers a turn back to conquest after 955 under new military leadership, in Cilicia, northern Syria, Cyprus, the Caucasus, and then Bulgaria, followed by a generation of prosperity, then a triple attack: Seljuk Turks from the east, Pechenegs over the Danube, Normans in southern Italy. Traditional accounts blame moral decay and civilian emperors who neglected the army. Kaldellis's first instinct is the opposite of the decline-and-fall reflex: a society without obvious structural collapse is hit by three different enemies it could not reasonably have prepared for. The Normans bring a grinding new Western style of war and, in his word, terrorism; the Pechenegs are nomadic horsemen; the Seljuks are horsemen backed by Central Asian manpower who terraform farmland into pasture. Internal trouble exists too, especially the end of a two-century Macedonian dynasty, childless old emperors buying support with exemptions and titles, then a budget crisis as armies have to be raised. The Romans survive by the skin of their teeth and rebuild, as they always have.

He does not see an oscillation so much as a short list of exogenous shocks: the 630s, the 1070s, 1204. Each causes massive territorial loss; none lasts that long; each time except the last, the Romans figure out how to hold a line and put society back on a path of slow growth. The question he wants his field to answer is whether to define the society by those brief outside disasters or by the centuries of endogenous consolidation. His answer is the latter. Nobody came from outside to help them rebuild. The point of no return is the early fourteenth century, somewhere between 1300 and 1350. Loss of Asia Minor around 1300 ends the two-wing strategy; civil wars, Serbian expansion, and the Black Death follow. In 1300 the state is a Balkan ribbon from Constantinople to the Adriatic, not much smaller than England and still raising plenty of gold. Then a rebel trades half the territory to the Serbs. After that, Kaldellis says, game over, even though the Turks take another century, delayed when Timur smashes their empire around 1400. By the 1340s it is finished. The cause is foreign invasion, nothing else. Lose often enough and there is nothing to fall back on.

Absent invasions it could not absorb, he is convinced the tax system, the way policy was made, and the way representation worked could have run another thousand years. The negative evidence is the point: no separatist movements among Roman provincials, except Bulgaria, conquered in 1018 and independent again in 1185; the center never loses the ability to tax; no widespread peasant war; no partition, as the Franks partitioned after Charlemagne; no warlord carving out a private state. Britain, Portugal, and Japan lasted in corners and on islands. This society sat "right in the middle of one of the main corridors of empire building and new religions in the world," the most dangerous neighborhood on the map. His two-part explanation is an extraordinary effort to persuade subjects they were being ruled on their behalf, mostly matched by action, plus a tight identity as Romans and Orthodox who did not want to live under people who were neither. They complained about taxes constantly, because complaining cost nothing and sometimes worked. Sufficient, not idyllic, is the word he wants.

The modern lessons are institutional and Machiavellian. Invest in arrangements that work for most subjects, explain why they cost something, and think in centuries; the Romans built roads and also the Christian church to last. Narrow the gap between rhetoric and action. In the society he studies, the army, the largest expense, mostly protected subjects, with occasional glory raids. American leaders, he says, proclaim breathtakingly lofty goals, including the claim that Afghanistan was a feminist war, while war technology is profitable; most of the world does not believe the stated aims. Vietnam forced the United States off conscription and delayed the next large war; losing a war in Constantinople brought criticism, not the charge that the army was being used for something else. He does not want to be ruled by emperors. Democracy's cost is ejecting leaders and their experts every few years. On human nature he sides with Thucydides against a postmodernism that treated history as incommensurate slices: people love, hate, scheme, and fail in every group of twenty or thirty; culture turns the dials from one to ten. A fourth phase of history after hunter-gatherers, agriculture, and industry will come, and people in it will still recognize families, armies, and burial of the dead. He is neither sunny nor despairing. He grew up in the 1980s expecting nuclear war while pop culture stayed cheerful, then watched the 1990s get gloomy after the Cold War ended. Mood and situation do not always match. Many problems, he thinks, are solvable if elites can be reined in.