Because the emperorship was not an office, nobody could inherit it, and every accession for three centuries was therefore an improvisation whose details reveal who actually held the power to decide.
The previous lesson ended with Augustus dead at Nola and Tiberius receiving powers that the Senate could not name. This lesson follows what happened at each of the next handful of transfers, because the sequence answers a question the constitution never addressed on paper. When an emperor died, three groups could plausibly claim a voice: the Senate, which had voted the powers; the praetorian guard, which was in the room; and the legions on the frontiers, which had the weapons. Between AD 14 and AD 193 all three tried, and the results were not close.
The family as a procedure
The first four successions were solved by blood, and by an ambiguity that suited everyone: if the ruler is not an officeholder but simply the greatest man in Rome, then his heir is a private matter, and private matters were settled in Roman law by wills and adoption.
Tiberius, adopted in AD 4, already held tribunician power and a proconsular imperium when Augustus died, so the transfer looked like a continuation. Gaius, whom the army had nicknamed Caligula, came in through Tiberius's will and the backing of the praetorian prefect. When Caligula was murdered in his palace in January 41 by officers of his own guard, the Senate met on the Capitol and debated restoring the Republic, and while it debated, the praetorians found Caligula's uncle Claudius in the palace, took him to their camp, and proclaimed him. Nero, thirteen years later, was produced at the camp gate by the prefect Burrus while Claudius's death was still unannounced and Britannicus, Claudius's own son, remained inside the palace.
Four transfers, none of them illegal, none of them following a rule, and only the first arranged in advance. Notice the direction of travel. The Senate is present at every one and decides none of them.
The price of a coronation
Claudius, who owed his position to the men who had carried him to the camp, paid them for it: Suetonius says 15,000 sesterces to each praetorian, the first time a Roman ruler had openly bought his accession from his own guard.
Example. A legionary was paid 900 sesterces a year. Work out what 15,000 sesterces represents, and what the transaction established.
Sixteen and two thirds years of a legionary's pay, handed over in one payment, to men who were already the best paid soldiers in the empire. As a bribe it is enormous; as a precedent it is worse. Two things were now true. First, a donative on accession had become the expected price, so every later emperor had to find that sum in his first week, whatever the state of the treasury. Second, and this is the part that mattered, the praetorians had learned that the death of an emperor was the most profitable event that could happen to them. An institution that profits from a vacancy will eventually create one, and within a century and a half the guard was doing precisely that. The Roman state had accidentally attached a large cash prize to regicide.
Now you. In 193, after the praetorians killed the emperor Pertinax, they announced that the position would go to whoever offered most, and Didius Julianus won by promising 25,000 sesterces a man. What does the auction actually demonstrate?
Answer
That the mechanism Claudius started had reached its logical end, and also that the guard had misjudged what it was selling. The auction is genuinely reported by Cassius Dio, who was a senator in Rome at the time and watched it, so this is not a hostile legend from a later century. What it demonstrates is that by 193 the praetorians believed the empire was theirs to sell, and what happened next demonstrates the limit of that belief: Septimius Severus marched on Rome from the Danube with sixteen legions, Julianus was executed after sixty-six days, and Severus dismissed the entire guard, stripped the men of their uniforms and re-formed it from his own legionaries. Nine thousand men in one city can choose an emperor only while no larger army objects. That correction had already been made once, in the year the next section is about.
The year of four emperors
Nero was the last of the Julio-Claudians, and when he lost the loyalty of the Spanish and Gallic commands in 68 there was no relative left to produce. He killed himself in June 68. What followed is the single most instructive year in Roman political history.
Galba, governor of nearer Spain, aged seventy-two, was proclaimed by his legions in April 68 and recognised by the Senate after Nero's death. He then made the one economy his position could not afford, refusing to pay a donative, with the line that he levied his soldiers and did not buy them. It was the correct constitutional sentiment and it killed him: on 15 January 69 the praetorians cut him down in the Forum and proclaimed Otho, who had been passed over as his heir.
Otho lasted ninety-one days. The armies of the Rhine had already proclaimed their own commander, Vitellius, on 2 January, before Galba was even dead, and their advance guard beat Otho at Bedriacum in northern Italy in April. Otho killed himself. Then, on 1 July, the prefect of Egypt had his legions swear to Vespasian, who commanded the war in Judaea; the Danube armies joined him, beat Vitellius at the second battle of Bedriacum in October, and took Rome in December in fighting that burned down the temple of Jupiter on the Capitol.
Example. List where each of the four was proclaimed and say what the pattern shows.
Galba in Spain, Vitellius on the Rhine, Vespasian in Egypt with the eastern and then the Danubian armies behind him, and only Otho in Rome itself, by the guard. Three of the four were made in frontier provinces by the men they commanded, and the one made in the city lasted three months. Tacitus states the conclusion in a single clause: the secret of empire had been let out, that an emperor could be made somewhere other than Rome. It is worth being precise about what the secret is. Not that armies mattered, which everyone had known since Sulla, but that the legal machinery in Rome would ratify whatever the strongest army had already decided, and that the decision could therefore be taken on the Rhine. From 69 onwards every serious political actor in the empire knows this, and the question of who commands which frontier becomes the central question of imperial politics.
Now you. Vespasian's soldiers proclaimed him on 1 July 69, but his reign was later dated from that day rather than from the Senate's recognition in December. Why does the choice of date matter?
Answer
Because it says where the reign was understood to begin, and the answer is the army camp, not the Senate house. The formal position was that an emperor's powers came from a vote of the Senate and people; the practical position was that the vote registered an accomplished fact, and the dating of the reign to the acclamation admits it in the official record, on coins and in documents, for anyone who cared to notice. Later emperors kept the same convention. It is a small piece of evidence and a solid one, because dating formulae are administrative rather than rhetorical: nobody was trying to make a point, which is why the point is trustworthy.
The nearest thing to a constitution
Vespasian's accession produced the one document that comes close to defining the position, and it survives because a bronze tablet was reused as an altar and rediscovered in the fourteenth century.
The lex de imperio Vespasiani is a law of AD 69 or 70 conferring the emperor's powers, and it works by citation: he is to have the right to make treaties as Augustus, Tiberius and Claudius had it, to convene the Senate as they had it, to extend the boundary of the city as Claudius had it, and so on through a list of precedents. The final surviving clause is the striking one. It authorises him to do whatever he judges to be in accordance with the advantage of the state, and declares valid anything done before the law was passed.
Two things follow. First, by 70 the position had become definite enough to legislate about, which it had not been in 14. Second, the definition is a list of what previous holders had done plus a clause of unlimited discretion, which is a description of custom rather than a constitution. That is what the Roman Empire had instead of a founding law for its entire existence.
Adoption, and why it looked like a principle
For eighty-four years after 96 the empire had five rulers, each of whom adopted his successor, and none of whom was succeeded by a son. Nerva, an elderly senator installed after the murder of Domitian, adopted Trajan, who commanded the army of Upper Germany, under pressure from the guard. Trajan's own adoption of Hadrian was announced by his widow as he lay dying in Cilicia and was doubted at the time. Hadrian chose Antoninus, requiring him to adopt in turn both Marcus Aurelius and Lucius Verus. Marcus and Verus then ruled jointly, the first time the position had been formally shared.
Gibbon called this the happiest period in the history of the world, and later writers have called the arrangement the adoptive principle, as though the Romans had solved succession by selecting on merit.
Example. Test that claim against the family circumstances of the five emperors.
Nerva had no children. Trajan had none. Hadrian had none. Antoninus Pius had two sons, both of whom died before he became emperor. Marcus Aurelius had a son who survived, Commodus, and he made him co-ruler at sixteen and successor at eighteen. The adoptive principle turns out to be a run of five rulers who had no surviving sons, followed immediately by one who did and who behaved exactly as a dynast. There is a real point underneath the correction: adoption was a genuine Roman legal instrument, respectable and normal, and it let the empire recruit its rulers from a wider pool than a bloodline would allow, which is why the second century looks well governed. But it was never a rule, no emperor ever declared it one, and it collapsed on first contact with a living heir. Beware of any pattern in Roman history that is visible only in a period when the alternative was unavailable.
Now you. Marcus Aurelius is often blamed for Commodus. Was the choice avoidable?
Answer
Barely, and the reason is structural rather than moral. A reigning emperor with an adult son faces this: if he passes the son over, he creates a rival with the best possible claim, around whom every disaffected officer in the empire can gather, and probably a civil war on his death. Passing over your own son is only safe if the son can be got out of the way, and the successful adoptive emperors were spared that decision by not having one. Marcus could see what Commodus was, and he could also see that the alternative to giving him the position was leaving him alive without it. This is the empire's permanent trap: it is a monarchy without a law of succession, so it can neither exclude an heir nor guarantee a good one, and the same feature that let it choose Trajan forced it to accept Commodus.
What the ritual settled
By the second century the making of an emperor had a shape: acclamation by soldiers, a donative paid to them, recognition by the Senate, and a vote of powers modelled on the tablet Vespasian's lawyers drew up. The order of those steps is the whole story, since the vote comes last and confirms rather than chooses.
That is not a small thing. A regime in which the army chooses and the civil authority ratifies is stable exactly as long as the army is unified, which for a century it broadly was: there was one civil war in 69, one in 193, and long stretches with none. It becomes lethal when several armies of comparable size disagree, and the third century will show what that looks like.
Which makes the army the institution to understand next. It decided who ruled, it consumed most of what the state collected, and it was the only Roman organisation present in every province. The next lesson is about what it actually consisted of: how many men, on what terms, standing where, and paid how.