In 305 CE, the Roman Emperor Diocletian did something that no emperor before him had done voluntarily: he resigned. He then retired to a palace he had constructed for himself on the Dalmatian coast — in what is now Split, Croatia — and, by his own account, spent his remaining years growing cabbages.
Photo: Diocletian, via images.fineartamerica.com
This detail, preserved by the historian Lactantius, has delighted readers for seventeen centuries. The man who had reorganized the Roman Empire, persecuted Christians with systematic ferocity, and divided imperial authority between four co-rulers apparently found genuine contentment in horticulture. When his former colleagues later begged him to return and stabilize the empire, which was fracturing badly without him, Diocletian is said to have replied that if they could see the cabbages he had raised with his own hands, they would not ask him to give up such happiness for the storms of power.
The story is almost certainly too neat. But the institutional problem it represents is entirely real, and entirely ancient.
The Problem Has Always Been the Same
Every civilization that has produced centralized power has eventually produced the same uncomfortable question: what is the correct procedure for a leader who will not, or cannot, relinquish what they have built?
This is not a question about ambition, though ambition is always involved. It is a question about identity. Power, exercised over a long period, does not merely describe a person's role. It constitutes their sense of self. The Roman general who has commanded legions for twenty years, the Chinese emperor who has governed from the Forbidden City for three decades, the family patriarch who built a business empire across half a century — these are not people who happen to hold power. They are people for whom power and personhood have become functionally indistinguishable.
The institution's problem is that it needs the transition. The individual's problem is that the transition requires them to become someone they have never been and do not know how to be.
Rome's Recurring Crisis
Rome's constitutional history is, in substantial part, a record of failed attempts to solve this problem. The republic had mechanisms for managing the end of official tenure — consuls served for one year, dictators for no more than six months — but those mechanisms depended on the willingness of the person holding power to honor the limit.
When Sulla marched his legions on Rome in 88 BCE, he was not merely conducting a coup. He was demonstrating that the republic's term limits were social contracts rather than physical constraints, and that a man with sufficient force of arms could simply decline to honor them. Julius Caesar drew the same lesson forty years later, with more lasting consequences.
The imperial system that emerged from the republic's collapse was, among other things, an attempt to institutionalize succession in a way that did not depend on the outgoing leader's virtue. It failed repeatedly and spectacularly. The problem was not the lack of formal mechanisms. It was that no formal mechanism can compel a man who controls the army to leave.
Diocletian's retirement was remarkable precisely because it was voluntary. He had the power to stay. He chose to go. The empire celebrated this as wisdom and then spent the next two decades demonstrating that his successors had absorbed none of the lesson.
The Chinese Solution: Captivity by Another Name
The Chinese imperial system, which ran with various interruptions for roughly two thousand years, developed a more systematic approach to the retired-leader problem — one that acknowledged, with considerable pragmatic honesty, that voluntary withdrawal could not be relied upon.
The institution of the Taishang Huang, or Retired Emperor, formalized the transition of power while preserving the outgoing emperor's status, his court, his ritual privileges, and, crucially, his physical confinement within the palace complex. A Retired Emperor was honored in every external respect. He was also, in practical terms, contained.
The system worked tolerably well when the transition was genuine — when an aging emperor stepped back in favor of a capable heir and the two reached a workable division of ceremonial versus actual authority. It worked considerably less well when the Retired Emperor retained political ambitions, when factions at court aligned with him against the reigning emperor, or when the reigning emperor grew impatient with the arrangement and resolved it through means that the historical record generally describes with diplomatic vagueness.
The psychological insight embedded in the Taishang Huang institution is worth examining directly. The Chinese court understood that the problem of the powerful former leader was not primarily a military or legal problem. It was an attention problem. A Retired Emperor who had sufficient ceremony, sufficient court, sufficient ritual significance to occupy his identity did not need to reach for political power to feel like himself. The gilded cage worked because it was genuinely gilded.
The Florentine Variation
The great merchant dynasties of Renaissance Italy confronted a secular version of the same problem, and their solutions were strikingly similar in structure if not in scale.
Cosimo de' Medici, in his final years, withdrew progressively from the daily management of Florentine affairs while never quite relinquishing the informal authority that his decades of accumulated relationships represented. His son Piero inherited the business and the political position, but operated for years in the peculiar condition of being nominally in charge of an institution whose most important relationships still ran through his father.
Photo: Cosimo de' Medici, via www.toshelper.com
This pattern — the founder who delegates without releasing, the patriarch who retires without stepping back — recurs so consistently in the records of medieval and Renaissance merchant families that it begins to look less like individual failure and more like a structural feature of how human beings experience the transition from building to having built.
The founder's identity is invested in the act of construction. Once the thing is constructed, the identity requires either continued construction — expansion, reinvention, new projects — or it requires the slower, more difficult work of building a self that does not depend on institutional relevance. History suggests that the latter is genuinely hard, and that most people, given the choice, prefer the former.
What This Tells Us About Ourselves
The modern American anxiety around retirement, succession, and relevance — the sixty-five-year-old executive who cannot quite bring himself to hand over the keys, the founder who returns to the company he was supposed to have left, the political figure who stages comeback after comeback long past the point of genuine usefulness — is not a product of contemporary culture's discomfort with aging, though that discomfort is real.
It is the same institutional crisis that produced Diocletian's colleagues begging him to return from his garden, that produced the Taishang Huang system with its carefully constructed ceremonial substitute for power, that produced generation after generation of merchant patriarchs who could not distinguish between the business and themselves.
The psychological core of the problem has not changed. Power, exercised long enough, stops feeling like something you do and starts feeling like something you are. The institution's need for transition and the individual's need for continuity of self are not compatible, and no formal mechanism — not term limits, not mandatory retirement ages, not carefully drafted succession plans — has ever reliably bridged that gap.
What history offers instead is a long catalogue of workarounds: the ceremonial role, the advisory position, the emeritus title, the carefully maintained fiction that the transition has occurred when everyone in the room knows it has not. These solutions are imperfect. They have always been imperfect.
Diocletian's cabbages remain the most honest answer anyone has ever given to the problem. Build something beautiful in a different direction, and tend it with your whole attention. It is not a solution that scales. But it is the only one in five thousand years of the historical record that seems to have actually worked.
