A theory says that history moves in a circle of roughly eighty to ninety years, divided into four quarters of about twenty. The last quarter is the rupture: the old order is torn down and a new one built, usually through war or revolution. This chapter draws that circle, shows all seven turns of it the authors claim, tests the arithmetic against their own dates, and asks the question most people come here with: we are eighty years after 1945, does that mean anything?
The word is old. The Romans called a saeculum the span of the longest life in a cohort, the time until everyone who personally remembered an event had died. William Strauss and Neil Howe took that word and gave it a mechanism: a saeculum of roughly eighty to ninety years, divided into four turnings of about twenty, each one shaped by the generation that grew up in the one before.
They describe it as the seasons of a year. The High is spring: institutions are strong, individualism is weak, the country is confident and conformist after the last rupture. The Awakening is summer: those same institutions come under attack in the name of personal and spiritual autonomy. The Unraveling is autumn: institutions are weak and distrusted, individualism flourishes, society atomises. The Crisis is winter: institutional life is torn down and rebuilt under a threat that everyone perceives as existential.
The engine is the hand-over between generations. Four archetypes follow one another in fixed order, and each is formed by the season it grows up in and forms the next season once it holds power. In this reading the circle closes at the moment when nobody in a position of authority personally remembers the last winter, which is roughly when the generation that lived through it has died out.
Each quarter is drawn to its real length, not to a tidy ninety degrees. Pick a cycle, then click a quarter.
The dashed marks at three, six and nine o'clock show where four equal quarters would end. Where a quarter reaches past its mark, that turning ran longer than the model says; where it stops short, it ran shorter. The circle is read clockwise from the top, and it closes where the last winter ends.
Each row is one archetype through a whole saeculum, beginning in the season it is born into. Read the last column: it is the whole theory in four cells.
The row runs across roughly eighty-five years, so its last cell already belongs to the next circle. Strauss and Howe derive the archetype purely from the year of birth, nothing else.
Grows up materially secure and morally certain, in a world that has just rebuilt itself and does not doubt its own institutions. That security is the precondition for what comes next: as a young adult in the Awakening this is the generation that turns on those institutions in the name of inner truth and personal autonomy.
Grows up under-protected while the adults are busy attacking institutions and finding themselves. Learns early that nobody is coming, becomes practical, sceptical and individually competent, and spends young adulthood in an Unraveling that rewards exactly that.
Grows up heavily protected in a society that is coming apart, because adults over-invest in children precisely when institutions are failing. Comes of age as the crisis breaks and does the collective work of it: builds, fights, mobilises, and pays the price in numbers.
Grows up over-protected in a dangerous time, learns to read the room before learning to speak in it. Comes of age inside the new order that others built and spends a working life refining, mediating and widening it rather than founding anything.
Strauss and Howe apply the pattern to seven saecula of Anglo-American history, from the late Middle Ages to the present. Each ring below is one of them, and each quarter is drawn to the length the authors themselves give it. The winter of each circle is in charcoal.
Read the shapes, not the labels. If the theory were as regular as its summary suggests, the seven rings would look alike. They do not. The first full circle takes 107 years, the Civil War circle takes 71, and its winter is a mere five years wide while the one before it ran twenty-one.
Charcoal marks the Crisis. The Late Medieval circle is incomplete in the authors' own table.
The point the theory rests on is that the last quarter is never quiet. In the Reformation circle the winter is the Armada Crisis of 1569 to 1594, the confrontation between Elizabethan England and Habsburg Spain. In the New World circle it is the run from King Philip's War to the Glorious Revolution, 1675 to 1704, which ends with a new constitutional settlement. In the Revolutionary circle it is the Age of Revolution, 1773 to 1794, from the Boston Tea Party to the young republic.
Then the pattern gets rougher. The Civil War winter lasts five years, 1860 to 1865, and the authors themselves call that saeculum an anomaly in which a generation was skipped. The next winter, the Great Depression and the Second World War, runs seventeen years from the crash of 1929 to 1946 and ends in the order most of us grew up inside: Bretton Woods, the United Nations, the welfare state, the Marshall Plan.
By the authors' dating the current winter began with the financial crisis of 2008 and should close around 2033. The events they file under it are the Great Recession, the pandemic and the war in Ukraine.
Measured with the authors' own dates the distance from the start of one crisis to the start of the next is not eighty years. It runs from 69 years, the Civil War to the Great Depression, to 106 years, the Armada to the Glorious Revolution. The Dependencies chapter charts those distances against the eighty-year claim.
The Late Medieval ring above is missing its High and its Awakening, and that is worth explaining because it is not a gap in history. Strauss and Howe's table simply begins in 1435. They start where they believe the Anglo-American thread they are following begins, so the first saeculum is truncated at the front, not incomplete in the world.
What lies immediately before, in English history, would fit the pattern loosely if one wanted it to. The Black Death from 1348 killed something between a third and a half of the population of Europe. From the 1370s John Wycliffe and the Lollards attacked the institution of the church in the name of personal conscience, which is what an Awakening is supposed to look like. The Peasants' Revolt of 1381 followed. Whether that is a High and an Awakening or simply the fourteenth century is interpretation, and it is ours, not the authors'. They make no claim about it, and we mark it as a reconstruction rather than adding it to their chart.
A useful test of any generational theory is whether its labels survive contact with real people. Below are figures from each of the crises in the table, with the year of birth, the generation Strauss and Howe put that year in, and the archetype that follows from it. The assignment is mechanical: the birth year decides, nothing else.
Read it as a test, not as a gallery. Some of the labels fit remarkably well. Others do not fit at all, and those are the more interesting ones.
Hindsight. Every name above was chosen because we already know how it ended. The same cohorts produced the people who lost, and far more people whose lives the crisis simply consumed. Richard III is in the list for exactly that reason: same year of birth, same generation, same archetype, the other side of the same battle.
The label is mechanical. The archetype follows from the birth year and from nothing else. That gives you Shakespeare as a Hero and Martin Luther King as an Artist, an adapter. Whoever takes the labels literally will find quickly that they do not explain individuals. At best they describe the position a cohort occupies when the crisis arrives.
The record is Anglo-American. Run the same arithmetic on Konrad Adenauer, born 1876, and you get a Prophet; on Ludwig Erhard, born 1897, a Nomad. Both rebuilt an order after 1945, so the assignment looks like it works. It says nothing about German history, which had its own ruptures in 1918, 1933 and 1945 that this table does not model at all.
The theory answers with a cohort, not a person, and that is either its honesty or its evasion, depending on how generous you feel. By its own table the Millennials, born 1982 to 2005, are the Hero archetype; in 2026 they are between 21 and 44 and are the young adults of a crisis that began in 2008.
It cannot tell you who the individuals will be, and neither can anyone else. In 1929 nobody was pointing at a 47-year-old governor with polio, and in 1936 nobody was pointing at a lieutenant colonel. Every name in the list above became obvious only afterwards, which is precisely why the list is not evidence.
It is also worth being clear that 'hero' here is a position in the cycle, not a moral rank. In every crisis above, the work was overwhelmingly done by people whose names are not recorded: the ones who kept the harvest in, ran the hospitals, held a line, or simply kept a family fed through a decade that had no intention of helping. The names are the visible tip of that.
Which makes the more useful question a personal one. Not who the hero is, but which position you occupy when it hits. The chart below is the theory's answer for the people alive right now.
Age of each living cohort in 2026, with the role the theory assigns it in the crisis.
This is the reading that carries the theory through social media, and it is worth taking seriously rather than waving away. It goes: the last great rupture ended in 1945, the cycle is about eighty years, so the next rupture is due about now. Everybody who remembers the war personally is gone; the brake is off.
What the theory actually says is close to that, but not the same, and the difference matters. Strauss and Howe do not date the next winter to 1945 plus eighty. They date the current winter to 2008, the financial crisis, and expect it to close around 2033. In his 2023 book Neil Howe puts it as passing through a great gate in history before the mid-2030s. On that reading we are not waiting for the rupture. We have been inside it for eighteen years.
So does 1945 plus eighty land anywhere meaningful? Test the rule the way it is meant: take the year the previous winter ended, add eighty, and see where that lands in the winter that followed.
Zero is eighty years after the previous crisis ended. Each bar is the crisis that followed, drawn from its start to its end, using Strauss and Howe's own dates.
The rule lands just before one crisis, inside three of them, and once, in 1874, nine years after the crisis had already ended. For the row we are living in it lands at about three quarters of the way through.
2008 to 2033 by Strauss and Howe, with Howe's own 2023 wording for the end.
The honest question is not whether a circle says a rupture is due. It is how tense the system measurably is, and those numbers exist.
Global debt stood at about $310tn, roughly 332% of world output, at the start of 2026, against $225tn before the pandemic; governments now account for around 40% of it, up from 35% in 2019. In the 2025 fiscal year the United States paid more than $1.1tn in net interest, more than it spent on defence, for the first time. In the same year the Uppsala conflict programme counted 65 armed conflicts, 13 of them wars and 8 conflicts between states, each the highest figure since the records begin in 1946, with about 244,600 battle deaths.
Every one of those is a condition, not a proof. They say the system is under strain; they do not say a clock is running. The difference matters, because a date invites waiting and a condition invites preparing.
The idea that an order forgets what it cost and then loosens until it breaks is not an American invention of 1997. It has been written down independently in at least four traditions, in different centuries and on different continents, and comparing them is the fastest way to see what is solid in the pattern and what is not.
What recurs across all of them is a mechanism, and it is always the same one: the people who remember the catastrophe die, the people who inherit the order do not know its price, and the order loosens until something breaks it. What does not recur is the clock. Every author who noticed the mechanism attached a different number to it, and the numbers are not close.
The sharpest early version, and the one closest to Strauss and Howe. In the Muqaddimah, written in 1377, the North African historian Ibn Khaldun builds his whole account of dynasties on asabiyya, the group solidarity that holds a people together and lets it take power.
His mechanism is generational and explicit. The first generation still carries the hardship of the desert: brave, cohesive, dangerous. The second has moved into settled comfort but still knows the first personally, so something of the old cohesion survives. The third has completely forgotten that there ever was a desert; solidarity is gone, and the dynasty cannot defend itself. He puts a generation at forty years and a dynasty at three generations, one hundred and twenty years, a few more or a few less.
That is the same engine six hundred and twenty years earlier, with a different number bolted to it. If the mechanism were as regular as either author claims, they would not disagree by forty years.
The Greek historian Polybius describes in Book VI of his Histories a cycle he calls anacyclosis: monarchy decays into tyranny, tyranny is replaced by aristocracy, aristocracy decays into oligarchy, oligarchy is overthrown into democracy, democracy decays into mob rule, and out of the chaos a single ruler rises again.
The driver is inheritance, exactly as in Ibn Khaldun: the founders earn their position, their children inherit it without having earned it, and abuse it. Polybius attaches no fixed length at all, which makes his version weaker as a prediction and more honest as a description.
The Chinese dynastic cycle is the same shape again, with a moral rather than a demographic engine. A founder receives the mandate of heaven, the dynasty consolidates and prospers, the court grows corrupt, taxes and population press on the land, natural disaster and famine arrive, rebellion follows, and heaven withdraws the mandate to give it to someone else. Mencius formulated the mandate in the Warring States period; the Han historian Ban Gu turned the cycle into the standard frame of Chinese historiography.
The lengths are the interesting part. Han, Tang, Song, Ming and Qing ran roughly 270 to 430 years each, three to five times the saeculum, and nobody in that tradition ever claimed they should be equal.
The longest record of the same shape belongs to Egypt, and it is worth looking at precisely because the numbers are so far from eighty. Egyptian history is conventionally divided into three long orders, the Old, Middle and New Kingdoms, separated by three Intermediate Periods in which central authority collapsed, local rulers took over, and in the second case foreign rulers, the Hyksos, took part of the country.
Order lasted between three hundred and five hundred years. Collapse lasted between one hundred and four hundred. The shape the theory describes is unmistakably there. The clock is not.
Conventional dating; figures vary between authorities. Charcoal marks the periods of collapse.
The same mechanism, described in four traditions, with numbers that are not close to each other.
Put the four side by side and the honest reading is hard to avoid. The mechanism is real and has been noticed independently by people who could not have read each other: a Greek in the second century BC, a North African in 1377, Chinese historians from the Han onwards, and two Americans in the 1990s. Something in the way societies forget is genuinely cyclical.
The numbers are not. One hundred and twenty years, eighty-five years, three hundred years, no number at all. If four careful observers of the same mechanism produce four incompatible clocks, the mechanism is the finding and the clock is the decoration. That is also why the useful conclusion in this chapter is about conditions and not about dates.
Take the theory at its word and measure it with its own dates. The gap between two crises runs from 69 years to 106 years, the mean is about 88, and the spread around it is roughly a fifth in either direction. On that scale a prediction is accurate to within a human working life, which is another way of saying it does not constrain anything.
The turnings themselves are worse. Crisis phases in the table run from 5 years, the American Civil War, to 29 years. Strauss and Howe handle that outlier by declaring the Civil War saeculum an anomaly in which one generation was skipped. That is the move a theory makes when the record disagrees with it, and it is the core of the criticism.
They are not fringe. Michael Lind called the predictions as vague as fortune cookies and the construction non-falsifiable. Interviewed by the New York Times in 2017, academic historians compared the book to astrology; David Greenberg of Rutgers called it crackpot. Jonathan Alter in Newsweek called the 1991 book an elaborate historical horoscope that will never withstand scholarly scrutiny. Frank Giancola found in 2006 that measurable generational differences are largely not borne out by empirical research. Eric Hoover objects that a single archetype per birth cohort is drawn from affluent suburban America and ignores class, income and origin. David Brooks, who likes the book, still called it not a good book when judged rigorously.
On the 2023 sequel, Francis Fukuyama in the New York Times called the thesis too neat and arbitrary, an updated Whig history. In City Journal, N. S. Lyons noted that Howe never says whose brighter tomorrow is coming, and devotes about four pages to the possibility that the outcome is bad.
And Peter Turchin, himself a defender of cyclical history, objects that the theory forces the record to fit by stretching in some places and cutting in others.
The whole table is Anglo-American. It is a history of England and then of the United States, and the turnings are drawn from that experience. The same years look entirely different from Vienna, Warsaw or Beijing. 1848 does not appear at all. 1918 and 1945 sit inside turnings defined by American politics rather than by the collapse of an empire on the Danube. For a company in Central Europe that is not a footnote: the crises that shaped this region are not the ones the circle is measuring.
There is a version of cyclical history that can fail, and the difference is worth knowing. Peter Turchin's structural-demographic theory names three measurable pressures instead of a mood: mass mobilisation potential from real wages relative to output, urbanisation and the share of young adults; elite overproduction, meaning more people trained and positioned for elite roles than there are roles; and state fiscal distress from public debt and trust in institutions. The three combine into a political stress indicator.
In 2010, when unrest in the West had been falling for a quarter of a century, that model forecast a rise in instability for the 2010s. Turchin and Korotayev published the retrospective in PLOS ONE in August 2020: anti-government demonstrations and riots in the United States and the United Kingdom had risen by roughly an order of magnitude after 2010. The forecast could have been wrong and was not.
That is the whole distinction. Turchin names indicators, a horizon and a way to be refuted. The Fourth Turning names a mood and a date range wide enough to accommodate almost any decade.
The circle is not a clock and should not be used as one. It is a decent description of something real, that societies forget and that forgetting has consequences, wrapped in an arithmetic that its own table does not support. Treat the mechanism as a question worth asking and the eighty years as a number that happens to be memorable.
Three books carry the theory, and a fourth carries the testable alternative. If you read only one, read the 1997 one for the argument and the 2023 one for what its author now claims.
The practical conclusion does not depend on believing in any cycle, which is what makes it worth having. Do not plan on a date. Plan on the conditions the theory describes and the numbers confirm: expensive money, contested trade, energy as a political price, and a decade in which the rules get rewritten.
Concretely, and each of these has a chapter of its own: a liquidity reserve measured in months of fixed costs, not in a ratio; a cost structure flexible enough to survive a bad year (→ Costs & leverage); dependencies you have actually mapped rather than assumed (→ Dependencies); no reliance on short-term refinancing and a clear view of supplier credit (→ Price & balance); and a realistic picture of what labour costs where (→ Work & taxes).
A theory that cannot be refuted cannot guide a decision either. The indicators can.
Presented as a contested theory, not as a description of the world. The dates and the event names are Strauss and Howe's own; the arithmetic on top of them is ours and can be recomputed from the table.