Peter Turchin's Cycles of History

[lpetrich]

Peter Turchin and Sergei Nefedov have published a book, Secular Cycles, in which they propose that large-scale human societies have cycles of growth and decay. He also has a popularized version: "War and Peace and War: The Rise and Fall of Empires" .

That's not a new sort of thesis, but they back it up with lots of evidence, like population, average height, and even counts of hoards of coins. In some premodern societies, people would bury their money, and since it was coins of precious metals, it would not corrode. If they got caught up in civil strife, like getting killed in it, then they would not be able to dig up those coins again.

He proposes this cycle:

• Integrative - centralized, unified elites, strong state, order, stability -- wars of conquest against neighbors
  • Expansion (Growth) - population increases
  • Stagflation (Compression) - population levels off, elites increase
• Disintegrative - decentralized, divided elites, weak state, disorder, instability -- civil wars
  • Crisis (State Breakdown) - population declines, elites continue, lots of strife
  • Depression - population stays low, civil wars, elites get pruned
• Intercycle - if it takes time to form a strong state

Era	Overall	Expansion	Stagflation	Crisis	Depression
Rome: Republic	350B - 30B	350B - 180B	180B - 130B	130B - 30B
Rome: Principiate	30B - 285	27B - 96	96 - 165	165 - 197	197 - 285
England: Plantagenet	1150 - 1485	1150 - 1260	1260 - 1315	1315 - 1400	1400 - 1485
England: Tudor-Stuart	1485 - 1730	1485 - 1580	1580 - 1640	1640 - 1660	1660 - 1730
France: Capetian	1150 - 1450	1150 - 1250	1250 - 1315	1315 - 1365	1365 - 1450
France: Valois	1450 - 1660	1450 - 1520	1520 - 1570	1570 - 1600	1600 - 1660
Russia: Muscovy	1460 - 1620	1460 - 1530	1530 - 1565	1565 - 1615
Russia: Romanov	1620 - 1922	1620 - 1800	1800 - 1905	1905 - 1922

Dates: B = BCE, (none) = CE

Here are the phases for the dominant states in western Europe until the late 19th cy. First Rome, then medieval German empires, then France.

Era	Integrative	Disintegrative
Republican Rome	350–130 BCE	130–30 BCE
Principate	30 BCE–165 CE	165–285
Dominate/Merovingian	285–540	540–700
Carolingian	700–820	820–920
Ottonian-Salian	920–1050	1050–1150
Capetian	1150–1315	1315–1450
Valois	1450–1560	1560–1660
Bourbon	1660–1780	1780–1870

[/td][/tr]

Here also for imperial China.

Dynasty	Integrative	Disintegrative
Western Han	200B - 10	10 - 40
Eastern Han	40 - 180	180 - 220
Sui	550 - 610	610 - 630
Tang	630 - 750	750 - 770
Northern Sung	960 - 1120	1120 - 1160
Yuan	1250 - 1350	1350 - 1410
Ming	1410 - 1620	1620 - 1650
Qing	1650 - 1850	1850 - 1880

In their book, PT and SN have counts of "instability" events, and they fit very well. European nations have 0.6 +- 0.06 instability events per decade in their integrative phases and 3.8 +- 0.5 in their disintegrative phases. For imperial China, the numbers are 3.4 and 11.3.

There is also a shorter-term cycle of violence that's apparent during the European disintegrative phases and a bit during some European integrative phases. It's not very apparent in China, however. It has a period of about 50 +- 10 years. PT calls it a "fathers and sons cycle", and he proposes a generational explanation. One generation revolts against perceived injustice, and its descendant generation does not think that the injustices it suffers is worth fighting about. But that's a more distant memory for the next generation, and its injustices may be worse. So this second generation revolts again, completing the cycle.

PT and SN do not pretend that this is a complete explanation; they note some events that don't fit in very well, like the mid first century crisis in the Roman Principiate, complete with a year of three emperors.

 

[Duke Leto]

Is this a joke? I'm sorry but I find the entire thing such a laughable pile of bullshit that it astounds me it could be published.

It's fucking around with terms that have different usual meanings than what they are using it for (ie. stagflation). I sense a great likelihood that they are going to try and apply this paradigm to the "stagflation" period in the USA from 1970-1984 and it will be a ridiculous shoehorn.

Looking at the major classifications of periods a fucking cat drawing notches on a timeline could have come up with something more meaningful.

The Late Roman Empire and the Merovingian Kingdom are charted in the same "Integrative" period? Are these guys on crack or just illiterate? The collapse of the Western Roman Empire is shoved in a period of "overall integration". Are you serious? The actual Valois dynasty started at the beginning of the Capetian disintegrative period. Are Spain and England included in these periodizations?

Without seeing a source table from where they are itemizing their "instability events" it's impossible to make a statement as to whether or not those statistics he cites are valid, and given that the periods have no evident periodicity to them, it looks like they just drew lines around events to get those averages.

It looks to me like this is the greatest case of theoretical overreach in Historical Analysis since Marx.

There definitely ARE macroeconomic cycles in the long term as population interacts with resources and ecology, but you're not going to get anywhere just labeling some periods integrative and disintegretive, you have to know what cause of changes in resource availability and ecology were. You can't expect to chart the rise and fall of empires like the orbits of the planets.

 

[lpetrich]

Secular Cycles - Peter Turchin, Sergey A. Nefedov - Google Books has much of that book's first and last chapters. You'll get a general idea of what he's arguing from there; he uses lots of statistics, like population figures and counts of strife events. Table 10.2, book page 310 gives the number of "instability events" for each phase of 7 of the cycles discussed in the book, and table 10.3, book page 311 does for for Imperial China.
Dynamics of political instability in the United States, 1780–2010 has graphs for ancient Rome and medieval France on page 2, and Imperial China on page 3.

As to the Roman Empire's decline being in an "integrative" phase, it was an integrative phase for the Eastern Roman Empire, better known to us as the Byzantine Empire, even if not for the Western Roman Empire.

 

[boneyard bill]

Secular Cycles - Peter Turchin, Sergey A. Nefedov - Google Books has much of that book's first and last chapters. You'll get a general idea of what he's arguing from there; he uses lots of statistics, like population figures and counts of strife events. Table 10.2, book page 310 gives the number of "instability events" for each phase of 7 of the cycles discussed in the book, and table 10.3, book page 311 does for for Imperial China.
Dynamics of political instability in the United States, 1780–2010 has graphs for ancient Rome and medieval France on page 2, and Imperial China on page 3.

As to the Roman Empire's decline being in an "integrative" phase, it was an integrative phase for the Eastern Roman Empire, better known to us as the Byzantine Empire, even if not for the Western Roman Empire.

I think Duke Leto makes a pretty strong case for this being largely a work of mental masturbation. They put the English Civil War right in the middle of the Tudor supremacy! The Stuarts, the House of Orange and the House of Hanover are still all part of this Tudor elite, not to mention Cromwell and the rising power of Parliament. The periods chosen for some of these cycles strike me as simply arbitrary.

Certainly, you can't study history without empirical evidence but trying to systematize all of this into a few statistical models in way too simplistic. Guess what? When countries fight wars, people get killed. So should we count the percent of the population that got killed and decide that a war was going on? Of course not, there's more to defining a war than counting dead bodies. After all, some wars have relatively few casualties.

Correlation does not prove causation. That is the Achilles heel of empirical methodology. At some point you have to apply logic. You have analyze the data and decide which was causative and which was effect. What developments were crucial and what developments were incidental.

 

[lpetrich]

They put the English Civil War right in the middle of the Tudor supremacy! The Stuarts, the House of Orange and the House of Hanover are still all part of this Tudor elite, not to mention Cromwell and the rising power of Parliament.

Let's see:

• The House of Stuart: 1567-1707
• The English Civil War: 1642-51
• Oliver Cromwell's rule: 1653-58
• The Glorious Revolution: 1688
• The House of Hanover: 1714-1901

Peter Turchin got the timing right: both the English Civil War and Oliver Cromwell's rule are in his Tudor-Stuart crisis period.

 

[Duke Leto]

I'm sorry, but if you are claiming there is a cyclical high level rhythm to history, and order to demonstrate this draw periods of non-uniform length to fit the oddities of individual countries and regions circumstances, than what you are doing is overfitting the data, and you do not, as it happens, have a sound theory. You have a series of ad-hoc rationalizations. Turchin is shamelessly overfitting.

More importantly, as noted above, this grand unified theory ignores the particular technological, ecological and macroeconomic factors in broad geographic areas in favor of a few master variables, so it's bloody well bound to get things wrong.

If an undergraduate or graduate history student had given this to me as a paper topic he'd have been unceremoniously flunked. (Had I stayed in academia.)

 

[boneyard bill]

They put the English Civil War right in the middle of the Tudor supremacy! The Stuarts, the House of Orange and the House of Hanover are still all part of this Tudor elite, not to mention Cromwell and the rising power of Parliament.

Let's see:

• The House of Stuart: 1567-1707
• The English Civil War: 1642-51
• Oliver Cromwell's rule: 1653-58
• The Glorious Revolution: 1688
• The House of Hanover: 1714-1901

Peter Turchin got the timing right: both the English Civil War and Oliver Cromwell's rule are in his Tudor-Stuart crisis period.

I wasn't arguing about what he said. I was arguing that it doesn't make any sense.