The Ghosts of Empires Past



Join 36.9K other subscribers

Apparently my blog posts on the historical roots of European dysfunction (see for example the last one, Visualizing Values Mismatch in the European Union) were noticed. I was approached by people who run Euromind and asked to contribute a short article to the edited volume they are putting together, titled Do Europeans Exist?  I did that and you can see it here:

Deep Historical Roots of European Values, Institutions, and Identities

In my previous post, I used data from the World Values Survey (WVS) to make a point about values mismatch that ensued when the core six countries, which originated the European Union in 1957, grew to the current 28 countries (soon to be 27 if Brexit really happens). Recollect that WVS has been collecting data on people’s beliefs in many countries since 1981. Researchers discovered that much of variation between populations of different countries can be mapped to just two dimensions: (1) Traditional values versus Secular-rational values and (2) Survival values versus Self-expression values. When values for each country in the sample are plotted in a two-dimensional space defined by these two axes, we have what is known as the Inglehart–Welzel Cultural Map.

Well, for the article that I contributed to Euromind, I wanted to explore the question of where do different values come from. Do they have deep historical roots? The argument I made was as follows:

Historical experience of living in the same state often results in the spread of common values, institutions, and identities among initially diverse groups. Elements of culture, including those that affect cooperation, change slowly, and often persist for long periods of time after the original empire has broken apart.

I then took the WVS data for European countries from the latest survey, and color-coded them by shared history within past states: the Carolingian, Habsburg, Ottoman, British, and Russian Empires. “Nordic” refers to the Danish and Swedish Empires (since Denmark at some points in historical time included Norway, Iceland, and a part of Sweden, while Sweden included Finland). Here’s what I got:

wvsAs the figure demonstrates, modern countries, which belonged to the same past and long-gone empire, cluster very closely together. There is little overlap. And when there is, it may reflect the influence of even more ancient empires. For example, Spain, Italy, Greece, and the Balkans were all core regions of the Roman Empire, and that’s where the most significant overlap (between the Habsburg and the Ottoman Empires) occurs.

This is really a remarkable result.

Notify of
Most Voted
Newest Oldest
Inline Feedbacks
View all comments

Why is Portugal coded as Habsburg? Union with Spain occurred in 1580, if memory serves, while the Austrian and Spanish portions of the Hapsburg dominions were split when Charles V abdicated a quarter century earlier.


60 years (1580-1640).


I like this approach, but can we say that the empire-clumping explains any more of the variance than say, geographical coding or other clumpings e.g. by dominant religion? Seems like we should be able to compare multiple multi-level regression models and see which carvings do the most explanatory work. Also, would be interesting to treat time under different empires as a variable, ie small amount of Hapsburgness for Portuagal, liberal doses of both Ottoman and Byzantium for the Balkans etc. The statistics would get gnarly but it’d be an interesting test of the ‘both were part of the (western) roman empire’-type anomaly-explanations here.


i have two question for you
1) it’ s is possible to make traslation in italian of same of your post for a not profit blog?

2) if i understand correcly your theory one of the trigger of a distructive cycle for agrarian societies is a malthusian trap. according to Dr Luois Arnoux the world net energy pro-capita growth by 321% after world war 2 and was 18GJ/head in around 1973, (actually 40GJ/head for the industrialised fraction of the global population at the time) now we have 4GJ/head . so it’ s possible that for industrial societies the malthus theory work also but with energy in place of food?


Food is a necessity for people to survive, energy isn’t. Lack of energy might lead to social upheavals but it’s not the same thing as the malthusian trap unless the energy shortage leads to food shortage and starvation.

And currently we are of course nowhere near a situation where energy per person would be crucially limited. There’s no link between population growth and energy availability either as far as I see.


Food is made from fossil fuels (oil for tractors and transportation, NG for fertilizers). Less fossil energy = less food, simple as that.

Loren Petrich

Energy works rather differently than food. I’ll give a very simplified historical summary.

Before the Industrial Revolution, nearly all of humanity’s energy sources were renewable ones: metabolic, biomass, wind, and water. The first two were limited by the amount of arable and grazable land, thus producing the aforementioned Malthusian constraint.

But existing renewable sources were insufficient to power the machines developed in the Industrial Revolution and their successors, so fossil fuels like coal and oil and natural gas were developed. Water was still used to generate electricity where it was practical to do so. But fossil fuels have a constraint. They are being consumed at something like a million times faster than they are formed, and I mean mean million literally. They are essentially long-ago biomass, mainly deposited over the last 300 million years or so. So unless we develop alternatives, industrial civilization will be kaput.

But we have done some of that, and we have made impressive progress with wind turbines and solar energy. Especially photovoltaic cells, which I’d thought of as a long shot before a few years ago. So humanity may well have a second age of renewable energy, one much less constrained than preindustrial sources.

Jesse Andrews

Regional splits would be great with this data. I would imagine Italy and to a lesser extent Spain would have a significant north/south split.


in italy local “original” population in northern regions are around 50%: the others came from southern italy and 20% from outside Italy: so if we keep in accunt only “original” local population maybe…but…


An interesting work on the north – south divide in Italy and how it affects how democracy works there, is Making Democracy Work by Putnam. Fascinating book on how the past affects the present, and how shared values affect how political institution work.

Loren Petrich

I think that it would be very interesting to evaluate different regions of the United States with tests like this one. Journalist Colin Woodard has written some very interesting books on American regionalism, “American Nations” and “American Character”.

In his first book, he argues that they have lots of differences in values, and that this was a result of which European settlers and settler descendants got there first. Later settlers tended to settle in regions that they were compatible with.

In that second book, he argues that an important struggle is between individual liberty and the common good. He points to Communism as common-good concern taken to extremes, and he points out the antebellum South as the opposite. Like the slaveowner who said about the Confederate Army requisitioning his crops that that made him feel like a slave.

Loren Petrich

To add to my previous comment, look at some recent election maps showing pro-Democratic and pro-Republican areas, and compare them to how the nation split during the Civil War. The Union states are nowadays mostly pro-Democratic, and the Confederacy states are nowadays mostly pro-Republican.


Even better: Look at the 1896 election map.
Even though they are pretty different, in some way, William Jennings Bryan and the Populists are pretty similar to Trump and current-day Trumpists.

Ross Hartshorn

I like it, but I agree with Carl that it would be good to see how much this improves on the results one might predict just based on geography. I do notice, however, that Poland is nowhere near Germany or Ukraine or Belarus on the values map, so it’s not all geography. Ireland and Iceland don’t look particularly similar, either. A more quantitative comparison of the two methods would be good, though.


Interesting to note that Poland is a corner of a Catholic bloc (Poland-Portugal-Spain-France-Belgium-Ireland with Italy in the middle) that is on the lower side of the Habsburgs and Carolingians and to the right of the Ottomans and left of the British,

Andrew Lancaster

For the results to be remarkable you should show that your empires match real history which they do not (e.g. Portugal), and secondly that the model things better than simple things like language groups or geographical regions or recent history such as length of time a country has been in the EU, or in the communist block etc.

Indeed, there was never a Nordic empire. It is simply a geographical region.

(To say that Nordic countries in various combinations were sometimes in various empires together, is true of any group of countries in any part of Europe. But even in this case, Russia would also have been in some of those “empires”.)


This is quite remarkable! Bowles comments on this phenomenon in ‘Microeconomics.’ I wonder how robust the results are to correcting for spatial auto-correlation (e.g. via the Dow (2007) and Eff and Dow (2009) models)


Perhaps it is time for America to admit that repudiating the Brest-:Litovsky treaty was wrong, since what Woodrow Wilson did was turning two strong countries into a dozen failed states.

Loren Petrich

That would have to have been the Treaty of Versailles. The Brest-Litovsk treaty was an earlier one, in which Lenin agreed to let the Central Powers have just about everything west of Russia’s present borders.

As to two strong countries becoming a dozen failed states, which ones do you mean? I’m guessing Germany and Austria-Hungary.

Germany lost a little bit of territory, but otherwise remained intact.

Austria-Hungary got broken up, but it was on the way to being broken up anyway, with its different nationalities wanting to become independent. Hungary had become autonomous, and before he was assassinated, Archduke Franz Ferdinand was working on making more nationalities autonomous, making A-H a federal state.

So Woodrow Wilson was not to blame.

giorgios papadopoulos

What i understand from the cultural map is that Greece as society, not as state, continues to be a society of the measure, rationalism and feeling toghether , heart and brain

  1. Home
  2. /
  3. Cliodynamica
  4. /
  5. Regular Posts
  6. /
  7. The Ghosts of Empires...

© Peter Turchin 2023 All rights reserved

Privacy Policy