Three commonwealths, thirteen facets
We identify the attributes of a peoplehood, measure them empirically, normalize them to a common scale, and cluster them. We test every result for robustness, and apply the method to European and European-derived populations.
Fifty-four regions, from Iceland to Russia and Israel, with the settler societies and Ibero-America, are scored on thirteen facets, each resting on a published measurement or a documented classification.
Every facet is received at birth; they differ in how they reach a person and what can happen afterwards. Ancestry is inherited in the body and cannot change, even in principle: measured as steppe, farmer and hunter-gatherer shares from ancient-DNA models, it is a fact of descent, which gene editing can alter in particular genes but not undo. Setting is the ecoregion of a region's homeland. It shapes peoples slowly and indirectly, through selection over millennia and through crops and kitchens over centuries, but it passes nothing directly to an individual. Eleven facets are transmitted, inherited by practice and kept only by practice, so they can be taken up by people not born to them and lost by people who were: language, confession, script, legal family, family and kinship, learning, civic trust, individualism, architecture, foodways and the festive calendar.
For every pair of regions, each facet gives a distance from 0 to 1. Facets are scaled to equal spread, and each facet's weight is discounted by how far it repeats the others, so that correlated facets share one voice. The combined distances are clustered by average linkage, which reproduces them with a cophenetic correlation of 0.84, and the result is tested against other weights, other linkage rules, perturbed and shuffled data, and held-out facets.
Results. The clearest division separates the two Hebraic units, Israel and the Ashkenazi diaspora, from all others. The next divides the remainder into a Latin Roman West of 36 regions and a Byzantine East of 16. The same three groups emerge with ancestry and setting given no weight. Built without any one facet and its close correlates, the groups still separate that facet's values for eleven of the thirteen facets, beyond what geographic distance explains; the exceptions are geographic descent and legal family. No shuffled dataset produces comparable structure (p < 0.005).
The cores are stable; the borders are not. Under 300 moderate re-weightings, each commonwealth overlaps its reference membership by 0.84 to 0.97 on average (Jaccard), but the exact three-way split recurs in only 88 draws. The regions that change sides are Croatia, Slovenia, Ibero-America, Romania, Poland and the Baltic states, where the church-related facets and the measured kinship, trust and individualism facets disagree. Clustered on ancestry and setting alone, the regions sort into climate bands rather than into the three groups; foodways is the only transmitted facet that tracks ancestry.
Choose a facet to recolour the map. Tap any region for its full profile. Grey land is not modelled. On Art, the map also marks over a hundred Gothic, Baroque and revival buildings still standing or standing as ruins; tap or hover a marker for its name and dates.
Israel is shown as a state, not as a stand-in for all Jews: about a fifth of its citizens are Arab, and most Israeli Jews are not Ashkenazi, so the Israel and Ashkenazim rows differ — compare them in Plate V. Grey around it, including the West Bank and Gaza, is not modelled.
For the settler societies, the Biological facet describes the founder populations from the British Isles and France, not today's far more mixed populations.
Tap a column heading to sort. The run count reports how many contiguous blocks each other column breaks into; a facet that agreed with the sort would sit at its floor, one block per category.
Biological cells read ancestry as steppe, farmer and forager percentages; Family cells the side of the Hajnal line (W, E or on the line), the kinship-intensity index (KII) where the Ethnographic Atlas covers the region, and the share of marriages between cousins, n/a where Bittles's compilation has no figure; Learning cells the year of the first university and the years of schooling in 1910; Trust and Individualism cells their survey scores; Foodways cells the share of alcohol drunk as beer, wine and spirits. The four measured facets are sorted, and their runs counted, by thirds of their score.
Before the facets are combined, each one draws its own map. Choose Each facet alone to see how the regions group on a single facet, at the number of groups that facet supports best. Choose Adding facets one at a time to watch the full result assemble: the facets go in one by one, and at every step the tree is read exactly as the final one is. Regions with a dashed edge are not yet where they end up.
Every facet clustered alone and cut at its best number of groups, ordered by how closely its own map matches the full result. Recomputed in your browser.
What each facet draws by itself. No single facet reproduces the full result, and the closest are not all the facets a reader might expect. Individualism, split into an individualist and a collectivist group, matches it best among the regions it covers — but most of the Orthodox world has no score, which makes agreement easy. The calendar draws the Hebraic pair and an Orthodox East. Confession, at its own best cut, sets apart only the Hebraic pair. Script draws five groups — Latin, Cyrillic, the three states that write in both, Greek and Hebrew — and puts Orthodox Romania and Moldova with the West, because they write in Latin letters. Family & kinship draws three bands: west of the Hajnal line, east of it, and a third group — the western Balkans and Israel — where every indicator a region has sits at the intensive end. Learning splits the early-university regions from the late ones and the small states. Biological isolates the three most farmer-rich populations, Sardinia, Israel and the Ashkenazim; Geographic descent draws climate bands; Legal family the common-law Atlantic and the Nordic countries; Foodways sets Israel apart. None of these has anything to do with the result.
Agreement with the full result, by adjusted Rand index, as the facets are added in two opposite orders: the facets least predictable from the others first, and the most predictable first. Both end at 1.
Order matters on the way, not at the end. Added most distinctive first, Learning and Civic trust sketch the West–East line, Language and Individualism sharpen it, and with the calendar in, six facets agree with the final result at four-fifths. Then Foodways pulls toward its north–south line, and Geographic descent and Biological dissolve the reading altogether; Confession, twelfth in, restores it exactly, and Script, added last, changes nothing. Added most repeated first, Script alone gives two-fifths of the result and Confession with it two-thirds; Art and Family bring it to nearly three-quarters, and the rest settle it gradually, with Learning, added last, completing it. Either way the end is the same, because the result is computed from all thirteen facets at once. The order shows which facets agree with it and which pull elsewhere.
Every region clustered on all thirteen facets together, by the method under Methodology. Shades within a colour are the finer sub-groups. Tap any region to see how far it sits from each commonwealth, which facets put it there, and how often it stays there when the method is varied.
For each commonwealth, every facet's share in separating it from the rest, largest first, with the values its members hold. Tap a region code to select it.
Latin Roman West, 36 regions: Scandinavia, the British Isles, the Low Countries, the German lands, Bohemia, Slovakia, Poland, the Baltic states, Hungary, France, Switzerland, Austria, Iberia, Italy and the settler societies. Byzantine East, 16: Russia, Belarus, Ukraine, Galicia, Romania, Moldova, Serbia, Bosnia, Montenegro, Macedonia, Bulgaria, Albania and Greece, with Croatia, Slovenia and Ibero-America at the seam. Hebraic Jews, 2: Israel and the Ashkenazi diaspora — two units, not all Jews.
The Hebraic pair splits off first, different from everyone on confession, script, calendar, art and ancestry at once; West and East divide next. Below that, every cut from four to ten groups scores about equally, so the data support several readings of Europe. The commonwealths are one.
The seams. On church, alphabet, art and calendar, Croatia, Slovenia and Ibero-America are as Latin as Poland; on kinship, trust, individualism and learning they sit nearer Serbia and Romania. They fall with whichever kind of facet gets more say.
Each facet's influence: its share of the variation in the combined distance, defined under Methodology. The last column assumes equal voices.
The church facets — Confession, Script, Art and the calendar — supply about two-fifths of the influence. Legal family, which agrees with almost nothing, has the most of any single facet. The four measured facets move the seams without redrawing the map.
One level down, Europe falls into ten groups: Nordic; Central Europe and the Low Countries; Romance Europe; the British Isles and settler societies; the Rus' lands; Danubian and South Slav; Greece and Albania; and Ibero-America, Israel and the Ashkenazim each alone. Their stability scores, in the cards above, run from firm to dissolving. Treat them as readings.
The structure is far beyond chance, survives every standard linkage rule, and predicts eleven of the thirteen facets it was built without. The cores hold; the borders move. Weighted on ancestry and setting alone, the map dissolves into climate bands. Tap any region for its own figure; the tests are read in full under Methodology.
After Dimitri Obolensky's The Byzantine Commonwealth (1971): peoples bound by a shared religious, legal and literary inheritance, not by a state or by blood. Under the Ottomans the Orthodox church went on administering the Balkans — Iorga's Byzance après Byzance — which is why Muslim Albania and Romance-speaking Romania sit in the Byzantine core.
Choose any two regions to set them side by side on every facet. The weighted distance is the D(i, j) the clustering uses, and each row shows what its facet contributes to it: the facet's own distance times its coefficient ω.
How the plates were produced and tested. Every number on this page is recomputed in your browser from the data in Plate II; formulas and full specifications are behind each toggle.
Fifty-four units, following cultural regions rather than states where the difference matters: Bohemia; Bavaria, Swabia, the Rhineland and North Germany; the Mezzogiorno and Sardinia; Galicia; Ulster; Québec. Boundaries are from Natural Earth. The settler societies and Israel appear as insets; Ibero-America and the Ashkenazi diaspora, which are not territorial, as swatches. The sub-national splits were drawn by hand before any clustering, which risks circularity, so the result was rerun without them. It holds, except at the Romance flank (see Testing stability).
The atlas divides Europe and its offshoots into fifty-four units. Units follow cultural regions rather than modern states where the difference matters: Bohemia rather than Czechia; Bavaria, Swabia, the Rhineland and North Germany; the Mezzogiorno and Sardinia split from northern Italy; Galicia from the rest of Ukraine; Ulster from the rest of Ireland; Québec from English Canada. Boundaries are drawn from Natural Earth, a free public-domain map dataset. Country outlines use its medium-detail edition; where a unit is part of a country rather than a whole one, its provinces are taken from the high-detail edition and merged into a single shape — Bavaria from the German states, the Mezzogiorno from the southern Italian regions, Galicia from three Ukrainian oblasts. The United States, English Canada, Québec, Australia, New Zealand and Israel appear as insets. Two units, Ibero-America and the Ashkenazi diaspora, are not territorial and appear as swatches rather than map regions.
How the sub-national regions were drawn, and why that matters. The splits inside Germany, Italy, the United Kingdom, Ukraine and Canada were drawn by hand before any clustering, following historic confessional, legal and linguistic boundaries that were already known to differ. That risks circularity: an atlas that claims to derive cultural regions should not assume them. Two consequences follow. First, the result was tested without them: leaving out each country's split regions, or all fourteen at once, leaves the remaining regions where they were, except that without Italy's three Spain and Portugal move (see Testing stability). These are also the regions that take their country's survey and food figures, so the same test covers that. Second, the regions themselves should be derived rather than drawn; that is the next revision.
Thirteen facets, in three kinds. Every one is received at birth: no one chooses their ancestry, their birthplace, their mother tongue or the church they are baptised into. They differ in how they reach a person and what can happen afterwards.
This is the test Heidegger sets out in Being and Time §74 — handed down rather than possessed, joined by taking up, losable — and it can be checked against the record: every transmitted facet has documented cases of adoption and of loss, and ancestry and setting have none. The clustering is run with all thirteen facets, and again with ancestry and setting at zero. Each facet rests on a published measurement or a documented classification, named in its category card.
| Facet | Kind | What is measured, and from where | Values | Heidegger's test |
|---|---|---|---|---|
| Biological | Ancestry | Ancient ancestral make-up: the share of steppe, Anatolian-farmer and hunter-gatherer descent, from published ancient-DNA models | Three percentages | Fails: settled at conception |
| Geographic descent | Setting | The biome and biogeographic realm in which most of the region's people are born, from the Udvardy–WWF ecoregion classification | Biome and realm | Fails: a person moves; nothing is handed down |
| Language | Transmitted | Shares of the population by first language, from censuses and surveys, grouped by branch | Twelve branches, as shares | Passes: taught, taken up, losable |
| Confession | Transmitted | Shares of the religious population by tradition (Pew Research Center 2010; censuses and surveys where finer) | Seven traditions, as shares | Passes: taken up or left by each person |
| Script | Transmitted | The alphabet in everyday public use, on its family tree (Daniels and Bright 1996) | Latin, Cyrillic, both, Greek, Hebrew | Passes: learned by every reader, changed by decree |
| Legal family | Transmitted | The family of the law in force, after JuriGlobe and Zweigert and Kötz; each card names the code and its model | Civil, Nordic, common, mixed, halakha | Passes: practised by each generation of jurists |
| Family & kinship | Transmitted | Kinship intensity from the Ethnographic Atlas (after Schulz and colleagues, 2019); the side of the Hajnal line; the share of marriages between cousins (Bittles) | Three indicators, each where it exists | Passes: kept or changed by couples and statute |
| Learning | Transmitted | Age of the region's first university; books per head, 1750–99 (Buringh and van Zanden); years of schooling in 1910 (Lee and Lee) | Three indicators, each where it exists | Passes: founded, purged, refounded |
| Civic trust | Transmitted | Share saying most people can be trusted (World Values Survey and European Values Study) | Per cent | Passes, barely: taught and lost by experience |
| Individualism | Transmitted | Hofstede's individualism index | 0 to 100 | Passes, uneasily: entangled with wealth |
| Art | Transmitted | Whether the Gothic-to-Baroque building tradition took root, evidenced by the standing buildings on the Plate I map | Both, Gothic only, Baroque only, revival, neither | Passes: commissioned or refused by each patron |
| Foodways | Transmitted | Alcohol drunk as beer, wine and spirits (WHO); fat calories from animal fat and vegetable oil; starch calories from wheat, maize, rice, other cereals and roots (FAO) | Three compositions | Passes: kept or dropped at each table |
| Festive calendar | Transmitted | The religious and seasonal public holidays set by statute for 2026; who brings the main presents; when the principal Christmas falls | Twenty-five holidays; seven gift-bringers; Eve, Day, Julian, none — as shares where a region keeps several | Passes: kept by households, changed by states |
The two kinds make each other over time. Descent carries culture: the steppe migrations brought Indo-European speech, and the Ashkenazim kept Yiddish and halakha by marrying within the community. Culture shapes descent: dairying selected for adults who can digest milk, which is why the trait is commonest in north-western Europe, and marriage rules decided who had children with whom, so that in the Netherlands Catholics and Protestants remain genetically distinguishable. What the atlas separates is the scale, not the origin. Within one life ancestry is settled, and only the transmitted facets can still be taken up or refused; the switch-off test asks which kind draws the present lines, not which came first. The data show the cycle where it should appear: Foodways is the one transmitted facet that tracks ancestry.
Laland, Odling-Smee and Myles, "How culture shaped the human genome", Nature Reviews Genetics (2010); Gerbault et al., "Evolution of lactase persistence", Philosophical Transactions of the Royal Society B (2011); Byrne et al., "Dutch population structure across space, time and GWAS design", Nature Communications (2020).
Birthright facets are settled by factors outside anyone's will. Most states confer citizenship on one of these grounds — by birthplace, as in the United States, or by parentage, as across much of Europe. Few extend it by descent from a people rather than from a citizen; Israel's Law of Return is the most prominent case.
Each facet gives every pair of regions a distance from 0, identical, to 1, as different as possible. Shares are compared by how much of one would have to move to make the other, and where the categories form a family tree, by how far apart on it they sit: two Slavic languages differ less than a Slavic and a Germanic one. Ancestry uses log-ratios, the correct geometry for shares that sum to one hundred.
| Both | Gothic | Baroque | Revival | Neither | |
|---|---|---|---|---|---|
| Both | 0 | 0.5 | 0.5 | 0.6 | 1 |
| Gothic | 0.5 | 0 | 0.7 | 0.5 | 0.5 |
| Baroque | 0.5 | 0.7 | 0 | 0.6 | 0.5 |
| Revival | 0.6 | 0.5 | 0.6 | 0 | 0.6 |
| Neither | 1 | 0.5 | 0.5 | 0.6 | 0 |
Two steps put the facets on one footing. Each facet is divided by its own spread, so that a unit of difference means the same on every facet. Then each facet's voice is divided by one plus its summed positive correlation with the others: facets that repeat one another share a voice, and a facet that agrees with nothing keeps a full one. The voices come from the data, not from a hand-set weighting. The combined distances are clustered by average linkage, which keeps them best of the standard rules (cophenetic correlation 0.84, against 0.79 for complete and single linkage and 0.63 for Ward's). The tree is read from the top until France, Russia and Israel fall in different groups: the three anchors name the groups and set the depth, and never move a region.
Every facet's distance runs from 0 to 1, but that does not make the facets equal. What a facet contributes to the clustering is its weight times how much its distances vary from pair to pair, and the facets vary very differently: Foodways gives a different distance for almost every pair but they bunch tightly, while Legal family takes a handful of values far apart. So, first, each facet is divided by its own spread, its standard deviation across the pairs where it exists, so that a unit of difference means the same on every facet.
The second step is the voice each facet gets, and it is where an atlas can quietly decide its own result. Version 1.3 weighted each facet by how much it agreed with the others, which rewarded repetition. Version 1.4 gave every facet an equal voice, which still lets several facets that record one event outvote the rest. Since version 1.5 the atlas discounts repetition. Each facet's voice is divided by one plus its summed positive correlation with the others, so facets that repeat one another share a voice, and a facet that agrees with nothing keeps a full one. No facet is singled out by hand: the correlations, and therefore the voices, come from the data. Equal voices and the version 1.3 weights are kept as tests.
The units are grouped by average-linkage hierarchical clustering: every unit starts alone, and the two closest groups merge again and again until one remains. Cutting that tree at different heights gives broader or finer groupings. No clustering rule is right in the abstract — every one returns groups, even from noise — so the choice is judged by two things: how faithfully the tree keeps the distances, and whether the groups survive tests they could fail. Average linkage keeps the distances best of the standard rules (a cophenetic correlation of 0.84, against 0.79 for complete and single linkage and 0.63 for Ward's method); it does not assume the distances are Euclidean, as Ward's method does; and it neither chains through outliers, as single linkage does, nor is ruled by them, as complete linkage is. The tests that could fail are under Testing stability. Step by step:
The correlation rfg between every two facets' distances, then each facet's spread sf, voice Vf, coefficient ωf and influence ef, with the influence each facet would have under equal voices for comparison, as this page computes them from the data. Correlations of 0.5 and above are in bold:
Agreement between two facets counts as corroboration only if they arrived by separate routes; if one carried the other, it is one fact recorded twice. Each of the seventy-eight pairs is classed from the sources as dependence, hypothesis, common cause, coincidence or independent. A facet is flagged when it agrees with another at r ≥ 0.5, or when more than half of it can be predicted from the rest. A flag prompts a decision (merge, split, or keep with a reason), not a verdict. The open flags are Confession with Script, Art and the calendar, which follow the churches, and Family & kinship with Confession, which is the Goody–Henrich hypothesis itself and so is kept apart to be tested.
Two facets that draw the same line can mean two different things. If they arrived by separate routes — a legal system and a building style, each carried by different people for different reasons — their agreement is corroboration, and it should count. If one carried the other, or both were carried by one institution, their agreement is the same fact recorded twice, and it should not. Statistics can measure the agreement; only history can say which kind it is. So every pair of the thirteen facets — seventy-eight pairs — is listed below with its correlation r, a Mantel permutation test of that correlation, and the link between the two as the sources describe it: dependence, where one is carried by the other or both by one institution; hypothesis, a causal link the literature proposes and this atlas can test but not prove; common cause, where both follow a third thing, usually geography or wealth; coincidence, agreement with no known mechanism; and independent.
The Mantel p-value is the share of 999 random relabellings of the regions that produce agreement at least as strong as the real one. It is computed after the page loads, so the column fills in a moment after the rest.
Each choice is varied in turn. A run reproduces the commonwealths only if all 1,431 pairs of regions are together or apart exactly as in the three-way split: one region changing sides fails it. The results are read below the table.
| Test | What is varied |
|---|---|
| Other voices | Equal voices; earlier weightings; transmitted facets only; without the two surveys; each facet left out |
| Birthright voice | The birthright share of the voice, from 0 to 1 |
| Random voices | 300 moderate and 300 unrestricted re-weightings |
| Group stability | How much of each group survives the moderate draws |
| Methods | Complete, Ward's and single linkage; an all-or-nothing distance |
| Data | Ancestry shares jittered; each region left out; hand-drawn regions left out |
| Noise | Each facet shuffled across the regions, 200 times |
| Prediction | Each facet and its close partners held out, then predicted, net of nearness |
Recomputing in your browser…
Chance and method. No shuffle comes near the real structure. Average, single and Ward's linkage draw the same West–East line, Ward's moving only Ibero-America; complete linkage, which fits the distances worst but one, also moves Israel and the Ashkenazim.
Prediction. The groups predict eleven of the thirteen facets held out of them, net of nearness. The two that fail are Geographic descent, a climate map, and Legal family, which crosses every other line.
Re-weighting. Across 300 moderate draws each commonwealth keeps most of its members in nearly every draw, but the exact three-way split reproduces in fewer than a third; the regions that move are Croatia, Slovenia, Ibero-America, Romania, the Baltic states and Poland.
Leaving facets out. Without the calendar the West–East line falls apart. Without Confession or Script alone it holds, because each carries most of the other. Without Legal family, Family & kinship, Learning, Individualism or Foodways, Croatia and Slovenia return to the West; without Art, Ibero-America does.
Hand-drawn regions. Rerun without them, every region stays put except at the Romance flank: without Italy's three, Ibero-America moves west; without all fourteen, Spain and Portugal move east.
Birthright alone. It groups temperate Europe with the Anglosphere, sets Ibero-America apart, and runs a Mediterranean cluster from Lisbon through Cagliari, Naples, Tirana and Athens to Jerusalem — Latin, Byzantine and Hebraic regions together, matching no church, alphabet, legal or language boundary.
Influence. Legal family leads because common law puts Dublin with Sydney and civil law puts Paris with Moscow, crossing every other line. Family, learning, trust and individualism have large voices but little influence: their variation is spread across every pair rather than concentrated on one boundary.
Nationality and peoplehood are easily confused, because both are usually acquired at birth. They differ in kind. Nationality is a legal status that a state confers: by birth on its soil (jus soli), by descent from a citizen (jus sanguinis) or by naturalisation, and that it can grant, withhold or withdraw by statute. The peoplehood this atlas measures is conferred by no one. It is what a population actually speaks, believes, builds, eats and keeps, and no decree can grant or revoke it. A state can make a citizen in a day; it cannot make a native speaker.
The atlas tests whether the two coincide anyway. Each region's citizenship law is coded on two measures that play no part in the clustering. Birth on the soil: 1 if birth on the territory confers nationality unconditionally, ½ if a parent must also have been born or settled there, 0 if it does not. Preference by origin: 1 if the law gives preferential access to people of the nation's ethnic origin who are not citizens' children, as Germany's Spätaussiedler rules, Greece's provisions for homogeneis, Hungary's simplified naturalisation of 2011 and Israel's Law of Return do; ½ for cultural affinity or discretionary provisions, such as Spain's two-year route for Ibero-Americans; 0 for none. Regions inside a country take its law. The Ashkenazi diaspora has no state and no citizenship law.
Recomputing in your browser…
Citizenship law follows other lines. Unconditional birthright citizenship belongs to the New World, the United States, Canada and Ibero-America, and it crosses the West–East line. Preference by origin belongs mostly to the kin-states of Central and Eastern Europe, whose borders moved beneath their peoples in the twentieth century, and to Israel. A peoplehood in this atlas spans many states; a state may hold more than one people, as Belgium, Switzerland and Israel do; and one people in the atlas, the Ashkenazim, has none.
Israel grants an automatic, near-immediate right of immigration and citizenship to Jews and their descendants (up to the grandchild level, plus spouses) worldwide under the Law of Return—making it one of the clearest and broadest ethnic/religious racist preference regimes for a diaspora.
Coded from the national citizenship acts as classified by the GLOBALCIT Citizenship Law Dataset (laws in force 1 January 2024) and the European Commission's Study on acquisition and loss of citizenship (EU-CITZEN, 2022), with Germany's 2024 reform applied; Joppke, Selecting by Origin (2005), on preference by origin.
The facets in no order of importance: the two birthright facets first, then the eleven transmitted ones, and last the Latin package that four of them were split from, with the test of whether it holds together. For each: what it measures, why it is in the atlas, the outside case that shows what it is doing, and the sources it rests on.
Ancient ancestral make-up: the blood half of blood and soil
The Biological facet is measured by one thing, ancient ancestry, taken from the published ancient-DNA models that decompose every European population into three founding migrations.
Almost everyone in Europe descends from the same three populations, in different proportions. Modern humans first reached Europe about 45,000 years ago, but those earliest groups left little or no ancestry in anyone alive today. Western hunter-gatherers, the first of the three, descend from populations that spread across the continent after the last Ice Age, from about 14,000 years ago. Anatolian farmers arrived out of what is now Turkey from about 7000 BC with agriculture, pottery and permanent settlement, and largely absorbed them. Yamnaya steppe herders arrived from the Pontic-Caspian grasslands from about 3000 BC with Indo-European speech, horses and wheeled carts, in a movement so male-biased that in Britain the Beaker horizon replaced roughly ninety per cent of the gene pool within a few centuries.
The mixture was essentially settled by 2000 BC. Everything the transmitted facets describe is Iron Age at the earliest and mostly medieval, which puts this facet three millennia too early to be causing any of it.
Fails. Geschick requires a heritage that can be taken up as a task and lost by not taking it up. Nobody can take up an ancestry, and nobody can lose one: it is settled at conception and carried whether or not anyone attends to it. Descent is a Kategorie applicable to things present-at-hand, and Dasein is not analysed through those.
Descent is the oldest answer to the question this atlas asks, and the one most often used to define a European peoplehood, so it has to be measured rather than assumed. Ancient-DNA studies now give that measurement in a form no earlier generation had: three founding populations, in proportions that can be compared region by region. The facet is in the atlas so that the claim that descent organises peoplehood can be tested on its best evidence.
Ancestry shares are compositional: they must sum to one hundred, so a rise in one component forces a fall elsewhere, and ordinary distances between percentages are distorted, worst near zero. The standard treatment, due to John Aitchison, is to take the centred log-ratio of the three shares and measure distance in that space; that is the Biological distance used in the clustering. For the maps, each region is placed on the main axis of variation in those log-ratios — in practice a gradient from farmer-rich to steppe-rich — and coloured along it, dark for farmer-rich, pale blue for steppe-rich.
An earlier version also measured phenotype, eye colour and stature, as a second sub-division. It was dropped because its figures were the softest in the atlas, and because stature measures nutrition as much as descent: Dutch men gained twenty centimetres in five generations on an unchanged gene pool.
The setting: the landscape a people's culture formed in, not a birthright of any person
The frame for this facet is Alfred W. Crosby's Ecological Imperialism: The Biological Expansion of Europe, 900–1900 (1986). Crosby asked why Europeans became the majority population in North America, the Southern Cone, Australia and New Zealand, but nowhere in tropical Africa or Asia. His answer was ecological, not cultural: Europeans succeeded where the land already resembled home — temperate forest and grassland in the same latitudes — and they did not arrive alone. They brought a "portmanteau biota": wheat and barley, cattle, sheep, pigs and horses, the weeds that followed the plough, and the Old World diseases that emptied the land ahead of them. Where that whole package could take root, it remade the landscape into what Crosby called Neo-Europes; where it could not, Europeans ruled but never became the people. The biome decided where Europe could be reproduced.
Each region is coded by the biome in which most of its people are born, not the biome that covers most of its area. Russia is mostly taiga by area, but most Russians are born in the mixed forests between Moscow and the Volga; Israel is mostly Negev desert, but most Israelis are born on the Mediterranean coastal plain. Every biome present in a region, taken from the ecoregion table, is listed in its region card. In the distance measure the biome carries three quarters of the weight — with related biomes, such as the two temperate forests, taiga and tundra, or Mediterranean scrub and desert, counted as half as far apart — and the realm carries the remaining quarter.
Crosby's own evidence also marks the limit of his thesis for this atlas. The biome explains where settlers could thrive; it does not explain what they carried. The Neo-Europes chose their biome by moving into it, and what crossed the ocean with the wheat and the cattle was the transmitted part — the language, the law, the church. The clustering bears that out: geographic descent agrees with biology, weakly with what a region eats, and with almost nothing else transmitted.
Fails, though at first it seems to pass. Anyone can move into a biome, own land in it and leave it, but each time the subject is the person, not a heritage. Land passes between generations as property, not as a possibility to be repeated. Moving into a landscape is residence, not appropriation: what a newcomer actually takes up is the language, law and faith already there. And a biome can be lost to the plough or to warming without anyone failing at anything. Nothing is handed down, taken up or lost as a heritage, which is where Heidegger's own later talk of rootedness strays from his criterion.
Soil is the other half of blood and soil, and the strongest version of the claim — Crosby's, that Europe could be reproduced only where the land resembled it — is an empirical one. Coding the biome where a people is born, from an independent scientific classification, lets the atlas ask whether the landscape organises anything a people hands down.
That is the strongest case anyone has made for soil as a condition of peoplehood, and the facet is built to test it. It asks where a people is born, measured not by borders but by the land itself, using the biogeographical system Miklos Udvardy drew up for UNESCO in 1975, as extended by the World Wide Fund for Nature. The Earth is divided into eight realms — the Palearctic, Nearctic, Neotropical, Australasian and others, each with its own lineage of plants and animals — and fourteen biomes, the kinds of living landscape a climate produces: temperate broadleaf forest, boreal taiga, Mediterranean scrub, steppe, desert. Within them sit some eight hundred ecoregions, from the Pannonian mixed forests to the Pontic steppe. On this facet the Anglosphere shares Europe's biome but not its realm: Crosby's Neo-Europes, visible on the map as the same forest on another continent.
At first glance the biome seems to pass all three of Heidegger's tests: anyone can move to the landscape they prefer, anyone can leave it, and land can be handed down. It fails every one of them, and for the same reason each time: what moves, owns and leaves is the person, not a heritage.
Transmitted, not possessed — fails. Land is the paradigm of what is possessed. It is held as private property under title, or as public land held by a state, a municipality or a crown, and when it passes between generations it passes as an estate: a thing, by inheritance law. The heritage of §74 is a store of possibilities to be repeated, not goods to be owned. What is genuinely handed down about land is the law of owning it — Roman dominium, title and conveyance, the public trust, the municipality as a legal person — and that belongs to Legal family, not to the soil.
Membership by taking it up — fails. Anyone can move to a biome, and whole peoples have: that is how the Neo-Europes were settled and how Florida and Arizona filled. But moving into a landscape is residence, not appropriation. The biome asks nothing of the newcomer and offers no possibility to repeat; a Finn who retires to Mediterranean scrub is not thereby Mediterranean. What a migrant actually takes up on arrival is the language, law and faith of the people already there — the transmitted facets.
Losable — fails, in the wrong sense. You can leave a biome, but then it is you who have gone, not a heritage that has lapsed; the biome carries on without you. And a biome can be lost without anyone leaving or failing at anything: the Pontic steppe was ploughed into cropland, Mediterranean forest degraded into scrub, and warming is moving the treeline and the edge of the steppe northward. Heidegger's losability is loss through failing to take a heritage up, a lapse into the anonymous "they". Loss that merely happens to people is not that.
So the biome imitates each criterion by changing its subject: a person can enter, own and leave a biome, but nothing is handed down, taken up or lost as a heritage. It stays present-at-hand. §74's Geschick has no place for soil, which is exactly where Heidegger's own later talk of rootedness strays from his criterion.
Fichte's criterion, and what happens when you run it
Fichte delivered the Addresses to the German Nation in Berlin in the winter of 1807 and 1808, with French troops occupying the city. The argument is almost always mis-remembered as a blood argument. It is a language argument, and he is explicit about it.
The claim: a people that has spoken its own tongue continuously since its origin possesses an Ursprache. Every abstract term in it still traces back through living metaphor to sensuous experience all its speakers have had. A people that adopted a foreign tongue instead — his case is the Romance nations taking Latin — received its abstract vocabulary pre-formed and severed at the root. Therefore the Germans are the Urvolk, and only a people with a living root can be genuinely philosophical and therefore genuinely free.
The tell is his own membership rule. In the Seventh Address the criterion stops being linguistic at the decisive moment: whoever believes in spirituality and its freedom, and wills its endless progress, is one of this people, wherever born and whatever tongue they speak; whoever believes in stagnation and mechanism is foreign, even if born in Germany. That is a metaphysical test, open to anyone, far closer to Heidegger's §74 than to anything in Gobineau.
The map colours each region by the branch of its majority language, which is why Ireland, Ulster and Scotland appear as West Germanic and Belgium, Luxembourg and Switzerland as multilingual. The distance uses every first language a region speaks, by its share.
Language is the criterion Fichte made central and the one most nationalisms since have used. It is transmitted, taught and losable, and it divides populations that every other facet unites, so it is the clearest test of whether a single cultural marker can define a people.
Every first language a region speaks, by its share of the population, from the latest census or survey: the CIA World Factbook's country figures, Statistics Canada's 2021 census for Québec and the rest of Canada, and Israel's 2011 social survey, in which 49 per cent of adults gave Hebrew as their mother tongue, 18 Arabic and 15 Russian. Where no shares are published, the official majority language counts as the whole. Languages outside these families — Basque, Romani, the Caucasian languages, recent immigrant languages — are left out and the rest re-summed. Two regions are then compared on the family tree: as far apart as the deepest split between their speakers, so that Belgium, three-fifths West Germanic and two-fifths Romance, sits between the Netherlands and France rather than apart from both, and Latvia, a third East Slavic, sits partway toward Russia.
Passes, and Fichte very nearly reaches the criterion himself. A language is handed down, taken up by each generation, and genuinely losable — Ireland lost most of a vernacular within three generations and has spent a century deciding whether to take it back up, which is Wiederholung in the strict sense. Where Fichte breaks from §74 is in claiming one particular inheritance is metaphysically privileged. Heidegger's apparatus has no resources for that claim.
The quarrels that split the inheritance
Confession divides Europe twice, and the two divisions are not the same kind of thing. The 1054 split turned on papal primacy and the Filioque — whether the Creed should say the Spirit proceeds from the Father and the Son. That line maps almost exactly onto the alphabet boundary and onto the old Latin Roman coding, because all three are the same event seen from different angles.
The Reformation is a different order of quarrel: scripture alone against scripture plus tradition, faith alone against faith formed by charity, two sacraments against seven. Arguments about how grace reaches a person, conducted in Latin by men trained in the same universities. Protestant polities kept canon-law marriage structure, corporate personality, the chartered university and the binding promise. The Greek Catholic value exists because of the Union of Brest in 1596: Byzantine rite, Latin jurisdiction. And the southeast corner of the map carries Islam — Albania's plurality and Bosnia's Bosniaks, the legacy of four centuries of Ottoman rule.
Religion is the most widely used marker of collective identity in European history, and the branch of faith — not merely the religion — is what divided Europe in 1054 and again in 1517. It records which church converted a people; the alphabet that came with it is the next facet.
Each region holds shares of the population that belongs to a religion: Latin Catholic, Greek Catholic, Protestant, Orthodox, other Christian, Muslim and Jewish. The national figures are the Pew Research Center's estimates for 2010; Germany's four regions are built from the Politbarometer's state figures for 2016, weighted by state population; Scotland, Northern Ireland and Québec take their censuses; Galicia takes the 2015 survey of its three oblasts, and the rest of Ukraine the national figure without them. People with no religion are left out, because the facet records which tradition a population belongs to, not how observant it is: most Czechs have no religion, and nine in ten of those who do are Catholic. The shares sit on a tree — Christianity, Islam and Judaism at the top, Western and Eastern Christianity below, Catholic and Protestant below that — so that Estonia, half Lutheran and nearly half Orthodox, sits between Finland and Russia, and Bosnia, 46 per cent Muslim, 39 Orthodox and 15 Catholic, sits between all three.
Passes. Confession is handed down through named institutions, taken up or not by each person and each generation, and losable in the most direct way any facet here is losable. Confession also shows the vacancy in §74 as clearly as anything could: the criterion cannot tell a Bavarian which of the two inheritances available in 1517 was properly his.
The letters a people reads and writes, and the church that brought them
In early medieval Europe almost everyone who could write was a cleric, and a people's first alphabet was the one its missionaries brought for scripture and liturgy. In 863 the Byzantine brothers Cyril and Methodius were sent to Moravia to preach to the Slavs in their own tongue, and built an alphabet for it — Glagolitic, from which their disciples in Bulgaria derived Cyrillic. The Latin church, meanwhile, carried the Latin alphabet with its liturgy. Where Rome won the conversion, Slavs write in Latin letters: Poles, Czechs, Slovaks, Slovenes, Croats. Where Constantinople won, they write in Cyrillic: Russians, Ukrainians, Belarusians, Bulgarians, Macedonians, Serbs. Moravia itself is the natural experiment: the same Slavs began with a Slavonic alphabet, the Frankish Latin clergy then won their church, and Czech has been written in Latin letters ever since.
The rule holds even across languages. Muslim Moriscos in Spain wrote Spanish in Arabic letters; the Karamanlides, Orthodox Christians who spoke Turkish, wrote Turkish in Greek letters; Jewish communities wrote German and Spanish in Hebrew script, as Yiddish and Ladino. And it follows the rite more precisely than the bishop: Galicia's Greek Catholics joined Rome and kept the Byzantine liturgy, and with it Cyrillic.
The exceptions are state projects, not drift. Romania, Orthodox, moved from Cyrillic to Latin letters in the 1860s to proclaim a Latin identity; Atatürk moved Turkish to Latin script in 1928; the Soviet state imposed Cyrillic on Moldova, and Moldova returned to Latin letters in 1989. So the alphabet is largely a mark of confession, and version 1.5 folded the two into one facet. Since then every facet's voice has been discounted for what it repeats, so the two share a voice without being merged, and kept apart, the places where they part company — Romania and Moldova, Orthodox in Latin letters; Serbia, Bosnia and Montenegro, in both alphabets; Galicia, Greek Catholic in Cyrillic — are visible on a map of their own.
The alphabet is the most visible boundary in Europe — a traveller meets it on the first street sign — and it records which church converted a people and what it has read and written since. It stands beside Confession rather than inside it because the two part wherever a state intervened, and the voices already discount what they share.
Each region is coded by the alphabet in everyday public use — on street signs, in newspapers, in school. Where two are in everyday use, as in Serbia, Bosnia and Montenegro, each counts half. Alphabets are compared on their family tree, by the same rule as languages and confessions: every alphabet in the atlas descends from the Phoenician, Latin, Cyrillic and Greek through the Greek alphabet and Hebrew through Aramaic, so two regions in different alphabets of the Greek branch are half as far apart as either is from Hebrew.
Put "archive" in place of "vocabulary" and Fichte's case transfers: a people that keeps its alphabet can read its own past, and one that switches is cut off from its written memory within a generation. Turks born a few years after the 1928 switch could not read their grandparents' documents. An alphabet is adopted or refused whole, at a datable moment, and it is defended at cost: the Lithuanian book smugglers carried Latin-letter books across the border for forty years rather than accept Cyrillic. But by Fichte's own continuity test every European script fails — Rome took its letters from the Greeks through the Etruscans, and Cyrillic was built for the Slavs by Greek missionaries — and an alphabet switched by decree cannot be constitutive, or Moldovans would have become a different people twice in fifty years.
Script is the sign of an inheritance more than its source, which is why its voice is shared with Confession's rather than counted in full.
Passes. An alphabet is handed down through schools and presses, taken up by every child who learns to read, and losable by decree in a single year, as Romania, Moldova and Turkey each found.
Where law comes from, and who is allowed to make it
The old Latin Roman facet asked which legal persons and obligations exist at all; this one asks how law gets made, and since version 1.6 it also carries what that facet meant by Roman-law contract, because the civil law is the Roman law's direct descendant. In the civil law tradition the code is primary, judges apply rather than make, and earlier decisions do not bind. In the common law tradition decisions are the law: precedent binds, judges make law incrementally, and no code purports to be complete.
From around 1500 the continent underwent the reception of Roman law. It largely stopped at the Channel, because the English royal courts had already spent three centuries building a working system out of writs and juries. Common law is a sibling inside the inheritance, not an exception to it. Mixed systems sit where the two met: Scotland, which kept a Roman-law foundation inside the United Kingdom, and Québec, which runs a French-derived Civil Code for private law and English common law for public law. The Nordic family is codified but grown from customary law rather than received Roman law, and halakha appears because the Ashkenazi row needs it: jurist-made law, with real courts, and no state behind any of it.
How law is made — by code or by precedent — is a separate question from which legal persons exist, and it divides regions that the churches unite: England from France, Québec from the rest of Canada. It is the atlas's test of whether two branches of one inheritance behave as one facet or two.
Each region is placed in its legal family after the University of Ottawa's JuriGlobe classification: civil law, common law, mixed, or, for the Ashkenazim, halakha, with the Nordic countries as a branch of civil law after Zweigert and Kötz. Each region's card on Plate I names the code in force and the model it was drafted on, as evidence. They show what the old Latin Roman facet missed: Roman law is not a Latin marker. Byzantium kept it in better condition than the West; Greece applied Harmenopoulos's Byzantine Hexabiblos as its civil law until 1946; and Russia's Civil Code of 1994 is a pandectist code drafted on German and Dutch models. What was Latin was the second jurisdiction and the university that taught the law, which the atlas now measures elsewhere.
Passes. A legal tradition is transmitted through named institutions, taken up by each generation of practitioners as a task, and genuinely losable — a jurisdiction that stops honouring precedent has left the common law however its statutes read.
Whom you may marry, where you live afterwards, and whether your cousins are your clan
The oldest argument about what made Europe different is about the family. Jack Goody argued in 1983 that the Western Church pressed rules on marriage that no other civilisation had: no marriage between cousins, at its height out to sixth cousins; no marriage to a dead husband's brother; no adoption of heirs; no second wife; and, from the twelfth century, marriage by the free consent of the two people, whatever their families wanted. The effect was to break up the extended kin groups that held land and loyalty. John Hajnal found its demographic trace in 1965: west of a line from Saint Petersburg to Trieste, Europeans married late, many never married at all, and couples set up their own households; east of it, as in most of the world, marriage was early, near-universal and into the husband's family. In 2019 Jonathan Schulz, Duman Bahrami-Rad, Jonathan Beauchamp and Joseph Henrich measured kinship intensity across the world's societies and linked it, country by country, to centuries of exposure to the Western Church and to individualism, trust and conformity today.
Canon-law marriage was the first of the four institutions of the old Latin Roman facet, and the one with the strongest modern evidence behind it. It is a facet of its own so that the claim can be tested rather than assumed: if the Church's marriage rules made the West, the family facet should draw the church line. It draws most of it, and leaves it in the Danube basin.
Three measurements describe a region's family system. Each becomes a score from 0 to 1 by where it falls between the lowest and the highest value any region has: score = (value − lowest) ÷ (highest − lowest).
Two regions differ on this facet by the average gap between their scores, counted over the measurements both of them have. A missing figure is left out, never guessed:
| Pair | Kinship intensity | Hajnal line | Cousin marriage | Family distance |
|---|
Each cell is the gap between the two scores; a dash means one of the pair has no figure, so that measurement is left out of the average.
What cousin marriage does to the result. Israel's 20 per cent sets the top of the scale, so every European country scores between 0 and 0.10 on it. Between two European regions, cousin marriage therefore moves the family distance by a tenth at most: the line between West and East is drawn by the Hajnal line and by kinship intensity, and cousin marriage chiefly sets Israel apart. It is also the thinnest of the three measurements, with a figure for 24 of the 54 regions.
| Country | Regions | Cousin marriages | Score |
|---|
Why a missing figure is not a blank. For north-western and central Europe the evidence that cousin marriage is rare is strong, even without a national survey:
For the Orthodox and Muslim south-east, Romania, Moldova, Bulgaria, Greece, Serbia, Montenegro, North Macedonia, Bosnia and Albania, and for Russia, Belarus and Ukraine, no comparable national survey was found, and the atlas makes no claim. Every country without a figure stays out of the cousin-marriage measurement rather than receiving an assumed value.
Three indicators, each where it exists, each put on a 0–1 scale by its range:
The distance is the mean difference over the indicators two regions share, so Bosnia and Montenegro, which have only the Hajnal line, are compared on that alone, and New Zealand and the Ashkenazim, which have none, are left out of this facet.
The Ethnographic Atlas codes the marriage form of 1,257 societies. Europe is the outlier:
| Region | Societies | Monogamous | Limited polygyny | General polygyny | Polyandry |
|---|---|---|---|---|---|
| Europe | 42 | 90% | 7% | 2% | 0 |
| Asia | 158 | 25% | 57% | 16% | 2% |
| Oceania and island South-East Asia | 156 | 21% | 55% | 24% | 1% |
| The Americas | 386 | 17% | 53% | 30% | 0 |
| Africa | 515 | 5% | 16% | 79% | 0 |
| All | 1,257 | 16% | 37% | 47% | 0.3% |
Monogamy as the rule for everyone, rather than the lot of men who cannot afford a second wife, is rare among human societies, and the Atlas's European societies are almost all monogamous. The exceptions are the Gheg of northern Albania, coded polygynous, and three peoples with occasional polygyny: the Bashkirs and Kazan Tatars, Muslims of the Volga, and the Karaim of Lithuania. Polyandry, several husbands to one wife, is rarer still: four societies in the whole Atlas. Cousin marriage tells the same story: in Bittles's figures every European and settler country lies under 2 per cent, against 20 to 65 per cent across much of the Middle East, North and West Africa and South Asia.
Passes, with an unusual edge. A household form is handed down as practice and as law, taken up by each couple, and losable: the zadruga was opened to division by statute, and the Irish marriage age moved by five years in two generations. The marriage pattern that defines the West was not inherited from any ancestor; it was imposed by the Church, resisted for centuries, and finally taken up.
The chartered university, and what came after it: books and schools
The chartered university — a self-governing corporation of masters and students with a licence to teach anywhere — is a Latin invention: Bologna from 1088, Paris about 1150, Oxford by 1167, then Salamanca, Prague, Kraków and Vienna before 1400. The Byzantine world had higher schools, some older than Bologna, but not the self-governing corporation, and Russia's first university dates from the eighteenth century. The facet asks how long the institution has stood in each region, and what followed it.
The chartered university was the third institution of the old Latin Roman facet. Measured separately, it tests whether the West's early universities still mark it off: they do in part, along the old line, but books and schooling follow the Reformation as much as Rome.
Passes. Learning is handed down through named institutions, taken up by each generation as a task, and losable: universities have been closed, purged and refounded — Prague's split by language in 1882, Kraków's professors deported in 1939 — and literacy has risen and fallen with the states that taught it.
Whether a stranger can be trusted: the habit corporations and courts depend on
Corporate personality, the second institution of the Latin package, lets strangers pool money, sue and be sued, and bind each other across generations. Such bodies work only where people expect strangers to keep their word. The facet measures that expectation directly: the share of people who answer that "most people can be trusted", rather than "you can't be too careful", in the World Values Survey and the European Values Study, the longest-running comparative survey of values.
Trust among strangers is what the corporation, the court and the contract of the Latin package presuppose, and what the kinship argument predicts the Church's marriage rules produced. It is in the atlas to test that: where the old Latin line and today's trust disagree — Belarus above France, Portugal below Austria — the facet says so.
The latest wave for each country, 2017–22 for most; Belgium, Ireland, Luxembourg and Moldova from 2010, Israel from 2004. Regions inside a country take its figure; Ibero-America is the population-weighted mean of seven countries; the Ashkenazim have no figure.
Passes, though it is the facet furthest from a heritage. Trust is taught by experience — by courts that enforce and officials who do not take bribes — and lost the same way. It is the clearest case in the atlas of a transmitted facet that is also an outcome of everything else.
The trait most often claimed for the West, measured
Individualism, in the sense psychologists measure, is how far people define themselves by their own attributes and choices rather than by the groups they belong to — family, clan, village — which look after them in exchange for loyalty. Geert Hofstede built the first comparative index from surveys of IBM employees in the late 1960s and 1970s, and it has been extended and replicated since. On it, the English-speaking world comes first by a distance, north-western Europe next, and the Balkans, Portugal and Latin America last.
Individualism is the trait most often claimed to define the West, and the one the kinship argument says the Church's marriage rules produced. It is in the atlas to test both claims. It follows the north-west to south-east gradient more than the church line, and on the headline voices it is one of the facets that moves Croatia and Slovenia east.
Hofstede's individualism index, 0 to 100, from his dataset of August 2015. Québec takes the French-Canadian score; the regions inside Germany, Italy and Britain take their country's; Ibero-America is the population-weighted mean of seven countries. There is no score for Iceland, Belarus, Ukraine, Moldova, Bosnia, Montenegro, Macedonia or Albania, nor for the Ashkenazim, and those regions are left out of this facet. The scores for the Balkans and parts of central Europe come from later replications and are rougher than the rest.
Keeping track of one's ancestors is a candidate trait too, and this atlas is itself an instance of it. It is left unscored because no source measures it comparably across countries: membership of genealogy sites follows marketing and language, and search interest is published only as relative indices. What can be measured is the record that makes genealogy possible, and it is unevenly old. England ordered every parish to keep a register of baptisms, marriages and burials in 1538; the Council of Trent ordered it for every Catholic parish in 1563; Sweden's Church Law of 1686 made its registers the most complete in Europe; Russia's metrical books began under Peter the Great in the 1720s. The Lutheran and Anglican state churches kept the fullest records, which is why Nordic and British family histories can usually be traced further back than most. A people's interest in its lineage is real; whether it marks the West, or only marks where the West kept paperwork, the atlas cannot yet say.
Passes, uneasily. A way of seeing oneself is handed down through families, schools and law, taken up or not, and losable. But individualism is the facet most entangled with wealth: richer countries score higher, and part of what the index measures is the century of economic history behind each country rather than a heritage.
Where the Gothic-to-Baroque building tradition took root
The facet codes one thing: whether the Gothic-to-Baroque building tradition took root in a region's own churches, town halls and palaces. Five values: both Gothic and Baroque; Gothic only; Baroque only, without a medieval Gothic behind it; revival only, where the style arrived as nineteenth-century Gothic Revival; and neither. Each region is coded by its core territory, and the exceptions are named below.
Revival only is about the medieval layer, not about whether Gothic or Baroque buildings stand there now: plenty do. The settler societies did not exist when Gothic was built, so what they have is a retrieval — Pugin's Westminster exported, St Patrick's Cathedral in New York, Washington National Cathedral, the Gothic quadrangles of their universities — a style consciously taken up again from a past the builders never lived in. The complication is Spanish colonial Baroque inside the modern United States — San Xavier del Bac in Arizona, the Texas missions, St Louis Cathedral in New Orleans — genuine Baroque, built in the Ibero-American and French traditions rather than the Anglo-American one, on land the United States later acquired.
Architecture was chosen over painting deliberately. Coding sacred images directly — naturalistic figure against icon against aniconism — would largely restate Confession and count the same event twice. Buildings are a separate record: they are expensive, durable, datable, and still standing to be checked.
A building tradition is expensive, durable and still standing, so it can be checked on the ground in a way few cultural markers can. Gothic and Baroque are the two styles that travelled with the Latin institutions, and whether each took root in a region's own churches and town halls records which inheritance reached it, and when. The buildings on Plate I are the evidence.
Passes. A building tradition is handed down through masons' lodges, workshops and academies, taken up or refused by each patron and each generation, and losable: the Dutch iconoclasm of 1566 stripped the churches of the Low Countries in weeks. The Anglosphere revival is the facet's clearest case of a heritage deliberately taken up again.
What a region actually eats and drinks, measured rather than coded
Every other transmitted facet is coded by judgement from the historical sources. This one is measured from official statistics, and it has three parts. Drink: the share of all alcohol drunk as beer, as wine and as spirits, from the World Health Organization's 2024 report, averaged over 2017 to 2019 — the measure behind the old "drinking belts" of Europe. Fat: the share of fat calories from animal fat — butter, lard, tallow — against vegetable oil, from the FAO's food balance sheets. Starch: how starch calories divide between wheat, maize, rice, other cereals such as rye, oats and barley, and potatoes and other roots, from the same tables. Each part is a composition, compared by the Hellinger distance, and the facet is their average.
The map colours each region along the main axis of the three together. One end is the Baltic table — rye, potatoes, animal fat and spirits, in Latvia, Poland and Lithuania. The other is vegetable oil with either wine, as in Italy and Iberia, or maize, as in Moldova, Bosnia and Ibero-America. Each region's card gives its own figures.
What people eat and drink every day is transmitted in households rather than institutions, is taken up or dropped by each generation, and is one of the first things anyone names when asked what makes their people theirs. Unlike every other facet it can be measured from official statistics rather than coded by judgement, so it adds an independent, numerical check on whether the institutional facets are the ones that matter.
Some of the best-known food markers belong to a single region and cannot be scored across fifty-four. The lye pretzel, the Laugenbrezel, is the bakers' guild sign of Bavaria and Swabia, and its long association with Lent shows how closely the southern German table was shaped by the church year. The meatless Christmas Eve supper — carp in Bohemia, Slovakia and Poland, the fish of the southern Italian Vigilia — is the same fast kept the night before the feast, and it belongs to the regions the festive calendar codes as keeping the Eve.
Passes, in the household rather than the institution. A cuisine is handed down at the table, taken up or refused by each generation, and losable: the potato arrived from the Americas and displaced older staples across the north within two centuries. What it cannot do is draw a sharp line. It changes by degrees across every border on the map, which is why it adds weight to no single boundary.
The holidays the law keeps, who brings the presents, and on which night the family sits down
The statute book. Half of the facet is the religious and seasonal public holidays each region's law sets for 2026, compared by the share two regions hold in common. Bavaria has nine, Greece eight including Clean Monday, Russia one — the Julian Christmas — and the United States one. Corpus Christi marks the Catholic lands, from Portugal and Croatia to Poland and Bavaria; Reformation Day, Slovenia and the German east; Saints Cyril and Methodius, Bohemia, Slovakia, Bulgaria and Macedonia; Midsummer, Finland, Sweden, the Baltic states and Québec; the Eids, Albania, Bosnia and Macedonia. In federations a holiday counts by the share of the population whose state keeps it, so the Reformation counts for most of north Germany and Corpus Christi for about a fifth of it.
The household. The other half codes two customs nearly every household keeps. The first is who brings the children's main presents: a Father Christmas figure — Santa Claus, Père Noël, the Scandinavian Tomte and Nisse, Finland's Joulupukki, Romania's Moș Crăciun, Bulgaria's Dyado Koleda; the Christ Child — the Christkind of southern Germany, Austria and Switzerland, Ježíšek in Bohemia, Jézuska in Hungary; St Nicholas on 5 or 6 December in the Low Countries and Galicia; the Three Kings on 6 January in Spain; Grandfather Frost at New Year across the former socialist east; or St Basil on New Year's Day in Greece. The second is when the principal Christmas observance falls: Christmas Eve or Christmas Day by the Gregorian calendar, or 6–7 January by the Julian.
The calendar is the most widely shared ritual in a household's year: who brings the presents and on which night the family sits down together. Both are transmitted, both are taught to children, and both have been changed by churches and states within living memory. It is in the atlas because it is the everyday trace of confession — and because, where it parts from confession, it shows what else a people hands down.
Regions that keep more than one custom are coded with shares: Ukraine half St Nicholas and half Grandfather Frost, half Gregorian and half Julian since 2023; Moldova between the two calendars; Bosnia's Catholics, Orthodox and Muslims at their own shares; Slovenia's three gift-bringers, Miklavž, Božiček and Dedek Mraz, a third each. For the customs, the distance is the share two regions do not have in common.
Italy is coded with Babbo Natale, who now brings most children's main presents on Christmas morning; the Befana, who arrives on 6 January with sweets in a stocking, is the older custom.
Folk dress is the other everyday marker most people would name, and it is left unscored because no source codes it consistently across the atlas. It shows the same pattern of loss and revival: Highland dress was banned by the Dress Act of 1746 and revived in the nineteenth century as Scotland's national costume; the Norwegian bunad is largely a national-romantic creation of the early twentieth century; the Ukrainian embroidered vyshyvanka has become a sign of national allegiance in the present one. Each is a heritage deliberately taken up again.
Passes. A festive calendar is taught to every child, kept or dropped by each household, and changed by churches and states: the Julian calendar in Ukraine was set aside by statute, and Grandfather Frost was installed by one. What it adds to the atlas is the everyday trace of confession — and, where it parts from confession, a second record of what a people hands down.
Four medieval legal inventions — did they travel together?
Until version 1.5 this was a facet: a judgement, coded by hand, of whether four legal institutions arrived in a region together — consent marriage under canon law, corporate personality, the chartered university, and the enforceable contract of the revived Roman law — scored inside, frontier or outside. It agreed with Confession & script and with Art more closely than any facet should, because it recorded much of the same event. Version 1.6 retires it as a facet and measures the four instead: canon-law marriage as Family & kinship; corporate personality, and the trust among strangers it depends on, as Civic trust; the chartered university as Learning; and Roman-law contract inside Legal family. Individualism is added as the trait the package is said to have produced. The old code is kept only to ask whether the parts still look like one package.
Partly. Family structure follows the old coding closely; individualism moderately; learning and trust weakly; legal family not at all, because the civil law descended from Rome is in force from Lisbon to Moscow. The parts agree with one another only weakly. The package reads better as a sequence than as a bundle: the Church's marriage rules came first and drew the line the old code drew; the university and the corporation followed the Church; and the trust and individualism measured today follow the family pattern and the north-west's long history of wealth as much as the church line. That is what moved Croatia, Slovenia and Ibero-America in version 1.6.
The four institutions, as the old facet described them:
The package is the most specific claim anyone has made about what set the West apart, and the strongest candidate for a transmitted inheritance that crossed oceans intact. Measuring its parts separately and testing whether they travel together is fairer to that claim than coding it as one thing, and it lets the parts disagree, which they do.
Being and Time §74 defines a collective by Geschick: a heritage handed down (Überlieferung) and taken over as a task through resolute repetition (Wiederholung). Not a property of a thing — a happening that people are inside of.
It is transmitted, not possessed. Gratian taught at Bologna, Bologna trained the canonists, the canonists staffed the courts, the courts made the rulings, the rulings were copied into the manuals that trained the next generation. You can name the teachers.
Membership is by taking it up. Peru enters in the 1530s by receiving the package and is fully inside; so do Philadelphia, Montreal and Sydney later. Birth confers nothing here.
It is losable — the test both birthright facets fail absolutely. A polity that abandons judicial independence, subordinates its universities, or stops enforcing contracts against the powerful has dropped out of the transmission whatever its population's genome says.
The honest limit. §74 shows this is a Geschick. It cannot show that it is yours — Heidegger never supplies a criterion for which heritage is properly one's own, and that vacancy is exactly what got filled by the state in 1933.
The Latin Roman West, the Byzantine East and the Hebraic Jews are emergent and quantifiable. Their cores persist under every test; their borders do not, and the seam regions go with whichever kind of facet gets more say (Plate IV). Every facet that draws them — a creed, a household form, a school, a habit of trusting strangers — can be received by someone not born to it, and lost by someone who was.
Sardinia, with almost no steppe ancestry, is fully inside the Latin West; so are Québec, Sydney and Philadelphia, for the same reason Kraków is. Lima and Mexico City took the Latin church, law and Baroque, yet on trust and individualism sit nearer Bucharest than Madrid. And what can be taken up can be let go: Fraktur abolished by memo in 1941, Moldova's alphabet switched twice in fifty years, Ireland's vernacular lost in three generations. Nothing here is owed by birth or closed by it. The inheritances are many, and each can be taken up, so anyone can join them in a combination no one else holds and make themselves an example of their own. In every age, each of us is headstrong enough to choose. Protect what you love, and take it up yourself.
Aristotle’s History of Animals keeps a strict order of inquiry: first the facts, the “what”—the parts of animals, their ways of life, their activities and characters—then the “why.” The differences and attributes of every animal must be grasped before their causes are sought (I.6, 491a). He distinguishes two kinds of knowledge: knowing that a thing is so (τὸ ὅτι, to hoti) and knowing why it is so (τὸ διότι, to dioti). The plates above are this atlas’s hoti: what peoplehood consists of, measured. What follows is its dioti—why this matters.
As the means of securing a rooted identity and safeguarding the distinctiveness of a people, love of one’s own serves a higher human purpose. It is an antidote to the ideologies that call themselves liberating but leave only dissolution, and they are easiest to see against the long rhythm of history.
Following Strauss and Howe, there are four secular-religious lodestones: each turning ends in a crisis that leaves a defining moral memory for the saeculum that follows. The first is the Revolution’s fight against unrepresented government (c. 1787–1866), the second the Civil War and slavery (c. 1867–1946), the third the Second World War, the Holocaust and the Holodomor (c. 1947–2026), and now, from about 2027, supplantation. The first three memories are true, and keeping them is owed. But a lodestone is meant to orient a people, not to hold it in place.
These events have hardened into civilizational dogmas of perpetual penance. They place the West, and since the war Germany above all, at the moral centre of the world, yet require it to remain spiritually disarmed. Its identity is permitted only as confession and its past only as indictment. That is a moral armistice, not a reconciliation, and a people that may only repent can never be at peace with itself or its neighbours.
The fourth crisis is one of supplantation across every register: ethnic, cultural, and environmental. Globalization does not merely move goods; it moves people, and the freedom to move without attachment produces nomads who carry no lasting obligation to the cultures they enter. Over time this dilutes inherited forms of life and, eventually, the ethnic substrates that sustained them. Yet the same process operates more quietly through commodities. Hybrid corn, bred so that it cannot reliably reproduce itself, must be repurchased each season from distant suppliers. The Holstein cow, optimized for industrial output, displaces local breeds that once suited particular landscapes and climates. These are not neutral efficiencies. They are sterile seeds. Just as the hybrid grain requires perpetual external replenishment, a globalized market requires populations that no longer know how to provision themselves—communities whose attention, taste, and sense of belonging have been converted into audited products. What cannot reproduce on its own must be supplied from outside. From this dependence a universal moral code arises. Once the customer base is planetary, moral language itself must become portable and frictionless. Native virtues—particular loyalties, inherited restraints, local hierarchies of value—are reread as obstacles or as structures of power to be dismantled. Campaigns of diversity, inclusion, and equity function as the ideological equivalent of the hybrid seed: they standardize the moral soil so that any population can, in principle, occupy any setting. The particular is treated as a defect to be corrected. The setting follows. Airports, logistics hubs, algorithmic feeds, and standardized urban cores begin to resemble one another more than they resemble the places that preceded them. When the material conditions of life are interchangeable and the moral vocabulary is interchangeable, the people who inhabit them become increasingly interchangeable as well. A globalized peoplehood is the terminal expression of the same logic that produced the sterile seed: a population whose continuity no longer depends on any particular place, custom, or inherited form of life. These forces are already visible in the fertility collapse of the advanced nations—Europe, China, Japan, and the Anglosphere above all. Populations that have lost the capacity for cultural self-reproduction, like seed that cannot set viable grain, do not replenish themselves. The pressures may prove difficult to reverse. Yet the clearer the chain becomes—sterile commodity, portable morality, interchangeable setting, deracinated people—the more deliberately its second-order effects can be refused. What cannot reproduce itself must either be continually imported or allowed to die. The choice is whether any given people will remain capable of reproducing its own forms of life, or whether it will accept the condition of the hybrid seed.
What joins penance and supplantation is a two-dimensional cosmopolitanism, or its alienated double, a nihilism that flattens every inherited form of meaning and offers nothing in its place. Yet humanity is only ever realised through the flourishing of distinct peoples, not their dissolution.
This atlas answers on its own terms. A people is not a bloodline but a heritage taken up as a task. The thirteen facets above measure that heritage, and none of it is owed by birth. Each part can be lost through neglect long before it is lost through numbers, and it can be recovered only by those who take it up again. That is the task of this turning: for Europe’s peoples to rediscover their shared inheritance and rebuild the communities that carry it. The plates are the map. The work is yours.
The remedy is not another doctrine but a different kind of person, and so it begins with education. Its task is to raise people who govern themselves: good at the root, who want what is right and do it of their own accord, without hesitating or wavering, expecting no reward and indifferent to the arithmetic of advantage. Such people are not made by exhortation. They are formed indirectly, by freeing the native drive to think for oneself. A school must therefore teach not what to think, but how to think.
Facts committed to memory matter less than the mind’s own play and the joy of learning for its own sake. That spontaneity, and the love that comes with it, is the sine qua non of moral life; it is the same love that, turned toward one’s own people, becomes patriotism. The true end of education is a student who can picture for himself the good order of a life and of a people, and who, seized by love for that picture, wills it and works to make it real.
Protect What You Love, don't kill what you hate.