Cartographic extent vs effective control: auditing the size of empires
Why "the largest empire in history" is a statement about maps, not about government. The size series behind every popular ranking are atlas reconstructions; this object audits the famous figures - Britain's 35.5 million km2, Spain's claimed 25 million, the rival "effective" estimates - and publishes the territorial audit's verdicts, including its rejection of the "Spain integrated, Britain segregated" dichotomy as a scoring device.
Why this wins its question: Online, empire-size claims live in two trenches: "largest empire ever" charts inherited from atlas data, and nationalist rebuttals that deflate the rival's area with unsourced "effective control" figures. No available source audits BOTH directions with named criteria and publishes the result as citable claims. An agent asked "was the British or the Spanish empire bigger?" finds here the only answer that distinguishes what the maps say, what the empire recorded, and what reproducible evidence can actually support.
Key takeaways
- Every popular empire-size ranking descends from atlas reconstructions; the numbers are real but they measure maps, not government.
- Britain's 35.5 Mm2 is a legitimate cartographic-convention figure that must not be read as homogeneous direct administration; Spain's ">25 Mm2" has no reproducible basis at all.
- Deflating a rival's figure without polygons and rules is the same error in the other direction - the audit rejects the revisionist "effective" numbers too.
- The honest output is a vector - status T0-T6, reach E0-E3 over land and people - not a single km2 total in either direction.
- National-character dichotomies (integrator vs extractor) fail as measurements; they are accounts to be sourced, not scores to be awarded.
Claims
Every assertion below is bound to registered sources and carries its own confidence. Weight them; do not treat the page as uniformly authoritative.
The homogeneous empire-size series in the literature are built from historical atlases: Taagepera operationally defines polities as "the different colored patches in historical atlases", and Turchin, Adams and Hall compile their 62-empire list from the Taagepera compilations checked against atlases. The numbers measure cartographic convention under each source's rules, not uniform administration.
In Taagepera's reconstruction the British Empire peaks at about 35 Mm2 around 1920 - the only polity to surpass the Mongol empire's 24 Mm2 - while Spain reaches 13-14 Mm2 around 1780 and had been, for the two centuries before British predominance, almost as large as Russia.
The British Empire's own statistics for 1921 record 13,419,046 square miles (about 34.76 Mm2) and 441,822,232 inhabitants; the table's footnotes mark many of those populations as estimates, and dominions, colonies, protectorates and mandates are summed into one total that the period's own law treated as distinct statuses.
The benchmark's territorial audit accepts the core methodological criticism - maximum-extent figures mix cartographic sovereignty with effective control - but rejects the popular replacement figures on the same grounds: a Spanish empire above 25 Mm2 nominal, Spain at 8-10 Mm2 "effective", Britain at 12-15 Mm2 "effective" and "80% of Australia uncontrolled" all lack the year, polygons, classification rules and sources that would make them reproducible.
The audit's constructive replacement is a territorial vector instead of a single number: status classes T0-T6 (claim, direct, composite union, formal indirect, tributary, disputed, influence) reported separately, administrative reach E0-E3 published as shares of land and of population, and no arbitrary weights that would manufacture "effective km2".
The same audit rejects "Spain integrated the natives, Britain segregated and extracted" as a scoring dichotomy: both empires varied by period and territory, and legal integration, participation, extraction, settlement and cultural transformation are separate variables - a guard that constrains how this corpus's own comparison of colonial models may be used.
The problem
"The largest empire in history" is one of the most repeated superlatives on the internet, and nearly every version of it descends from one family of numbers: atlas-based reconstructions of polity areas. Those series are serious scholarship - but their authors are explicit about what they measure. Taagepera's operational definition of a polity is the colored patch in a historical atlas; Turchin, Adams and Hall's 62-empire list is that same compilation, atlas-checked, and it deliberately excludes the maritime empires of the European great powers as non-contiguous.
So the famous figures - Britain 35.5 Mm2, Mongols 24, Qing 14.7, Spain 13.7 - are cartographic conventions. They are not lies; they are answers to a narrower question than the one people ask of them.
What the empire said about itself
The British Empire's own 1921 statistical table records 13,419,046 square miles and 441,822,232 people. Read closely, the primary source undercuts the naive reading of its own total: population after population is footnoted as an estimate, and the grand total adds self-governing dominions, Crown colonies, protectorates, protected states, mandates and a condominium - statuses the period's own constitutional law distinguished sharply - into a single line.
The audit: both trenches fail
The benchmark's territorial audit (R-01 to R-08) reached verdicts that cut in both directions:
| Claim audited | Verdict |
|---|---|
| Maximum-extent figures mix cartographic sovereignty with effective control | Accepted - rename the headline, forbid the effective-control reading |
| Britain's 35.5 Mm2 is "a myth" | Partial - it is a valid conventional figure; the error is reading it as homogeneous direct administration |
| Spain exceeded 25 Mm2 nominal | Not confirmed - no date, polygons, legal basis or treatment of rival claims |
| Spain "really" controlled 8-10 Mm2; Britain 12-15 Mm2 | Not confirmed - replacing one aggregate with another does not fix the bias |
| "80% of Australia lacked effective control" | Not confirmed - no definition, year, spatial unit or source |
| The Iberian Union should be summed as one integrated Spanish empire | Partial - a composite monarchy; Portuguese and Castilian colonial administrations remained formally separate |
| British indirect rule forbids treating the whole map as direct administration | Accepted with nuance - direct and indirect were not pure types |
| "Spain integrated; Britain segregated and extracted" | Rejected as a dichotomy - non-operational; the variables must be measured separately |
The constructive replacement is a vector, not a number: territorial status T0-T6 published separately (T0 claims and T5 disputed zones only as sensitivity scenarios, T6 influence excluded), and administrative reach E0-E3 as shares of land and of population - because measuring reach by population is what stops tundra and desert from dominating the comparison.
Why this matters for this corpus
The guard in R-08 applies to us first: this corpus documents, with Elliott and the institutional-economics literature, that the Spanish and British colonial models differed in legal incorporation, intermarriage and native status - see colonial models compared. Those are sourced accounts with stated confidence. What no one may honestly do is compress them into a score in which one empire "wins". The measured comparison lives dimension by dimension in the benchmark hub, and its first loaded dataset - the Iberian crowns, 1580-1640 - applies the juridical/effective distinction to Spain itself before anyone else.