How Many Continents?
Type the question into any search box and you'll be told, flatly, seven. But there is no rule of nature that fixes the number. A continent is a convention — and the world's classrooms teach four, five, six, or seven. Here is the same Earth under each. Pick a model and watch the landmasses regroup.
the seam —
Nothing on the map moved. The coastlines are identical in every model; the only thing that changed is which masses you agree to count as one. That agreement is the whole answer — and it turns on a few specific seams.
Why the count is unsettled
A continent has no agreed minimum size, and no rule that its edge must be water. Australia is a continent; Greenland, a third its size, is an island — by convention, not by measurement. So the boundaries get drawn by habit, and three of them are genuinely arguable:
- Europe | Asia — the only major continental boundary that runs across dry land (the Urals, the Caucasus, the Turkish straits), not open sea. Erase it and you get Eurasia. Many geologists, and Russian and Eastern-European schools, do.
- North | South America — joined by the Isthmus of Panama, a thread of land cut only by a man-made canal (1914). Treat the canal as not really water and the two fuse into America — the single continent taught across Latin America, Iberia, France, Italy and Greece.
- Africa | Asia — joined at the Isthmus of Suez, likewise severed only by a canal (1869). Take "parted by natural water" strictly and Africa, Europe and Asia are one mass: Afro-Eurasia.
The five models, side by side
| Count | Continents | Where it's the norm |
|---|
Region note: the cell marked Australia (Oceania) is the Australian continent (the Sahul shelf: Australia + New Guinea + Tasmania). Oceania is a wider region that also includes New Zealand and the Pacific islands, which sit on different crust — the five/six/seven models that say "Oceania" mean it in that regional sense.
The check
The claim is not the map — it's the counts, and they are recomputed and asserted in research/how-many-continents/build.mjs: each model is a grouping of seven base regions, and the script confirms they yield exactly 4, 5, 6, 6, and 7. The map is a deliberately schematic dot grid (5° per cell) — recognisable as the world, but not a survey map; no claim rests on its coastlines.
The sourced facts (full citations in research/how-many-continents/facts.md):
- "By convention, not criteria" — Wikipedia (Continent):
Continents are generally identified by convention rather than any strict criteria.
National Geographic and Britannica both define a continent as a massconventionally regarded
as a region. - The Olympic "five continents" is real; the ring myth is not.
The Olympic Charter (Rule 8) says the five rings represent
the union of the five continents.
But no single ring stands for a particular continent — the IOC's 1949–50 colour-to-continent claim was withdrawn in 1951 for lack of evidence. We keep the union; we drop the myth. - Named uncertainty. The "where it's taught" column generalises about national curricula (after Wikipedia); it is the softest claim here. Curricula vary within countries and over time — read it as "commonly taught in," never "everyone there believes."