The Axiom You Can Refuse

Euclid's Elements opens with five postulates. Four are short and obvious: a line can be drawn between two points, a line can be extended, a circle can be drawn with any centre and radius, all right angles are equal.

The fifth is long, awkward and phrased in terms of angles adding to less than two right angles. The version usually quoted is Playfair's: through a point not on a given line, there is exactly one line that never meets it.

Everyone could see the fifth was different. It looked like something that ought to be provable from the other four, and Euclid himself seems to have felt it, since he avoids using it for as long as he possibly can. For two thousand years geometers tried to supply the missing proof.

They all failed, and they failed in an instructive way: every apparent success turned out to assume something equivalent to the postulate itself. That the angles of a triangle sum to two right angles. That parallel lines stay the same distance apart. That similar triangles of different sizes exist. Each is a restatement, and using one to prove the fifth is arguing in a circle.

Girolamo Saccheri, an Italian Jesuit, took the last serious run at it by contradiction. Assume the postulate false, derive an absurdity, done. He worked through consequence after consequence, building an elaborate structure, and in 1733 published it as *Euclid Freed of Every Flaw*, declaring at the end that he had reached a conclusion repugnant to the nature of a straight line.

He had not reached a contradiction. He had derived a substantial portion of hyperbolic geometry and refused to recognise it.

A century later, three people saw it. Gauss worked it out and did not publish, telling a correspondent he feared the outcry of the Boeotians. János Bolyai, a young Hungarian officer, published in 1832 as an appendix to his father's textbook, having written to him that out of nothing he had created a strange new universe. Nikolai Lobachevsky in Kazan published first, in 1829, in Russian, in a provincial journal, and was largely ignored.

Deny the fifth postulate and nothing breaks. You get a different geometry, internally consistent, in which through a point there are infinitely many lines missing a given line, triangles have angles summing to less than 180 degrees, and there are no similar figures of different sizes: a triangle's angles determine its size.

The other direction gives spherical geometry, where every pair of great circles meets and triangle angles exceed 180 degrees. A triangle with a vertex at the north pole and two on the equator has three right angles.

'Consistent' was a claim that needed proving, and it was, by building models. Beltrami, Klein and Poincaré constructed hyperbolic geometry inside Euclidean geometry: take the interior of a disc, call arcs meeting the boundary at right angles 'lines', and all the hyperbolic axioms hold. Any contradiction in the new geometry would therefore be a contradiction in the old one. Relative consistency, which is how such questions are still settled.

The philosophical damage was extensive. Kant had taken Euclidean geometry as the clearest example of knowledge that is both necessary and about the world; synthetic a priori, built into how we must perceive space. Now there were several geometries, all consistent, and the question of which describes the room you are sitting in had become empirical.

Gauss's other contribution made the empirical question answerable. His theorema egregium shows that curvature is intrinsic: a creature confined to a surface can determine whether it is curved without any access to a surrounding space. Measure a large triangle and add the angles. On a sphere you get more than 180 degrees, and the excess is proportional to the area.

Which is what Riemann generalised to any number of dimensions in 1854, and what Einstein needed sixty years later. General relativity says mass curves spacetime, and objects follow the straightest available paths through the curved geometry, so light bends near the sun, and it was measured bending in 1919. Mercury's orbit precesses by an amount Newtonian gravity could not account for, and the curvature accounts for it exactly.

None of that could have been written in a century that believed Euclid's axioms were facts about space rather than one option among several. The two thousand years of failure were not wasted. They were the slow discovery that the postulate is a choice, and that the universe had made a different one.