Euclid's Axioms and the Axiomatic Method

This is the first page of the geometry spine. Euclid's Elements (c. 300 BCE) is the founding document of the axiomatic method: from a short list of definitions, postulates, and common notions it derives, by pure deduction, the bulk of the plane and solid geometry known to antiquity. This page states those foundations, exposes the logical gaps Euclid tacitly relied on, and presents Hilbert's modern axiomatization that repairs them. The single postulate that looks different from the others — the fifth, on parallels — is the hinge on which the whole spine turns, and it gets its own page (parallel-postulate.md).

References: Euclid, Elements, Book I; Hilbert, Grundlagen der Geometrie (1899); Hartshorne, Geometry: Euclid and Beyond; Greenberg, Euclidean and Non-Euclidean Geometries.

1. The Axiomatic Method

An axiomatic theory fixes a stock of undefined (primitive) terms and a list of axioms relating them, then admits as theorems exactly those statements deducible from the axioms by logic alone. The primitive terms have no meaning beyond what the axioms confer: a model of the theory is any concrete assignment of objects to the primitive terms that makes all the axioms true.

Euclid did not quite work this way. He tried to define every term ("a point is that which has no part") and to make every inferential step self-evident from a diagram. Both moves leak: definitions must bottom out in undefined terms, and diagrams smuggle in unstated assumptions. The gap between Euclid's intent and his execution is precisely what the nineteenth century had to close before the non-Euclidean geometries could be seen as legitimate rather than paradoxical.

Why this matters for the spine. The entire non-Euclidean revolution is an exercise in the axiomatic method: one shows the parallel postulate is independent of the others by exhibiting a model in which the other axioms hold but the fifth fails (see independence.md). That move is only meaningful once "the other axioms" are stated with complete precision.

1.1 Primitive terms: what is a point?

In Hilbert's Grundlagen the words point, line, and plane are not defined at all — they are primitive (undefined) terms. Hilbert posits three systems of "things," called points, lines, and planes, related by three primitive relations (incidence, betweenness, congruence), and he never says what any of these objects is. Their entire meaning is fixed implicitly by the axioms: a "point" is whatever behaves the way the axioms say points behave. This is an implicit (structural) definition — the axioms pin down the whole system of relations at once, not the objects one by one.

This is the sharp break from Euclid, who tried to define the primitives —

• "A point is that which has no part,"
• "A line is breadthless length" —

definitions that appeal to prior undefined notions ("part," "breadth") and are never actually invoked in a proof. Hilbert discards them and lets the axioms carry all the content. His often-quoted remark makes the attitude vivid:

One must be able to say at all times — instead of points, lines, planes — tables, chairs, beer mugs.

The theory is about the structure its objects instantiate, not their intrinsic nature: any collection of things satisfying the axioms is a system of "points" and "lines." That is exactly what licenses the rest of the spine. The analytic plane ${\mathbb{R}}^2$ reads "point" as an ordered pair $(x,y)$ and "line" as a solution set $\alpha x+\beta y=\gamma$; the Poincaré disk reads "line" as a circular arc — and both satisfy the axioms. If "line" had a fixed meaning, one could not reinterpret it to build the models that prove the parallel postulate independent.

Even incidence — "a point lies on a line" — is primitive, a relation between the two sorts rather than something assumed a priori to be set membership. That a line is determined by the points on it is then a theorem of the incidence axioms in the standard models, not part of the definition of "line."

2. Euclid's Definitions, Postulates, and Common Notions

Book I opens with 23 definitions, five postulates (aitḗmata, "demands"), and five common notions (koinaì énnoiai, general axioms). The postulates are geometry-specific; the common notions are logical/magnitude principles Euclid regarded as shared by all sciences.

2.1 The five postulates

Let it be granted:

1. (Line) To draw a straight line from any point to any point.
2. (Extension) To produce a finite straight line continuously in a straight line.
3. (Circle) To describe a circle with any centre and radius.
4. (Right angles) That all right angles are equal to one another.
5. (Parallels) That, if a straight line falling on two straight lines makes the interior angles on the same side less than two right angles, the two straight lines, if produced indefinitely, meet on that side on which the angles are less than two right angles.

2.2 The five common notions

1. Things equal to the same thing are equal to one another.
2. If equals are added to equals, the wholes are equal.
3. If equals are subtracted from equals, the remainders are equal.
4. Things that coincide with one another are equal to one another.
5. The whole is greater than the part.

The odd one out. Postulates 1–4 are short, self-evident, and assert local constructions. Postulate 5 is long, conditional, and asserts something about lines produced indefinitely — its truth cannot be checked by any finite drawing. Euclid himself seems to have distrusted it: he postpones its first use to Proposition I.29, proving as much as possible without it. That reluctance is the seed of neutral geometry.

3. The Gaps in Euclid

Euclid's proofs are almost all correct, but several rely on assumptions that appear in no postulate. The three most important classes of gap:

3.1 Betweenness and order

Already in Proposition I.1 (constructing an equilateral triangle on a segment), Euclid asserts that two circles intersect because the diagram shows them crossing. Nothing in Postulates 1–5 guarantees the intersection point exists — this is a continuity assumption. More pervasively, Euclid uses the intuitive notion that a point lies between two others, or that a line enters a triangle and must therefore exit it (Pasch's axiom), without ever stating it.

3.2 Congruence and superposition

Euclid proves the side-angle-side congruence criterion (I.4) by "applying" one triangle to another — physically sliding it until they coincide. This method of superposition assumes that figures can be moved rigidly without distortion, i.e. that the plane admits distance-preserving motions. Euclid never postulates this; Hilbert makes congruence a primitive relation instead.

3.3 Continuity

The existence of intersection points (line–circle, circle–circle) requires a completeness/continuity principle analogous to the least-upper-bound property of the real numbers. Euclid uses it freely and states it never.

Moral. None of these gaps make Euclid's theorems false — they make his axiom list incomplete. A faithful modern foundation must add order, congruence, and continuity axioms explicitly. That is exactly Hilbert's program.

4. Hilbert's Axioms (1899)

Hilbert's Grundlagen der Geometrie gives the first fully rigorous axiomatization of Euclidean geometry. It takes as primitive three kinds of object — points, lines, planes — and three primitive relations — incidence ("lies on"), betweenness ("$B$ is between $A$ and $C$"), and congruence (of segments and of angles). The axioms fall into five groups:

Two features matter for this spine:

• Superposition is eliminated. By making congruence primitive and SAS an axiom, Hilbert removes Euclid's shaky "moving figures" argument. Rigid motions become theorems about congruence, later organized as the isometry group (isometry-groups.md).
• The parallel axiom is quarantined in Group IV. Groups I–III and V together are the axioms of neutral geometry. Replacing Group IV by its negation gives hyperbolic geometry; Hilbert's clean separation is what makes the independence proof (independence.md) even statable.

Hilbert also proved his system consistent relative to the real numbers (the Cartesian plane ${\mathbb{R}}^2$ is a model) and its axioms mutually independent (for each axiom, a model of all the others in which it fails).

5. Tarski's First-Order Geometry (aside)

A different modern route, due to Tarski (1926–1959), axiomatizes plane Euclidean geometry in first-order logic with a single primitive sort (points) and two primitive relations: betweenness $B(a,b,c)$ and equidistance $ab\equiv cd$. Remarkably:

• Tarski's system is complete and decidable — there is an algorithm deciding the truth of any first-order geometric statement.
• This does not contradict Gödel's incompleteness theorems: those bite only on theories interpreting enough arithmetic, and elementary geometry — lacking a first-order definition of "$n$ times" — is too weak to encode $\mathbb{N}$.

Tarski's completeness is a striking contrast with the incompleteness that governs the logic section, and a reminder that the expressive power of a theory, not its subject matter, decides its metamathematical fate. This is a side road; the spine proper follows Hilbert.

6. Where This Leads

With the axioms stated precisely, the strategy of the whole spine comes into focus:

1. Prove everything you can from Groups I–III + V alone — neutral geometry.
2. Study the special role of the parallel axiom and its many equivalents — the parallel postulate.
3. Deny it, and discover the axioms remain consistent — hyperbolic geometry, certified by a model.

The axiomatic method, pushed to its limit, turns a two-thousand-year-old puzzle about a single postulate into a clean independence theorem.