Geometric Constructions

⚡ Key Takeaways

Geometric constructions are the art of drawing precise figures using only an unmarked straightedge and a compass. The central result: a length is constructible if and only if it can be obtained from unit length 1 by a finite sequence of additions, subtractions, multiplications, divisions, and square roots. This governs which figures are drawable (bisected angles, regular pentagons) and which are impossible (trisecting an arbitrary angle, squaring the circle). Codified by Euclid (~300 BCE); the impossibility theorems were settled by Wantzel (1837) and Lindemann (1882).

💡 The Core Insight

Every straightedge-and-compass step intersects either two lines, a line and a circle, or two circles. Lines give linear equations; circles give quadratic equations. So each new point's coordinates satisfy at most a quadratic over the field of numbers you already have. Therefore constructible numbers are built by square roots only — never cube roots, never transcendental numbers. Constructibility is a question about powers of 2.

DefinitionConstructible Point, Line, and Circle

Begin with a set of given points (by convention, at least two, defining unit length). The basic operations are: 1. Straightedge: draw the line through two already-constructed points. 2. Compass: draw the circle centered at one constructed point passing through another (or with radius equal to a constructed segment). A point is constructible if it arises as an intersection of two such lines/circles after finitely many steps. A real number $\alpha$ is a constructible number if a segment of length $|\alpha|$ can be constructed from a unit segment.

TheoremField Characterization of Constructibility

A real number $\alpha$ is constructible if and only if there is a finite tower of fields

4

\quad \mathbb{Q} = F_0 \subset F_1 \subset \cdots \subset F_n, \qquad [F_{k+1}:F_k] = 2, \qquad \alpha \in F_n

Derivation hint: each construction step adjoins at most one square root (from [3]), so the degree of each extension over the previous is 1 or 2. Consequently, if $\alpha$ is constructible then its degree over $\mathbb{Q}$ is a power of $2$ :

5

\quad [\mathbb{Q}(\alpha):\mathbb{Q}] = 2^m \quad \text{for some } m \ge 0

Derivation hint: by the tower law $[\mathbb{Q}(\alpha):\mathbb{Q}]$ divides $[F_n:\mathbb{Q}]=2^n$ , so it is itself a power of 2.

TheoremConstructible Numbers Form a Field Closed Under Square Roots

The set $\mathbb{K}$ of constructible numbers is a field: if $a,b \in \mathbb{K}$ then so are $a\pm b$ , $ab$ , $a/b$ (for $b\neq 0$ ), and $\sqrt{a}$ (for $a\ge 0$ ). Each operation has an explicit ruler-and-compass construction.

ProofSquare Roots Are Constructible — the Geometric Mean

We show $\sqrt{a}$ is constructible from a segment of length $a$ and a unit segment. 1. On a line, mark points $A$ , $B$ , $C$ so that $AB = a$ and $BC = 1$ , with $B$ between $A$ and $C$ . Then $AC = a+1$ . 2. Construct the midpoint $M$ of $AC$ (perpendicular-bisector construction) and draw the semicircle with diameter $AC$ , radius $\tfrac{a+1}{2}$ . 3. Erect a perpendicular to $AC$ at $B$ , meeting the semicircle at $D$ . 4. By Thales' theorem, angle $ADC = 90^\circ$ , so triangle $ADC$ is right-angled at $D$ . 5. The altitude $BD$ from the right angle satisfies the geometric-mean relation $BD^2 = AB \cdot BC = a \cdot 1 = a$ . 6. Therefore $BD = \sqrt{a}$ . $\blacksquare$

∎

Figure

The geometric-mean construction of √a: the altitude of a Thales semicircle

ProofImpossibility of Doubling the Cube

The classic problem asks: given a cube of side 1 (volume 1), construct the side of a cube of volume 2. 1. The required side length is $\alpha = \sqrt[3]{2}$ , the real root of $x^3 - 2 = 0$ . 2. The polynomial $x^3 - 2$ is irreducible over $\mathbb{Q}$ (by Eisenstein's criterion at $p=2$ , or because it has no rational root). 3. Hence the minimal polynomial of $\sqrt[3]{2}$ has degree 3, so $[\mathbb{Q}(\sqrt[3]{2}):\mathbb{Q}] = 3$ . 4. But by Equation [5], any constructible number has degree a power of 2 over $\mathbb{Q}$ . 5. Since $3$ is not a power of $2$ , $\sqrt[3]{2}$ is not constructible. The cube cannot be doubled. $\blacksquare$

∎

mathematician · Pierre Wantzel | 1814–1848

A French mathematician who, at age 23 in 1837, published the first rigorous proof that doubling the cube and trisecting a general angle are impossible with straightedge and compass — problems open since antiquity. He reduced each to a question about whether a number satisfies an irreducible cubic, and showed that constructibility forces degree to be a power of two. Wantzel also clarified which regular polygons are constructible, completing the analysis begun by Gauss. Despite this landmark, his work was long overshadowed; he died at 33, reportedly exhausted by overwork and irregular habits.

🧠 Intuition

Think of your starting points as the numbers you already "own": just $0$ and $1$ . Each tool lets you do bookkeeping with them. A straightedge plus a compass can copy lengths, so you can add and subtract. Using similar triangles (parallel-line constructions), you can scale, which gives multiplication and division. The one genuinely new power is the compass's ability to draw a circle — and circle-meets-line is a quadratic, which is exactly a square root. So the constructible world is "the rationals plus all the square roots you could ever want, nested as deeply as you like." Cube roots and $\pi$ live outside this world because they cannot be reached by stacking square roots. The deep lesson: a question that looks purely about drawing is secretly a question about which numbers you can write using $+,-,\times,\div,\sqrt{\ }$ .

ExampleBisecting a 60° Angle to Get 30°, and Why 20° Fails

Given: A constructible $60^\circ$ angle (built from an equilateral triangle).

Bisecting (possible): 1. With vertex $V$ , draw an arc cutting both rays at $P$ and $Q$ (equal radii, so $VP = VQ$ ). 2. Draw equal-radius arcs centered at $P$ and $Q$ ; they meet at $R$ . 3. Line $VR$ bisects the angle, giving $\boxed{30^\circ}$ .

Trisecting (impossible): Trisecting $60^\circ$ requires constructing $\cos 20^\circ$ . The triple-angle identity gives

6

\quad \cos 3\theta = 4\cos^3\theta - 3\cos\theta

Derivation hint: expand $\cos(2\theta+\theta)$ using sum and double-angle formulas. Setting $\theta = 20^\circ$ and $x = \cos 20^\circ$ , with $\cos 60^\circ = \tfrac12$ :

7

\quad 8x^3 - 6x - 1 = 0

Derivation hint: substitute $\cos 60^\circ=\tfrac12$ into [6], multiply through by 2. This cubic has no rational root (rational-root test on $\pm1,\pm\tfrac12,\pm\tfrac14,\pm\tfrac18$ all fail), so it is irreducible over $\mathbb{Q}$ and $[\mathbb{Q}(\cos20^\circ):\mathbb{Q}] = 3$ . Since $3$ is not a power of $2$ , $\cos 20^\circ$ is not constructible. Trisecting a general angle is impossible. $\boxed{\text{30° yes, 20° no}}$

ExampleConstructing a Length of √6 from Unit Segments

Given: A unit segment of length $1$ , and the ability to lay off integer multiples.

Goal: Construct a segment of length $\sqrt{6}$ . 1. Lay off $AB = 6$ and $BC = 1$ on a line, so $AC = 7$ . 2. Midpoint $M$ of $AC$ : center of a semicircle of radius $3.5$ . 3. Perpendicular at $B$ meets the semicircle at $D$ . 4. By the geometric-mean theorem, $BD = \sqrt{AB \cdot BC} = \sqrt{6 \cdot 1}$ .

Check numerically: $BD = \sqrt{6} \approx 2.449$ . As a sanity test, $BD^2 = 6.000$ , and $AB\cdot BC = 6\cdot 1 = 6$ . ✓

\boxed{BD = \sqrt{6} \approx 2.449 \text{ units}}

This confirms $\sqrt{6}\in\mathbb{K}$ , consistent with $[\mathbb{Q}(\sqrt6):\mathbb{Q}]=2$ , a power of $2$ .

⚠️ Common mistake

"Impossible" means impossible exactly, not approximately. You can trisect any angle to ten decimal places with ruler and compass by clever approximation. The impossibility theorems say there is no finite exact construction. Likewise "squaring the circle is impossible" does not mean you cannot build a square almost-equal in area — only that you cannot build one of area exactly $\pi r^2$ .

⚠️ Common mistake

The tools are idealized. The straightedge is unmarked — you may not slide it to mark off a measured distance. Allowing a marked ruler (a neusis construction) genuinely increases your power and does let you trisect angles and double the cube. The impossibility results are about the specific Euclidean toolkit, not about geometry in general.

Counterexample

A heptagon (7 sides) cannot be constructed, but a 17-gon (17 sides) can. Intuitively one might guess "more sides = harder," but constructibility depends on the arithmetic of the side count, not its size. A regular $n$ -gon is constructible iff $n = 2^k p_1 p_2 \cdots p_m$ where the $p_i$ are distinct Fermat primes ( $3, 5, 17, 257, 65537$ ). Since $7$ is not a Fermat prime, the heptagon is impossible; since $17$ is, Gauss's 17-gon is constructible — a fact he was so proud of he requested it on his tombstone.

connection

Geometric constructions sit at the meeting point of geometry and algebra. The characterization theorem is a gateway to Field Extensions and Galois Theory, where degree-towers and group theory explain every impossibility uniformly. The square-root constructions rest on Similar Triangles and Thales' Theorem. The constructible angles connect directly to Roots of Unity and Cyclotomic Polynomials, since a regular $n$ -gon is built from the $n$ th roots of unity. The impossibility of squaring the circle depends on Transcendental Numbers, specifically the transcendence of $\pi$ .

🔬 Where this appears

⏳ History

- ~440 BCE: Hippocrates of Chios squares certain lunes (crescents), the first nontrivial constructible curved-region results, raising hope of squaring the circle.

- ~420 BCE: The three classical problems crystallize: doubling the cube (the Delian problem), trisecting the angle, squaring the circle.

- ~300 BCE: Euclid's Elements</strong> codifies construction methods in Books I–IV and VI, establishing the straightedge-and-compass framework as a logical system.

- ~250 BCE: Archimedes gives a neusis (marked-ruler) trisection, showing the problems yield to stronger tools.

- 1796: Carl Friedrich Gauss, aged 19, constructs the regular 17-gon and characterizes constructible polygons via Fermat primes.

- 1801: Gauss states the full polygon criterion in Disquisitiones Arithmeticae; he proved sufficiency.

- 1837: Pierre Wantzel proves the impossibility of doubling the cube and trisecting the general angle, and the necessity of the Gauss condition.

- 1882: Ferdinand von Lindemann proves $\pi$ is transcendental, settling the impossibility of squaring the circle once and for all.

- 20th century: Origami (Huzita–Hatori axioms, 1980s–90s) and other extended toolkits show how relaxing the rules restores solvability of the cubic problems.

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also