Chapter VII: Geometrical Constructions Based Upon Axioms I–V — Notes

Historical, Philological & Cross-Repository Commentary

GEOMETOR Commentary & Cross-Repository Notes

1. From Streckenübertrager (1899) to Eichmass (Later Editions) & geometor/elements

In geometor/elements and Euclid’s Book I, straightedge (Postulates 1–2) and collapsible/modern compass (Postulate 3, I.2–3) generate quadratic field extensions via line-circle and circle-circle intersections (Euclid I.1, I.22). Why does Chapter VII replace the full compass with a transferer of segments (Streckenübertrager in 1899; Townsend: “transferer of segments”)?

  • Because Axioms `I–V` (without *Vollständigkeit*) never mention circles intersecting circles! Axiom IV 1 only guarantees laying off a given segment $AB$ along a given straight line from a point $A'$.

  • In later editions (e.g., 7th ed., 1930), Hilbert sharpened this instrument even further to the fixed-unit gauge (Eichmass), which can only lay off a single fixed unit segment $1$ along any ray—because once parallel lines (Problem 2) are constructible, any segment ratio can be transferred via parallels using a single fixed unit!

  • Look at fig_gauge_constructions.svg (§ 36, Problem 5): to drop a perpendicular without drawing intersecting circles (Euclid I.11–12), Hilbert lays off four equal radii $AB = AC = AD = AE$ from $A$ along the base line and two oblique rays through $A$; by Thales’s semicircle theorem (Euclid III.31), $\angle BDC$ and $\angle BEC$ are right angles, so $BD$ and $CE$ are two altitudes of $\triangle BCF$ meeting at its orthocenter $H$, making the third line $FH$ strictly perpendicular to $BC$ using straightedge and segment-transferer alone!

2. Birth of Real Algebraic Geometry & Hilbert’s 17th Problem (§§ 37–39)

Sections 37–39 bridge synthetic constructive geometry (geometor/elements) with deep algebraic number theory:

  • Why Compass > Segment-Transferer (`§ 37`): Every coordinate in $\Omega$ (§ 9) is built from sums of squares $\sqrt{\alpha^2 + \beta^2}$, so all algebraic conjugates of every number in $\Omega$ are real (totally real). By contrast, a circle can intersect a line whenever $\alpha > 0$ numerically, even if an algebraic conjugate of $\alpha$ is negative! Thus $\sqrt{2\sqrt{2} - 2}$ (constructible by intersecting a unit circle with $x = \sqrt{2} - 1$) has imaginary conjugate $\sqrt{-2\sqrt{2} - 2}$ and cannot lie in $\Omega$!

  • Hilbert’s 17th Problem (`§ 38`, `Theorem 43`): Proving Theorem 43 (every positive-definite univariate polynomial over $\mathbb{Q}$ is a quotient of sums of squares) led Hilbert at the end of § 38 to ask whether every positive-definite rational function in $n$ variables over $\mathbb{Q}$ is a sum of squares of rational functions—posed one year later in Paris as Hilbert’s Seventeenth Problem (1900) and solved affirmatively in 1927 by Emil Artin via Artin–Schreier real closed fields!