Unit III Commentary & Notes: Projective Purity, Golden Ratio & Triangle Synthesis¶
Cross-Corpus Commentary Linking Euclid, Pappus, ETC, Mascheroni & Steiner
1. Spontaneous Emergence of Compass-Only and Straightedge-Only Optimality¶
One of the deepest discoveries in Unit III is that unrestricted Géométrographie—seeking only to minimize \(S\) when both straightedge and compass are freely available—spontaneously selects:
Pure compass (\(l_1 = l_2 = 0\)) for circle inversion (Problem XXXVIII: \(5C_1 + 3C_3, S = 8\), identical to Mascheroni’s Libro V, § 86, mascheroni-1797) and point reflection (Problem XVIII bis: \(S = 6\));
Pure straightedge (\(m_1 = m_2 = m_3 = 0\)) for circle polarity (Problem XXXIX: \(12R_1 + 7R_2, S = 19\), identical to Steiner’s § 10, steiner-1833) and harmonic conjugation (Problem XLVI: \(9R_1 + 6R_2, S = 15\), from Pappus’s Synagoge VII, pappus-synagoge).
This reveals a fundamental duality between transformation group and optimal instrument: conformal/inversive point maps minimize \(S\) on the compass, whereas projective/harmonic collineations and correlations minimize \(S\) on the straightedge!
2. Algebraic Anatomy of Lemoine’s 1-Line Golden Section (Problem XXXVI)¶
In Euclid’s Elements II.11 (euclid-heath; see and ), dividing \(AB\) of length \(2a\) in extreme and mean ratio constructs a square \(ABDC\) on \(AB\), bisects \(AC\) at \(E\), extends \(CA\) to \(F\) with \(EF = EB = a\sqrt{5}\), and constructs square \(AFGH\) cutting \(AB\) at \(H\). Even stripped to its barest textbook form, three straight lines are drawn.
In Lemoine’s Problem XXXVI (\(2R_1 + R_2 + 9C_1 + 6C_3, S = 18\)), only one straight line—the perpendicular bisector \(x = a\) through midpoint \(M\) of \(AB\)—is ever traced. Placing \(C = (a, 2a)\) on that single line via \(M(AB)\) creates the un-drawn Pythagorean hypotenuse \(BC = \sqrt{a^2 + (2a)^2} = a\sqrt{5}\) directly between existing points \(B = (2a, 0)\) and \(C = (a, 2a)\). Sweeping \(B(C)\) onto the carrier line at \(E = (2a - a\sqrt{5}, 0)\) yields:
which arc \(A(ME)\) transfers onto \(AB\) at \(X\). Every quadratic irrational in \(\mathbb{Q}(\phi) = \mathbb{Q}(\sqrt{5})\) is extracted via compass distance between determined points without tracing hypotenuse \(BC\)!
3. Mannheim and Brocard’s Objections Answered (pp. 57–59)¶
Two celebrated historical exchanges in Lemoine’s Considérations générales illuminate Unit III:
Reply to Amédée Mannheim on the Radical Axis (pp. 57–58): Mannheim objected that in descriptive geometry lectures one defines the radical axis of two exterior circles by joining the midpoints of two common tangents. Lemoine calmly evaluates Mannheim’s construction: two common tangents plus two midpoints cost \(18R_1 + 9R_2 + 19C_1 + 12C_3\) (\(S = 58, E = 37\), 9 lines, 12 circles!), whereas sweeping two arbitrary intersecting circles (\(2C_3\)) costs \(10R_1 + 5R_2 + 2C_3\) (\(S = 17, E = 10\))—more than a threefold reduction in both effort and alignment error!
Reply to Henri Brocard on :math:`omega, omega’` (pp. 58–59): Brocard wrote to Lemoine defending the symmetric beauty of intersecting the three adjoint circles tangent to the sides of \(\triangle ABC\) at its vertices. Lemoine prints the exact accounting: Brocard’s symmetric construction costs \(22R_1 + 11R_2 + 24C_1 + 17C_3\) (\(S = 74, E = 46\)), whereas Lemoine’s Problem LV obtains both \(\omega\) and \(\omega'\) in \(12R_1 + 6R_2 + 12C_1 + 8C_3\) (\(S = 38, E = 24\)), demonstrating once and for all that symmetric algebraic specification is distinct from optimal constructive synthesis.