Mathematical Summary & Symbolic Architecture

Exact Symbolic Projective Geometry, Veronese Invariants & 60-Line Web

From Two Straight Lines to Any Conic Section

Thirteen centuries after Pappus of Alexandria proved his Hexagon Lemma (Synagoge VII.139) for six points distributed along two straight lines $a$ and $b$, sixteen-year-old Blaise Pascal published a single printed broadsheet in February 1640—Essai pour les coniques—announcing a breathtaking generalization: the two lines can be replaced by any conic section.

Algebraically, Pappus’s theorem and Pascal’s theorem are two faces of one projective law. Six points $P_k = (x_k, y_k)$ lie on a planar curve of degree two—whether a smooth conic $\mathcal{C}: A x^2 + B x y + C y^2 + D x + E y + F = 0$ (circle, ellipse, parabola, or hyperbola) or a degenerate conic $L_a(x, y) \cdot L_b(x, y) = 0$ (Pappus’s pair of lines)—if and only if their $6 \times 6$ Veronese matrix drops rank ($\det V_6 = 0$):

\[\begin{split}\det \begin{bmatrix} x_0^2 & x_0 y_0 & y_0^2 & x_0 & y_0 & 1 \\ x_1^2 & x_1 y_1 & y_1^2 & x_1 & y_1 & 1 \\ \vdots & \vdots & \vdots & \vdots & \vdots & \vdots \\ x_5^2 & x_5 y_5 & y_5^2 & x_5 & y_5 & 1 \end{bmatrix} = 0.\end{split}\]

In geometor.pascal.projective.are_coconic, this exact SymPy rank test unifies smooth circles and ellipses with degenerate configurations under a single algebraic predicate, while Colin Maclaurin (1720) and William Braikenridge (1733) proved the converse: if the three opposite-side intersections of a hexagon are collinear, its six vertices necessarily lie on a conic.

Theorem (Pascal’s Hexagrammum Mysticum, 1640)

Let $P_0, P_1, P_2, P_3, P_4, P_5$ be any six distinct points on an irreducible conic $\mathcal{C}$ in $\mathbb{P}^2$, ordered in a cyclic hexagon $\langle P_{c_0}, P_{c_1}, P_{c_2}, P_{c_3}, P_{c_4}, P_{c_5} \rangle$. Then the three intersections of opposite sides,

\[M_{03} = [P_{c_0} P_{c_1}] \cap [P_{c_3} P_{c_4}], \quad M_{14} = [P_{c_1} P_{c_2}] \cap [P_{c_4} P_{c_5}], \quad M_{25} = [P_{c_2} P_{c_3}] \cap [P_{c_5} P_{c_0}],\]

are collinear along a single line $p$—the Pascal line of the inscribed hexagram.

The Mystic Star Hexagram on the Pythagorean Circle

To explore Pascal’s theorem in 100% exact rational arithmetic ($\mathbb{Q}^2$) without floating-point approximations, geometor.pascal.projective.circle_stereographic_point maps any rational parameter $t \in \mathbb{Q}$ onto the unit circle $x^2 + y^2 = 1$ via Weierstrass stereographic projection:

\[P(t) = \left(\frac{1 - t^2}{1 + t^2},\; \frac{2t}{1 + t^2}\right) \in \mathbb{Q}^2.\]

Seeding preset="hexagrammum" with $t = \left(0,\, \frac{1}{3},\, 1,\, 3,\, -2,\, -\frac{1}{2}\right)$ places six cardinal and $3\text{-}4\text{-}5$ Pythagorean vertices around all four quadrants of $x^2 + y^2 = 1$:

\[P_0 = (1, 0),\quad P_1 = \left(\frac{4}{5}, \frac{3}{5}\right),\quad P_2 = (0, 1),\quad P_3 = \left(-\frac{4}{5}, \frac{3}{5}\right),\quad P_4 = \left(-\frac{3}{5}, -\frac{4}{5}\right),\quad P_5 = \left(\frac{3}{5}, -\frac{4}{5}\right).\]

Tracing the non-convex mystic star cycle $\langle P_0, P_2, P_4, P_1, P_3, P_5 \rangle$ produces two interior meets and one compact exterior meet, all strictly rational:

\[M_{03} = \left(\frac{2}{5},\, \frac{3}{5}\right), \qquad M_{14} = \left(-\frac{3}{10},\, \frac{1}{10}\right), \qquad M_{25} = \left(\frac{9}{5},\, \frac{8}{5}\right),\]

aligned along the exact Pascal line $25x - 35y + 11 = 0$:

Pascal's Hexagrammum Mysticum (1640) on the Unit Circle ( O P0 ): x**2 + y**2 - 1 = 0 [ P0 P2 ]: -x - y + 1 = 0 [ P2 P4 ]: 9*x/5 - 3*y/5 + 3/5 = 0 [ P4 P1 ]: -7*x/5 + 7*y/5 + 7/25 = 0 [ P1 P3 ]: y - 3/5 = 0 [ P3 P5 ]: 7*x/5 + 7*y/5 + 7/25 = 0 [ P5 P0 ]: -4*x/5 + 2*y/5 + 4/5 = 0 [ M03 M14 ]: x/2 - 7*y/10 + 11/50 = 0 O = (0, 0) P0 = (1, 0) P1 = (4/5, 3/5) P2 = (0, 1) P3 = (-4/5, 3/5) P4 = (-3/5, -4/5) P5 = (3/5, -4/5) M03 = (2/5, 3/5) M14 = (-3/10, 1/10) M25 = (9/5, 8/5) O P0 P1 P2 P3 P4 P5 M03 M14 M25
step Inscribe the six Pythagorean rational points P0, P1, P2, P3, P4, and P5 around the unit circle 𝒞 centered at O $= (0, 0)$.
step Trace the self-intersecting star hexagram P0 P2 P4 P1 P3 P5 and intersect the first opposite side pair P0 P2 ($x + y = 1$) with P1 P3 ($y = \frac{3}{5}$) at M03 $= \left(\frac{2}{5}, \frac{3}{5}\right)$.
step Intersect the second opposite pair P2 P4 ($3x - y + 1 = 0$) with P3 P5 ($5x + 5y + 1 = 0$) inside the circle at M14 $= \left(-\frac{3}{10}, \frac{1}{10}\right)$, and the third opposite pair P4 P1 ($5x - 5y = 1$) with P5 P0 ($2x - y = 2$) outside the circle at M25 $= \left(\frac{9}{5}, \frac{8}{5}\right)$.
step Draw the Pascal line p ($25x - 35y + 11 = 0$): $25\left(\frac{2}{5}\right) - 35\left(\frac{3}{5}\right) + 11 = 10 - 21 + 11 = 0$, $25\left(-\frac{3}{10}\right) - 35\left(\frac{1}{10}\right) + 11 = -11 + 11 = 0$, and $25\left(\frac{9}{5}\right) - 35\left(\frac{8}{5}\right) + 11 = 45 - 56 + 11 = 0$!

Universal Symbolic Proof & Conic Invariance

Why did Pascal need only prove his lemma for the circle in order to establish it for every ellipse, parabola, and hyperbola? Because any non-degenerate conic is the central projection of a circle from the apex of Apollonius’s cone, and incidence ($P \in \ell$) is preserved under every projective collineation $T \in \mathrm{PGL}_3(\mathbb{R})$.

  1. Universal 6-Parameter Proof: verify_pascal_symbolic() parameterizes six generic points $P_k = [1 - t_k^2 : 2 t_k : 1 + t_k^2]$ over the polynomial ring $\mathbb{Z}[t_0, t_1, t_2, t_3, t_4, t_5]$, computes $M_{03}, M_{14}, M_{25}$ via homogeneous 3D cross products, and expands $\det[M_{03}, M_{14}, M_{25}]$—which cancels identically to zero.

  2. General Ellipse Verification: Applying the affine transformation $(x, y) \mapsto (2x, y)$ in preset="ellipse" maps the Pythagorean hexagram onto the non-circular ellipse $\mathcal{E}: \frac{x^2}{4} + y^2 = 1$ ($x^2 + 4y^2 - 4 = 0$), stretching the three Pascal intersections to $M'_{03} = \left(\frac{4}{5}, \frac{3}{5}\right)$, $M'_{14} = \left(-\frac{3}{5}, \frac{1}{10}\right)$, and $M'_{25} = \left(\frac{18}{5}, \frac{8}{5}\right)$ along $25x - 70y + 22 = 0$:

Pascal's Hexagrammum Mysticum on the Ellipse x^2/4 + y^2 = 1 ( O P0 ): x**2/4 + y**2 - 1 = 0 [ P0 P2 ]: -x - 2*y + 2 = 0 [ P2 P4 ]: 9*x/5 - 6*y/5 + 6/5 = 0 [ P4 P1 ]: -7*x/5 + 14*y/5 + 14/25 = 0 [ P1 P3 ]: y - 3/5 = 0 [ P3 P5 ]: 7*x/5 + 14*y/5 + 14/25 = 0 [ P5 P0 ]: -4*x/5 + 4*y/5 + 8/5 = 0 [ M03 M14 ]: x/2 - 7*y/5 + 11/25 = 0 O = (0, 0) P0 = (2, 0) P1 = (8/5, 3/5) P2 = (0, 1) P3 = (-8/5, 3/5) P4 = (-6/5, -4/5) P5 = (6/5, -4/5) M03 = (4/5, 3/5) M14 = (-3/5, 1/10) M25 = (18/5, 8/5) O P0 P1 P2 P3 P4 P5 M03 M14 M25
step Compare the stretched ellipse ℰ ($\frac{x^2}{4} + y^2 = 1$) with the unit circle above: every chord, opposite meet M'03, M'14, M'25, and Pascal line p' transforms covariantly while preserving exact three-point collinearity.

The Complete 60-Line Hexagrammum Mysticum Web

On two lines (Pappus), permuting the vertices yields $3! = 6$ hexagons and 6 Pappus lines concurring in 2 Steiner Star points. On a smooth conic, however, no three of the six vertices are collinear, so every cyclic ordering of $P_0, \dots, P_5$ up to dihedral rotation and reversal ($D_6$, order 12) inscribes a valid hexagon—producing:

\[\frac{6!}{2 \cdot 6} = \mathbf{60\text{ Pascal lines}}!\]

In the nineteenth century, Jakob Steiner (1828), Julius Plücker (1830), Thomas Penyngton Kirkman (1850), Arthur Cayley, and George Salmon uncovered an astonishing, self-dual incidence architecture hidden inside these 60 lines. Running enumerate_hexagrammum_mysticum(preset="hexagrammum") constructs and verifies every element of this web in exact homogeneous coordinates with zero combinatorial degeneracies:

Exact Projective Census of Pascal’s Hexagrammum Mysticum (preset="hexagrammum")

Geometric Stratum

Count

Incidence Law & Discoverer

Conic Chords $[P_i P_j]$

15

$\binom{6}{2} = 15$ secant lines joining pairs of the 6 conic vertices

Pascal Diagonal Points

45

$\frac{1}{2}\binom{6}{2}\binom{4}{2} = 45$ intersections $[P_i P_j] \cap [P_k P_l]$; 4 Pascal lines pass through each

Pascal Lines $p_H$

60

$6! / 12 = 60$ lines, each carrying 3 of the 45 diagonal points (Pascal, 1640)

Steiner Points $G$

20

60 Pascal lines concur 3-by-3 at 20 Steiner points—2 per complementary triangle pair (Steiner, 1828)

Kirkman Points $K$

60

60 Pascal lines also concur 3-by-3 in a second way at 60 Kirkman points—3 on each Pascal line (Kirkman, 1850)

Plücker Lines $\ell_{\mathrm{Pl}}$

15

20 Steiner points lie 4-by-4 along 15 Plücker lines—3 through each Steiner point (Plücker, 1830)

Bi-Triangle $S_3$ Concurrency: How Pappus’s Two Star Points Generalize

How do Steiner’s 20 points on a smooth conic relate to the 2 Star points $(S_+, S_-)$ of the Pappus–Steiner theorem? Partition the six conic vertices into two complementary inscribed triangles—say $\triangle_A = (P_0, P_2, P_4)$ and $\triangle_B = (P_1, P_3, P_5)$. There are $\frac{1}{2}\binom{6}{3} = 10$ such complementary triangle pairs.

Fixing $\triangle_A$ and permuting $\triangle_B$ by $\sigma \in S_3$ generates 6 alternating hexagons $\langle P_0, B_{\sigma_0}, P_2, B_{\sigma_1}, P_4, B_{\sigma_2} \rangle$—the exact smooth-conic counterpart of the six $S_3$ Pappus hexagons! Just as on two lines:

  • The 3 even ($\sigma \in A_3$) Pascal lines concur at the even Steiner point $G_+ = \left(\frac{78}{25},\; \frac{1}{25}\right)$.

  • The 3 odd ($\sigma \in S_3 \setminus A_3$) Pascal lines concur at the odd Steiner point $G_- = \left(\frac{498}{1565},\; \frac{281}{1565}\right)$.

Across all $10$ triangle partitions, $10 \times 2 = \mathbf{20}$ Steiner points arise; when the conic degenerates into two straight lines $a \cup b$, only the single partition separating $a$ from $b$ survives without collapsing onto the carrier lines, leaving Pappus’s two Star points $(S_+, S_-)$!

Bi-Triangle S_3 Pascal Lines Concurring at Steiner Points G+ and G- ( O P0 ): x**2 + y**2 - 1 = 0 [ P0 P1 ]: -3*x/5 - y/5 + 3/5 = 0 [ P0 P3 ]: -3*x/5 - 9*y/5 + 3/5 = 0 [ P0 P5 ]: 4*x/5 - 2*y/5 - 4/5 = 0 [ P2 P1 ]: 2*x/5 + 4*y/5 - 4/5 = 0 [ P2 P3 ]: 2*x/5 - 4*y/5 + 4/5 = 0 [ P2 P5 ]: 9*x/5 + 3*y/5 - 3/5 = 0 [ P4 P1 ]: -7*x/5 + 7*y/5 + 7/25 = 0 [ P4 P3 ]: -7*x/5 - y/5 - 1 = 0 [ P4 P5 ]: y + 4/5 = 0 [ M0143 M2145 ]: 49*x/5 + 28*y/5 - 154/5 = 0 [ M0145 M2143 ]: -147*x/65 - 427*y/195 + 217/195 = 0 [ M0341 M2345 ]: x - 4*y + 2/5 = 0 [ M0345 M2341 ]: -3*x - y + 47/5 = 0 [ M0541 M2543 ]: -39*x/10 - 33*y/10 + 123/10 = 0 [ M0543 M2541 ]: -83*x/30 + 19*y/30 + 23/30 = 0 [ S1 S2 ]: -1092*x/7825 - 21924*y/7825 + 4284/7825 = 0 O = (0, 0) P0 = (1, 0) P1 = (4/5, 3/5) P2 = (0, 1) P3 = (-4/5, 3/5) P4 = (-3/5, -4/5) P5 = (3/5, -4/5) M2145 = (18/5, -4/5) M2305 = (2, 2) M0145 = (19/15, -4/5) M2143 = (-12/13, 19/13) M2503 = (1/4, 1/4) M0341 = (2/5, 1/5) M2105 = (6/5, 2/5) M0345 = (17/5, -4/5) M2341 = (12/5, 11/5) M0541 = (9/5, 8/5) M2103 = (4, -1) M2541 = (3/10, 1/10) M2301 = (4/7, 9/7) S1 = (78/25, 1/25) S2 = (498/1565, 281/1565) O P0 P1 P2 P3 P4 P5 S+ S−
step Partition the conic vertices on 𝒞 into complementary triangles $\triangle_A =$ (P0, P2, P4) and $\triangle_B =$ (P1, P3, P5), joined by their nine cross-chords.
step Trace the three even ($A_3$) Pascal lines p012 ($49x + 28y = 154$), p120 ($15x + 5y = 47$), and p201 ($39x + 33y = 123$), concurring at G_+ $= \left(\frac{78}{25},\, \frac{1}{25}\right)$.
step Trace the three odd ($S_3 \setminus A_3$) Pascal lines p021, p102 ($5x - 20y + 2 = 0$), and p210 ($83x - 19y = 23$), concurring inside the circle at G_- $= \left(\frac{498}{1565},\, \frac{281}{1565}\right)$, and join the pair along G_+G_-.

Golden Hexagrammum Mysticum in $\mathbb{Q}(\phi)$

By Michel Chasles’s Theorem (1865), four points $P_0, P_1, P_2, P_3$ on an irreducible conic $\mathcal{C}$ possess an intrinsic conic cross-ratio $(P_0, P_1; P_2, P_3)_{\mathcal{C}}$: for any fifth observer point $V \in \mathcal{C}$, the pencil of four rays $[V P_0], [V P_1], [V P_2], [V P_3]$ has the exact same cross-ratio, independent of $V$.

Because Weierstrass stereographic projection is a birational isomorphism $\mathbb{P}^1 \to \mathcal{C}$, seeding preset="golden_conic" with the golden stereographic chain $t = (0,\, 1,\, \phi,\, \phi + 1,\, -1,\, -\phi)$ inscribes six exact $\mathbb{Q}(\sqrt{5})$ points on $x^2 + y^2 = 1$:

\[P_0 = (1, 0), \quad P_1 = (0, 1), \quad P_2 = \left(-\frac{\sqrt{5}}{5},\, \frac{2\sqrt{5}}{5}\right), \quad P_3 = \left(-\frac{\sqrt{5}}{3},\, \frac{2}{3}\right), \quad P_4 = (0, -1), \quad P_5 = \left(-\frac{\sqrt{5}}{5},\, -\frac{2\sqrt{5}}{5}\right).\]
  1. Chasles Golden Conic Cross-Ratio: Calling conic_cross_ratio(P_0, P_1, P_2, P_3, V) from either $V = P_4$ or $V = P_5$ simplifies identically to the Golden Ratio $\phi = \frac{1 + \sqrt{5}}{2}$!

  2. Emergent Golden Section on the Mystic Hexagram: Constructing the inscribed hexagram on cycle $\langle P_0, P_2, P_5, P_1, P_3, P_4 \rangle$ yields exact opposite-side intersections:

    \[M_{03} = \left(\frac{5 - 4\sqrt{5}}{11},\; \frac{7 + \sqrt{5}}{11}\right), \qquad M_{14} = \left(-\frac{\sqrt{5}}{5},\; 0\right), \qquad M_{25} = \left(-\phi^{-1},\; -\phi\right) = \left(\frac{1 - \sqrt{5}}{2},\; -\frac{1 + \sqrt{5}}{2}\right).\]

    Running geometor.divine (find_golden_sections) on this model automatically discovers an emergent golden section: the Pascal intersection $M_{25} = (-\phi^{-1}, -\phi)$ extends the conic chord $[P_4 P_0]$ ($y = x - 1$) in the exact Divine Proportion! Because $x(P_0) - x(P_4) = 1 - 0 = 1$ and $x(P_4) - x(M_{25}) = 0 - (-\phi^{-1}) = \phi^{-1}$ along the $45^\circ$ chord, $|P_0 P_4| / |P_4 M_{25}| = |M_{25} P_0| / |P_0 P_4| = \phi$ identically:

Golden Hexagrammum Mysticum in Q(phi) — Emergent Section / M25 P4 P0 / ( O P0 ): x**2 + y**2 - 1 = 0 [ P0 P2 ]: -2*sqrt(5)*x/5 + y*(-1 - sqrt(5)/5) + 2*sqrt(5)/5 = 0 [ P2 P5 ]: x + sqrt(5)/5 = 0 [ P5 P1 ]: x*(-1 - 2*sqrt(5)/5) + sqrt(5)*y/5 - sqrt(5)/5 = 0 [ P1 P3 ]: x/3 - sqrt(5)*y/3 + sqrt(5)/3 = 0 [ P3 P4 ]: 5*x/3 + sqrt(5)*y/3 + sqrt(5)/3 = 0 [ P4 P0 ]: -x + y + 1 = 0 [ M03 M14 ]: x*(sqrt(5)/11 + 7/11) + y*(-5/11 + 9*sqrt(5)/55) + 1/11 + 7*sqrt(5)/55 = 0 O = (0, 0) P0 = (1, 0) P1 = (0, 1) P2 = (-sqrt(5)/5, 2*sqrt(5)/5) P3 = (-sqrt(5)/3, 2/3) P4 = (0, -1) P5 = (-sqrt(5)/5, -2*sqrt(5)/5) M03 = (5/11 - 4*sqrt(5)/11, sqrt(5)/11 + 7/11) M14 = (-sqrt(5)/5, 0) M25 = (1/2 - sqrt(5)/2, -sqrt(5)/2 - 1/2) O P0 P1 P2 P3 P4 P5 M03 M14 M25
step Project the golden parameter chain $t = (0, 1, \phi, \phi + 1, -1, -\phi)$ onto the unit circle 𝒞 to obtain P0, P1, P2, P3, P4, and P5, where $(P_0, P_1; P_2, P_3)_{\mathcal{C}} = \phi$.
step Trace the hexagram P0 P2 P5 P1 P3 P4 and locate the three exact $\mathbb{Q}(\sqrt{5})$ opposite-side intersections M03, M14, and M25 $= (-\phi^{-1}, -\phi)$ on Pascal line p.
step Inspect the emergent geometor.divine golden section M25 P4 P0 along chord P4 P0: $|P_0 P_4| = \sqrt{2}$ and $|P_4 M_{25}| = \phi^{-1}\sqrt{2}$, dividing the extended conic chord in the exact Golden Ratio $\phi = \frac{1 + \sqrt{5}}{2}$.