{
  "id": "pascal-theorem",
  "title": "Pascal's Theorem",
  "lang": "en",
  "source_lang": "en",
  "summary": "Pascal’s theorem states that the three intersections of opposite sides of a hexagon inscribed in a conic are collinear.",
  "category": "mathematics",
  "subcategory": "geometry",
  "type": "concept",
  "aliases": [
    "Pascal’s hexagon theorem",
    "Hexagrammum mysticum theorem"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/pascal-theorem/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/pascal-theorem/",
      "json": "https://aiwiki.page/zh/pascal-theorem.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Pascal’s theorem is a fundamental result in projective geometry: if six distinct points lie on a nondegenerate conic, the three intersections of opposite sides of the hexagon they define lie on a single straight line. This line is called the Pascal line of the ordered hexagon. The theorem applies to a circle, ellipse, parabola, or hyperbola, with parallel sides handled by intersections at infinity. It is named after Blaise Pascal, who published the result in 1640."
    },
    {
      "heading": "Statement and notation",
      "text": "Let \\(A,B,C,D,E,F\\) be six distinct points on a nondegenerate conic. Write \\(AB\\) for the complete line through \\(A\\) and \\(B\\), rather than merely the segment joining them. Define\n\n\\[X=AB\\cap DE,\\qquad Y=BC\\cap EF,\\qquad Z=CD\\cap FA.\\]\n\nThen \\(X,Y,Z\\) are collinear.\n\nThe pairs \\(AB,DE\\), \\(BC,EF\\), and \\(CD,FA\\) are the opposite sides of the hexagon \\(ABCDEF\\). Its vertices need not occur in that order around the conic: the hexagon may be self-intersecting, and the intersections may lie outside the figure.\n\nThe order of the six points matters. Reordering them generally produces a different Pascal line. For six points in a generic configuration, the possible orderings yield 60 distinct Pascal lines. This count comes from identifying orders related by cyclic rotation or reversal:\n\n\\[\\frac{6!}{6\\cdot 2}=60.\\]\n\nSpecial configurations can make some of these lines coincide."
    },
    {
      "heading": "Projective interpretation",
      "text": "In Euclidean geometry, parallel lines do not meet at a finite point, so a statement about intersections requires qualifications. Projective geometry adds points at infinity: all lines with the same direction meet at the corresponding point at infinity.\n\nConsequently, Pascal’s theorem has a uniform projective statement, including configurations with parallel opposite sides. If one of \\(X,Y,Z\\) is at infinity, the line through the other two has the direction represented by that point. If two are distinct points at infinity, their Pascal line is the line at infinity, and the third intersection is also at infinity.\n\nThe theorem concerns only incidence—whether points lie on lines or conics—not distances, angles, or lengths. Its conclusion is therefore preserved by projective transformations. Although ellipses, parabolas, and hyperbolas differ in an affine or metric setting, a nondegenerate real projective conic with real points can be transformed projectively into any other such conic."
    },
    {
      "heading": "An algebraic proof",
      "text": "A concise proof uses homogeneous equations and the fact that a nondegenerate conic can be parametrized by quadratic expressions.\n\nWork in the real or complex projective plane. Let the conic have equation\n\n\\[Q=0,\\]\n\nwhere \\(Q\\) is a homogeneous quadratic polynomial. For any two vertices \\(U,V\\), let \\(\\ell_{UV}=0\\) be an equation of their joining line. Form the cubic polynomials\n\n\\[F=\\ell_{AB}\\ell_{CD}\\ell_{EF}, \\qquad G=\\ell_{BC}\\ell_{DE}\\ell_{FA}.\\]\n\nBoth vanish at all six vertices.\n\nUnder a projective parametrization of the conic, \\(F\\) and \\(G\\) restrict to homogeneous polynomials of degree six in two parameters. Each has precisely the six distinct parameter values corresponding to \\(A,B,C,D,E,F\\) as its zeros. They are therefore proportional on the conic: for a suitable nonzero scalar \\(\\lambda\\),\n\n\\[F-\\lambda G=0\\]\n\nat every point of the conic.\n\nA polynomial vanishing identically on a nondegenerate conic is divisible by its defining quadratic. Hence\n\n\\[F-\\lambda G=Q\\,L,\\]\n\nwhere \\(L\\) is a nonzero homogeneous linear polynomial.\n\nAt \\(X=AB\\cap DE\\), both \\(F\\) and \\(G\\) vanish, so\n\n\\[Q(X)L(X)=0.\\]\n\nThe point \\(X\\) cannot lie on the conic: otherwise the line \\(AB\\), already meeting the conic at \\(A\\) and \\(B\\), would have a third distinct intersection with it. Thus \\(Q(X)\\ne0\\), giving \\(L(X)=0\\). The same reasoning applies to \\(Y\\) and \\(Z\\).\n\nAll three points therefore lie on the line \\(L=0\\), proving the theorem.\n\nThis proof expresses a connection between an incidence theorem and the factorization of polynomial equations, a recurring theme in algebraic geometry."
    },
    {
      "heading": "Converse and construction of conics",
      "text": "A standard converse states that if six distinct points, with no three collinear, have the three opposite-side intersections collinear, then the six points lie on a conic.\n\nTo see the construction behind this result, take five points \\(A,B,C,D,E\\) in general position. They determine a unique nondegenerate conic. Compute\n\n\\[X=AB\\cap DE\\]\n\nand choose a line \\(m\\) through \\(X\\). Set\n\n\\[Y=m\\cap BC,\\qquad Z=m\\cap CD,\\]\n\nthen construct\n\n\\[F=EY\\cap AZ.\\]\n\nWhenever these intersections give a nondegenerate configuration, the resulting point \\(F\\) lies on the conic through the original five points. The line \\(m\\) is the Pascal line of \\(ABCDEF\\).\n\nThus Pascal’s theorem provides both an incidence test for six points to lie on a conic and a straightedge construction of further conic points from five known ones."
    },
    {
      "heading": "Related theorems and limiting cases",
      "text": "Brianchon’s theorem\n\nUnder projective duality, points and lines exchange roles, as do joining and intersection. Pascal’s theorem corresponds to Brianchon’s theorem:\n\nIf the six sides of a hexagon are tangent to a nondegenerate conic, the three lines joining opposite vertices are concurrent.\n\nThe duality can be expressed using the polarity associated with a nondegenerate conic. A point on the conic corresponds to its tangent line, while Pascal’s collinear intersection points correspond to Brianchon’s concurrent joining lines.\n\nPappus’s theorem\n\nPappus’s theorem is the corresponding incidence result for a degenerate conic consisting of two distinct lines. Take \\(A,C,E\\) on one line and \\(B,D,F\\) on the other. The intersections\n\n\\[AB\\cap DE,\\qquad BC\\cap EF,\\qquad CD\\cap FA\\]\n\nare collinear, provided the configuration is well-defined.\n\nIn the ordinary real or complex projective plane, Pappus’s theorem can therefore be viewed as a degenerate version of Pascal’s theorem. The nondegenerate-conic proof above does not transfer unchanged, because the quadratic equation now factors into two line equations.\n\nTangent versions\n\nWhen two adjacent vertices approach the same point along a smooth conic, their joining secant approaches the tangent at that point. Taking such limits gives versions of Pascal’s theorem involving tangent lines and fewer than six distinct vertices.\n\nThese statements require the limiting interpretation. Simply replacing two distinct vertices by one in the original notation leaves their joining line undefined."
    },
    {
      "heading": "Historical context and significance",
      "text": "Pascal presented the theorem in his 1640 Essai pour les coniques, written when he was sixteen. His work on conics was influenced by Gérard Desargues and the emerging projective approach to geometry. The expression hexagrammum mysticum, or “mystic hexagram,” is associated with the hexagon configuration and its later-developed incidence structure.\n\nPascal’s theorem became a central example of a geometric fact governed by incidence rather than measurement. Its many Pascal lines led to further investigations of configurations of points and lines. In modern treatments, it connects synthetic geometry, homogeneous coordinates, polynomial methods, and geometric constructions.\n\nThe theorem’s exact collinearity condition should be distinguished from numerical testing. For approximate coordinates, an almost-zero collinearity determinant may reflect either a nearly valid configuration or numerical error; the theorem itself concerns exact incidence."
    }
  ],
  "links": [
    "projective-geometry",
    "conic-section",
    "circle",
    "ellipse",
    "parabola",
    "hyperbola",
    "blaise-pascal",
    "euclidean-geometry",
    "polynomial",
    "algebraic-geometry",
    "pappus-theorem",
    "homogeneous-coordinates",
    "determinant"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}