On the reconstruction problem for Pascal lines (Q1991090)

If \((A, B, C, D, E, F)\) is a sixtupel of points on a conic, the six points \(AE \cap BF\), \(BD \cap CE\), \(AD \cap CF\) lie on a straight line, the sixtuple's \textit{Pascal axis.} Permutations of the points give rise to a configuration of 60 Pascal axes, known as ``Pascal hexagram'' or ``hexagrammum mysticum''. This paper shows that it is possible to reconstruct the six points in a unique way from the Pascal axes \[ (A,D,B,E,C,F),\;(A,C,F,E,D,B),\;(A,D,F,E,C,B),\;(A,B,C,F,D,E). \] The roles of the four sixtuples are not symmetric. Arranging the entries in a \(2 \times 3\) matrix, the columns of the first three sixtuples all give rise to the partition \(\{A, E\} \cup \{C, D\} \cup \{B, F\}\) while the last sixtuple gives \(\{A, F\} \cup \{B, D\} \cup \{C, E\}\). The first three sixtuples allow to reconstruct the conic chords \(AE\), \(CD\), and \(BF\) but are not sufficient for a proper assignment of endpoints. The proof shows that coordinates of \(A\), \(B\), \(C\), \(D\), \(E\), and \(F\) actually rational functions of the given Pascal axes' line coordinates. This not only implies uniqueness but also suggest existence of a (yet unknown) elementary geometric construction for them. The non-trivial proof avoids computer algebra. Instead it reduces the reconstruction problem to transvecant identities that are then shown to be true by a graphical calculus for binary forms.