Hilbert's twofold overshift (the case $\kappa=2$ of the 1887 invariant-theoretic program): the eigenform theory of binary forms of odd order, and the deferred most general degeneration of the $\kappa=1$ theory

David Hilbert's 1887 Habilitationsschrift (Math. Ann. 28, 381-446) introduced a general invariant-theoretic framework for binary forms: chains of Überschiebungen (transvections, the classical SL2-equivariant differential pairings of two binary forms) applied to a generic covariant, with the length of the chain called the characteristic number $\kappa$. The paper carried the framework out only for $\kappa=1$, a single overshift against one base form $f$ of even order $n=2i$: there the degree-preserving operator $\varphi\mapsto(f,\varphi)_i$ has a determining equation whose coefficients are invariants of $f$, whose roots are irrational invariants, and whose eigenforms are irrational covariants, complete with a four-case parity analysis, a degeneration theory, a geometric interpretation as the fundamental configuration of a collineation, and worked examples for orders 2, 4, 6 and 8. The paper closes by explicitly posing the case $\kappa=2$ - the twofold overshift $\varphi\mapsto(f,(f,\varphi)_{i_1})_{i_2}$, which is what binary forms of odd order require - as an investigation that "will play a role no less important" than the case $\kappa=1$, and separately defers the most general degeneration (allgemeinste Ausartung) of the $\kappa=1$ eigenform system. The $\kappa=2$ theory has never been constructed: Hilbert's own later work treats odd degree by other routes (skew invariants, counting), and the modern line engages only the $\kappa=1$ even-degree invariants. The problem is to construct the twofold-overshift theory - determining equation with invariant coefficients, parity structure, eigenforms as irrational covariants, degeneration dictionary, geometric interpretation, low-order examples - and to complete the deferred degeneration analysis.

Background

Classical invariant theory studies binary forms - homogeneous polynomials $f(x_1,x_2)$ of order (degree) $n$ in two variables - together with their invariants: polynomial functions of the coefficients that change only by a power of the substitution determinant under a linear change of the variables. Covariants are the relative of invariants whose values are again binary forms (the form itself, its Jacobian with a covariant, and so on). The operation at the center of this problem is the transvection (German Überschiebung, literally "overshift"), $(f,\varphi)_r$: the bilinear differential pairing of two binary forms obtained by applying $r$ mixed polar operators $\partial x_1\partial y_2-\partial x_2\partial y_1$ to $f(x)\,\varphi(y)$ and then identifying the two pairs of variables, where SL2 (the group of $2\times2$ matrices of determinant 1, the effective group behind all linear changes of two variables) acts on everything. Transvections realize the projection maps in the Clebsch-Gordan decomposition (the classical splitting of a tensor product of $\mathrm{SL}_2$ representations into irreducible summands) and were the basic computational tool of the classical theory.

In 1887 Hilbert published his Königsberg Habilitationsschrift (the German post-doctoral qualifying thesis), "On a general viewpoint for invariant-theoretic investigations in the domain of binary forms". Its "general viewpoint" is the study of chains of overshifts: starting from a generic covariant $\varphi$ of order $v$, one applies successive transvections with the base form(s), $\varphi\mapsto(f,\varphi)_{i_1}\mapsto(f,(f,\varphi)_{i_1})_{i_2}\mapsto\dots$, with the transvection numbers chosen so that the final form of the chain has the same order $v$ as $\varphi$ - the source's condition (2): the sum of the orders of the base forms used in the chain equals twice the sum of the transvection numbers - while the intermediate forms may have different orders; the number $\kappa$ of steps is the characteristic number of the chain. The paper develops only the case $\kappa=1$. Then the base form $f$ must have even order $n=2i$, and the degree-preserving operator $\varphi\mapsto(f,\varphi)_i$ on the $(v+1)$-dimensional space of order-$v$ forms yields: (i) a determining equation $A(\lambda)=0$ of degree $v+1$ whose coefficients are invariants $J_g$ of $f$ and whose roots $\lambda$ are irrational invariants of weight $i$ (quantities algebraic, not rational, over the field of invariants); (ii) a system of eigenforms constructed as bordered determinants (determinants of the eigenvalue system enlarged by one extra row and one extra column carrying the variables of two copies of the form), which are irrational covariants of $f$; (iii) a four-case parity analysis (the Hauptfälle, by the parity of $i$ and $v$) showing the roots pair as $\pm\lambda$ or double; (iv) a degeneration theory (Ausartungen) of the eigenform system, in three genera, used to characterize special position of the base form; (v) a geometric interpretation: the eigenform system defines a collineation (projective linear map) of the space of order-$v$ forms with invariant points and a fundamental configuration whose degenerations mirror those of the base form; and (vi) worked examples for orders 2, 4, 6 and 8.

The paper ends with two explicit deferrals. First, at the end of the closing section: the case $\kappa=1$ "had the consequence of excluding forms of odd order beforehand; this circumstance alone points to the necessity of investigating the further special case $\kappa=2$, which for binary forms of odd order will play a role no less important than the case $\kappa=1$ treated above for forms of even order". For odd $n$ no single transvection with $f$ preserves degree - two steps with $i_1+i_2=n$ are the minimal construction inside the program - so the twofold overshift is the canonical odd-order counterpart of Hilbert's operator. Second, at the end of the degeneration section: the satisfactory treatment of the most general degeneration (allgemeinste Ausartung) of the $\kappa=1$ eigenform system "must be reserved for a later occasion", the paper claiming only to have laid down the guiding viewpoints and the essential research tools.

The subsequent history bypassed both deferrals. Hilbert's own follow-ups go elsewhere: his 1888 paper on pencils of binary forms with prescribed functional determinant uses an adjacent theorem, and his 1893 paper on full invariant systems handles odd order by a different route - the skew invariant of the quintic, whose square is rational in the even invariants, and Cayley-Sylvester counting (the classical generating-function count of invariants by degree) for the degree of the invariant field - not through overshifts; his Chicago congress survey does not mention the overshift program at all, and nor does van der Waerden's afterword (Nachwort) to the collected volume. The published literature contains no treatment of $\kappa=2$. The modern line of descent engages precisely the $\kappa=1$ objects: Dixmier (1985) conjectured that Hilbert's invariants $H_{d-2,2},\dots,H_{d-2,d-1}$ - the coefficients of the determining equation, equivalently traces of powers of the $\kappa=1$ operator - form a homogeneous system of parameters (an algebraically independent set of invariants over which the invariant ring is a finite module) for the invariant ring of the binary $d$-ic when $d\equiv0\pmod 4$; Littelmann and Procesi (1990) studied the same invariants for Poincaré-series purposes (the Poincaré series being the generating function of the numbers of invariants in each degree); Abdesselam (2019) proved algebraic independence of a subfamily and analyzed why Dixmier's conjecture is hard. Adjacent modern theories - the associated forms of Alper and Isaev (which assign to a form with nonvanishing discriminant a dual form via apolarity), systems-of-parameters degree computations of Brouwer, Draisma and Popoviciu, the tensor-eigenvector literature, and recent applications of classical transvections to Galois-theoretic machine learning - are all distinct from the requested odd-order eigenform theory. The $\kappa=2$ theory has therefore never been constructed.

Problem Statement

Hilbert's 1887 paper solves its general overshift problem only in the characteristic number $\kappa=1$, which forces the base form to have even order, and it explicitly leaves two pieces of the program undone. The problem is to complete both pieces of the program: the twofold-overshift theory of Part A (deliverables (a)-(e)) and the deferred most-general degeneration of Part B.

Part A. Let $f$ be a generic binary form of odd order $n$ over an algebraically closed field of characteristic zero (the complex numbers suffice), and let $(\cdot,\cdot)_r$ denote the $r$-th transvection. For every choice of positive integers $i_1,i_2$ with $i_1+i_2=n$, the twofold overshift

$T_f:\varphi\mapsto(f,(f,\varphi)_{i_1})_{i_2}$

is a degree-preserving linear operator on the space $S_v$ of binary forms of each order $v$. Develop the invariant-theoretic theory of $T_f$ in the same sense in which the source develops the $\kappa=1$ theory, delivering in particular:

(a) the determining equation $\det(\lambda\,\mathrm{id}_{S_v}-T_f)=0$ on $S_v$: proofs that its coefficients are invariants of $f$, determination of their degrees and weights, and analysis of the arithmetic of its roots (the analogue of the $\pm\lambda$ pairing and doubled roots of the four Hauptfälle);

(b) the eigenforms: their construction, their nature as irrational covariants of $f$, and the linear relations among the quadratic covariants of the eigenform system - by the quadratic covariants of the eigenform system are meant the covariants of degree two in the eigenforms, i.e. the pairwise transvections $(\varphi^{(\mu)},\varphi^{(b)})_r$ of the eigenforms; in the source's $\kappa=1$ theory (Section 3, GA pp. 59-61) relation (41) gives the vanishing of $(\varphi^{(\mu)},\varphi^{(b)})_v$ unless $\lambda^{(\mu)}=(-1)^v\lambda^{(b)}$, table (42) lists the generally nonvanishing pairwise transvections in each of the four Hauptfälle, and the identities (43) express their product, up to sign, as $N\,\Phi^2$ ($N$ a product of binomial coefficients, $\Phi$ the functional invariant of the eigenform system);

(c) the degenerations (Ausartungen) of the eigenform system - coinciding eigenvalues, vanishing of the covariant determinant - and the dictionary between these degenerations and the geometry of $f$, such as multiple linear factors of $f$;

(d) the geometric interpretation: the collineation of the projective space of order-$v$ forms induced by $T_f$, its invariant points and its fundamental configurations;

(e) worked examples for low odd orders, at least the cubic and the quintic, exhibiting the new invariants and covariants explicitly.

Part B (deferred in the same paper). Give the promised detailed treatment of the most general degeneration (allgemeinste Ausartung) of the $\kappa=1$ eigenform system: the source distinguishes three genera of degeneration, treats the first two completely, and for the third - the most general - states only the characteristic criteria and the higher-type continuation, expressly reserving the satisfactory elaboration of all corresponding circumstances for a later occasion.

References

1. D. Hilbert, 'Ueber einen allgemeinen Gesichtspunkt für invariantentheoretische Untersuchungen im binären Formengebiete', Mathematische Annalen 28 (1887), 381-446, DOI 10.1007/BF02440000, https://doi.org/10.1007/BF02440000.
2. D. Hilbert, Gesammelte Abhandlungen, Zweiter Band (Algebra, Invariantentheorie, Geometrie), Springer, Berlin, 1933; the 1887 Habilitationsschrift is reprinted, with editorial footnotes marked as such, on pp. 38-101, and the volume contains B. L. van der Waerden's 'Nachwort zu Hilberts algebraischen Arbeiten'; digitized at the Göttinger Digitalisierungszentrum under PPN237833719, https://gdz.sub.uni-goettingen.de/id/PPN237833719.
3. D. Hilbert, 'Ueber Büschel von binären Formen mit vorgeschriebener Functionaldeterminante', Mathematische Annalen 33 (1888), 227-236, DOI 10.1007/BF01443853, https://doi.org/10.1007/BF01443853.
4. D. Hilbert, 'Ueber die vollen Invariantensysteme', Mathematische Annalen 42 (1893), 313-373, DOI 10.1007/BF01444162, https://doi.org/10.1007/BF01444162.
5. J. Dixmier, 'Quelques résultats et conjectures concernant les séries de Poincaré des invariants des formes binaires', Séminaire d'algèbre Paul Dubreil et Marie-Paule Malliavin, Paris 1983-1984, Lecture Notes in Mathematics 1146, Springer, Berlin, 1985, pp. 127-160, DOI 10.1007/BFb0074537, https://doi.org/10.1007/BFb0074537.
6. P. Littelmann and C. Procesi, 'On the Poincaré series of the invariants of binary forms', Journal of Algebra 133 (1990), no. 2, 490-499, DOI 10.1016/0021-8693(90)90284-u, https://doi.org/10.1016/0021-8693(90)90284-u.
7. A. Abdesselam, 'An algebraic independence result related to a conjecture of Dixmier on binary form invariants', Research in the Mathematical Sciences 6 (2019), article 26, DOI 10.1007/s40687-019-0189-x, https://doi.org/10.1007/s40687-019-0189-x; preprint arXiv:1903.11147, https://arxiv.org/abs/1903.11147.
8. J. Alper and A. Isaev, 'Associated Forms in Classical Invariant Theory', arXiv:1308.6624, https://arxiv.org/abs/1308.6624.
9. A. E. Brouwer, J. Draisma and M. Popoviciu, 'The degrees of a system of parameters of the ring of invariants of a binary form', Transformation Groups 20 (2015), no. 4, 953-967, DOI 10.1007/s00031-015-9335-8, https://doi.org/10.1007/s00031-015-9335-8; preprint arXiv:1404.5722, https://arxiv.org/abs/1404.5722.
10. J. Mezinaj, 'From Polynomials to Databases: Arithmetic Structures in Galois Theory', arXiv:2511.16622, https://arxiv.org/abs/2511.16622.