Strong minimality and geometric triviality of generic differential equations (Conjectures 1.5 and 1.7 of DeVilbiss–Freitag)

DeVilbiss and Freitag proved that a generic algebraic differential equation of order $h\ge 2$ — a linear combination of all monomials of degree at most $d$ in $x,x',\ldots,x^{(h)}$ with differentially independent transcendental coefficients — is strongly minimal whenever $d\ge 2(h+2)$, and later work on the degree of nonminimality lowered this to $d\ge 6$. Their paper poses two sharper conjectures that remain open: (Conjecture 1.5) every generic equation of fixed order and degree greater than one is strongly minimal, the residual cases being degrees $2$ through $5$; and (Conjecture 1.7) any two distinct solutions of such a generic equation are differentially algebraically independent, i.e. the solution set is geometrically trivial in the strongest form, which is open at every degree. The problem asks to prove or refute these conjectures in the differentially transcendental coefficient setting; the constant-coefficient autonomous analogue was settled affirmatively by Jaoui.

Background

Work takes place in the theory $\mathrm{DCF}_0$ of differentially closed fields of characteristic zero: a differential field carries a derivation $\delta$ (an additive map satisfying the Leibniz rule), and $\mathrm{DCF}_0$ is the model companion of the theory of such fields, in which solution sets of polynomial differential equations are studied by model-theoretic means. A differential polynomial $f$ in one differential indeterminate is a polynomial in $x,x',\ldots,x^{(h)}$ over the base differential field; its order is the largest $h$ for which $x^{(h)}$ occurs, and its degree $d$ is its total degree as an ordinary polynomial in the variables $x,x',\ldots,x^{(h)}$. Its zero set $Z(f)$ is the set of all solutions in the differentially closed field. Following DeVilbiss and Freitag, $f$ is called generic of order $h$ and degree $d$ if it is a linear combination of all monomials of degree at most $d$ in $x,x',\ldots,x^{(h)}$ whose coefficients are differentially independent transcendental elements — informally, a completely random nonlinear equation with no special algebraic structure among its coefficients.

A definable set is strongly minimal if every definable subset, allowing parameters, is finite or cofinite. For the solution set of an order-$h$ equation this is equivalent to a transcendence dichotomy: for every solution $a$ and every differential field $L$ containing the coefficients, the transcendence degree of the differential field generated by $a$ over $L$ is either $0$ or $h$. Strongly minimal sets are the atoms of geometric stability theory and are classified by the Zilber trichotomy: each is non-locally modular (nonorthogonal to the field of constants), locally modular and nontrivial (controlled by an abelian variety not descending to the constants), or trivial. A strongly minimal set is geometrically trivial (also called disintegrated) when any forking dependence among finitely many generic solutions already appears between some pair; for a differential equation the strongest form of triviality is that any two distinct nonalgebraic solutions are differentially algebraically independent over the coefficient field, meaning that no nonzero differential polynomial in two differential indeterminates with coefficients in that field vanishes on the pair.

Interest goes back to Poizat (1980), who showed that the nonconstant solutions of $x\cdot x''=x'$ form a strongly minimal set and asked whether general nonlinear equations of every order $n\ge 2$ behave the same way. Model-theoretic analyses of classical families — Painlevé equations, Schwarzian triangle equations, the differential equation of the modular $j$-function — confirmed the pattern case by case, and Jaoui proved that a general order-two equation of degree $d\ge 3$ over constant coefficients is strongly minimal and disintegrated. DeVilbiss and Freitag then developed a general sufficient criterion combining the degree of nonminimality of a type — the least $k$ such that a Morley sequence (a sequence of independent realisations) of length $k$ already witnesses a nonalgebraic forking extension of the type — with a linearization of the equation through Kolchin's differential tangent space. Their Theorem 1.3 shows that a generic equation of order $h>1$ and degree $d$ is strongly minimal whenever $d\ge 2(\mathrm{nmdeg}(p)+1)$ for the generic type $p$ of its solution set; with the bound $\mathrm{nmdeg}(p)\le\mathrm{RU}(p)+1$ of Freitag and Moosa this yields strong minimality for $d\ge 2(h+2)$, answering Poizat's question for sufficiently large degree and nonconstant (differentially transcendental) coefficients.

Two later developments sharpen the picture but leave the conjectures below open. First, Freitag, Jaoui, and Moosa proved that every finite-rank complete type in $\mathrm{DCF}_0$ has degree of nonminimality at most $2$, so the same argument gives strong minimality of generic equations for every $d\ge 6$, independent of order; the bound is sharp for arbitrary types, and this route cannot certify degrees below $6$. Second, Jaoui proved that a generic system of $n\ge 2$ first-order polynomial equations of degree $d\ge 2$ over constant coefficients is strongly minimal and geometrically trivial, settling the autonomous (constant-coefficient) reading of Poizat's question at all degrees — but by invariant-variety and foliation-singularity arguments that use autonomy and constant coefficients essentially, and do not transfer to the non-autonomous setting with differentially transcendental coefficients in which the two conjectures are posed.

Problem Statement

Prove or refute the two conjectures posed by DeVilbiss and Freitag for generic differential equations — equations whose coefficients are differentially independent transcendental elements:

(1) (Conjecture 1.5) For every fixed order $h\ge 2$ and every degree $d\ge 2$, if $f$ is a generic differential polynomial of order $h$ and degree $d$ (a linear combination of all monomials of degree at most $d$ in $x,x',\ldots,x^{(h)}$ with independent differentially transcendental coefficients), then the zero set $Z(f)$ is strongly minimal.

(2) (Conjecture 1.7, stated in the source as 'any two solutions') For every such generic equation, any two distinct nonalgebraic solutions of $f$ are differentially algebraically independent over the coefficient field.

Strong minimality is already known for $d\ge 6$ at every order $h\ge 2$, so part (1) currently reduces to the degrees $d\in\{2,3,4,5\}$; part (2) is open for every $d>1$, including the degrees where strong minimality is already known.

References

1. Matthew DeVilbiss, James Freitag, Generic differential equations are strongly minimal, Compositio Mathematica 159(7):1387–1412 (2023), DOI 10.1112/s0010437x23007212, https://doi.org/10.1112/s0010437x23007212 (arXiv:2106.02627, https://arxiv.org/abs/2106.02627)
2. James Freitag, Rémi Jaoui, Rahim Moosa, The degree of nonminimality is at most 2, Journal of Mathematical Logic 23(3), article 2250031 (2023), DOI 10.1142/S0219061322500313, https://doi.org/10.1142/S0219061322500313 (arXiv:2206.13450, https://arxiv.org/abs/2206.13450)
3. James Freitag, Rémi Jaoui, Rahim Moosa, When any three solutions are independent, Inventiones mathematicae 230(3):1249–1265 (2022), DOI 10.1007/s00222-022-01143-8, https://doi.org/10.1007/s00222-022-01143-8 (arXiv:2110.08123, https://arxiv.org/abs/2110.08123)
4. James Freitag, Rahim Moosa, Bounding nonminimality and a conjecture of Borovik–Cherlin, Journal of the European Mathematical Society 27(2):589–613 (2023), DOI 10.4171/jems/1384, https://doi.org/10.4171/jems/1384 (arXiv:2106.02537, https://arxiv.org/abs/2106.02537)
5. Rémi Jaoui, Generic planar algebraic vector fields are strongly minimal and disintegrated, Algebra & Number Theory 15(10):2449–2483 (2021), DOI 10.2140/ant.2021.15.2449, https://doi.org/10.2140/ant.2021.15.2449 (arXiv:1905.09429, https://arxiv.org/abs/1905.09429; the arXiv preprint carries the shorter title 'Generic planar algebraic vector fields are disintegrated')
6. Rémi Jaoui, On the density of strongly minimal algebraic vector fields, preprint (2023, revised November 2025), arXiv:2301.06362, https://arxiv.org/abs/2301.06362
7. Guy Casale, Matthew DeVilbiss, James Freitag, Joel Nagloo, Strong minimality of triangle functions, preprint (2022), arXiv:2205.08684, https://arxiv.org/abs/2205.08684
8. James Freitag, When any four solutions are independent, preprint (2025), arXiv:2510.08387, https://arxiv.org/abs/2510.08387
9. James Freitag, Thomas Scanlon, Strong minimality and the j-function, Journal of the European Mathematical Society 20(1):119–136 (2017), DOI 10.4171/jems/761, https://doi.org/10.4171/jems/761
10. James Freitag, Rémi Jaoui, David Marker, Joel Nagloo, On the Equations of Poizat and Liénard, International Mathematics Research Notices 2023(19):16478–16539 (2023), DOI 10.1093/imrn/rnac322, https://doi.org/10.1093/imrn/rnac322 (arXiv:2201.03838, https://arxiv.org/abs/2201.03838)
11. Christine Eagles, Léo Jimenez, Internality of autonomous algebraic differential equations, preprint (2024, revised July 2026), arXiv:2409.01863, https://arxiv.org/abs/2409.01863