In function-field arithmetic, the Carlitz exponential exp⁡C\exp_C plays the role of the classical exponential, and Denis (1995) proposed a Schanuel-type conjecture for it over the completed algebraic closure C∞\mathbb{C}_\infty of Fq((1/θ))\mathbb{F}_q((1/\theta)). Pellarin (2017) extended exp⁡C\exp_C to complete difference fields Ks\mathbb{K}_s built from Tate algebras in variables t1,…,tst_1,\dots,t_s, and proposed a stronger, operator-theoretic Schanuel-type conjecture (Conjecture 18 of the arXiv version, Conjecture 3.14 of the published version): whenever f1,…,fn∈Ksf_1,\dots,f_n\in\mathbb{K}_s are Fq(t1,…,ts)[θ]\mathbb{F}_q(t_1,\dots,t_s)[\theta]-linearly independent, at least nn of the 2n2n elements f1,…,fn,exp⁡C(f1),…,exp⁡C(fn)f_1,\dots,f_n,\exp_C(f_1),\dots,\exp_C(f_n) are regularly μ\mu-independent over an explicitly defined countable base field LsL_s, where regular μ\mu-independence means non-vanishing of all analytically regular μ\mu-difference polynomials. The paper proves implications but no new case beyond those already known for Denis' conjecture, and the conjecture — already open for s=0s=0, where it reduces to Denis' conjecture — has neither been proved nor refuted in the subsequent literature.

Background

Let Fq\mathbb{F}_q be the finite field with q=peq=p^e elements, A=Fq[θ]A=\mathbb{F}_q[\theta], and K=Fq(θ)K=\mathbb{F}_q(\theta). Completing KK at its infinite place, then taking an algebraic closure and completing again, yields the field C∞\mathbb{C}_\infty, the function-field analogue of C\mathbb{C}. The Carlitz module endows any AA-algebra BB with an AA-module structure in which multiplication by θ\theta acts through the Frobenius twisting operator τ(b)=bq\tau(b)=b^q via Cθ(b)=θb+bqC_\theta(b)=\theta b+b^q. The Carlitz exponential is the entire, Fq\mathbb{F}_q-linear function exp⁡C(z)=∑i≥0di−1zqi\exp_C(z)=\sum_{i\ge0}d_i^{-1}z^{q^i}, where did_i is the product of the monic elements of AA of degree ii (an analogue of qi!q^i!); it uniformizes the Carlitz module, and its kernel is the rank-one AA-module π~A\widetilde{\pi}A, where the Carlitz period π~\widetilde{\pi} is transcendental over KK (Wade, 1941). Denis (1995) conjectured the direct Schanuel analogue: if u1,…,un∈C∞u_1,\dots,u_n\in\mathbb{C}_\infty are AA-linearly independent, then K(u1,…,un,exp⁡C(u1),…,exp⁡C(un))K(u_1,\dots,u_n,\exp_C(u_1),\dots,\exp_C(u_n)) has transcendence degree at least nn over KK. This remains open; the known cases are the Hermite–Lindemann analogue for the Carlitz exponential (Wade, and in general a theorem of Yu on Drinfeld modules, the higher-rank generalizations of the Carlitz module), the Lindemann–Weierstrass analogue for uiu_i algebraic over KK (Thiery), and the case where all exp⁡C(ui)\exp_C(u_i) are algebraic (Papanikolas, via Tannakian duality for Anderson tt-motives — certain finite-rank difference modules that play the role of motives in this setting).

Pellarin's paper moves the problem into a several-variable rigid-analytic setting. For s≥0s\ge0, the Tate algebra Ts\mathbb{T}_s is the completion of C∞[t1,…,ts]\mathbb{C}_\infty[t_1,\dots,t_s] for the Gauss norm (the norm given by the maximum of the norms of the coefficients), and Ks\mathbb{K}_s is the complete field of generalized formal series ∑iciθ−i\sum_i c_i\theta^{-i} (ci∈Fqac(t1,…,ts)c_i\in\mathbb{F}_q^{ac}(t_1,\dots,t_s), well-ordered rational support), containing C∞(t1,…,ts)\mathbb{C}_\infty(t_1,\dots,t_s). The map μ:x↦xp\mu:x\mapsto x^p extends to a continuous, Fp(t1,…,ts)\mathbb{F}_p(t_1,\dots,t_s)-linear automorphism of Ts\mathbb{T}_s and Ks\mathbb{K}_s, making them difference fields; since τ=μe\tau=\mu^e, the Carlitz exponential extends by exp⁡C=∑i≥0di−1μei\exp_C=\sum_{i\ge0}d_i^{-1}\mu^{ei} to a map Ks→Ks\mathbb{K}_s\to\mathbb{K}_s with kernel π~ Fq(t1,…,ts)[θ]\widetilde{\pi}\,\mathbb{F}_q(t_1,\dots,t_s)[\theta] (Anglès–Pellarin–Tavares Ribeiro). The base field is Ks=Fp(θ,t1,…,ts)K_s=\mathbb{F}_p(\theta,t_1,\dots,t_s).

On such difference fields, algebraic relations are encoded by μ\mu-polynomials: elements of the ring L[X‾]μ=L[Xi,μ(Xi),μ2(Xi),… ]L[\underline{X}]_\mu=L[X_i,\mu(X_i),\mu^2(X_i),\dots] over a μ\mu-difference subfield L⊂KsL\subset\mathbb{K}_s, evaluated at tuples f‾\underline{f} by replacing each μj(Xi)\mu^j(X_i) with μj(fi)\mu^j(f_i). Because μ\mu acts as the pp-th power on C∞\mathbb{C}_\infty, every element of C∞\mathbb{C}_\infty is transformally μ\mu-algebraic (its μ\mu-iterates are its pp-powers), so the naive notion of difference-algebraic independence (Levin's transformal independence) is useless for arithmetic statements. Pellarin therefore isolates a class of exceptions: a nonzero μ\mu-polynomial PP is analytically critical if its zero variety in Ksn\mathbb{K}_s^n contains a rigid analytic immersion of the unit polydisc (a convergent power-series map with everywhere invertible Jacobian), and analytically regular otherwise. Every μ\mu-polynomial of depth zero (an ordinary polynomial) is analytically regular, and so is every tame μ\mu-polynomial, one in which no exponent of any μj(Xi)\mu^j(X_i) reaches pp. Elements f1,…,fn∈Ksf_1,\dots,f_n\in\mathbb{K}_s are regularly μ\mu-independent over LL if P(f1,…,fn)≠0P(f_1,\dots,f_n)\ne0 for every analytically regular μ\mu-polynomial over LL; this implies both ordinary algebraic independence and tame μ\mu-independence, and the reverse implications are false. Finally, LsL_s denotes the smallest μ\mu-difference subfield of Ks\mathbb{K}_s containing KsK_s and all elements regularly μ\mu-algebraic over KsK_s (a single element is regularly μ\mu-algebraic over KsK_s if some nonzero analytically regular μ\mu-polynomial in one variable with coefficients in KsK_s vanishes at it, and regularly μ\mu-transcendental otherwise); it is countable and plays the role that Q\mathbb{Q} plays in Schanuel's conjecture. The interest of the setting is illustrated by the Anderson–Thakur function ω=exp⁡C(π~/(θ−t))\omega=\exp_C(\widetilde{\pi}/(\theta-t)), the generator of the kernel of Cθ−tC_{\theta-t} in T1\mathbb{T}_1 (where Cθ−t(b)=(θ−t)b+τ(b)C_{\theta-t}(b)=(\theta-t)b+\tau(b) is the Carlitz-type operator with θ\theta replaced by θ−t\theta-t): it satisfies the tame relation μe(ω)=(t−θ)ω\mu^e(\omega)=(t-\theta)\omega, so it is tamely (hence regularly) μ\mu-algebraic, while being transcendental over K(t)K(t) — exactly the phenomenon the refined notion is designed to accommodate.

Problem Statement

Prove or disprove Conjecture 3.14 of Pellarin's paper "On a variant of Schanuel conjecture for the Carlitz exponential" (Conjecture 18 in the numbering of arXiv:1610.04048): for every prime power q=peq=p^e, every integer s≥0s\ge0, every n≥1n\ge1, and all elements f1,…,fnf_1,\dots,f_n of the complete difference field Ks\mathbb{K}_s that are linearly independent over Fq(t1,…,ts)[θ]\mathbb{F}_q(t_1,\dots,t_s)[\theta], at least nn among the 2n2n elements f1,…,fn,  exp⁡C(f1),…,exp⁡C(fn)f_1,\dots,f_n,\;\exp_C(f_1),\dots,\exp_C(f_n) are regularly μ\mu-independent over the countable base field LsL_s (the smallest μ\mu-difference subfield of Ks\mathbb{K}_s containing KsK_s and all elements regularly μ\mu-algebraic over KsK_s); that is, some nn-element subtuple among these 2n2n elements is such that no analytically regular μ\mu-polynomial in nn variables with coefficients in LsL_s vanishes at it. A complete affirmative answer includes, as the special case s=0s=0, Denis' Schanuel-type conjecture for the Carlitz exponential over C∞\mathbb{C}_\infty, itself open; a single explicit counterexample configuration (one qq, one ss, one linearly independent tuple for which every nn-element subtuple of the 2n2n elements is regularly μ\mu-dependent over LsL_s) refutes the conjecture.

References

  1. F. Pellarin, "On a variant of Schanuel conjecture for the Carlitz exponential", Journal de Théorie des Nombres de Bordeaux 29 (2017), no. 3, 845–873. DOI: 10.5802/jtnb.1004, https://doi.org/10.5802/jtnb.1004 (arXiv:1610.04048, https://arxiv.org/abs/1610.04048)
  2. L. Denis, "Indépendance algébrique et exponentielle de Carlitz", Acta Arithmetica 69 (1995), no. 1, 75–89. DOI: 10.4064/aa-69-1-75-89, https://doi.org/10.4064/aa-69-1-75-89
  3. L. I. Wade, "Certain quantities transcendental over GF(p^n,x)", Duke Mathematical Journal 8 (1941), 701–720. DOI: 10.1215/s0012-7094-41-00860-8, https://doi.org/10.1215/s0012-7094-41-00860-8
  4. A. Thiery, "Indépendance algébrique des périodes et quasi-périodes d'un module de Drinfel'd", in: The Arithmetic of Function Fields, de Gruyter (1992), 265–284. DOI: 10.1515/9783110886153.265, https://doi.org/10.1515/9783110886153.265
  5. M. Papanikolas, "Tannakian duality for Anderson–Drinfeld motives and algebraic independence of Carlitz logarithms", Inventiones mathematicae 171 (2008), no. 1, 123–174. DOI: 10.1007/s00222-007-0073-y, https://doi.org/10.1007/s00222-007-0073-y
  6. C.-Y. Chang and J. Yu, "Determination of algebraic relations among special zeta values in positive characteristic", Advances in Mathematics 216 (2007), no. 1, 321–345. DOI: 10.1016/j.aim.2007.05.012, https://doi.org/10.1016/j.aim.2007.05.012
  7. G. Anderson and D. Thakur, "Tensor powers of the Carlitz module and zeta values", Annals of Mathematics (2) 132 (1990), no. 1, 159–191. DOI: 10.2307/1971503, https://doi.org/10.2307/1971503
  8. B. Anglès, F. Pellarin and F. Tavares Ribeiro, "Arithmetic of positive characteristic L-series values in Tate algebras", Compositio Mathematica 152 (2016), no. 1, 1–61, appendix by F. Demeslay. DOI: 10.1112/s0010437x15007563, https://doi.org/10.1112/s0010437x15007563
  9. A. Maurischat, "Prolongations of t-motives and algebraic independence of periods", Documenta Mathematica 23 (2018), 815–838. DOI: 10.25537/dm.2018v23.815-838, https://doi.org/10.25537/dm.2018v23.815-838 (arXiv:1710.08252, https://arxiv.org/abs/1710.08252)
  10. O. Gezmiş, "Taelman L-values for Drinfeld modules over Tate algebras", Taiwanese Journal of Mathematics 23 (2019), no. 4. DOI: 10.1007/s40687-019-0181-5, https://doi.org/10.1007/s40687-019-0181-5
  11. O. Gezmiş and F. Pellarin, "Trivial multiple zeta values in Tate algebras", International Mathematics Research Notices 2022 (2022), no. 18, 14319–14383. DOI: 10.1093/imrn/rnab104, https://doi.org/10.1093/imrn/rnab104 (arXiv:2008.07144, https://arxiv.org/abs/2008.07144)
  12. K. Chung, T. Ngo Dac and F. Pellarin, "Universal families of Eulerian multiple zeta values in positive characteristic", Advances in Mathematics 422 (2023), Article 109003. DOI: 10.1016/j.aim.2023.109003, https://doi.org/10.1016/j.aim.2023.109003 (arXiv:2111.06973, https://arxiv.org/abs/2111.06973)
  13. F. Pellarin, "The analytic theory of vectorial Drinfeld modular forms", Memoirs of the American Mathematical Society (2025). DOI: 10.1090/memo/1581, https://doi.org/10.1090/memo/1581 (arXiv:1910.12743, https://arxiv.org/abs/1910.12743)
  14. P. Kowalski, "Positive characteristic Ax–Schanuel", Model Theory 3 (2024), 685–700. DOI: 10.2140/mt.2024.3.685, https://doi.org/10.2140/mt.2024.3.685 (arXiv:2311.00546, https://arxiv.org/abs/2311.00546)
  15. D. Matsuzuki, "On algebraic independence of Taylor coefficients of certain Anderson–Thakur series", arXiv:2404.16427 (2024), https://arxiv.org/abs/2404.16427
  16. T. Hasegawa, "Explicit formulas for the exponential and logarithm of the Carlitz–Tate twist, and applications", Journal de Théorie des Nombres de Bordeaux 36 (2024), no. 1, 251–291. DOI: 10.5802/jtnb.1278, https://doi.org/10.5802/jtnb.1278