A Diophantine approach to the three and four exponentials conjectures (Q2397077)

The so-called three (resp. four) exponentials conjecture states that there is no non-trivial relation of the form \((\log \alpha)^2=\gamma \log \beta\) (resp. \((\log \alpha)(\log \beta)=(\log \gamma)(\log \delta)\)) with algebraic numbers \(\alpha, \beta,\gamma,\delta\). The author relates these statements with a conjecture of Wilkie related with the o-minimal structure \({\mathbb{R}}_{\mathrm {exp}}\), thanks to the following characterisation of definable sets in \({\mathbb{R}}_{\mathrm {exp}}\), due to \textit{A. J. Wilkie} [J. Am. Math. Soc. 9, No. 4, 1051--1094 (1996; Zbl 0892.03013)]. A subset \(X\) of \({\mathbb{R}}^n\) is definable in \({\mathbb{R}}_{\mathrm {exp}}\) if and only if there exists a polynomial \(P\) in \(2(n+m)\) variables \(t_1,\dots,t_{n+m},y_1,\dots,y_{n+m}\) such that \(X\) is the image, under the projection \({\mathbb{R}}^{n+m}\to {\mathbb{R}}^n\) on the first \(n\) components, of the set \[ \{(x_1,\dots,x_{n+m})\; \mid \; P((x_1,\dots,x_{n+m},e^{x_1},\dots,e^{x_{n+m}}))=0\}. \] Wilkie's conjecture is the following statement: given \(X\subset{\mathbb{R}}^n\) which is definable in \({\mathbb{R}}_{\mathrm {exp}}\) and given \(F\subset{\mathbb{R}}\) a real number field of degree \(f\), there are constants \(c_1=c_1(X,f)\) and \(c_2=c_2(X)\) such that, for any \(T\ge e\), the number \(N(X^{\mathrm{trans}},F,T)\) of elements in \(X(F)\) of height \(\le T\), which do not belong to any semi-algebraic connected subset of positive dimension contained in \(X\), is bounded by \(c_1(\log T)^{c_2}\). Here, the height is the exponential of the logarithmic height. This conjecture for \(F={\mathbb{Q}}\) is Conjecture 1.11 in [\textit{J. Pila} and \textit{A. J. Wilkie}, Duke Math. J. 133, No. 3, 591--616 (2006; Zbl 1217.11066)]. It is extended to any real number field in [\textit{J. Pila}, Sel. Math., New Ser. 15, No. 1, 151--170 (2009; Zbl 1218.11068)]. The conjecture that \(c_2\) should not depend on \(F\) is Conjecture 1.4 of [\textit{J. Pila}, Ann. Inst. Fourier 60, No. 2, 489--514 (2010; Zbl 1210.11074)], where the surface \(\{(x,y,z)\in{\mathbb{R}}^3\; \mid \; x,y,z>0,\; (\log x)(\log y)=\log z\}\) is studied. See also [\textit{L. A. Butler}, Bull. Lond. Math. Soc. 44, No. 4, 642--660 (2012; Zbl 1253.03063)]. For the three exponentials conjecture, the author introduces the following threefold in \({\mathbb{R}}^4\): \[ X_3=\{(x,y,z,t)\in{\mathbb{R}}^4\; \mid \; x,y,z>0,\; t\not=0,\; (\log x)(\log y)=t\log z\}. \] He proves that if, for every number field \(F\subset{\mathbb{R}}\), we have \(N(X_3^{\mathrm{trans}},F,T)=o((\log T)^3)\), then the real case of the three exponentials conjecture is true, and conversely, that if the real case of the three exponentials conjecture is true, then, for every number field \(F\subset{\mathbb{R}}\), \(X_3^{\mathrm{trans}}(F)=\emptyset\). For the four exponentials conjecture, the author introduces the following threefold in \({\mathbb{R}}^4\): \[ X_4=\{(x,y,z,w)\in{\mathbb{R}}^4\; \mid \; x,y,z,w>0,\; (\log x)(\log y)=(\log z)(\log w)\}. \] He proves that if, for every number field \(F\subset{\mathbb{R}}\), we have \(N(X_4^{\mathrm{trans}},F,T)=o((\log T)^4)\), then the real case of the four exponentials conjecture is true, and conversely, that if the real case of the four exponentials conjecture is true, then, for every number field \(F\subset{\mathbb{R}}\), \(X_4^{\mathrm{trans}}(F)=\emptyset\).