On the greatest common divisor of \(u-1\) and \(v-1\) with \(u\) and \(v\) near \(\mathcal S\)-units (Q2581030)

Two integers are said to be multiplicatively independent if they are not both integral powers of the same integer. It was shown by \textit{Y.\ Bugeaud, P.\ Corvaja}, and \textit{U.\ Zannier} [Math.\ Z. 243, No.1, 79--84 (2003; Zbl 1021.11001)] that if \(a\) and \(b\) are multiplicatively independent integers then for every \(\varepsilon > 0\) the inequality \(\gcd(a^n-1,b^n-1) < \exp(\varepsilon n)\) holds for all but finitely many positive integers \(n\). The author generalizes this result by showing that if \(f\), \(f_1\), \(g\), and \(g_1\) are non-zero polynomials with integer coefficients and \(a\) and \(b\) are multiplicatively independent integers, then for every \(\varepsilon > 0\) the inequality \(\gcd(f(n)a^n+g(n), f_1(m)b^m+g_1(m)) < \exp(\varepsilon \max\{m,n\})\) holds for all but finitely many pairs of positive integers \((m,n)\). The authors actually deduce this from a more general statement involving \({\mathcal S}\)-units and logarithmic heights. The result on \(\gcd(a^n-1,b^n-1)\) obviously follows by taking \(f=f_1=1\), \(g=g_1=-1\), and \(n=m\). The proof makes use of Schmidt's Subspace Theorem as well as linear forms in logarithms, and the author notes that bounds on the largest exceptional pair are not effectively computable.