Quantum \((abc)\)-theorems (Q1976813): Difference between revisions

We recall that, very loosely speaking, the \(abc\) theorem for relatively prime polynomials \(a\), \(b\), and \(c\) not all constant says that if \(a+b=c\), then not too many of the zeros of \(a\), \(b\), and \(c\) can have high multiplicity (one says a polynomial \(f\) is `constant' if \(f'=0\); note the subtlety in characteristic \(p\)). Here the author replaces `zeros of high multiplicity' by `sets of zeros in progression' -- that too in a fairly generalised sense: thus one may have a `multiset of zeros in progression', namely a divisor (in the sense used in algebraic geometry \(\alpha+f\alpha+\cdots+f^{m-1}\alpha\), where \(f\) is a map of the set of potential zeros into itself. Very familiar examples include `running powers' \((x)^h_n=x(x+h)\cdots(x+(n-1)h)\) and the `\(q\)-version' of \(x^n\), namely \((1-x;q)_n=(1-(1-x))(1-(1-x)q)\cdots(1-(1-x)q^{n-1})\). The author provides less familiar generalisations.