The denomination of key signatures of musical works does not always reflect the actual reality. For example, Beethoven‘s Kreutzer Sonata, Op. 47, is referred to as an A major work due to the tonality of the introduction and the finale, whereas the sonata-allegro part of the 1st movement is in A minor. A slightly different case is Schumann‘s String Quartet, Op. 41 No. 1, where the introduction also determines the alleged tonality of A minor, regardless of the fact that the first sonata-allegro is in F major. Even more curious is the key of Schubert‘s Impromptu, Op. 90 No. 4, notated and generally given as A flat major, even if the music clearly begins in A flat minor.
Isn’t it a bit awkward to only look at a piece’s beginning when identifying its overall key? To me it seems reasonable to consider the main section of a movement more relevant than an introduction, as long as the latter only switches between major and minor modes—such as Mendelssohn‘s Rondo capriccioso, Op. 14 (correctly termed an E minor work, even if its introduction is in E major), or Dvořák‘s Eighth Symphony, Op. 88 (a G major work with a 1st-movement introduction in G minor). According to that principle, we should speak of the Kreutzer Sonata as an A minor composition. On the contrary, the abovementioned Schumann remains an A minor work, as indicated by three of its four movements, and despite the 1st movement moving to the submediant. What do you think?