From “Notes on a Complete Recording of Beethoven’s Piano Works,” included in Music, Sense and Nonsense:
My work on the Beethoven sonatas took five and a half years. One of the crosses the artist has to bear is that the date of a recording is so rarely indicated on the record sleeve. He is all too easily blamed, or, almost worse, praised for interpretations that have lost some of their validity, at least as far as he himself is concerned. People expect an artist to develop, and yet they are only too ready to impale him, like an insect, on one of his renderings. The artist should have the right to identify his work with a certain phase of his development. It is only the continuous renewal of his vision—either in the form of evolution or of rediscovery—that can keep his music-making young.
The recordings of Beethoven’s variation works, with the exception of the Diabelli Variations, were made in three stages between December 1958 and July 1960. There followed, at the turn of 1960–61, the last five sonatas together with the Fantasy Op 77. In March 1962 I played the sonatas Op 31, Nos. 1 and 2, Op. 57 and Op. 90; in June and July of that year all the remaining sonatas between Op. 23 and Op. 81a. The early sonatas from Op. 2 to Op. 14 were recorded in December 1962 and January 1963 (by coincidence my work on the thirty-two sonatas was finished on my thirty-second birthday). Finally in July 1964 I played the miscellaneous pieces and the greatest of all piano works: the Diabelli Variations.
I recall a cold winter morning in a rather dilapidated Baroque mansion in Vienna; the logs in the fireplace of the hall where we recorded crackled so loudly that we had to throw them out of the window into the snow.