ABSTRACT
Sometimes a mistake, or a demonstration that does not work, can actually enhance a lecture-but only if it occurs spontaneously; it should never be deliberate. I suppose it works on the same principle as that of the trapeze artist who falls off the first time he attempts a trick and gets a much enhanced round of applause when he succeeds the second time. One classical example of this occurred during the recording of one of the 1971-2 Christmas Lectures. In one lecture, which was all about the interactions of technology and music, I used a bank of synthesisers, which were then brand new. My intention was to synthesise a few bars of a Bach Two-part Invention live, in order to demonstrate the principle, and in particular the sequencer, or memory, which, because it stored control signals rather than wave forms, could play back at any speed without change of pitch and with any desired timbre. The plan was to play the first part slowly with one finger on the keyboard, then to use the sequencer system to play this back and add the second part with a different timbre, and then to play both parts back together and to speed it up to normal tempo. I succeeded in storing both parts and then instead of pressing the replay button I pressed the erase button, instantly destroying the whole piece. I quickly explained to the audience and then repeated the whole thing, this time successfully. I assumed that the television producer would edit out the mistake-but he insisted that it made much better television to leave it in, thereby enhancing the dramatic impact of the final result.
