I WAS walking in the street, wondering whether the tunes that occur to one were determined by associations other than their verbal content, and I said to myself that hitherto I had always been able to find a verbal association to explain why a tune had occurred to me.
A few moments later I caught myself humming a tune with which I could find no association. What was it? Oh, of course, one of Mendelssohn’s Songs without Words. The association was simply a continuation of what I had been thinking about—my preconscious had supplied a contradiction and called on me to explain it. For, of course, there are songs without words (i.e. tunes with no verbal accompaniment, which often occur to me, symphonies, sonatas, etc.). But I said to myself at once that, just as in this instance the occurrence to my mind of the ‘Song without Words’ had turned out to have a meaning, so there would turn out to be a demonstrable meaning in other similar cases, or because of some other temporal, spatial, or causal association with the wordless tune.
But I am not prepared to deny that there may also be purely musical associations. If I start humming a tune, very soon another, based on pure similarity, occurs to me. I am very musical, but unfortunately no musician. The laws of musical association will have to be established by a psycho-analytically trained musician. Probably a rhythm corresponding to one’s mood is often sufficient to cause a wordless tune to ‘occur’ to one. Sometimes a spirited waltz (in my case) means: ‘I am so happy that I should like to dance.’ (Unfortunately this does not happen very often.) The rhythm of the tunes that occur to me generally corresponds exactly with the degree of my cheerfulness or gloom.
In my pre-analytic period I formed a theory about the striking tone-painting in Wagner’s operas. I said to myself that every idea, every word, every situation (on the stage, for instance) awakens a certain feeling in people, to which there must correspond a neuro-physical process (oscillation) of definite quantitative proportions (wave lengths, building waves into complicated systems, rhythms, etc.); and that music, by the combination of sounds and sound-sequences, must be able to create acoustic structures having the same quantitative relations as those of the nervous oscillations. That was why on the one hand music associated itself with ideas and moods, and on the other why ideas and moods associated themselves with music. Music was really only the product of emotion; man went on varying the notes until they corresponded to his moods; a natural musical instrument (the organ of Corti) and its connexions with the central nervous system were the regulating factors in musical production.
Since analysis and reading the work of Kleinpaul I have dropped the whole of this fantasy. I now believe it to be probable that music (like speech) is only a direct or indirect representation or imitation of (organic or inorganic) natural sounds or noises, and as such is in a position to awaken moods and ideas just as any natural sound is.
For a musical tune to occur to one two things are generally necessary:
1 Posthumous paper. Published in German: Bausteine IV. (1939). First English translation.