A ‘lied’ or ‘mélodie’ is a little synthetic artwork in which poetry and music are fused into an artistic whole. You’re supposed to listen text in hand, unless you’re able to comprehend the words directly from the singer’s mouth. But I’ve often caught myself listening casually, enjoying ‘art songs’ in ignorance of the poetry. As, no doubt, many listeners do. A guilty pleasure, that as an aesthetic experience is evidently incomplete, but may be intense all the same.
Sometimes ignorance of the text is an advantage. It allows us not to be distracted by poetry which is inferior to the music, or by mismatches between music and text – as may occur in songs that are otherwise quite successful.
A more interesting aspect of this casual listening experience is that it may direct our attention to the question how much of the poetry is actually in the music. Eliminating the content of the poem from our awareness does not convert a song into ‘pure music’ (if there is such a thing). Much of the musical structure is determined by the syntax and prosody of the text. Along comes a great deal of the poem’s meaning, in a broad sense.
Meaning exists in layers – from the sum of word meanings to the more vague dimensions of mood and atmosphere, created by choice of words, rhythm, imagery, etcetera. What generally matters most is not the poem’s paraphrasable content (what it is about), but an experience of a particular quality that is expressed or evoked.
This experience is realized in the music as a sort of enhanced rendering of the poem. The music lends an audible voice to the ‘I’ that is usually contained, explicitly or implicitly, in lyrical poetry. The accompaniment may have an equal share in this, and occasionally represent more concrete phenomena (rustling leaves, a horse’s gallop, a spinning wheel).
Much of the pleasure of this kind of casual listening derives from (subconsciously) guessing and experiencing these less concrete levels of meaning. This is only possible with a careful and detailed vocal interpretation. We are dependent upon the details of vocal delivery, including the music of vowels and consonants. A singer who sacrifices enunciation to an exquisite overall sound robs the music of much of its meaning.
We (listeners and composers) don’t like singing without a text. It is a curious fact, not at all self-evident and rarely discussed, that textless song is quite exceptional in the European tradition, and probably worldwide. The vocalise as a textless composition for voice with accompaniment has originated as an exercise in the early nineteenth century. It maintained itself as an educational genre in the early twentieth century through the efforts of Amédée-Landély Hettich, a teacher at the Paris Conservatoire. Hettich’s Répertoire Moderne de Vocalises-Études (Leduc, 1907-1938) includes compositions written for the Conservatoire’s exams by Dukas, Roussel, Honegger, Villa-Lobos and Messiaen, among others. The Italian publisher Ricordi has followed with Vocalizzi nello stile moderno (1929-1933). Ravel’s Vocalise-étude en forme habanera (1907) is one of the earliest and finest in the French collection, and one of the very few textless songs that have a stable place in the concert repertoire, alongside Rachmaninoff’s vocalise (1915), Villa Lobos’ Bachianas Brasileiras nr. 5 (1945), and such oddities as Reinhold Glier’s Concerto for Coloratura Soprano (1943).
The vocalise is almost always sang on the vowel /a:/, which is the most open and comfortable throughout the registers. Though nothing would forbid a composer to use other vowels or nonsense syllables – as in the improvisation practice of jazz singers (scat).
Another restrictive feature of the genre is a preference for a moderately slow tempo and elegiac mood (think of Rachmaninoff), though the virtuosic valse chantée is a secondary favourite (Glier’s concerto combines the two). In concert practice the genre is almost entirely limited to female voices. These are typically cast in roles similar to that of nineteenth century ballerina’s: nymphs and sirens, female beings that are either sexually immature or perniciously seductive. Another favourite model is, of course, the nightingale.
Maybe that a protracted a-a-a-a tends to sound like a desperate cry for help. As if the singer is slowly drowning.
Another factor may be that we humans are not merely animals with language, but language junkies. We like speech and talking for the sake of talking, even passively. The popularity of talk shows on radio and television can hardly be due to an information addiction; most of what goes on there is uninformative and readily forgotten. It is, rather, a manifestation of a talking addiction. Hearing other people talking may give the listeners the feeling of being safely embedded in a human, social environment. Whenever we hear a human voice, what we expect and desire is speech.
wpDiscuzMy chamber opera Pedrillo Botón (2009) includes an aria-vocalise which should be alternately sung on /a:/ (legato) and on scat syllables of the singer’s invention (staccato). The mezzosoprano represents a belcanto-crazy Prime Minister’s Wife, who makes selfies while singing. The vocal part is in this recording given to the clarinet.