Jerome Rothenberg‘s site UbuWeb includes, among many other interesting-sounding articles, an excerpt from Bronislaw Malinowski‘s The Language of Magic and Gardening (Coral Gardens and Their Magic, volume 2), which discusses the peculiar use of words in magic spells and the problems involved in translating them:
If the main principle of magical belief is that words exercise power in virture of their primeval mysterious connexion with some aspect of reality, them obviously we must not expect the words of Trobriand magic to act in virtue of their ordinary colloquial meaning. A spell is believed to be a primeval text which somehow came into being side by side with animals and plants, with winds and waves, with human disease, human courage and human frailty. Why should such words be as the words of common speech? They are not uttered to carry ordinary information from man to man, or to give advice or an order. The natives might naturally expect all such words to be very mysterious and far removed from ordinary speech. And so they are to a large extent, but by no means completely. We shall see that spells are astoundingly significant and translatable and we shall also see why this is so.
But the fact remains that unless the reader is forewarned that a great deal of the vocabulary of magic, its grammar and its prosody, falls into line with the deeply ingrained belief that magical speech must be cast in another mould, because it is derived from other sources and produces different effects from ordinary speeech, he will constantly be at cross-purposes with the principles according to which the translation of magical utterance has to proceed. If the ordinary criteria of grammar, logic and consistency were applied, the translator would find himself hopelessly bogged by Trobriand magic.
Take the very first formula, for example. This is a direct address to ancestral spirits—a man-to-man communication we might say; hence in parts it is lucid and grammatical. And then comes the sentence: ‘Vikita, Iyavata, their myth head is.’ After much consultation with informants and etymological research in their company, I had to conclude that in no sense can these words be set equivalent to any ordinary prose sentence. The meaning of the magical expression is simply the intrinsic effect which, in native belief, it exerts on the spirits and indirectly on the fertility of the soil. The commentaries of the natives, however, reveal the mythological references connected with the names Vikita and Iyavata. Those who are versed in the magical tradition of this spell can interpret the significance of these words and tell us why they are ritually effective.
In what way, then, can we translate such a jumble of words, meaningless in the ordinary sense? The words are supposed to exercise a mystical effect sui generis on an aspect of reality. This belief is due to certain properties and associations of these words. They can therefore be translated in one sense and in one sense only: we must show what effect they are believed to produce, and marshal all the linguistic data available to show how and why they produce this effect.
To take another example, the exordium of the most important spell:
Vatuvi, vatuvi, vatuvi, vatuvi. Vitumaga, i-maga.
Vatuvi, vatuvi, vatuvi, vatuvi. Vitulola, i-lola.
The better one knows the Trobriand language the clearer it becomes that these words are not words of ordinary speech. As actually recited in the spell they are pronounced according to a special phonology, in a sing-song, with their own rhythm and with numerically grouped repetitions. The word vatuvi is not a grammatical form ever found in ordinary speech. The compounds vitulola, vitumaga, are again weird and unusual; in a way, nonsense words. Words like vatuvi or the root lola are clipped; but there are other words which are compounded, built up, developed…
Thus all magical verbiage shows a very considerable coefficient of weirdness, strangeness and unusualness. The better we know the Trobriand language the more clearly and immediately can we distinguish magic from ordinary speech. The most grammatical and least emphatically chanted spell differs from the forms of ordinary address. Most magic, moreover, is chanted in a sing-song which makes it from the outset profoundly different from ordinary utterances. The wording of magic is correlated with a very complicated dogmatic system, with theories about the primeval mystical power of words, about mythological influences, about the faint co-operation of ancestral spirits and, much more important, about the sympathetic influence of animals, plants, natural forces and objects. Unless a competent commentator is secured who, in each specific case, will interpret the elements of weirdness, the allusions, the personal names or the magical pseudonyms, it is impossible to translate magic. Moreover, as a comparison of the various formulae has shown us, there has developed a body of linguistic practice—use of metaphor, opposition, repetition, negative comparison, imperative and question with answer—which, though not developed into any explicit doctrine, makes the language of magic specific, unusual, quaint.
(Via wood s lot.)