Language, Magic and Power

David Ellis ('08), Brown University

This essay was written by the author for Prof Mathiesen's UC082, "Magic in the Middle Ages," Brown U., 2004.

In this paper, I intend to describe the unique characteristics of magical language, and primarily to investigate how differences in a community’s linguistic environment might affect the power of its magical practices. I begin with an inquiry into the nature of magical naming, and how things named in the context of magic are transformed. This is followed by a discussion of an ‘ideal’ magical language, and of what characteristics of language are most essential to its magical power. I hope to detail a magico-linguistic hermeneutic, covering many facets of language and magic, and of the power that they generate in combination.

The power of naming in the context of magic is mirrored in ancient mythology. The original human language, spoken before the fall of the Tower of Babel, is portrayed in mythological tales – such as the Epic of Gilgamesh – as immensely powerful. According to many religious traditions, including Christianity, Hinduism and Islam, God used a single word (or Gk. λογος - "Logos") to create the cosmos. Similarly, utterances in humans’ Proto-Language brought into existence or changed the nature of the things they named. These effects of language could be described, respectively, as "fiatic" and "Orphic." Fiatic linguistic magic includes "the creation of the signified by its signifier," whereas Orphic magic "[enchants] an already existing entity by means of vocal material" (Idel, p. 23). The power of human language was at its height in the Garden of Eden, when Adam was instructed by God to name what he had created – in some sense, to finish or confirm the act of creation.

Modern languages as used in magic retain some of this power or control over objects or beings they name. The name is a symbol that provides a reference to a physical thing, or to an abstract, imaginal, or imaginary one, and allows the one who uses this name to affect its referent. According to the magical Law of Relevance, "Any two objects which are or have ever been relevant to one other are symbolically equivalent in magic" (Mathiesen). So, a name is equivalent to the thing named – but perhaps one language is better able to capture some specific meanings or categories of meaning than others. This equivalence of sign and signified indicates that one cannot exist without the other. "The unnamed is the ignored" (Fauquié, p. 405). A thing without a name will cease to exist, because it cannot be referred to.

The power of linguistic naming can be described through performativity and interpellation. Performativity is defined by J. L. Austin in his 1955 lecture series, "How to Do Things With Words." In terms of a performative utterance, "the uttering of the words is, indeed, usually a, or even the, leading incident in the performance of the act…, the performance of which is also the object of the utterance…" (Austin, p. 8). Austin’s examples of performative verbs include "to bet" and "to promise." In the context of magical spells and incantations, verbs such as "to invoke" and "to consecrate" might be more common.

Any statement of the form "I hereby…" is explicitly framed as a performative; the speaker is declaring that the utterance itself has accomplished the act it describes. In some cases, such a performative utterance is immediately followed by supporting actions or statements: for example, a magician might say "I hereby consecrate this ground and invoke the power of the runes inscribed about my magic circle." He might then sprinkle a powder, potion or simple water around the circle, similar to the consecration of a Catholic altar with holy water, and pronounce the names of the runes he had inscribed on the earth.

The naming of the runes consists of uttering the phonemes they represent, and is in this sense merely a phonetic act, or the uttering of certain noises. (Austin, p. 95) On the other hand, this could be considered a phatic act, since the magic-user’s rune system would have a vocabulary (its runes) and possibly a grammar, or set of rules that determine the meaning or result of any combination of runes. In either case, the pronouncing of rune-names and the water-sprinkling complete and validate the effect of the performative utterance or rhetic act. Words serve a variety of functions within magic, but each of these is crucial to the success of a spell or ritual. Magic could not exist without the foundation of human speech and communication. The power of words (vis verborum) within magic is described by Cornelius Agrippa as follows.

Words...[carry] with them not only the conception of the mind, but also the virtue of the speaker with a certain efficacy unto the hearers, and this oftentimes with so great a power, that oftentimes they change not only the hearers, but also other bodies, and things that have no life. [p. 211]

Agrippa does not distinguish between types or categories of words, but attributes this power to any spoken word or sign. In his commentary on Agrippa’s occult philosophy, Christopher Lehric notes that

The crux of the Agrippan approach to vis verborum is the division we have come to expect: insofar as words are treated as sound or noise, they have a natural power; insofar as they are intelligent language requiring a rational interpreter, they are celestial [p. 60]

There is a distinction between words that carry meaning – are used in a phatic or rhetic act – and those that are meaningless, and constitute a phonetic act. A magical word whose meaning is not defined within a natural human language, but which might still have a function, would have its effects restricted to the natural realm, as opposed to the celestial.

According to Agrippa, "humans are divided into natural, celestial, and divine portions, in strict microcosm of the tripartite universe" (Lehrich, p. 61). Man’s mental and spiritual powers can be divided among the three spheres as follows: "the senses are natural, reason celestial, and intellect divine." (Lehrich, p. 61) This suggests that gibberish, a constructed magical language, or a ‘word’ consisting of patterned phonemes without meaning, could affect only the senses, and though such sounds might be used in illusions or performance magic, they would never have true performative power. Examples of magical words that could be considered gibberish include the simple "Abracadabra," and a multidimensional palindrome like the Latinate word-cube "sator arepo tenet opera rotas." According to Keith Dickson’s analysis of magical mumbojumbo, "all such devices aim to reconstitute the Sign as concrete matter, to make it opaque, to transmute bodiless sound into plastic artifact, all the better to control it by refiguring its contours" (Dickson, p. 160).

Interpellation is also a type of performative utterance, and in this sort of performativity, the social body (or any animate thing) is called into being, or constructed, through naming. Interpellation can be described as "the power of discourse to enact what it names" (Butler, p. 187). As noted above, in the context of magic, the name is equivalent to what it names, and a name naturally holds power over its referent, as do any who know the name. Magical language is largely performative, but how do these utterances themselves exert such power? This can be answered in part in that "for discourse to materialize a set of effects, ‘discourse’ itself must be understood as complex and convergent chains in which ‘effects’ are vectors of power" (Butler, p. 187). These vectors then, in magical discourse, could be the relations between utterances (chants, incantations, etc.) and ritual actions. Because each statement or action is dependent on what has preceded it, which may in turn depend on one or more previous acts – which could include utterances – the power of the complete magical ritual is constituted in an intricate combination or compounding of the success of its components.

The language of magic thus functions within discourse, and can therefore have performative effects on things and beings, even molding their very nature. But if, as is often the case, the magician is performing a ritual incantation alone, in private, how does the language used become discourse? Perhaps any language, magical or otherwise, holds the power to construct discourse, and therefore to affect the things it names. "[W]hen a word is pronounced, its vibrations have the ability to bring forth to the mind an ideal form, ‘la notion pure’, of the object in question" (Hampton, p. 25). The speaker can easily modify certain variable qualities of the instantiation of this object in the minds of those who hear the words, but the ideal form, at least as conceived by Plato, is immutable.

However, magic seems to require of its language that communication and communion be somehow achieved. Performativity serves to establish communion, in that the speaker interacts with the things he names in a magic spell, and it may also open a channel of communication with these named entities. This communication may simply be the utterance itself, viewed as an imperative or directive to a summoned or invoked being, or as a constative that changes some aspect of an object by communicating its resulting state. On the other hand, if discourse is interpreted as a body of knowledge within a field, then magical discourse per se might be thought of as a collection of spells and incantations, or "phrases of power," but not the individual instances or iterations of these. The discourse of magic would, in this sense, consist of ideal or prototypical versions of utterances, of incantations, invocations, imprecations and summonings, and even of the complete spells, sacrifices and rituals these could constitute. Language therefore cannot create magical effects outside the realm of discourse, nor, it seems, can language even exist without being included in some form of discourse. Unarticulated thought, whether in one’s native tongue, another natural language, or in some form of mentalese could constitute a type of discourse, but performative effects can only result from discursive utterances – not from thoughts, ideas, or mental reasoning.

There might be an ideal of magical language – similar, perhaps, to language before the Tower of Babel – that holds all the discursive power of magic, and from which each utterance of a magical poem, text, spell, incantation, or even a simple name used in magic derives its power. Are some languages more readily and easily adaptable to use in spells and other magic, and therefore closer to the ideal language for magical purposes? Bronislaw Malinowski wrote that "powerful magical language is distinguished by a very high coefficient of weirdness" (Malinowksi, p. 220). One’s native language is not strange enough to be very effective in magic, so most successful magicians, sorcerers, witches and other magic-users find other languages for their magic.

In many communities, several languages are spoken and understood by many of the inhabitants. These are often related, or derived from a common ancestor or proto-language. The language used for magical spells and incantations in these communities is rarely the vernacular; it is often a language of learning, used particularly by scribes or in universities, and in other scholarly pursuits. Sometimes, magic might use a constructed language from the mind of an individual or a relative few, rather than a language that naturally evolved as a population used it to communicate and commune. A variety of options – infinitely many, in fact – exist from which a magic-user must select a language to use.

A spell might incorporate several tongues: a runic language; an ancient or esoteric tongue that divine powers might more likely respond to; and the local vernacular for commentary, instructions to assistants, and anything else that need not be in a specific, more specialized language. The use of multiple magical languages within a single ritual, and possibly different languages – within the same community or in two separate ones – between instances of a spell, brings up several questions. One might wonder whether it is possible to translate a name. If so, the translation would probably lose at least some of the power the original had in relation to its referent. If not, do names exist supra lingua, or beyond language?

The answer may be a combination of these. Some names, those derived from other lexical entries (vocabulary items) in a particular language, can be translated by mirroring the derivational process in the target language. Many names have equivalents in other languages, particularly closely related languages, which differ primarily in that each conforms to the phonotactic constraints of the language it is used with, since speakers’ conceptions of natural and acceptable sound patterns depend on these, and they often constrain the speakers’ ability to produce certain sounds and sound sequences.

However, there are names that are impossible to translate satisfactorily. These include "true names" and "soul names," which are often linked more directly with a being. A human might have a "soul name;" its existence, nature and properties are to some extent determined by the person’s culture and the social context. A place, natural spirit, fairy, daimon, god, goddess, and a variety of other things might each have a "true name." This is also dependent on culture; for instance, within Aboriginal Australian culture, a person’s ancestors would include a proto-human hero and, in some sense, every natural thing (terrain, rock, tree, bush, etc.) he had touched, each of which has a name known only to those who are related: descended from the same hero. Any such name – whether "true name," "soul name," or of another, similar category – would give its user greater power over the entity than any other way of referring to it, such as a translated or alternate name, or a descriptive phrase. Any name for an entity would, by definition, have the same referent, but the relevance of each might vary.

Can relevance and reference of a word be assigned, as in a word translated into a constructed language, or even an improvised language, used only in the context of a single instance of a rite or ritual? Referents of pronouns are assigned largely by the context, and each reference exists in the mind of the speaker and addressees; if there is ambiguity, the referent could be different in each person’s mind. Relevance is also a mental construct, so it can be assigned to novel words. This happens when new vocabulary is formed in a dialect, resulting from innovation – usually within a relatively small speech community – or when a loan word is taken from a language spoken nearby. Loan words often provide ways of referring to something that had been outside the scope of the language’s lexicon: being named and brought into scope is necessary for this entity or set of entities to exist in the minds of those within the speech community.

Words allow a speaker to interact with the world, and his specific environment within it. "The connection between men and the world is always mediated by the word" (Fauquié, p. 405). One use of the word is to describe one’s internal state (psychological, emotional, mental, etc.) to another person. Can the magic-user’s intent be conveyed despite meaningless syllables, in glossolalia or other purely phonetic acts, or only because of meaningful utterances composed of morphemes found in a natural or constructed language? "For the magician… deals out verbal elements of the abracadabra, sesame, hocus pocus type, that is, words the function of which is not ‘meaning’ in the ordinary sense, but a specific magical influence which these words are believed to exercise" (Malinowski, p. 214). So, not only names can exert power, but also words of this other, magic-specific type.

Such a phrase does not hold power over a set of entities it names, because it is merely a phonetic act. So, what is its effect, and how does it derive power? Perhaps these words function in part by "[supplanting] the conventional production of meaning with some other kind of semiosis" (Dickson, p. 154). The meaning produced by such a word is not a reference, but a way to exert direct control over the world by subverting the border between signifier and signified. The word holds power because it is difficult to grasp its meaning, which is therefore not fixed, but adaptable. This is more obvious with patterns – often palindromes – "that build repetitively upon a simple dominant syllable" (Dickson, p. 158), or sequences that are distortions of common words. For example, by transposing letters in pairs (an effect along the syntagmatic axis of language): "powerful" becomes "opewfrlu."

The following is a common palindrome in magic practiced in the language of ancient Greece, "in which the Greek vocalic inventory is first arranged as a palindrome (α-ε-η-ι-ο-υ-ω-υ-ο-ι-η-ε-α), [then] incrementally reduced to the tonic long Ω-sound by eliminating the initial and ultimate phoneme in each series…" (Dickson, p. 158). Magicians can be creative with sounds, combining them in intricate patterns, and often modifying the patterning with each utterance. A magician also acts as lexicographer, in that he assigns meanings, or functions, to these novel words. Within magic, "the meaning of any significant word, sentence or phrase is the effective change brought about by the utterance within the context of the situation to which it is wedded" (Malinowski, p. 214). Each word a magician speaks during his spell or ritual can have both a traditional meaning and this sort of functional meaning. Seemingly meaningless words therefore lack only part of this dual semiosis.

What other sorts of words might exist? The Enochian language, which angels purportedly transmitted into the minds of Dr. John Dee and Sir Edward, is perhaps a mysterious combination of constructed and "supernatural" language. The angels’ "speech" does somehow have characteristics that can be mapped onto human phonology, and the grammar of the Enochian language conforms with the universals of natural human languages. However, some of the vocabulary, and perhaps much of the grammar imparted to these two men may have been affected by their native language, English, and other information in their minds. Perhaps these men had constructed the language themselves, and the angelic transmission was a hoax. Or, their mortal minds could have been unable to process or understand angelic language directly, so it was somehow converted. Or, angelic language could be a sort of telepathy or annotated mentalese – meaning thoughts are communicated directly, but with some linguistic form that would account for the non-English characteristics.

Whatever the form or nature of Enochian itself truly is, we can see that a non-natural language can have any of a wide range of structures. But what are the magical effects of linguistic variety within natural languages? Linguistic differences occur on two distinct axes: the syntagmatic, along which transposition and metonymy function; and the paradigmatic, with substitution and metaphor. The syntagmatic axis is parallel to time, and so affects the ordering and dependencies of phonemes and morphemes. Magical power can be drawn along this axis through contiguity or contagion – the association of objects in close proximity. The paradigmatic axis is perpendicular to time, and deals with the form of individual units of sound or meaning. Objects with similar form are associated along the paradigmatic axis, and therefore equivalent for magical purposes.

Differences between distant or unrelated languages, like Finnish and Telegu, tend to be primarily lexical, and along the paradigmatic axis. In such languages, word order and morphological systems – rules for combining morphemes, or units of meaning - are also often disparate, and these grammatical differences are along the syntagmatic axis. Languages of the same family or subgroup, like Finnish and Estonian, tend to have nearly identical grammatical structure, and many cognates: related or similar words. Which differences between languages are significant to the power of magic? Do differences on one axis or the other more greatly affect the magical power available to a language’s speakers? Or is the power of a magical language always and only relative to the speaker’s native language? If some hypothetical creature had an unnatural native language, by human standards, but could communicate (telepathically, perhaps) with communities of human magic-users, would there be any way to predict which magical language would seem most powerful from this objective perspective?

I would posit that the depth and complexity of a language’s syntax (including phrase structure) and morphology, and the breadth of its phonological inventory, which could be described in combination as the "richness" of this language’s structure, might contribute to its magical power. The vocabulary, or concepts identified by this language, would likely have a greater effect on its magical power. A deficient language, in which a concept or shade of meaning is unutterable (has no lexical representation), would be dramatically less effective in the performance of certain magical rites that require reference to this meaning. Certain concepts and phrases might be realized more powerfully in one language than another.

An entire spell or incantation could be alliterative and rhythmic – and therefore magically powerful – in one language, but relatively unremarkable in translation. The following example is an ancient Karelian poem in the Finnish language: a spell (manaus or loitsu) against snake venom.

Kyy, kylmänahka kavala Treacherous and cold-skinned viper,
vihollinen viirusilmä slithering and slit-eyed fiend,
maan nuora kanervakarva heather-colored belly-crawler,
maan on sykkä synyntäsi learn now of your contemptible extraction,
tiijä sykkä synyntäsi hear and know your lowly provenance:
maan kavala kasvantasi Earth it was who first uncoiled you,
maast on muutkin matoset as it did much crawling vermin,
kirjavatkin käärmehet even many-colored serpents.
Enpä tijä karvoasi As for what your proper hue be
mikä karva lienetkää I can't say, nor does it matter
oo, vaikka yheksikarva, if you were nine different colors,
mustakarva, marmiakarva, whether you are black or greyish
vaikka vaskenkarva or perchance a shade of copper.
Pirun riivattu purja Evil, stinging devil's minion,
et very miun verellä never shall my blood refresh you,
et liho miun lihalla nor my flesh sustain your body.
sie sahaselkä sihisijä Hissing ghoul with jagged backside,
pitkähammas paholainen long-fanged, vicious, wicked creature,
mäne metsähän mäelle find a hillside in the forest,
pakene pajun vesoille hide amongst the tender willows,
luikerra kivien kolohon slink into a stony hollow,
musta mato mää mullan alle      creep, black worm, into a burrow
vie vaiva männessäs and take my affliction with you,
kipu kova kulkeissas carry off the pain I suffer
noille tappotanterille to those killing fields of battle,
vainovaivoille sijoille to the very sites of warfare.
puhista purema paha Cleanse the grievous wound you gave me,
poita myrkky myötämöiten rid my veins of this your venom.
eläkä enää ehätä Henceforth do not hasten thee hither,
tule näille nurkiloille never wend your winding way here,
ruma roikale luikertele be, foul thing, forever banished.

This traditional poem (or runo) was artfully translated by Kirsi Kähkönen, but some essential characteristics of the original were inevitably lost in translation, and so its magical power and effectiveness as an incantation are dramatically reduced. Whereas the Finnish is beautifully constructed, the English text seems somewhat awkwardly phrased. For example, "maast on muutkin matoset" could be glossed as "[from the earth]-[are]-[also, even more]-[worms]." Finnish grammar allows for the particular ideas in this poem to be more elegantly phrased than any English translation. The consistent and beautiful alliteration in the Finnish is only occasionally shadowed in the English, as in "slithering and slit-eyed…." This is not a problem with this particular translation, but a difficulty with translating magical or poetic language in general. Each language has different vocabulary (and even phonotactic constraints), so alliteration is almost impossible to reproduce in any translation – particularly if the translator tries to preserve the meaning of each line or phrase, and the metrical form of the entire piece. These attributes that generate magical power depend on the phonetic form of an utterance, and each language uniquely constrains the forms in which a given meaning can be conveyed, It is therefore impossible to maintain the power of magical language in translation.

Translation of magic can never be wholly successful, because the source language’s power cannot be preserved, and also because certain specific languages are more suitable for a given magic-user. Magical language is unique in that it must be strange to those who hear it, and even the one who speaks it. This weirdness is necessary to substitute function for ordinary meaning, and to effect changes in the minds of everyone involved. Since humans perceive the world indirectly, through cognitive processes, these effects can also modify properties of the world around the participants. A magic-user can control any object or entity he can name with sufficient magico-linguistic power. Any community can have a powerful magical language, since its speakers can construct their own – by transforming an old, esoteric dialect or neighboring tongue, or creating a language completely from scratch.

However, it seems that even a constructed language could have only finite distinct possibilities for expressing a given thought, and only a narrow range of meaning can be associated with a given phrase. The infinite range of possible meanings to be expressed within the context of magic cannot be captured by a uniformly powerful set of phrases; some must lack the characteristics necessary to produce magical power. Or, if all phrases within the language were powerful, it could not be sufficiently complex to express – without irresolvable ambiguity or semantic overlap – every meaning that could be conceived by humans’ combinatorial cognitive processes. Therefore, although a magic-user can construct a language so that it is strange from his perspective, and also seems weird to those who share his native tongue, this language cannot achieve ultimate magical power for every phrase. Because the power of a language depends on the context, there can be no single ideal magical language.


Agrippa, Henry Cornelius. Three Books of Occult Philosophy. ed. Donald Tyson. Minnesota : Llewellyn Publications, 1993.

Austin, J. L. How to do things with Words. London : Oxford University Press, 1962.

Butler, Judith P. Bodies that Matter: on the Discursive Limits of "Sex". New York : Routledge, 1993.

Dickson, Keith M. Ritual Semiosis Mumbojumbo: Magic, Language, Semiotic Dirt. American Journal of Semiotics, 1994, 11, 1-2, 151-172.

Fauquié, Rafael. The Power of the Word. Thesaurus, 1993, 48, 2, May-Aug, 405-410.

Idel, Moshe. On Talismanic Language in Jewish Mysticism. Diogenes, 1995, 43, 2(170), summer, 23-41.

Lehrich, Christopher Ian. The language of demons and angels: Cornelius Agrippa’s occult philosophy. Leiden, The Netherlands : Brill, 2003.

Malinowski, Bronislaw. The Meaning of Meaningless Words and the Coefficient of Weirdness. From Coral Gardens and Their Magic, 1935, vol. 2, "The Language of Magic and Gardening," 213-222.

Mathiesen, Robert. The Laws of Magic. 1999.

Morris, D Hampton. The Creative Word: Mallarme and Esoteric Theories of Language. The University of South Florida Language Quarterly, 1980, 18, 3-4, spring-summer, 25-26, 32.

Todorov, Tzvetan. The discourse of magic. L'Homme, 1973, 13, 4, Oct-Dec, 38-65.

Jean Baudrillard

Last modified 10 March 2005