The “Book of Enoch”, the Angelic Alphabet and the “Real Cabbala” in the Angelic Conferences of John Dee (1527–1608/9)

D espite the wide range ofhis scholarly interests, John Dee is probably best known (at least in the popular imagination) for a series of seances or magicai "actions" conducted during the 1580s and possibly earlier, in which he believed himselfto have established contact with angels. The angels entrusted him with esoteric truths in order that he might play a crucial role in the coming Apocalypse and the subsequent redemption of mankind. Vital to this project was the revelation of a holy book written in a previously unknown language (supposedly the language of the angels and of Adam), as well as a magicai alphabet into which the text of the book was to be transcribed. The book was supposed to contain prophecy, the interpretation of which would restare human knowledge to the state from which it had fallen over the course ofhuman history. Dee kept meticulous records of his actions, some of which were published by Meric Casaubon in the mid-seventeenth century (Dee, 1659), probably with the intention of discrediting Dee as either a conjurar of demons or the dupe of his medium, the disreputable Edward Kelley (French, 1972:4-19). Diaries covering an earlier period (March 1582-May 1583) were subsequently discovered hidden in a chest along with several other works, and eventually carne into the possession of Elias Ashmole (1617-1692) (Harkness, 1996:717-718; Peterson, 2003:48-49). These are bound in Sloane MS. 3188, an edition of which was recently published by Joseph Peterson (2003). The present discussion is based primarily on this manuscript, as it is here that Dee relates the revelation of the "angelic book" and the first part of its contents (the remainder, written up by Kelley ata later date, is in Sloane MS. 3189). A series of prophetic "calls" known as the 48 Claues angelicce (Sloane MS. 3191) form a further important source of material in the angelic language. Because the Claues are accompanied by an English gloss, they seem to offer a key to the mysterious language of the angels, and have consequently been of particular interest to la ter generations of magicians (Magickal Review website ). Donald Laycock' s ( 1994) "dictionary" of angelic words and their English equivalents is based exclusively on the Claues angelicce, rather than on any analysis of the unglossed material. It is not my intention here to fill this gap by analysing the angelic text in detail, and I restrict my comments about the language to the first part of the "angelic book". The early text is of especial interest because it differs in nature from the later material (i. e., the contents of Sloane MS. 3189) and reveals the process of development of the work. This

material and the angelic alphabet-which are analysed in more detail below (sections 4 and 5) -yield valuable insights into the nature of Dee' s language project and its relationship to other Renaissance magicai movements, in which the imminence of the Apocalypse and the urgent need to regain lost knowledge emerge as key themes.

The "Book of Enoch" or Liber Logaeth
The holy book dictated to Dee by the archangel Raphael is composed of 49 leaves, each folio ofwhich contains a table or grid of 49x49 cells. The diaries in Sloane MS. 3188 describe and record the contents of the first leaf (two folios, i.e., a total of (2x49x49) = 4802 cells). Here, each cell was to be filled with a number of letters, usually representing a word ( although in some cases a cell contains two words, o r a word may be spread across several cells ). ln the diaries, however, the lines are simply written out "at large" as they were dictated. The pattem changes in the last 9 lines of the second folio; these have only one letter per cell and are laid out in a grid in the manuscript (Sloane 3188 fol.85v; Peterson (2003:343) has not drawn the grid).
The first leaf is far from perfect in its structure. For each folio, 12 of the 49 lines contain more or fewer than 49 words; the shortest has 35, the longest 59. Leaves 2-48 (leaf 49 was never revealed to De e) are written in table form (Sloane MS. 3189), with one letter per cell, like the final lines of the first leaf. All of the material is written in the Roman rather than the angelic alphabet. Despite repeated admonitions from the angels to "leme those holy letters ... in memory" (Peterson, 2003 :274), neither Dee nor Kelley seems to have been willing to do so, let alone to undertake the laborious task of transcription.
The identities of the book and of the angelic language are rather complex and warrant some discussion at this point. Both are associated with the prophet Enoch and the lost book of scripture attributed to him, the text of which was known in De e' s lifetime only through a quotation in the New Testament: Enoch the seuenth from Adam prophesied before of such, sayinge. "Beholde, the Lorde shall come with thousands of saynetes, to geue iudgement against all men, and to rebuke all that are ungodly amonge them of all their ungodly dedes, which they haue ungodly committed, and of all their cruell speakynges, which ungodly sinners haue spoken against hym. (Jude 1: 14-15 1 ) The book quoted in the Epistle to Jude is known today as 1 Enoch, an Ethiopic version of which was "rediscovered" by European scholars in the eighteenth century (Barker, 1988:5-16). According to tradition, Enoch was able to communicate with the 1 I have quoted the text from the 1539 Great Bible. Dee preferred to read Scripture in Latin; he did not own an English Bible, and when he cites Scripture in English, the translation is his own (Roberts and Watson, 1990:27 angels; he was translated to Heaven without dying (Hbr. 11 :5); and h e is supposed to be one of the two prophets who fight against the Beast in Revelation (11:3-12), the other being Elijah (the prophets are not named in the text of Revelation itself). Later commentators often refer to De e' s angelic book as the "Book of Enoch", an identification which is not clear from the early diaries. Dee was certainly aware of the lost scripture, about which he questions the spirit called 11 in the action of 18 April 1583 (Peterson, 2003:354-355). The references to Enoch and his book in this conversation are ambiguous, however. Clearer evidence can be found in a diary entry from 10 April 1586 (Josten, 1965:247-255), written during Dee's traveis on the Continent, when h e and Kelley enlisted the angels' help in avoiding the unwelcome attention of the Catholic authorities. ln a list of notebooks and manuscripts which they were instructed to destroy (some of which were miraculously restored at a la ter date), Dee mentions one that "contained that wisdom and science, with which Enoch (by God's will) was imbued" (Josten, 1965:249-according to Josten, the item in question is Sloane MS. 3189). Later, Dee refers to another document (which he was instructed not to bum), as "a small part of the book of Enoch which Thou [Dee's unidentified divine interlocutor] hast given me" (Josten, 1965:254). These references imply that Dee did believe that h e had in his possession at least a part of the lost Book of Enoch, which can be identified with the angelic book revealed in 1583.
To complicate matters further, the book is referred to elsewhere by titles in the angelic language. ln the action of 6 April 1583, Kelley is shown the book with a covering of blue silk, upon which in gold lettering is written Amzes naghézes Hardeh, which apparently means, "the universall name of him that created universally be praysed and extolled for euer." (Peterson, 2003:325). Ashmole inferred that this was the title of the book (Peterson, 2003 :x), although this is not clear from the diary itself. The title later given (at least, to the material contained in Sloane MS. 3189) by the angels themselves is Logaeth, "which in your Language signifieth speech from GOD" (Dee, 1659: 19).
Since the book is identified with Enoch, the angelic language is often referred to by later writers (though never by Dee) as "Enochian". Confusingly, Peterson (2003:37) calls the language of Logaeth "Enochian", but elsewhere implies that this label should be applied to a distinct language which appeared in la ter works (2003 :32). Laycock makes a clearer distinction between the earlier angelic language of Logaeth and the "true" Enochian language of the Claues, which he suggests is somehow generated from the earlier material (1994:29-35, 39-44). He does not suggest any mechanism to link the two, however, and his criteria for distinguishing them from one another are largely impressionistic.

Cryptography and Cabbala
The practice of manipulating the alphabet for magicai purposes has a history stretching back into antiquity, but in the sixteenth century it gained new impetus from the adaptation of Jewish Cabbalistic techniques by Christian writers like Johannes Reuchlin (1455-1522) and Heimich Comelius Agrippa (1486Agrippa ( ?-1535 (Scholem, 1974: 196-201;Maxwell-Stuart, 1999;Léon-Jones, 2006: 149-151). These techniques, which are themselves at least partly based on Greek alphabet-magic, promised access to higher knowledge through an understanding of the correspondences between letters, numbers, and elements of the natural and supernatural worlds. N ames and sounds were central to magicai operations, the principie being that by discovering the name with which a thing was originally created by God, it was possible to influence that thing (be it a part of the natural world, a human being, or a spirit). ln his Conclusiones sive theses DCCCC (1486;Kieszkowski, 1973), Giovanni Pico della Mirandola (1463-1494) included 26 conclusions on magic, in which he stated that "Sounds and words have efficacy in a magicai operation because that by which Nature works magic . . . is the voice of God" (translated by Maxwell-Stuart, 1999: 147).
Magicai writings of the Renaissance period contain an abundance of lettertables, many of which employ shifts or specific permutations of the alphabet based on Cabbalistic techniques for discovering holy names (Reeds, 2006: 196-201). Dee's interest in alphabetic gymnastics of this kind is well known, and forms part of the background to his ambitious Monas Hieroglyphica, published in 1564 (Josten, 1964). ln this highly influential work he asserts that there is, beyond the Hebrew Cabbala, a "real Cabbala" based on the pure quality of number. This is applicable not only to Hebrew but to all languages and scripts; a fact which Dee offers as proof that "the sarne most benevolent God is not only [the God] of the Jews, but of all peoples, nations, and languages" (Josten, 1964: 133). He claimed that his "monad" (a complex figure based on the symbol for Mercury -see fig.l) could be used to demonstra te Figure l: Dee's monad that the forms and sequences of the letters of "all three" alphabets (Latin, Greek and Hebrew), as well as the symbols and concepts of astrology and alchemy, had a rational basis. What Dee claimed to have achieved, some fifteen years before he began trying to contact angels, was a means of unifying all knowledge based on a Pythagorean understanding of number as "the indisputable, fundamental component of what exists" (Léon-Jones, 2006: 150). This mystical fascination with number was well established in Renaissance occult philosophy -among his conclusions on mathematics, Pico asserted that "Through numbers is to be had a way to the investigation and understanding of all knowable things" (Kieszkowski, 1973:74;my translation). Agrippa was similarly enthusiastic: The Doctrines of Mathematicks are so necessary to, and have such an affinity with Magick, that they that do profess it without them, are quite out of the way, and labour in vain, and shall in no wise obtain their desired effect. (Agrippa, II.i (1650: 167)) The manipulation of letters on mathematical principies naturally calls to mind the art of cryptography, and it is no surprise to find Dee taking an interest in texts which deal with ciphers as well as magic. ln 1563 he went to considerable lengths to obtain a manuscript copy of Trithemius' Steganographia, a work written c.l500 but first published in 1606 and placed on the Index Librorum Prohibitorum shortly afterwards because it (apparently) advocated the practice of angel-magic (Bailey, 1879;Reeds, 1998). The work outlines procedures for sending secret messages by invoking certain angels with formulae written in what appears to be an unknown language, e.g. "Padiel aporsy mesarpon omeuas peludyn malpreaxo ... " (Caramuel, 1635:24). As many readers realised, these formulae contain ordinary texts hidden steganographically, with the names ofthe angels (or rather, the order ofthese names in a list given in the first chapter) providing the key to the recovery of the text. ln the above example, Padiel is the second of the angels listed, and the hidden message is revealed by reading every second letter of every second word. So, from "padiel aPoRsY mesarpon oMeUaS peludyn mAlPrEaXo ... " we can extract the Latin plaintext primus apex. Some of the hidden messages also use simple substitution ciphers (Reeds, 1998:3).
There has been much debate about the true nature and purpose of Steganographia -whether it is "really" a treatise on cryptography disguised as an occult work, ora manual of angel-magic disguised as a book of cryptography. The former position is generally favoured by scholars, perhaps under the influence of Juan Caramuel (1608-1682) and other seventeenth-century redactors ofthe work, who were keen to dissociate Trithemius from the practice of sorcery. Many modem writers have followed Walker (1958:86-90) in arguing that the presence of the cryptograms in books I and II of Steganographia demonstrates that the author' s purpose here was cryptological, while the third book is genuinely occult as it does not contain this kind of material. However, more recent work (Reeds, 1998) shows that book III does in fact contain ciphers, although the key is based on a table of numbers rather than on the names of angels; Walker' s position is therefore no longer tenable.
It seems to me that there is little point in being dogmatic about Trithemius' intentions; I see no reason why the work must be understood to be about either cryptography or magic, but not both. As has been mentioned, Renaissance magic utilises Cabbalistic techniques of letter substitution and transposition -the sarne operations used in cryptography -for generating magicai names, words and formulae. It must also be remembered that in the sixteenth century, mathematics (which underlies all such techniques) was viewed with deep suspicion -at least by the uneducated (French, 1972:5) -and it was regarded by magicians as fundamental to their operations, as the quotes from Agrippa and Pico illustrate. Furthermore, there was in this period no formal division between the occult sciences and what we would now regard as "modem science": the boundaries between alchemy and chemistry, astrology and astronomy, mathematics and mathesis (the number mysticism which fascinated Dee) were blurred or absent. The impulse of scholars like Dee was not to discrimina te but to synthesise, and ultimately to uncover the mystical key which would unite all the sciences.
There remains an unanswered question in respect of the invocations in Steganographia: What is the source of the nonsensical text which covers the hidden message? Regardless of whether it functions purely as "noise" to conceal a message or is intended to have some genuine magicai effect, it must have been generated somehow; perhaps it is connected with the angel-names themselves, or with the numbers appearing alongside them in the list indicating the times and seasons govemed by those angels. Trithemius makes no special claims about it ( e.g., that it is the Adamic language ); but the fact that he uses this sort of material to conceal the message, rather than an innocuous-looking piece of Latin, would seem to suggest that this pseudo-language is considered to be an appropriate form for magicai incantations (see section 5, below).
Given De e' s interest in the book ( which is evident from the great effort which he undertook to obtain a copy (Bailey, 1897:402)) and the superficial resemblance between his angelic texts and Trithemius' invocations, it is conceivable that there is some link between the two. Admittedly, the resemblance does not necessarily go any further than that both are ostensible nonsense purporting to be some kind of special, magicai language associated with angels. Laycock states that Trithemius' mystical language "in some ways resembles Enochian" (1994:57), but he does not elaborate. The possibility naturally presents itself that, if messages are hidden in the invocations of Steganographia, perhaps Logaeth also contains cryptograms.
The general consensus is that this is not the case. Laycock alludes to unfruitful attempts by modem occultists to apply Trithemius' methods to Dee's angelic text, though again he does not go into any detail. Reeds (2006: 197) notes that the tables of Logaeth do not conform to any mathematical pattem (in contrast to those of Steganographia and other magicai texts ). Moreover, the evidence of the diaries suggests that Dee was entirely sincere in his belief in Logaeth as a text of prophecy written in a genuine "angelic" language; it is not simply a "noise" channel concealing a message. Finally, the diary entries indicate that the book was dictated spontaneously by Kelley while in a trance, so any secret messages hidden within it would be Kelley' s, not Dee' s, and the ciphertext would have to have been either prepared and memorised in advance of the action or else generated ex tempore during it. I hesitate to suggest that either of these feats is impossible, but they would certainly require some prodigious skill on Kelley' s part. On balance, it is probably safe to reject the idea of Logaeth as cryptography.

The angelic alphabet
The alphabet revealed by Raphael on 26 March 1583 (Peterson, 2003:268-271)  . The values of the letters (that is to say, their equivalents in the Roman alphabet) were not given directly to Dee by the angels; at the end ofthe diary entry in which the letter-names and the first line of the text are dictated, Dee writes out the angelic letters with their Roman equivalents and a marginal note "Thus I deciphred them after a day or two or three" (Peterson, 2003 :271 ). H e does not explain how the deciphering was carried out. The first two lines of the text were dictated letter by letter, then read out in full; Dee could have deduced the letter values from the pronunciation, as all of the letters except Q are present in these two lines. He would have to have overcome some discrepancies between spelled and pronounced forms; when Dee points these out to Raphael, the archangel becomes decidedly peevish (Peterson, 2003 :288-296). Laycock does not have a great deal to say about the alphabet, though he makes much (far too much, in my view) of a slight stylistic resemblance to Ethiopic, and even goes so far as to speculate that Dee might have had in his possession a copy of the Ethiopic Book of Enoch (Laycock, 1994:28) -a suggestion for which there is no supporting evidence whatsoever. This text was scarcely known outside Ethiopia until the mid-eighteenth century -not, as Laycock states, the seventeenth. Here Laycock has confused the "discovery" of the Ethiopic Enoch by James Bruce with the publication of fragmentary quotations from a Greek version of 1 Enoch by J. Scaliger in 1658 (Barker, 1988:8-11). There are earlier reports of contact with the Ethiopic text -Guillaume Postei wrote in his De originibus (1553) that in Rome he had met an Abyssinian priest who showed him a copy and explained its contents (Schmidt, 1922:50). Pico is said to have owned a copy ofthe book ofEnoch, though there is no direct evidence that this was the case, and Schmidt's claim (1922:46-47) that this might have been the Ethiopic rather than a Hebrew or Greek version is unfounded. It seems highly unlikely, therefore, that even so remarkable a library as Dee's would by chance have contained a copy which mysteriously made its way from Africa to England and of which there does not appear to be any trace in De e' s writings, the (admittedly incomplete) catalogues ofhis books (Roberts & Watson, 1990), or in fact anywhere else. He may well have been aware of Postel's anecdote about the Abyssinian priest, since Dee did own a copy of De originibus (no. 868 in Roberts and Watson's edition ofDee's catalogue).
With regard to the inventory of letters, the angelic alphabet is simply an alterna te Roman alphabet ( although there are only 21 characters, in comparison with the usual 23-letter Roman alphabet-the redundant letters K and Y are omitted). Dee reports that after Raphael dictated the letters by name, "there cam two lines and parted the 21 letters into 3 partes, eche being of 7" (Peterson, 2003:270). Each of these groups of 7 letters is a reordering of the equivalent sequence in the 21-letter Roman alphabet, with one exception-O is in the third group and Q in the second 2 : Roman (with 0/Q ABCDEFG HILMNPQ ORSTUXZ exchanged) 2 Given that Q is the only letter not present in the first two lines of the text, this discrepancy may be significant. If we hypothesise that the threefold division of the Roman alphabet is to be preserved and that Dee made an error in his decipherment, groups 2 and 3 of the angelic alphabet would read MIHLPON and XQRZUST. Translated into numbers, as on the following page, the order of group 3 is unaffected and group 2 would have a sequence 4213765. Even with this alteration, the hypotheses presented below would still be unproductive.

I Angelic IBCGDFAE I MIHLPQN I XORZUST
This correspondence suggests that the angelic alphabet may have been generated algorithmically from the Roman alphabet in some way dependent on the threefold division. As a first hypothesis, let us suppose that each group of seven Roman letters undergoes some specific permutation in order to generate the equivalent group of angelic letters. If this is so, when we replace the letters of the groups with the numbers 1-7 (based on the conventional order -A=l, B=2 ... G=7; H=l, 1=2 ... Q=7; 0=1, R=2 ... Z=7), the number sequence should be the sarne for each ofthe three groups. It is immediately apparent that this is not the case: Another possibility is that each of the three group orderings is generated from the preceding one, the first group being generated from the Roman ordering, i.e.: The progression from one sequence to the next can be expressed in terms of the difference between the input and output for each number in the sequence (in other words, by subtracting modulo 7). For example, the first digit of group 1 is 2, while the first digit of the Roman ordering is 1, which gives us a difference of 1. The results of this procedure are as follows:

Roman
Roman -7 Group 1: 1140125 Group 1 -7 Group 2: 2616060 Group 2 -7 Group 3: 261463 6 If this hypothesis were correct, each of these three sets of differences should be identical, since the algorithm goveming the procedure would be the sarne in each case. Although the hypothesis as it stands is evidently false for the whole set, it does appear to work for the first three letters of the transitions between the groups ( though not for Roman -7 Group 1 ). The transition BCG -7 MIH is, in terms of modular arithmetic, equivalent to the transition MIH -7 XOR. It is highly tempting to conclude that this correspondence has some significance; however, it is equally plausible ( and, since I have been unable to develop any consistent solution on this basis, perhaps more so) that it is merely a chimaera.
The division of the alphabet into three produces another noteworthy feature which might be the result of deliberate organisation. Within each group of 7 letters there is one which occupies the sarne ordinal position in both the Roman and angelic alphabets. These are D (group 1, position 4), I (group 2, position 2), and U (group 3, position 5). These three letters together spell the Latin word diü "by day; for a long time". If the pattem is intentional, diü could perhaps be a reference to the coming of the day of redemption after long ages of mankind's suffering; though this connection is tenuous at best. A more promising altemative would be to construe the "stationary" letters as D and the Roman numeral IV -D being the fourth letter not only of the Roman and angelic alphabets, but also of the Greek and Hebrew; and, of course, Roman D is homophonous with the philosopher' s name. ln his diaries, notes and elsewhere, Dee usually signs his name with a Greek 1' 1. His letter to Emperor Rudolf II ( 17 August 15 84) mentions that "[I am symbolized by] the fourth letter of each of the three [se. the Hebrew, Greek, and Latin] alphabets" (Josten, 1964:93, n.50), which Dee links to the fact that Rudolf is the fourth Holy Roman Emperor with whom he has been involved.
A further possibility is that all three letters are supposed to stand for Roman numerais, giving DIV = 504. This number has various properties that might appeal to a student of mathesis. It is the product of 7 and 72, both of which are prominent in mystical and occult tradition (for example, Cabbalists say that God has 72 names, of which 7 are especially important; the Septuagint (the Greek translation of Hebrew scriptures) was supposed to have been produced by 72 scholars; there are 7 planets in astrology, with whose influences the 7 metais of alchemy are associated, and which are said to be govemed by 7 angelic intelligences; and, of course, the angelic alphabet contains groups of 7 letters ). 504 is also a multiple of the total number of letters in the angelic alphabet (21x24=504), and a multiple of the Pythagorean sum of 7 (1+2+3+4+5+6+7=28; 28xl8=504). A set of 7 elements, such as the letter-groups of the angelic alphabet, can be combined in 5040 different ways, or 504 multiplied by the Pythagorean sum of 4 (1+2+3+4=10), which plays an important role in the Monas Hieroglyphica (Josten, 1964:209). It is doubtful whether Dee would have been aware of the general principie that a set of n elements has n/ permutations; work on permutations and combinatorics had been done by Chinese and Indian mathematicians, but was not known in the West until the publication of Leibniz' De arte combinatoria (1666) (Cooke, 2004:215). On the other hand, it is no accident that the Monas contains 24 theorems; these are based, according to Léon-Jones (2006: 150) on the 24 Metatheses (i.e., permutations) of the Pythagorean Quatemary (4! = 24). Dee may have also known, therefore, that the Septenary has 5040 Metatheses.
A further connection between this number and the numerology of the Monas can be found in that work's seventeenth theorem (Josten, 1964: 172-175;Léon-Jones, 2006: 154). Here Dee synthesises the conclusions of earlier theorems to produce the significant number 252, which he associates with the philosopher's stone and "which number we recommend beginners in the cabbala to explore, for we can deduce it from our premises in yet two other ways, here omitted for the sake of brevity" (Josten, 1964: 175). This number, which was evidently of considerable Cabbalistic value, can be doubled to produce 504.
These and other correspondences are terribly beguiling, and are precisely the sort of thing that would have appealed to Dee, with his love of number and his belief in it as the key to the "real Cabbala". The letters DIU offer us an opportunity to link Dee's own name and identity to his earlier work in the Monas Hieroglyphica, to the Pythagorean Quatemary, to the lapis philosophorum, and to the holy alphabet and the revealed book of prophecy itself, the book which Dee believed was to play a crucial role in the coming Apocalypse. Entertaining as it is to speculate that these letters form a key to the ordering of the alphabet, I have had no success in identifying any way to use this hypothetical key to link the 7-letter groups either to one another or to their corresponding groups in the Roman alphabet. If the pattems identified here are genuinely involved in the construction of the angelic alphabet ( and I doubt that this is the case), I am forced to admit defeat, at least for the time being, in my efforts to determine the method used.
The ordering of the angelic alphabet does have some similarities to the "three alphabets" known to Dee. The sequence BCGD at the beginning is reminiscent of the Greek and Hebrew ordering BGD (BíL1, ~n); the final pair ST is also final in Hebrew (tl.m); and the vowels appear in the sarne order as in the Roman alphabet, AEIOU. That the alphabet begins with B, the second letter of the Roman, Greek and Hebrew alphabets, may be connected with the Cabbalistic significance ofthis letter (bet) as the first letter of Genesis 1:1 (bereshit elohim "God created"); in this connection, Léon-Jones notes that from a Pythagorean perspective, "two is the first number, since one is the basis of number" (2006:151). There is clearly some common ground, but no systematic relationship with any of the three "natural" alphabets is evident. ln summary, the angelic alphabet contains various pattems which appear to suggest some method in its arrangement, but I have been unable to uncover any consistent organisation. The pattems may simply be consequences of the fact that Dee and Kelley were working in an environment filled with altemate alphabets of various kinds. Surrounded as they were by tables of commutation such as those found in Agrippa's De occulta philosophia (1650[1533) or the anonymous Soyga (Reeds, 2006), it is quite plausible that they might unconsciously have produced an alphabet with structural similarities to others, natural and artificial. The DIU pattem is harder to explain. I suspect that it is no more than a happy coincidence or an artefact of the transposition; but it may yet tum out to have some greater significance.
Much as Dee loved magicai alphabets and alphabet-manipulation, he was not necessarily always committed to the complete exposition of an ide a. For example, in the preface to the Monas Hieroglyphica he asserts that "Reasons must be given for the shapes of the letters, for their position, for their place in the order of the alphabet ... , for their numerical value" (Josten, 1964: 123) and goes on to discuss the mysteries of the alphabet at some length, claiming that his theorems justify the forms and positions of the letters. ln the body of the work, however, he offers explanations for the letters X and V and L only (Josten, 1964:159, 169-173); since these explanations rely on the values of the letters as Roman numerais, it is difficult to see how the principie might be extended to the rest.

The angelic language
I do not propose here to present a detailed analysis of the angelic "language" itself. From my examination ofthe first leaf ofthe Liber Logaeth, Iam inclined to agree with Laycock's conclusion (1994:29-35) that it displays a number of features characteristic of glossolalia (Goodman, 1972;Williams, 1981: 169-191)0 Firstly, the phonology and phonotactics of the angelic utterances present few difficulties to an English speaker. This behaviour conforms to what we would expect from an English-speaking glossolalist. (The text does contain a few unusual consonant clusters ( eogo excol-phagmartbh (Laycock, 1994:33;Peterson, 2003:315); note that <ph> is pronounced [f], as in English and Latin, and is used interchangeably with <f> in several places )o Secondly, the angelic text exhibits repetition, rhyme and alliteration throughout. Some lines are highly repetitive, eogo umas ges umas umas ges gabre umas umáscala um 'phazes umphagám (Laycock, 1994:33;Peterson, 2003:312-313)0 Others contain "paradigmatic" variation of the sort found in glossolalia: compare Kelley' s quamsa ol danfa dot santa on anma (Laycock, 1994:33;Peterson, 2003:339)  Over the course of the revelation, the individual utterances ( the lines of the angelic book) become noticeably briefer. The second page of the first leaf contains many places where a single word is spread across several cells (indicated by underscoring in the manuscript), and later lines are composed largely of monosyllableso ln the last 9 lines of the second page, polysyllabic words reappear, but there are fewer of them as the lines fall into the pattem of one letter per cell, with word boundaries indicated by commas in the manuscript. The final line consists of only 6 words, ganfúmarabómonah, gástages, órdolph, naqas, orgemvah, noxad (Peterson, 2003:343)0 One is left with the impression that Kelley's utterances became gradually more like fluent glossolalia and less like (simulated or real) language; or perhaps he was simply tiring of the exercise and eager to reach the endo Occasionally, the angels offer a tantalising glimpse into the meaning of the text by revealing the translation of a word or phraseo English or Latin equivalents ( often complex phrases) are given for just 28 words of the first leaf, eogo, argedco "with humility we call you, with adoration of the Trinity" (Peterson, 2003:31 O, no266)0 Only one piece of explicitly grammatical information is given -the word Befes is supposed to represent the vocative case of Befafes ( the name of an angel mentioned earlier in the diaries)o Throughout the process of revelation, hints like these are dropped from time to time, with nothing of substance ever being revealedo The information given is enough to persuade Dee of the book's importance, and to encourage the expectation that the full meaning would be made known when the time was right.
Another of the many unanswered questions surrounding Dee's angel-magic is that of why angels should be expected to speak a tongue unlike any human languageo Most Renaissance authors on the subject of communication with angels and demons assumed either that it was silent (impressions being transferred directly between minds) or that the spirits used the languages of their mortal interlocutorso ln the sixteenth century it was widely believed that the language of Adam and of God was Hebrew, "because that was and carne from heaven o o o and seeing all tangues have, and do undergo various mutations, and corruptions, this alone doth alwaies continue inviolated" (Agrippa, III.xxiii (1650:412-413)). This belief is implicit in Cabbalistic magic, an axiom of which is that any created thing (including angels) can be influenced by invoking its original name. Angel-names in Jewish Cabbalistic texts (carried over into Christian Cabbala) usually consist ofthe Hebrew word for whatever power or quality is the angel's province, with the suffix -el "God", e.g. Raphael "physician ( or medicine) of God", cf. Heb. rafa' (~~1) "to heal".
On the other hand, there does seem to have been an established practice of using exotic pseudo-language in magicai incantations such as those found in Steganographia. ln his 26 conclusions on magic, Pico states: Though the authors of the manifestos identify their philosophy with the wisdom of Adam, Moses and Solomon, the magicai language which they claim to know is not that of Adam and Enoch, but a new one constructed by the members of the Fratemity. The authority claimed by the Rosicrucian authors for their books and their secret language is based on an appeal to logical and philosophical perfection, rather than to divine origin. The Rosicrucian notion of a "magicallanguage" expresses an idea more closely related to other seventeenth-century language projects like those of George Dalgarno (c.l616-1687) or John Wilkins (1614-1672) (Salmon, 1979:97-126;Shumaker, 1982: 132-172). Nevertheless, the influence of Dee is evident, and he was believed by many seventeenth-century occultists to have been a member of the (probably fictitious) order. ln 1652 Ashmole received a communication from a Mr. T ownesend who claimed that De e was a member of "ye Brotherhood of ye R. C R." (quoted by French, 1972:14), and that he "was accused to haue stoll'n the booke he owned called Monas Hyeroglifica out of All Sowles Colledg in Oxford" ( quoted by Josten, 1966:603, n.2). Townesend's allegation of theft is intended to imply that the Monas is older and therefore more distinguished than if it were accepted as De e' s own work. It is not impossible that, in Townesend's mind or perhaps even in those of the manifesto-writers, the Monas is to be identified with the mysterious "book M.", said to have been written by the founders of the Fratemity some time in the fifteenth century (Yates, 1972:242). Ifthis was Townesend's belief, then it would have been necessary for him to deny Dee's authorship.

Conclusion
Much as I would like to be able to claim a breakthrough in "deciphering" the alphabet or the text of Logaeth, my explorations have led me to conclude that there is probably no cipher to be broken. Beguiling as they are, the structural pattems in the angelic alphabet lack consistency, and I suspect that they reflect the organisation of other alphabets with which Dee and Kelley were familiar in the writings of magicians and occult philosophers such as Agrippa. Given the method of revelation, it is to be expected that the alphabet would emerge from Kelley's subconscious (unless we choose to ascribe it to a genuinely supernatural source ), and so would bear similarities to other alphabets without having any coherent structure of its own. The DIU pattem is highly attractive; nevertheless, I suspect that it is nothing more than an illusion, albeit an extremely fortuitous one. If it does represent a key to the arrangement of the angelic alphabet, I am at a loss to explain how it might work.
Although De e' s actions have parallels with earlier magicai operations, his acquisition of what he believed was a book of prophecy in a special holy language stands out from the activities of other occult philosophers of his day, and forms a vital part ofhis lifelong search for the "real Cabbala" ofthe Monas. It was a commonplace of Renaissance theology that the state of human knowledge had been in continuous decay since the Fall (Harkness, 1996:727;2006:277-278), and students of the occult sciences hoped to rediscover what had been lost in order to hasten the redemption of mankind. Dee, like many of his contemporaries, believed that the last days of the world were at hand -a belief expressed frequently in the pronouncements of the angels. The revelation of the Liber Logaeth made Dee the custodian of precious knowledge and confirmed his central role in the eschatological drama presented by the angels (Harkness, 1996:732-733). ln the action of 5 May 1583, Uriel explains the significance of the book in terms that leave no room for doubt: Oute of this, shall be restored the holy bokes, which haue perished euen from the begynning, and from the first that liued. And herein shalbe deciphred perfect truth from imperfect falshode, True religion from fals and damnable errors, With all Artes; which are propre to the use of man, the first and sanctified perfection: Which when it hath spred a While, THEN COMETH THE ENDE. (Peterson, 2003:395. Dee's emphasis)