THE CODEBREAKERS

by DAVID KAHN

  In their discussion of William Stone Booth’s “string cipher” (which, incidentally, bears no relation to the string cipher for which it was named), the Friedmans pointed out time and again how the enigmatologists stretch, ignore, or break their own rules when these stand in the way of a needed enigmaduction. They cast a skeptical eye upon the many anagrams of the longest word in Shakespeare, found in the Clown’s remark {Love’s Labour’s Lost V. i), “I marvell thy M[aster]. hath not eaten thee for a word, for thou art not so long by the head as honorificabilitudinitatibus: Thou art easier swallowed then a flapdragon.” Sir Edwin Durning-Lawrence enigmalyzed it to the Latin Hi lu-di F Ba-co-nis na-ti tui-ti or-bi (“These plays, F. Bacon’s offspring, are preserved for the world”). The Friedmans proved these offspring to be illegitimate by citing a quantity of other, equally valid anagrams of this word, one of which, by its very presence, hints that Dante, who died two hundred years before Shakespeare was born, may have rather literally ghost-written the Bard’s plays: Ubi Italicus ibi Danti honor fit (“Where there is an Italian, there honor is paid to Dante”).

  Some of the gayest moments in cryptology come when the Friedmans disprove some enigmaplans by showing that they are so loosely constructed that they can render multiple enigmaductions. In the poem “To the Reader” underneath the famous Droeshout portrait of Shakespeare in the First Folio, one Edward D. Johnson saw a symmetrical diagram of 22 letters that when rearranged spelled, for him, the 25-letter ejaculation Fr Bacon author author author. The triple repetition must have bolstered his confidence, for he issued a challenge: “If after checking the signatures … the reader is still of the opinion that they are all accidental, the writer would ask him to try a small experiment. Let him take from any book, ancient or modern, 20 consecutive lines of prose or poetry, place the letters in a Table, and then try to see if he can make up any word out of the letters the same distance apart in the text in the form of a chain.”

  The Friedmans found it “hard to resist this courteous request. We decided to use the text of one of Johnson’s own examples; and the poem ‘To the Reader’ divulged the message ‘No kidding, Francis Bacon: I wrote these plaies!—Shakespeare.’ … Our message is nearly twice the length of Johnson’s; it is a complete sentence; and it uses each letter of the diagram once and only once. But the disadvantage of this ‘method’ comes out very clearly here. Since our chosen letters do not have to ‘appear’ in their correct ‘order’ (i.e., we can arrange them any way we please), there may be several alternative ‘messages’ to choose from: amongst them, one (giving a very different sense to the pattern) runs: ‘No kidding! I, Francis Bacon, wrote these Shakespeare plaies.’ This alone is enough to show that Johnson’s method is worthless as a piece of cryptography.”

  Indeed it is. Virtually all the Baconian enigmaplans suffer from the grave weakness of multiple answers, as the Friedmans amply demonstrate. This fault instantly vitiates any alleged method of secret communication. For such a method, though intended to be secret, is first and foremost a method of communication. Of what value is it if the encipherer can never be sure that the message he puts in will be the one the decipherer will get out? If Walter Raleigh inserted into Julius Caesar a message beginning Dear Reader: The Asse Will Shakespeare brought William Hatton down to his grave to prove Bacon’s authorship, as the economist Wallace McC. Cunningham asserted he did, and the Friedmans, attempting to extract it by the rather vague rules that Cunningham said Raleigh used, read instead Dear Reader: Theodore Roosevelt is the true author of this play … Friedman can prove that this is so by this cock-eyed cypher, what good did all of Raleigh’s work do?

  Practical considerations such as that rarely trouble the Baconians. Nor do they often deign to answer these criticisms. When they do, their defense usually resembles a remark that Arensberg made. The Friedmans had just used his enigmaplan to enigmalyze The author was William F. Friedman from one of Arensberg’s own books to show him the invalidity of it all. Replied he with equanimity: “What you have done does not disprove the presence of the sentence The author was Francis Bacon which I found in The Tempest” But the sentence is not present in The Tempest. Arensberg imposed it upon the thousands of letters that make up the play. It is like looking up at the hundreds of stars in the night sky and projecting upon a few neighboring ones the image of some mythic hero or animal. Orion and Pegasus exist only in the mind of the beholder, as does Arensberg’s sentence. The proof is that other minds, like the Friedmans’, may organize other patterns.

  Enigmatology resembles nothing so much as the Rorschach tests given by psychologists, in which a subject tells what he sees in an ink blot. The blot is formless, of course, and so anything that the subject reports can come only from within himself. The test thus discloses a great deal to the psychologist. To the enigmatologists, the Shakespearean plays may serve as a kind of literary Rorschach—and the snobbery and the fantasies of incest and adultery that appear in many enigmaductions may themselves be revelatory. This may partly explain the great emotional involvement of Baconians in their theories. Certain it is that to think that these mental pictures, whether inspired by ink blots or stars or letters, exist in external reality is to be out of touch with that particular area of reality.

  One system of steganography, and one alone, of all those applied by the Baconians to the First Folio, is valid. It has a special attraction, for it was invented by Francis Bacon himself “in our youth, when we were at Paris,” or some time between the ages of 15 and 18 during his service under the English ambassador from 1576 to 1579.

  He alluded to it in 1605 in his Of the proficience and advancement of Learning, divine and humane. In discussing “cyphars,” he wrote, in a statement that has become classic in cryptology, that “the vertues of them, whereby they are to be preferred, are three; that they be not laborious to write and read; that they bee impossible to discypher; and in some cases, that they bee without suspition. The highest Degree whereof, is to write OMNIA PER OMNIA [anything by anything]; which is undoubtedly possible, with a proportion Quintuple at most, of the writing infoulding, to the writing infoulded, and no other restrainte whatsoever.” He did not expand this ellipsis until the publication, in 1623, of De Augmentis Scientarum, an enlarged Latin version of the Advancement. This influential volume, whose classification of the sciences shaped man’s view of human knowledge for nearly two centuries, subsumed “ciphers” under writing, which he considered a branch of grammar, which in turn formed “the organ” of the “traditive doctrine, which takes in all the arts relating to words and discourse.” From here Bacon ascended the epistemological ladder via rungs of logic, faculties of the soul, human philosophy, knowledge of man, and philosophy to encompass at last all learning. At 62, he felt his own system “a thing that yet seemeth to us not worthy to be lost” and so gave it place in this grand design.

  He began by replacing the 24 letters of his alphabet (Bacon naturally used both i and j and u and v interchangeably) by permutations of two symbols, A and B, taken five at a time:

  a AAAAA

  b AAAAB

  c AAABA

  d AAABB

  e AABAA

  f AABAB

  g AABBA

  h AABBB

  i ABAAA

  k ABAAB

  l ABABA

  m ABABB

  n ABBAA

  O ABBAB

  p ABBBA

  q ABBBB

  r BAAAA

  s BAAAB

  t BAABA

  v BAABB

  w BABAA

  x BABAB

  y BABBA

  z BABBB

  Thus but becomes AAAAB BAABB BAABA. He needed five places because two things taken five at a time produce 25 or 32 permutations, whereas taken four at a time they yield 24 or only 16, too few for all the letters. Modern terminology would call this a quinquiliteral binary alphabet, and modern notation would replace the A’s and B’s by O’s and l’s, so that d would be 00011, but Bacon called it “bi-literal,” and the name has stuck
. He went on: “Neither is it a small matter these Cypher-Characters have, and may performe: For by this Art a way is opened, whereby a man may expresse and signifie the intentions of his minde, at any distance of place, by objects which may be presented to the eye, and accommodated to the eare: provided those objects be capable of a twofold difference onely; as by Bells, by Trumpets, by Lights and Torches, by the report of Muskets, and any instruments of like nature.” The raising of a torch could signify A, its lowering B. The teletypewriter utilizes Bacon’s binary principle by sending five marks or spaces within a given time to represent a letter (though its equivalents differ from his).

  The conversion of a message to biliteral form is only the first step in Bacon’s scheme to write “without suspition.” This scheme requires a cover-text. Now, among the objects “capable of two differences” are faces of printing type. These come not just in two but in scores of different styles, usually named for their designers, as Caslon, Baskerville, Bodoni, Garamond, each with its roman and its italic. Bacon’s system can be most clearly illustrated however, by using the roman and the italic forms of the present typeface as if they were different faces. The A’s of the hidden message become roman letters in the cover-text, and the B’s of the hidden message become italic letters in the cover-text. Thus, the cover-text “Do not go till I come” would represent the hidden message fly, which is AABAB ABABA BABBA, by setting the D and the O in roman, the N in italic, the O in roman again, the T in italic, and so on, like this:

  AA BAB AB ABAB A BBAD-

  Do not go till I come

  The cover-text says exactly the opposite of the hidden message; it is, of course, entirely independent of it, and this is what Bacon means when he refers to writing “omnia per omnia,” or anything by anything.

  This example is not very subtle because of the obvious difference between the roman and italic type styles. Bacon’s original suggestion was, however, not to use two strongly contrasting forms of the same typeface, but to use two separate typefaces with only slight differences between them, one face for the A’S, the other for the B’s. If these two resemble one another as closely as do, for example, Caslon and Baskerville, the ordinary reader might never even suspect the presence of the two. The decipherer, of course, would have to observe some very fine differences in the shading and curvature and dimensions of, say, the lowercase r’s in the two faces so that he will be able to distinguish the A r from the B r.

  One of Francis Bacon’s own examples of his biliteral cipher, in which plaintext Fuge (“Flee”) is concealed under covertext MANERE TE VOLO DONEC VENERO (“Stay till I come to you”). Note the difference between the A-form and the B-form e’s in Manere.

  Yet the procedure is valid. If the encipherer prevails upon a friendly, patient printer to set type following copy suitably marked, the message that he inserts will be the one that the decipherer extracts. There will be no ambiguities, no looseness, no multiplicity of decipherments. What is commonly called the “Bacon biliteral cipher” is therefore in no way an enigmaplan, but a true system of steganography, and an exceedingly clever and useful one at that. Technically, a message in it would fall into the category of the semagram, for its substitutes are not really the letters of the cover-text but the forms or shapes of those letters.

  It was only natural that Baconians should turn to their hero’s own system in an attempt to prove their case. The first enigmaduction based on it to see print applied it, not, however, to the First Folio but to an “uncouth mixture of large and small letters” that uncertain tradition records as the inscription of the original Shakespeare tombstone at Stratford. The carving of the epitaph seemed made to order for the Baconian biliteral:

  In 1887, Hugh Black took the lowercase letters as A-forms and the uppercase as B-forms. Regarding the G’s of digg as lowercase and the TY as a single capital, he produced saehrbayeeprftaxarawar. “To an ordinary person,” commented the Friedmans, “the resultant message would be enough to prove that there is no cipher being used. The difference between the ordinary person and the Baconian is, shall we say, one of degrees of persistence and ingenuity.” Black arranged his text in a special formation, drew a line to divide it into two parts, anagrammed one part into Shaxpeare and the other into Fra Ba wrt ear ay, and confidently interpreted the latter to mean “Francis Bacon wrote Shakespeare’s plays.”

  Black’s work was emended and expanded by one Edgar Gordon Clark. He enigmalyzed Fra Ba wryt ear. A A! Shaxpere and Fra Ba wrt ear. HzQ Ay A! Shaxpere out of the tombstone. These, to him, meant “Francis Bacon wrote here. Aye Aye! Shakespeare” and “Francis Bacon wrote here. His cue. Aye Aye! Shakespeare.” Among the many other tombstone enigmatologists was Ignatius Donnelly, who had so thoroughly misconstrued the biliteral early in his Baconian career that he turned it into his number cipher, but later learned enough so that he could mismanage it. Taking advantage of a “double-back-action quality” that Bacon, the inventor of the system, had neglected to mention, repeating some groups and omitting others, and anagramming, he cheated his way to a superb proof of Baconian authorship. And there have been others, equally valid.

  In 1899, the same year that Donnelly’s effort appeared, there was published the first report of a Baconian cipher message hidden in the way that Bacon himself had suggested—in printed books. This was The Bi-literal Cypher of Sir Francis Bacon Discovered in his Works and Deciphered by Mrs. Elizabeth Wells Gallup. A Michigan high school principal, then 51, Mrs. Gallup was an honest, gentle, religious woman who had studied at the Sorbonne and at the University of Marburg. Bacon interested her and Dr. Owen’s “word cipher” attracted her and she worked with him on it. She evidently accepted his results, for her own paralleled them, but rejected his method, for she began her own search for a message based on the biliteral cipher.

  Struck by the variations of type used to set the First Folio, she studied the printing with a magnifying glass to see if these differences represented Bacon’s use of the biliteral cipher. Since the differences are sharpest in the italic letters, she tried first to decipher the Prologue to Troilus and Cressida, which is almost entirely in italic in the First Folio. Slowly and painstakingly, she drew from that page and from others bits and pieces of a sensational life story of Bacon quite similar to the biography enigmalyzed by Owen, and soon found that it was continued, also in biliteral form, in the books of Marlowe, Jonson, Spenser, Burton, Peele, and Greene, all of whom served, as in Owen’s work, as masks for Bacon. The tale skipped about from place to place; a sentence broken off in one book resumed in another; and the substance repeated over and over again, as if Bacon were making sure that at least one message would be found.

  Mrs. Gallup found the nub of the story in the “Catalogue,” or table of contents, of the First Folio:

  F. Bacon.

  Queene Elizabeth is my true mother, and I am the lawfull heire to the throne. Finde the Cypher storie my bookes contain; it tells great secrets, every one of which, if imparted openly, would forfeit my life.

  Bacon’s father was Robert Dudley, the Earl of Leicester, Mrs. Gallup discovered, and she drew forth a hair-raising tale of how Elizabeth, “unwilling in th’ seventh month to proclaim herself a woman wedded and pregnante,” gave birth to the unwanted child whom she later entrusted to Nicholas Bacon for upbringing: “… she who bore me, even in the hour of my unwelcom’d coming, outraging every instinct of a naturall woman, in the pangs and perills of her travail cherisht one infernal purpose. ‘Kill, kill,’ cried this madden’d woman, ‘kill.’ ”

  The publication of the book stirred great interest, as do nearly all reports of new “proofs” of Baconian authorship of Shakespeare. A second edition appeared in 1900 and a third in 1902. Mrs. Gallup was berated; she replied mildly. The most sensible statement in the uproar was that the Baconian controversy had been shifted to new ground. In 1907 she sailed for England to look for manuscripts that her readings told her were in Canonbury Tower in London, where Bacon once lived, or in his country home at Gorhambury. But the
first was reconstructed, the second in ruins and, no more than Dr. Owen, who was digging at the same time, did she find any manuscripts.

  After her return to the United States a few years later, she was hired by George Fabyan to work at his Riverbank Laboratories in Geneva, Illinois, and “decipher” the manuscripts there, with the assistance of a staff and photographic equipment for enlarging the letters of the First Folio. The wealthy Fabyan hoped to be revered as a literary pioneer after the expected Baconian victory; he had financed Owen’s excavations and had been introduced to Mrs. Gallup by a mutual friend. In his campaign to “sell” her work to the academic world, he invited prominent scholars to Riverbank at his expense, fed, housed, and entertained them at his villa, treated them on their first day to a well-organized lantern-slide lecture on the biliteral cipher, urged them to observe the staff at work and to talk with Mrs. Gallup, and reiterated the need to keep an open mind. Mrs. Gallup remained at Riverbank until a few years before her death in 1934 at 87; during all those years she never produced another “decipherment.”

  The Laboratories’ staff members were told, however, that she was working on Bacon’s New Atlantis. Among her assistants was William F. Friedman. Though head of the Riverbank Department of Genetics, he helped make the photographic enlargements. Instead of clarifying the slight differences between the A and B forms of the letters, however, the enlargements obscured them, because many proved to be the result of damaged type, imperfections in the paper, poor presswork, or ink-spread around the actual printed letter.

  Nevertheless, the work went on. Elizebeth Smith, the future Mrs. Friedman, collated the A and B assignments of the other workers on the basis of a tally and attempted to read the secret message. When, as invariably happened, she failed to get more than a word or two, she brought the text to Mrs. Gallup, who produced extensive readings with little apparent effort. When Miss Smith would say, “But you must have changed some of the assignments,” Mrs. Gallup would point out that the group had missed the position of the dot over an i or some similar minutia. Miss Smith, who at first admired Mrs. Gallup’s acute facility in extracting intelligence from what she herself could see only as gibberish, found her admiration turning to “uneasy questioning, and then to agonizing doubt, and then to downright disbelief. I can state categorically that neither I nor any other one of the industrious research workers at Riverbank ever succeeded in extracting a single long sentence of a hidden message; nor did one of us so much as reproduce, independently, a single complete sentence which Mrs. Gallup had already deciphered and published.”