[Elgar's cipher table]

1970-71, previously unpublished 

Typescript by courtesy of Dr. Andrew Lamb


In his last published essay on “Elgar’s Enigmas” (M&L 1997), Eric Sams wrote:


His famous cipher message to Dora Penny […] is dated 1897; my own 1970 solution and comments […] remain uncontroverted. In 1971 I inferred its complete cipher table, which has remained unpublished as too technical. But it was, as I had said, based on Elgar's various names and sobriquets; and I noted with interest the new feature that its top line spelt ODIN.


It seems that, following his articles for The Musical Times in 1970 [“Elgar’s Cipher Letter to Dorabella”, “Variations on an original theme (Enigma)”, “Elgar’s Enigmas”] on Elgar’s use of cipher, Eric Sams had deduced the cipher table behind Elgar’s message to Dora Penny. He offered a further article to that same journal’s editor, Stanley Sadie, who “[took] the view, quite justifiably, that it would have on his circulation somewhat the effect of a tourniquet” (letter to Andrew Lamb of 17th May 1972). Eric Sams therefore sought to summarise the whole subject in an article for the American Saturday Reviewwhose music editor Irving Kolodin had in 1952 sought solutions to Elgar’s Enigma. Then, it seems, Eric Sams’s “nerve failed” (letter to Andrew Lamb of 17th May 1972).




    With his letter to Andrew Lamb of 17th May 1972 he enclosed a draft typescript of the unpublished article. I’m much obliged to Dr. Lamb for the courtesy of a copy of the unpublished material, which is of course of great value and worthy of posthumous publication.

    What follows is: a) a typewritten note (in brackets) by Eric Sams related to the projected introduction to  the new Elgar essay; b) the core of the essay, as enclosed in  the letter to Dr. Lamb of 17th May 1972; c) the cipher table [ ex. 3], which was enclosed in a previous letter to Dr. Lamb of 2nd February 1972 (“…the enclosed diagram […] represents what I should have seen earlier about Elgar’s cipher message, or rather non-message, to Dora Penny.”); d) a sample from a note- book of Elgar’s, enclosed in the letter of 17th May (“the cipher diagrams are, as conjectured, clearly based on the Ogham primer in the Book of Ballymoteso at least I was right about that” see Eric Sams’s lasMT Elgar article, “Elgar’s Enigmas”).


[I’m obliged to Dr. Lamb for checking the transcription and for providing the  explanatory introduction - EB]




(an introduction gives the background, ?quotes Saturday Review articles 1953, reproduces cipher text in facsimile).



     An 87-symbol cipher ought to be child’s play. This one isn’t; so there must be a trick. Elgar was obsessed by trick spelling; “excuse” was “xqqq”, while “score” became “ckor”, “skore”, “skorh”, “skowre”, “skourrghe”, “csquorr”, “skourghowore”, and “ssczowoughohr”. So perhaps his symbols stand for sounds as well as letters? Try it and see what happens.

    Ordinary methods yielded ex. 1. Each letter looked plausible in at least two places, and was then written in wherever its symbol occurred (leaving the rarest ones blank)




- B S C – A H C M – N E W L E – E E A B B E L O – O I N T

H E W A H C M – C S E E S A I W H E N E A R T W L O M M O N T

S T A R T S L R – I T S E A R W T I C B I T A C L O C


    Any cryptanalyst might get that far – and give up in despair. But if we think of this as a new language (Elgarese) the sounds begin to make sense. The first two blanks suggest the vowels in OBSCURE; one neutral, the other like EW in FEW. Next we expect R – and find AH. But this is literally another way of saying the same thing, and a very Elgarian way too (like Q-CUE). Suppose one joke is that AH or AR or letter R or vowel A (as in DARK) or word ARE are all interchangeable, like OW or OR or letter O or vowel O or word OH, or any other of the sounds in Elgar’s SCORES?

    With that idea in mind, look at the blank in the third line. It would follow the sounds L and AR; it occurs only once. How about X used as KS? Elgar himself did this, with “Thanx for the indeks”. The word LARKS was a favourite expression of his; it meant fun and games, such as playful puzzles. He used it in a letter about Dora Penny.

    If the symbols can mean vowel-sounds, the third blank in the top line might be long I, to take MY NEW. And the blank in the middle line could be long A, as in MAKES – if C equals K. To Elgar, it did, not only in “skore” but in “sarkazzum”, “koda” and “kompoliments”. So the last four symbols could read CLOCK or CLOAK. Next we recall Elgar’s whimsical way of writing W as a vowel, as in “crwwwd”, meaning “crude”. In his handwriting W looked like the Greek letter omega (which was formed from OO). He could not read Greek; but he knew its letters and used them for comic effect. The fourth blank in the top line suggests LETTER. The missing double T might be jokingly represented by one Greek letter, namely Pi (compare Elgar’s own way of writing double capital T).

    His own use of Greek letters included AI, pronounced like long I in English. So perhaps SAI says SIGH in cipher. E.E. was Elgar’s name to his friends and himself; and so on. We seem to be reading his letters in more ways than one. Already we can translate 77 out of 87 symbols from Elgarese into English, via his own known vocabulary. The result (beginning with STARTS, which seems reasonable) is shown in ex. 2. Now only the ten points numbered for reference remain obscure. They might all just be mistakes, and the message would still be almost readable. However, I think we can explain them all.



                                1          2                       3                   4



 5                 6         7                 8                                      9     10



    At 1, 4 and 8 the symbol which is E in eight other places makes no sense. Yet each has an obvious meaning – CH, RS, and YOU; or X, Z and V in Greek. Just another coincidence? If not, then one o two other symbols might mean more than one letter. If the M-symbol also means G, and in Greek (where double G is pronounced NG-G) then the answer is not LOMMON, point 9, but LONG GONE. Point 2 would read BUT for BIT if I were U – as when Elgar referred to his friend Griffith as Gruffyd. Similarly with point 6 – for OIN read OWN (which most Englishmen now, and certainly Elgar then, would pronounce O-UN). The fifth blank in the top line of ex 1 is the same as the third, ie the long vowel I. I OWN is of course still current (if old-fashioned) English for “I confess” or “I admit”. Finally the W-symbol may have been used just once for D, point 7, giving I OWN THE DARK MAKES E.E. SIGH WHEN YOU ARE TOO LONG GONE. Even with no special reason for the equations W/D, I/U and M/G, they seem plausible; and as we shall see there is a special reason, which applies with equal force to all three.

   The idea of Greek may help with point 5. Perhaps Elgar has written Greek letters (alpha, beta) below English ones in a cipher-table. Again, we shall see how plausible that is. For the moment there are only two points left. The second C in OBSCURC looks as if it means S. This too, as we know, is an Elgarian usage – “score” is “ckor”. But it may be just a mistake. Last comes the final T. This too may be a null. Or it may mean what it says, namely


            “I own the dark makes E.E. sigh

            When you are too long gont”.


If so, there would have to be a good reason – for example a joking reply to something said or sung in the same metre. Perhaps not


            “Be sure Sears Roebuck can supply

            Whatever you may want”


   if only because that is not (yet) part of a song or hymn or poem. We need something on the same lines but more memorable. The possibilities are very few, but almost nothing else rhymes. Then suppose the echo was not rhyme but assonance, “sigh/gont” chiming with (say) “got/sign”? As perhaps in


            “Should old acquaintance be forgot

            And days of old lang syne?”


- an amusing and memorable way for friends recently parted to say “be sure to come and see us again soon”?

   It was Irving Kolodin who in these pages 17 years ago first linked the cipher with the Enigma Variations – to which the usual answer has always been Auld lang syne as the unheard melody. Elgar said to Dora Penny “I thought you of all people would guess it”. And later on she did. Many musicians also accept that Elgar had Friendship in mind, because he knew and loved Sir Thomas Browne’s Religio Medici, which mentions Enigmas in connection with Friendship and Music. It also contains the train of thought “cloak, chaos, obscure, letters, alphabet, Greek, dark, sigh, absence” as in the present solution, in a dozen pages preceding the Enigma/Friendship passage.

   The message remains cryptic. But so was Elgar, even en clair. For example a postcard to another friend ran “What is the boon of far Peru?”. Perhaps the friend understood it; I own it leaves me in the dark. The Penny postcard seems much less obscure. Elgar was keenly interested in cryptography; he had in fact read a magazine article about an indecipherable cipher, with a challenge to readers to solve it. Naturally he wished to tease and baffle and intrigue a young woman admirer with a riddle of his own. When they next met he could show her how, despite appearances, his symbols could make some sort of sense. To entertain them both, the message would touch on topics they had discussed together. Auld lang syne is readily explicable. But why bring in Religio Medici? That masterpiece is famed for its theme of toleration of Catholics by Protestants, eg of the Elgars, by the Pennys. Then why choose a cipher-system so akin to the ancient Oghams, thought to be the secret script of Druids? Elgar had just begun his cantata Caractatus, which is about Druids; and he told Miss Penny about its plot. The whole idea of mysterious, complex and wistful amusement is entirely Elgarian. “Smiling with a sigh” was his own phrase for many of the most memorable and moving moments in his own music. A phonetic cipher for Dorabella would also be a personal expression of feeling, composed of sound-symbols. She never asked what it meant; he never explained; a triumph of British reserve. But the whole world can hear the meaning of the intermezzo Dorabella in the Enigma Variations.

     If Elgar’s cipher was self-expressive, we would expect it to be built up from his own names. At the time, these were not only E.E. but “Edu” (for Eduard) as well as Elgar. Dora Penny was told about “Edu” and sworn to secrecy. In the suggested decipherment, ex. 1, D and U are represented by the simplest (one-arc) symbols. So are L and G. This suggests a beginning. We know that Elgar loved rectilinear patterns, and that he was passionately devoted to crosswords. Write down D and U, L and G, with their symbols, in squares; complete the symbol-pattern, symmetrically; add the two and three arc symbols; write in the other letters; and the result spells ELGAR (see ex. 3 below)


Ex 3


    Furthermore, it looks (by coincidence?) exactly as if he had decided to add his middle name William to the centre of the pattern thus formed. Add W to the D-square; then I goes in the U-square. The L is already present, in ELGAR; so is the A. That leaves M, which presumably follows the A, in the same square. But that won’t do, because it allots letter M to the cipher symbol which looks like M. So that letter goes into the G-square instead. That process gives the equations W/D,I/U, M/G, already suggested for completely different reasons. As a bonus, it brings together the letters I,G,M,A – which may have suggested the end of a word and the beginning of an idea.

   Next, other vowels and letters would be added; perhaps at random, perhaps on some plan. However it was done the 24 possible symbols of Elgar’s system are not enough to cover all the possible sounds. Even with few letters sharing the same symbol there would still be some left homeless at the end of the process. The stragglers would be likely to be letters at the end of the alphabet; say V X and Z. They would be put in a square with some lower room to let, eg where the English letter had the same form as the Greek one, which would not then need to be written in. E is the obvious example. And when V, X and Z were written in below, they would be treated as jocular Greek; Nu, Chi, and Zeta. Q, one might say, E.D.

    All this adds up to ex 3 (which omits the Greek letters for the sake of simplicity; the blank-square symbols were no used in the cipher message, but may have meant eg theta or ypsilon). On any reckoning, Elgar went to a great deal of trouble to devise and compose a message which was never understood or explained. But what if that cipher-table served another purpose?

   Dr. Percy Young’s standard biography tells us that Elgar used a music cipher; the names of people he disliked were thus consigned to the Demon’s Chorus in The Dream of Gerontius. An Oxford professor of music, Sir Jack Westrup, has suggested that Elgar used cipher in the Enigma Variations.

   Perhaps interested readers would like to consider on what lines (or spaces) ex 3 might make a music cipher?




Appendix [d) Elgar’s cipher diagrams, see introduction]