The lot contained a rather distressed (“binding broken, pages loose, wear and some damp-staining“) first edition copy of Dora Penny’s (1937) “Edward Elgar: Memories of a Variation” apparently from Dora Penny’s own library (“D.M.P. 1937”, though by then the final ‘P’ then stood for ‘Powell’, her married surname), along with various photographs of Elgar and his coterie all “captioned by Dora in blue ink”.
Oh, And A Micro-Cryptogram, Too
Also included was a small fragment written by Elgar, noting that he “wrote to [musical instrument dealer] Hill offering to purchase Gagliano [violin]”, on a ~5.5cm x ~7cm piece of paper that had “traces of mounting to verso”. This is what it looks like (image taken from Sothebys’ site):
The expected price was £600-£800, but the actual hammer price was £1750.
What seems to have raised the level of buyer interest was the presence of a single three-letter cryptogram repeated six times (though with the first time crossed out), with three of the instances preceded by a ‘£’ sign. Given that this is arguably the shortest cryptogram I’ve yet posted here, I thought it was well worth dubbing it a “micro-cryptogram”.
But… what could this be? And, most importantly, might we be able to crack it?
Ah, It’s Also In His Diary
Fortunately, Elgar historians and biographers got there first (after a fashion). For if you turn to p.158 of Jerrold Northrop Moore’s (1984) “Edward Elgar: A Creative Life”, you will discover that the same micro-cryptogram appears in his diary: “[56] Elgar’s diary entry [for buying the eighteenth-century Gagliano violin] follows the £-sign with three squiggled marks“. (Figuring that out took me all of thirty seconds, much of which was spent trying to unwedge the copy of “A Creative Life” from the back of the bookshelf it was sitting on.)
So it would seem that what was on sale at Sothebys contained something like pen-trials or rehearsals on a scrap of paper for the same three-glyph cryptogram he added to his diary. Moreover, if we could discover by other means what price the Elgars paid for the Gagliano violin in 1891, then it seems we would be able to solve the cryptogram.
However, having now spent significantly more than thirty seconds trying to determine this (with no success), all I can do is throw it open to you all. How many pounds did the Elgars pay for their Niccolò Gagliano violin in 1891? Find that out and you presumably will have solved possibly the shortest genuine historical cryptogram ever. 🙂
Absence of Provenance Is Not Evidence Of Providence
All the same, I have to say it seems odd to me that the (normally very thorough) Sothebys people failed to pick up on this connection with Elgar’s diary. The catalogue entry for the preceding Lot 91 (Lot 91) was much more their normal style, with a rock-solid provenance (“From the collection of Edward Speyer, to whom Elgar gave these manuscripts“): unsurprisingly, that went for a handsom £72,500 (close to the middle of their estimated range).
So… what was the difference with Lot 92? What was its provenance? I can’t help but wondering whether the “binding broken, pages loose, wear and some damp-staining” condition of Dora Penny’s own copy of her book might be trying to tell us, along with all the photographs hand-annotated by her.
You see, there is one person who could very easily have been the source for this: Dora Penny herself (albeit indirectly).
When I tried to trace the history of the Dorabella Cipher itself a few years ago, I found that it had been part of a a sizeable set of Dora Penny’s Elgar-related papers, that had been presented to the Royal College of Music Library “by Mr and Mrs Claud Powell [in] 1986”. However, as an RCM archivist I talked with told me, several boxes of this Elgar material were somehow lost (possibly in Leeds?) while being transported to London, and that was the last that was seen of them.
What, then, are the chances that one or more of these cartons ended up in someone’s slightly damp garage for the next thirty years, and that this rather poor condition copy of the book is the first sight anyone has seen of these since 1986? Perhaps the seller didn’t want to be identified for that reason, in which case it could easily be why the lot was clearly marked as “sold not subject to return”, and without a hint of a flicker of a provenance.
Even So, Does It All Add Up?
Even if the above will turn out to be the story behind this item, I have to say that the picture as a whole still doesn’t quite ring true to me.
Put simply, I would be a little surprised if Dora Penny had had reason to mount this poor scraggly piece of writing on her wall. After all, she had the Dorabella Cipher itself: this micro-cryptogram is surely very much its poor relation, as well as being unprepossessingly tiny.
Might it be that the person who owned this had had it mounted on his or her wall in their study, sitting next to the Dorabella Cipher itself? What an incredible story that would be! Well… something to think about, anyway. 🙂
On p.114 of Jerrold Northrop Moore’s weighty “Edward Elgar: A Creative Life”, the author notes that Elgar’s enciphered “Liszt fragment” had been decoded (in 1977, according to Anthony Thorley whose decryption it was) to read:
Gets you to joy, and hysterious
Well… it’s certainly a claim, even if ‘hysterious’ is a made-up word found nowhere else. And one of the (cryptologically, at least) interesting aspects that link this Liszt fragment and Elgar’s Dorabella Cipher is that while both of them seem unlikely to have employed complicated cipher systems, for all of that both also seem improbably hard nuts to crack. You’d certainly need a sweet nutcracker to achieve it. [*]
I’ve discussed Elgar’s Liszt fragment before, written in the left margin of an 1885/1886 Crystal Palace Saturday Concert Programme:
The cipher on its own looks like this (sorry, but I don’t have a better scan):-
It’s not a great scan, certainly: but given that though the dash looks as though it is meant to sit at the end, and there are several half-space-sized gaps, it looks as though we might be able to transliterate this as:
ABC DECFGB HID CBJKDK
What should be immediately apparent is that there is no obvious way to convert this 3 + 6 + 3 + 6 = 18 letter cryptogram into Thorley’s 25-letter “Gets you to joy, and hysterious”, without a singularly large floor space for mental acrobatics to bounce around on. (If that’s what you want to do, feel free to go ahead.)
And yet, what we undeniably have with the Liszt fragment that we don’t seem to have with the (much later) Dorabella Cipher is context, specifically a musical context. And here I can’t help but notice not only that the Liszt ciphertext seems to have been written in sets of three or sets of six, but also that the music it sits besides also has a very strong emphasis on triplets, groups of three notes.
Moreover, the 18-letter group is written immediately beside an 18-note line of music, “No. 6 Allegretto Pastorale”. Might the first be enciphering the other in some way?
I can’t see any obvious cryptographic connection myself here, but I was somewhat surprised to find that nobody had apparently suggested this at least as a reasonable possibility for the Liszt fragment, far more so than for the Dorabella Cipher. (Plenty of people [e.g. Javier Atance, etc etc] have suggested that the Dorabella Cipher might be enciphering music, but that’s another story entirely).
As regular Cipher Mysteries readers will know, Edward Elgar’s most famous ciphertext is the Dorabella Cipher, a tiny cryptographic walnut that continues (more than a century after it was produced) to defy all cryptologic jackhammers sent to crack it.
A confusing part of the mystery is that in the 1920s, Elgar (re-)used the same pigpen-style “rotating-3” alphabet as a simple substitution cipher in two pages of his notebooks (the “MARCO ELGAR” / “A VERY OLD CYPHER” / “DO YOU GO TO LONDON TOMORROW?” page reproduced here. These messages are easily deciphered (helpfully, the key is reproduced on the page itself), so the two ciphertexts appear to have nothing to do with each other apart from their shared alphabet.
Furthermore, what I didn’t realise until last year (but never quite got round to mentioning here) is that Elgar re-used these same cipher shapes in two other places. Firstly, the “Liszt fragment” in the margin of an 1885/1886 Crystal Palace Saturday Concert Programme, which is basically a string of the same rotating-3 alphabet:
Tony Gaffney thinks that this is trivially solvable by using one of the “clock-face” diagrams from the 1920s notebook, but I think this mainly depends on whether you think the kind of language Tony believes the main Dorabella cipher was enciphering had the same kind of allusive & abbreviated private codewords as this (short) message.
But there is also one other place we know of in Elgar’s notes where these distinctive shapes appear, and it is in a decidedly cryptographic context. In 1896, the Pall Mall Magazine published an article with an “uncrackable” code challenge for readers, one based on the Russian Nihilist cipher. However, Elgar took such great delight in cracking it that he later had it painted on the floor: he also wrote his explanation on nine cards (“the Courage card set”).
What is odd is that ten of the rotating-3 alphabet letters appear on the first card of this set. The description given of these is of the set of eight rotating triple cup shapes, followed by an upward-facing double cup and finishing with an upward facing single cup. I haven’t seen an image of this in context, but Christian Schridde reconstructed one for one of his very readable “NumberWorld” blog posts on the Dorabella Cipher:-
In Christian’s third Dorabella post, he muses on the timeline aspects: that the Liszt fragment is more than a decade earlier than Dorabella, while the “MARCO ELGAR” page is more than two decades later. By contrast, the ink was only just dry on the Courage card set when the Dorabella cipher itself was written.
Schridde therefore wonders whether the Dorabella cipher might use some element of the Nihilist cipher (which, itself, is a kind of bodged-together mix of Polybius square and Vigenère cipher): certainly, this cipher must have been almost as fresh in Elgar’s mind as the paint on his wooden floor.
On the one hand, the fact that there are ten rotating-3 symbols on the card seems to vaguely imply that Elgar was trying to see a way of mapping a set of digits [0-9] onto his rotating-3 alphabet: which, given that the Nihilist cipher system enciphers individual plaintext letters as 2- or 3-digit numbers, does seem oddly coincidental.
And yet there is no obvious sign of digit pairing or grouping, which you’d perhaps expect if you were seeing something Polybius-style or perhaps groups of digits: the ABAB-style length pattern at the start of the Dorabella (2323121312…) quickly disappear, replaced by quite different structures (e.g. 22222 and 111111 on the second line).
At one point, I also wondered whether the sequence on the first Courage card might in fact be a kind of length-10 sequence to offset the Dorabella symbols, that I previously called a “rotating pigpen”, e.g.
Well, given that this is a reasonable possibility, I tried out a few variants of it in Excel… but nothing plausible-looking jumped out at me, which was basically what happened when I tried a similar rotating pigpen in C code hack before.
I really don’t know what’s going on with all Elgar’s cipher bits, particularly the Liszt fragment. Of his three undecrypted mini-ciphertexts, the Dorabella is in some ways the least odd – it at least looks like a proper cryptogram with proper-looking statistics, something to get your cryptologic teeth into. The mystery deepens! 🙂
Researching cipher mysteries is almost always ponderous and frustrating: it will doubtless take all of 2021 for the work I put in to the WW2 pigeon cipher and the Bernardin Nageon de l’Estang letters during 2020 to bear fruit.
Yet what Dave Oranchak’s recent epic crack of the Z340 tells us is that if we can identify any unsolved cipher’s single most telling feature and then doggedly pursue that to its logical extreme, we stand a chance of toppling that particular Colossus (in the Classical sense of the word).
For the Z340, that telling feature was that taking every 19th character from the cipher yielded statistically anomalous results. And we now know that this was because a central part of the Z340’s cipher system was a “knight’s move” transposition step (i.e. two steps along and one step down).
So my two main challenges this year are (a) to stay laser-focused on the telling features presented by different cipher mysteries, and (b) to find new ways to pursue these telling features all the way to their logical conclusion.
Specifically…
1. Voynich Manuscript
In my mind, there’s something really wrong with Voynichese. Specifically, even though the differences between Currier A Voynichese and Currier B Voynichese run really deep, nobody seems to be talking about this.
Let’s compare a couple of lines from f1v (Herbal A)…
Voynich linguists typically try to downplay the differences between the two, but… really? What similarities there are tend to be either at the (low) level of (verbose cipher-like) groups (e.g. aiin, ar, al, etc) or purely positional (line-initial “p-“, word-final “-y”, line-final “-m”, etc). Even really common features like qo- are used very differently in A and B.
So, even though A & B seem to share a common framework, beneath that framework there seems to be more dividing them than joining them. And I think I’ve been guilty in not separating out A and B from the framework they share more clearly: we’ve probably all been guilty of that to some degree.
My first challenge for 2021 is therefore to look at Currier A and Currier B with fresh eyes. What do the two share, and how do they differ? Though I can’t yet properly express this, it feels as though we’ve been building our theories about Voynichese on sand, and the answers may be much simpler than we’re allowing ourselves to see.
2. Voynich Manuscript (Again)
One thing that popped up during 2020 was Antonio Averlino’s herbal. If you recall, having published The Curse of the Voynich in 2006, I was surprised to find out two years later that Antonio Averlino had his own herbal.
Thorndike’s “Science & Thought” quoted Giovanni Michele Alberto in MS Ashburnham 198, fol.78r: “Sed et Antonius Averlinus Philaretus lingua vernacula scripsit eleganter.” So it would seem that Filarete had written on plants “elegantly in the vernacular tongue”.
It’s been a while since I last picked up the Filarete trail (which I’d worked pretty much to death back in 2006), so what I’d like to do this year is to go a-hunting for Filarete’s vernacular (i.e. Tuscan Italian) book on plants. This would involve drawing up a list of Tuscan herbal mss dating to around 1450-1460 (which surely can’t contain more than 40 or 50 possibilities), then reducing it down, and finally closely examining that which remaineth.
It’s a plan, at least. 🙂
However, because MS Ashburnham 198 isn’t visible online at the Biblioteca Medicea Laurenziana (please correct me if I’m wrong!), I still haven’t seen Alberto’s quotation in its full context. This annoys me because I don’t know whether Alberto was referring to an illustrated herbal or a purely textual herbal. As a consequence I don’t yet even know what kind of book to go looking for here. But hopefully I will soon…
(Note that Thorndike’s Chapter XII was “Revised and enlarged” [p.195] from the version that appeared in The Romanic Review, Vol. XVII, No.3, July-September 1926, so the latter is unlikely to have any extra information.)
3. Dorabella Cipher
I’ve recently been corrected by Cipher Mysteries commenter John Rehling, who took me to task for numerically over-egging Keith Massey’s theory about the Dorabella Cipher. Thanks to the magic of the binomial expansion, the chances of 13 cipher shapes out of 87 being immediately followed by their flipped version is in fact a mere 1 in 20815. (!)
On balance, I’m now wondering whether this is no more than a sign that the set of mirrored pairs at the end of the second line is just filler / padding, i.e. that Massey’s conclusion is still correct, but only in a very local and limited way.
However, if that is true, then the long stretches of the Dorabella Cipher that contain neither vertical E-shapes nor downward slanting E-shapes then become markedly more problematic. So I continue to think that there’s something deeply artificial about this cryptogram that messes up all our statistical analyses.
So I therefore need to have something of a Dorabella rethink in 2021. :-/
I just wish that the person who secretly owns the Dorabella Cipher (and who I can’t help but suspect sold a small piece of their Elgariana at Sotheby’s in 2016) would come forward, perhaps via a trusted third party. I believe that shining a simple UV light (even a bicycle marking light) on it might reveal Elgar’s real solution – and how good would that be?
There are plenty of things about Edward Elgar’s Dorabella Cipher that rarely appear in the countless gosh-wow sites that feature it on the web. And arguably one of the biggest of these is its timeline.
1886: The Liszt Fragment
The earliest instance we know of where Elgar used the ‘Dorabella’ shapes to write something down was when jotting something in the left margin of a programme for a Liszt concert at the Crystal Palace (10th April 1886). The best quality image of this fragment appears on p.134 in Craig Bauer’s magisterial “Unsolved!”, which I reproduce here:
This contains a fair few repeated shapes, which would be good grist for the cryptanalytic mill were the fragment not so darned short:
Though Anthony Thorley claimed to have ‘decrypted’ this fragment in (or before) 1977 as “GETS YOU TO JOY, AND HYSTERIOUS”, this looks just plain wrong to Bauer (and to me). This is not only because none of the repeated letter usages line up, but also because it’s basically the wrong length (Thorley’s phrase is 25 letters long, while the fragment is made up of eighteen shapes plus a terminal dash).
If this Liszt fragment is a cipher, I’m sure Big Data people wouldn’t have to try hugely hard to build up a list of all 18-letter English letter sequences with the same aBCDeCfgBhiDCBijkl pattern. Perhaps looking for matches for the 13-letter stretch from BCD to DCB might be a productive exercise?
At the same time, it is tempting to wonder whether Elgar was using these shapes as some kind of idiosyncratic musical notation. However, even though the eighteen-glyph-plus-hyphen fragment appears in the margin beside an eighteen-note arpeggiated melody (Liszt’s “allegretto pastorale” motif, which appears as “an independent episode” according to the programme notes), it has none of its musical symmetry.
So: even though the Liszt fragments looks as though it really ought to be a simple cipher (and, moreover, a simple cipher that Elgar had without any real doubt used many times before), none of the claimed decryptions put forward for it make much sense. It’s possible Elgar was using his own brand of self-pleasing nonsense verbiage but… this is as far as we can get.
1897: The Dorabella Cipher
According to her 1937 book “Edward Elgar: Memories of a Variation”, a young Dora Penny’s first met Edward Elgar on 6th December 1895. Elgar’s wife was an old friend of Dora’s stepmother, and so the couple had come to visit. Elgar and Dora talked, but not about music: rather, he wanted to know about Wolverhampton Wanderers (the club was close to the Penny’s house).
All the same, he did sit down at the piano in the drawing room before luncheon, where Dora turned over the pages for him. This proved to be a challenge, as “[w]hen it came to playing from his own manuscripts you often saw nothing but a few pencilled notes and a mark or two, when he was playing something tremendous – full orchestra and chorus perhaps“, though over time she did become “rather clever at it“.
Elgar first got to see a football match with Dora on 17th October 1896. He subsequently “was much taken with the names of some of the players – particularly Malpas. […] I have known him say when we met: ‘There you are. How’s Malpas?’ – a question I was not always able to answer.“
Her book reproduces a letter she received from Elgar with a distinctive red ‘E’ seal: all of which I think gives as close a representation of the likely content of the Dorabella Cipher as can reasonably be hoped for:
Forli Malvern March 4 [1897] Dear Miss Penny Here is some locomotive learning; so much nicer than mouldy music. Alice tells me you are warbling wigorously in Worcestor wunce a week (alliteration archaically Norse). I am very glad, but on second thoughts, as I have never heard you sing I am not sure: but perhaps some day if you are not rushing away I might arrange to show you over the Cathedral organ, K. John’s tomb and the Dane’s skin: (the Dane is dead). By the way I have taken to ‘die-sinking’ as a recreation: here on the back of this is my parcel-post seal: I have to register all my MSS & they will not give a receipt unless they are sealed: so I put this on that my works may be Esily distinguished. Kindest regards to everybody Believe me sincerely yours EDWARD ELGAR
The Dorabella Cipher is dated July 14 [18]97 and, if you haven’t already seen it a thousand times or more, looks like this:
This was “the third letter [Dora Penny] had from him, if indeed it is one“: so the March 4 letter was only one of two Elgar had previously written to Dora.
As to what the Dorabella Cipher says: I’ve previously (in 2013) speculated whether the first two words might be, just as with the March 4 letter, FORLI MALVERN. And the obvious suggestion that Elgar might have also included the phrase “How’s Malpas?” is entirely possible, though untested.
All the same, I’d point out that the general character of the glyph shapes seems to change on the third line. That is, the shapes lose their variety, and become visually monotonous, bland, repetitive, even dull. It’s as though Elgar kind of lost momentum, and stopped wanting to sustain the joke. Much as I have suggested with the famous unsolved Zodiac Z340 cipher (where the top half and bottom half have different statistical profiles / patterns), I do wonder whether we might be seeing two different things grafted together here, i.e. that the third line is quite different in nature from the first two. Just a thought.
Note: it was September 1898 when Edward Elgar first called Dora Penny “Dorabella” (as a quotation from Mozart’s Così fan tutte): so the one word we should not expect to see in the Dorabella Cipher is ‘Dorabella’.
1924 or later: the Marco Elgar Cipher
Yet another place where the rotating e/ee/eee letter-shapes appear in Elgar’s papers is where he uses it as a simple pigpen-style cipher:
This we can date as having been written not before 1924, because the plaintext refers to “MARCO ELGAR”, the name of Elgar’s beloved spaniel, and who was born on 27th May 1924 (a picture of his grave is here).
While the most obvious interesting thing about this it doesn’t work for the Dorabella Cipher, there is something about this sheet that gives me the impression that what Elgar is trying to do is to reconstruct his cipher system. It is hardly a coincidence, I would say (apologies to Thomas Ernst) that another phrase enciphered on this same page is “A VERY OLD CYPHER”.
Given the roughly thirty years’ difference between the Dorabella Cipher and the Marco Elgar cipher (and the absence of any other similar letter-shapes in Elgar’s generally quite well-preserved writings), perhaps it was something he amused himself with as a young man, but which he had by the age of about 70 (he was born in 1857) just plain forgotten.
Perhaps the circular shape on this page is some kind of E-based mnemonic (i.e. that the letters of the alphabet were arranged around), but which had slipped his mind. Certainly, you can see the letter E concealed in it without much difficulty, so perhaps that was part of the game?
Undated: The Cryptogram Card
Our final Elgarian cipher shapes first appeared in Craig Bauer’s “Unsolved!”. These are on a card marked “Cryptogram” (hence “The Cryptogram Card”), but are undated:
Though the writing is tiny, there are two main runs of eee-shapes: in the one just above the word “Cryptogram”, the triple curve shape rotates around, as if (as Craig points out) it is doing a gymnastic forward roll. In the run just below the word “Cryptogram, the three sizes of right facing ‘e’ appear in descending order, followed by the next rotation round. There are also a couple of cipher letters at the top.
What we see here are more like pen trials than cryptograms: so in almost all senses there’s really nothing of importance here.
Recent Dorabella Theories
Plenty of clever people – not just Eric Sams and Tony Gaffney – have already put forward their thoughts about (and their attempted decryptions of) the Dorabella Cipher. Needless to say, not more than one of them can be right at the same time. 🙂
But the list of attempts to explain it keeps getting longer. When Klaus Schmeh blogged about the Dorabella in 2018, one of the commenters (Thomas Ernst) put forward – at some length – his notion that Dora Penny might herself have faked all Elgar’s ciphers. This is an interesting suggestion: Dora certainly had full access to Elgar’s archives for decades, so clearly had opportunity – and I can see what he’s getting at when he draws a parallel between Dora Penny’s book and Bettina von Arnim’s (1835) “Goethes Briefwechsel mit einem Kinde”, which contains numerous stories about Goethe falsely adjusted to bring Bettina herself into the foreground (think of it as a kind of literary Forrest Gump).
But I think Ernst is being far too literal when he draws negative conclusions from the way that the post-1924 Marco Elgar page “A VERY OLD CYPHER” alphabet does not work for the Dorabella Cipher. His reasoning (unless I’ve misunderstood it) is that because the two are inconsistent, at least one is not genuine. And he then goes on to argue that if one is not genuine, there’s no reason to think that they are both not genuine.
I would agree that the two are indeed inconsistent. However, the rather different inferences I draw from the Marco Elgar page are that (a) in it, Elgar gives the impression that he was trying to reconstruct a cipher system he had used as a much younger man; and (b) there was some kind of underlying symmetry to the letter-to-glyph assignments in that cipher that he simply could not remember.
So, although there are good reasons we should all be aware of the inconsistency between accounts, Ernst’s move to a full-fat hoaxed-by-Dora theory seems somewhat pessimistic and extravagant to me.
As an aside, I think it would be a good exercise to analyze the Liszt Fragment and the Dorabella Cipher to see if they are consistent or inconsistent with each other (e.g. by comparing letter contact tables etc).
Another commenter (ShadowWolf) used the same Klaus post to put forward his/her own Dorabella decrypt (which, perhaps almost inevitably, involves a cipher-style first pass and a this-is-what-Elgar-really-meant-by-that interpretational second pass, Eric Sams-style):
Plaintext: PBS AFT DALYRENCE MEET B BECO YOUR IDEDTD ALWASE E STUNDER E THINC OLL OR IS IT HIS CH GUISE THNIC ABU IT ACOA
Message: Problems after dalliance meet is because your identity always a stutter I think all or is it his charming guise? Think about it acolyte.
In a similar vein, Cipher Mysteries readers may possibly recall that I posted about Allan Gillespie’s Dorabella Cipher theory back in 2013: his (somewhat hybridized) theory was that the plaintext began “ForlE Malvern Link”, the encipherment used a Vigenère cipher system, but (not entirely unlike Thomas Ernst’s theory) it had been “concocted by someone other than Elgar (possibly in the run-up to WWII when GC&CS were recruiting; possibly with Dora Powell’s connivance, more likely not)“.
Another Dorabella solver is Mark Pitt, a Cleveland police officer “with an MA in crime patterns” who has already had the oxygen of publicity in the Times, Daily Mail, Daily Telegraph and Guardian (to name but four). Pitt has also claimed to have decrypted the Liszt fragment: his solution for both seems to be based around Schooling’s cipher that Elgar famously cracked, with the key “PRUDENTIA”. I suspect Pitt has a (not very active) Twitter account, but that’s just my hunch. A (paywalled) Telegraph article on him from early 2019 is here.
The Two Massey Observations
Finally, a very different take on the Dorabella Cipher has been put forward by Keith Massey in an 11-minute YouTube video from 2017 (but which I only stumbled upon recently), based on two very specific observations.
His first observation is that the Dorabella Cipher contains long sequences of glyphs where no two adjacent glyphs have the same number of loops. Specifically, the first line has a sequence with 12 loop-number-differing glyphs in a row; the second line has a sequence with 9 loop-number-differing glyphs in a row; while the third line has a sequence with 8 loop-number-differing glyphs in a row; all three sequences are near the start of their respective lines. Massey’s control experiments (two of them, which one might reasonably argue is a little bit lightweight) each yielded a single maximum of only 5 or 6 loop-number-differing glyphs in a row. (A more Oranchek-esque researcher would surely have done the experiment by anagramming the Dorabella a billion times over, but I suspect the results would have been not wildly dissimilar.)
Massey’s second observation is that the Dorabella Cipher contains way too many pairs of opposed symbols (i.e. where a glyph is immediately followed by a glyph with its same basic shape but where that second shape is rotated by 180 degrees). Massey calculates that this should on average ~5.3 times for 87 characters, but it instead occurs 12 times.
If we assume all 24 shapes are equally likely to occur (which isn’t true) across the Dorabella’s 87 characters, the probability of 13 exact opposites occurring is ((1/24)^13)*((23/24)^(87-1-13)), which Google tells me is 5.1046414e-20 (i.e. about one in twenty billion billion). Again, a more realistic (i.e. Oranchakian) way of doing this would be to anagram the ciphertext billions of times and see how often twelve exact opposites occur (i.e. using the actual distribution rather than an ideal [perfectly flat] distribution). My prediction here is that the probability there would still be no higher than one in a billion billion, so I believe this too is likely to be a statistically significant result.
Massey thinks the final nail in the Dorabella’s cryptological coffin is that these two patterns don’t overlap: he believes that
Massey’s overall conclusion is that that Elgar created the Dorabella Cipher as nonsense text to resemble a ciphertext as a joke on Dora Penny, but that this nonsense text eventually escaped to become a joke on all of us.
Even if you disagree with the full strength of his conclusion, I suspect these two Massey observations will prove difficult for anybody proposing a simple MASC as a solution (albeit typically with an interpretational second phase) to satisfactorily account for what we see in the Dorabella Cipher.
Thoughts and Conclusions
I have to say that I’m very largely with Keith Massey here, insofar as he is pointing out statistical features of the Dorabella Cipher that are highly improbable. It is almost impossible not to see that these sit awkwardly with the traditionalist (one might call it ‘Samsian‘) reading of the cryptogram’s system as a pigpen-style simple substitution cipher applied to an idiosyncratic Elgarian nonsense-wonsense text. It would be good if Massey’s observations were to be confirmed in a more statistically robust manner, but I would be surprised if the actual results proved to be vastly different.
My own suspicion (just as in 2013) remains that the Dorabella Cipher may turn out to be a stegotext visually concealing a guessable personal message (e.g. “FORLI MALVERN”) rather than a cryptotext mathematically concealing a plaintext. And I believe this is far from inconsistent with Massey’s observations, though only for the left hand half of the three lines.
But even so, I’m really not at all convinced that his observations hold true for the Liszt Fragment, which I believe was written in the same “VERY OLD CYPHER” that Elgar was trying to reconstruct in the Marco Elgar page.
So there is perhaps still work to be done on a genuine Elgar cipher here, even if Massey has indeed managed to nail down the Dorabella MASC coffin (and all credit to him if he has!).
As promised a long while back (i.e. before I got caught up in pirate history minutiae, etc), I had some interesting emails from Cheltenham music teacher Allan Gillespie, describing his claimed decryption of Elgar’s well-known Dorabella Cipher.
Allan’s starting point seems to have been my hunch that the Dorabella’s first two words were likely to be “Forli, Malvern”, a modest little seed which he then grew out into his own complete decryption.
Specifically, he claims that it’s a vaguely Vigenère-like polyalphabetic cipher, with the key sequence AIUEGSOLXMKWCQZTDPBNYHFR rotated right by five places every eight plaintext characters, i.e.
AIUEGSOLXMKWCQZTDPBNYHFR - for characters #1 to #8
NYHFRAIUEGSOLXMKWCQZTDPB - for characters #9 to #16
ZTDPBNYHFRAIUEGSOLXMKWCQ - for characters #17 to #24
MKWCQZTDPBNYHFRAIUEGSOLX - for characters #25 to #32 GSOLXMKWCQZTDPBNYHFRAIUE - for characters #33 to #39 (etc)
Furthermore, Allan claims (I think) that the output from this gets mapped onto Elgar’s rotated-3 alphabet via this second table (which he presented in a transposed form to make it look as though the keyword was “HAUNTED” [+Y], but it’s actually no more than a monoalphabetic substitution alphabet):-
... N. NE E. SE S. SW W. NW
u.. A. N. E. Y. T. H. D. U.
uu. G. F. ?. R. M. I. ?. Z.
uuu ?. ?. L. B. S. ?. O. C.(Unplaced letters: K P Q W X)
Undo these two stages (he says) and you get a plaintext of:-
ForlE, Malvern Link
A. and Dai’s qko [=quick opinion?]
Met St Stephen ‘eighty six.
Wed at Brompton Oratory but owed takC Mogul ob’d.
He further believes the Dorabella cipher was “concocted by someone other than Elgar (possibly in the run-up to WWII when GC&CS were recruiting; possibly with Dora Powell’s connivance, more likely not)“.
Having said all that, I should add that I’m not entirely sure how serious Allan is about all this; and, moreover, the likelihood that Elgar would have used a messed-up Vigenère in combination with a second substitution stage seems to me to be as close to zero as makes no odds. But all the same, I’ve tried to reproduce Allan’s claim here as clearly as I can, just in case someone else wants to try to reproduce his results.
As you probably already guessed, I’m almost completely sure (as I indeed wrote to Allan at the time) that this “sits in the esteemed and excellent company of those such as Eric Sims and Tony Gaffney who have tried to solve the Dorabella’s cryptographic mystery rather at the expense of its historical mystery“. That is, neither the details (in the allegedly derived cleartext) nor the methodology (that Allan believes to have been used to encrypt the message) cast any light on Elgar, Dora Penny, their relationship, or any reason that such a devilish complex cipher system and linguistically idiosyncratic message would have been appropriate or even useful.
Allan response was that by replying in this way, I was (entirely unsurprisingly) acting in the same way that other cipher mystery establishment figures do, by working hard to “resist any attempt by an outsider to knock down [the establishment’s] battlements”.
Gosh darn it, but doesn’t it just turn out he’s got me bang to rights there? I indeed spend three nights a week chairing a secret cryptographic cabal downstairs at the Athenaeum Club library (or, failing that, Westminster School’s dining hall next to the Abbey) that decides how to misdirect plucky independent codebreakers away from the heretical and uneasy truths behind cipher mysteries. This website is, of course, simply part of our community outreach programme: and let’s face it, when the obfuscating powers of the NSA, GCHQ, and the Bilderberg Group get combined in this way, what chance do all you ordinary people stand, hmmm?
I was writing up a recently-claimed Dorabella Cipher decryption just now, when an incoming email clattered noisily out of the pneumatic mail tube and into my mahogany in-tray. Nicely, it contained a link to a new Dorabella Cipher article by San Francisco writer Mark MacNamara in online magazine Nautilus, jauntily entitled “The Artist of the Unbreakable Code” (i.e. Edward Elgar).
Given that I exchanged some Dorabella-related emails with MacNamara back in his summer research phase, it was no great surprise to discover – as Bill Walsh and others have kindly pointed out during today – that my, errrm, “stego-Bella” suggestion gets a short mention there. 🙂
Regardless, MacNamara covers Elgar’s enigmatic ground at a fair old pace, and works through Tim Roberts’ and Tony Gaffney’s claimed decryptions, along with their angry annoyance (if not outright outrage) at having the ridiculous stuffed shirts of the Elgar Society turn down their decryptions. Really, who were mere musicologists to tread so heavily on the toes of such ingenious and hard-working code-breakers? etc etc.
Of course, Cipher Mysteries regulars will already know what I believe: that Roberts, Gaffney and even Eric Sams produced attempts that were cryptologically clever at the expense of being historically and practically unsound. For me (and it’s just my opinion), any proposed solution should go some way towards explaining not only the message (the crypto mystery) but also the reason or necessity for the cryptographic wrapper (the historical mystery). The practical problem with these three claimed decrypts is that they are as impenetrable unenciphered as ciphered: which is also presumably why people have rarely enciphered alchemical texts. Or legal contracts. Or legislation.
Will we ever see a Dorabella decrypt that is both cryptologically sound and, as the Elgar Society required for their £1500 lucre-pile prize, “glaringly obvious”? I think it is entirely true that such a criterion is both foolishly idealistic and cryptographically inappropriate for judging most ciphertexts, so I am somewhat sympathetic towards Tony Gaffney’s condemnation. But all the same, I really don’t think our Tone has cracked this particular curate’s egg of a cipher just yet, hen’s shells or no. Perhaps hen’s teeth might be closer? 🙂
Anyway, I rather liked MacNamara’s article, and would recommend it to you with only a few minor corrections:-
(1) Elgar only called Dora Penny “Dorabella” after 1897
(2) The cipher isn’t too short to analyse – in fact, simple substitution ciphers are usually breakable with roughly 30 characters (and this has 87). With a good guess and a bit of luck, you may need only 20 characters, or even 15. Which is why it’s so odd we can’t crack it – really, if it were simple we should have more than enough “depth” to crack it.
(3) The cipher doesn’t strictly “defy” frequency analysis – it’s letter frequencies are what they are. In fact, frequency analysis makes it seem even more likely to be a simple substitution cipher. Rather, the Dorabella Cipher defies its own strong resemblance to a simple substitution cipher.
(4) Elgar not only sent Dora Penny no other ciphers (either before or after), but they never talked about ciphers in their relationship that spanned many decades.
(5) It;s not really accurate to say that I have yet “come to believe” my whole stego-Bella hypothesis. Rather, I have come to disbelieve most of the presumptions that other people have built their own theories upon: and the stego-Bella thing is just my first proper attempt to think outside the generally-accepted Dorabella crypto box. It’s early days, but we shall see where it all eventually leads…
In the red corner we have #1 codebreaking musicologist Eric Sams: and in the blue corner, historical mystery specialist Beatrice Gwynn! Who will be the winner in tonight’s Dorabella Cipher Ultimate Smaaaaackdooooown?
A frisson of crypto excitement ripples through the crowd as Eric Sams rises to his feet. He’s humming to himself, rhythmically pounding his gloves, and with a gleam in his eyes that’s well-earned: his 1970 Musical Times article has been in the bibliography of nearly every Dorabella Cipher article that followed. Sams certainly looks in cracking form: his article Cryptanalysis and Historical Research from Archivaria 21 (1985-1986) casts light on how he decrypted the shorthand used by William Clarke (secretary to Cromwell’s army), the shorthand used by Sir John Thompson (Prime Minister of Canada between 1892 and 1894), etc. He’s got power, reach and stamina, normally an unstoppable combination in this game, as you can also see from his many articles on cryptography at the Centro Studi Eric Sams.
But his opponent tonight, Beatrice V. Gwynn, has many tricks up her fighting sleeves, and perhaps her decades of experience looking at mysteries rather than histories will guide her to victory here. She co-authored a 1977 book on the Phaistos Disc; proposed a theory on the Voynich Manuscript (it’s apparently a hygiene manual written in left-right-mirrored Middle High German, but let’s not hold that against her); and even wrote a book on the evidence used to convict Alger Hiss (“Whittaker Chambers: The Discrepancy in the Evidence of the Typewriter”). Sure, she may not have Sams’ raw cryptanalytical clout, but perhaps she can match him for reach and stamina.
The referee’s in the centre, the seconds have left the ring and… Rrrrrround One begins. The two fighters eye each other up warily over their gloves, waiting to see who will make the first move. And it is Sams who strikes first, whipsmart and sharp as ever:
STARTS: LARKS! IT’S CHAOTIC, BUT A CLOAK OBSCURES MY NEW LETTERS, α, β
BELOW: I OWN THE DARK MAKES
E. E. SIGH WHEN YOU ARE TOO LONG GONE.
Gwynn reels on the ropes, punch-drunk from the sheer interpretativeness of Sams’ claimed decryption. But she quickly collects herself, before launching her own cryptological counter-attack (in her article “The Elgar Cipher”, The Elgar Society Newsletter 1975):-
PINCH FROM TOBACCO
DORA A RIDDLE NOON
SILENT ALONE TIME
Sams is rocked on his heels by the power of the blow (though he must surely be wondering what happened to the rest of the letters – let’s just say it’s a long story). But he soon powers back with what he thinks is a knockout blow – an unpublished 1972 article containing his raw decrypt (i.e. without his wobbly interpretation on top):-
STARTS. LARKS! IT’S EXOTIC BIT A CLOK OBSCURC MY NEW LETTEE
AB BELOW. I OIN THE WARK MAKES E.E. SIGH WHEN E ARE TOO LOMMONT.
Even though Gwynn deftly dodges the wildness of this codebreaking haymaker, Sams has a follow-on jab – pages from Elgar’s archives where he uses the same shapes to encipher “MARCO ELGAR” (the name of his dog), “A VERY OLD CYPHER”, “DO YOU GO TO LONDON?” and “THE GOLD IS BURIED UNDER THE KITCHEN FLOOR” (only kidding!).
Surprised by the primary evidence, Gwynn drops to her knees on the canvas and gets up again quickly, only to receive a standing count from the referee. As the bell rings for the end of the contest, Sams punches the air in victory and the cheer of “E-ric / E-ric / E-ric” fills the arena air. Gwynn looks desolate: have all her years of effort and striving really been trumped by Sams’ nonsensical-sounding decrypt?
At last the judges pass the result to the referee, who announces it as… A DRAW – neither side managed to land a clean punch. But wait… in all the chaos, a haunting melody starts to fills the room, quieting the crowd. It’s Javier Atance playing his claimed solution (direction 1/2/3/4/5/6/7/8 = do/re/mi/fa/so/la/te/do, 1/2/3 humps = natural / flat / sharp) on a distant organ. But the sound system is quickly unplugged and the pandemonium resumes… in fact, will there ever be an end to it?
* * * * * * * * * * * *
To me, the Dorabella Cipher is a bit like the “Voynich Lite” – even though they both closely resemble simple monoalphabetic ciphers, they both employ one or more tricks that make utter fools of those who seek to break them in a classical kind of way.
Sams was keenly aware of this, and even points out that taking a traditional crypto approach will get you only as far as something like this:-
– 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
(or, as Decrypto 8.5 suggests, “TD INVISHE CREE ENT TEROOM HAVECNVISIDEEDNMCVEHEN YACROSSO HAD ANY A DRY MADENY CAMIT MAN IRO I”).
That’s the cryptanalytical mystery: while at the same time, the historical mystery is that Edward Elgar and Dora Penny never spoke of ciphers before or after this. What conceivable rationale would he have had for sending her a near-unbreakable cipher (disguised as a perfectly breakable cipher) in only his third ever letter to her? How can we find a solution to both of these mysteries at the same time?
I’ve been reading more about La Buse & Le Butin, and I have to say I’m not hugely impressed by the research that has been done into either. More books on 18th century corsairs are (as Eddie Elgar might have said) ‘winging their woundabout way’ to Cipher Mysteries Mansions; but if what I’ve seen so far is any guide, I’ll be no less confused in a year’s time.
But really, I think that good historical research is painfully easy to spot, as it combines:-
(1) an appreciation of primary sources (or at least early secondary ones);
(2) a healthy scepticism towards the mythology built up around events and objects; and
(3) empathy towards the people involved (but without a lot of modern back-projection).
Even though we now arguably have better access to primary (or at least closely contemporary) historical sources than ever before, few historians now seem to have the knack for dealing with them properly. Perhaps this is from the slow-motion death of taught codicology and palaeography; or perhaps it’s from the way many of them seem eager to lock themselves into a tightly-specialized silo without no obvious broader-brush historical context or framework to bounce their research against. I guess you’ll have your own thoughts on this, it’s not exactly front page news.
Similarly, the guff that Internet sites pass off as “history” tends to be even more romantic and speculative than even Victorian historians ever managed. In particular, cipher mysteries are so plagued by this rot that I now routinely tell people it’ll take me at least a month to separate what’s real from what’s Maybelline in any new cipher strand – the whole “La Buse / Le Butin” thing is simply the exemple du jour of what is a miserable and much larger trend.
But to my hay-fevered eyes, it’s arguably empathy that I find most obviously lacking. The people of the past aren’t cut-out stick-figures jerking on a historian’s Punch-and-Judy stage, they were real people stuck in uncertain situations, operating blind of their actions’ future consequences. Their decisions were often (quite literally) life-and-death ones; so reducing past lives to mere critical reading textual exercises misses the point.
For me, empathy is that which transcends the details and defies the scepticism: it’s the negentropic force that gives History back the three-dimensionality stripped away by temporal distance, and that pulls the fragmentary pieces together into a sensible whole. Yet… I just don’t see who gives a monkey’s about empathy any more.
Despite The Dorabella Cipher‘s brevity, its link to composer Sir Edward Elgar (who wrote it) has brought it a cult following over the years. Like other unbroken ciphers, it has appeared as a mysterious motif in TV plays, novels, and even recently in a children’s book (The Orphan of the Flames).
At first sight, it looks to be merely a straightforward simple substitution cipher of the kind that pen, paper, and an agile mind should crack relatively quickly. But what is mystifying is that even though Elgar apparently used precisely the same pigpen-like (3 sets of 8 orientations each) cipher alphabet elsewhere in his writings and notes, the letter-for-symbol replacements he used there make no sense when applied to his Dorabella Cipher. The key seems to match the lock, but doesn’t open the gate.
Moreover, given that the ciphertext’s statistical distribution sits awkwardly with those of natural languages, code-breakers’ numerous attempts to shoehorn their preferred substitutions into the cipher’s three short lines come across as clunky and false (at best). Worst of all, I’m sorry to say that even prolific cipher-solver Tony Gaffney’s ingenious and elegantly-structured decryption failed to please pretty much anyone apart from him.
However, the upside to all that grim cryptanalysis is the indisputable truth that Elgar messed around with language quite a lot, typically in a playful and mischievous way. In general, he loved subverting the rules of language, speech and music, which arguably culminated in his famous Enigma Variations, which some people like to call ‘musical cryptograms’ because many lightly parody (for example) various close friends’ speech and laughter rhythms.
Yet what has long tipped my own judgment against the Dorabella Cipher’s being a cipher of any sort is that by 14th July 1897 (the date of the note), Elgar (who wrote the note) hadn’t known Dora Penny (to whom or for whom the note was written) very long at all; and they never communicated in any kind of cipher before or after that date. But even so, my opinion was no more than a hunch, based only on various modern references on Elgar’s life I’d read… not very satisfactory, but that’s how these things tend to go.
Anyway, having spent far too long reading and relying on secondary sources on this particular cipher mystery, a few weeks ago I decided to instead go right to the source of the story – Dora Penny’s book “Edward Elgar: Memories of a Variation” (I bought a copy of the 1946 second edition, which has rather more information about the Enigma Variations than the first edition), written under her married name “Mrs Richard Powell”.
What I read there only served to strengthen my historical argument against The Dorabella Cipher’s being a cipher at all. Elgar and Penny first met on 6th December 1895, and the cipher was only the third letter Elgar ever wrote to Dora (if indeed, as she points out, it is a letter at all). (Also, he only started calling her “Dorabella” in 1898, so there’s a case to be made that its name isn’t chronologically accurate… oh well!) From all I could see, it would defy common sense if he had sent her something written in an deliberately intractable cipher: no matter how much of a fascination he personally had with such things, cryptography of any sort was not a discussion subject the two friends seemed to have shared at all.
And yet what we see does so resemble an enciphered cryptogram, a paradox which ultimately gives it its place at the Cipher Mysteries top table: for it really ought to be a simple cipher, but it surely is not one. And I find it hard not to hear Elgar’s voice saying to Dora Penny exactly what he said to her about the Enigma Variations (one of which is ‘hers’) – that surely she “of all people” would be able to unwrap its central mystery, its hidden themes. Wouldn’t his cipher, too, be steganography – hidden in plain sight?
As to the content of the note, I don’t believe that the newly married Elgar would have sent Dora Penny, for all the fun they had together (going out to the races, seeing Wolverhampton Wanderers, reading maps, flying kites, etc) a love letter. So in all probability, I think that what we are looking at here is a three line note or letter from him to her, in broadly the same joking and playful manner that he adopted in his other letters to her (though probably not as Byzantine in lexicographical complexity as later letters would become), regardless of the particular manner in which that effect is achieved.
The only other clue I have to offer is that in July 1897, the Elgars were living in a house called “Forli” (named after the talented Renaissance painter Melozzo da Forli, who incidentally gets mentioned a few times in Elizabeth Lev’s rather good The Tigress of Forli) in Malvern in Worcestershire. And so I wondered whether “Forli” and/or “Malvern” might be effective as cribs into the cryptogram, for Elgar would typically head even very short notes with his current address (several of which are charmingly reproduced as inserts in Dora Penny’s book). OK, it’s not quite “HEILH ITLER” at the start of Enigma messages, but you gotta work with what you’ve got, right? 🙂
And so with all these fragmentary clues in mind, I stared and stared and stared at the Dorabella Cipher, trying to see what Elgar (mistakenly) thought Dora Penny would see straight away. And then I stared somemore. After a (fairly long) while, here’s what I noticed:-
Essentially, I suspect that Elgar was so certain that Dora Penny would know what he would be saying in a short note that all he felt he needed to do was to write the general form of the words (even presented in the form of a ciphertext-like medium) and she would still be able to ‘read’ them. [Unfortunately, this proved not to be true!] So, I believe that what we are looking at could well be more like Elgar’s improvised steganographic attempt at a mind-reading trick than a traditional ciphertext per se. Such a process would (probably) produce something like what we see: a non-mathematical stegotext that fails to have the kind of rigorous statistical profile that “proper” ciphers would.
I’m the first to admit that it’s far more of a wobbly observation and a loose speculation than a rigorous proof: but what I’m proposing is that the Dorabella Cipher could turn out to be a quite different class of object from that which code-breakers have been trying (unsuccessfully) to crack. It’s not the end of the road here, but it might possibly be the very start of one… hopefully we shall see! 🙂
