Here’s a delightful little cipher story that has so far defeated the NSA’s cryptologists. Can you do better?
In 1819, Lady Magdalene De Lancey, whose first husband Colonel Sir William De Lancey had died at Waterloo shortly after their honeymoon, was busy being wooed by the (presumably) no less dashing Captain Henry Harvey. While the latter settled his affairs in preparation for their marriage, she assailed him with letters nearly every day. What is nice here is that her lovesick correspondence was retained in a family archive, where a few years ago it was found by David Miller while researching his (2000) book Lady De Lancey at Waterloo. But what is curious is that some of the letters contain short sections apparently in cipher…
Transcriptions of coded phrases used by Magdalene Hall De Lancey when writing to her fiancé Henry Harvey.
Note: the letters are in the possession of Philip Davies and have been transcribed by David Miller and Sally Smith.
Letter #6: 4/5 February 1819
I told you to the Tytlers, & George St but I find myself invariably the worse of it – so I refuse all without exception of any but Geo: St – such kind parents deserve a little sacrifice – I dine there today – My Mother talks to me wt such delight of all my prospects shtesirreshteltdtyetoogdterldcofcshtsr glateshedeshdgtrshitdhlyskbtwisterhdgthis Oh I forgot I forgot – I beg 10000 pardons – (it was a joke)
[Alternative transcription: Mteirrethteltdtyetoogterloefeshtrplatesheideshdytehtdhlyskhtwiserhyithis
]Letter #9: 8/9 February 1819
Lying down in the carriage is not comfortable — nor to be wished as we must not think of travelling at night — I know where I shall lye when I need rest, & I do not care for any one else you know. Oh that we were in the carriage, anywhere! This delay, this probation thatlydsrsofseedltilingsorgtrdbghiliglingsolersafftchiy Gfetdjsitlotltertelrleslity
pdgrebrditlosefdrelrsethligksofillttle’selrg but we shall be much much happier for the delay — we shall know each other thoroughly — we shall begin with an acquaintance of years, from the interest & constancy of our present communications —Letter #12: 14 February 1819
You must allow me, henry, to look up to you as to a being superior in as much as is good for me – do not fear, my beloved – at moments when we are playful & alone & relaxed & easy I shall be your equal altogether & as fearless & familiar & impertinent as a spoilt child, but this is not at all incompatible with deference for your opinion & respectful attention to your wishes ridthjdbgdtehtetsisovdlgansdtofdrhtghldsdbtdilijche-
soerdreisgdrdtatderihjsthit’sIBitbdersjchrtdthsdpsprotis&shojosbtjeexptrtiyleitltsbijte besides, Henry, I cannot lean unless I feel the power of my support – & I positively insist on leaning on you – I am all feminine – I have no independent powers, naturally – they have been forced up & called forth by circumstances & they have, like other unnatural & forced cultivation left the stem weaker than ever – I am as a honeysuckle which creeps & scrambles all over the tree near which it takes root –
On the one hand, it’s completely plausible that these are in cipher. David Miller wonders whether Magdalene met General Sir George Scovell at Waterloo, the man famously responsible for cracking Napoleon’s codes. (Yes, Scovell could possibly have given Magdalene some kind of cryptographic tip for keeping her correspondence secret… but all the same, that does seem a touch heavy-handed to me.)
On the other hand, they may not be in cipher at all. Sally Smith (whose book on Lady Helen Hall [Magdalene’s mother] is due next year) suggests that these odd little sections might simply be textual expressions of Magdalene’s frustration at the limitations of polite language, and that her husband-to-be would understand completely what she was alluding to in context. That is, he would have known from context what was frustrating his wife-to-be without her actually having to name names. Lady Magdalene does quickly follow the cipher-like sequence in the first letter with “(it was a joke)“, so perhaps this is the safest interpretation?
All the same, I do get an odd sense of things poking through the mix (and it’s not just the few actual words that are embedded in the ciphertext-like sequences), and of sense rather than merely nonsense. The first sequence appears to relate to what Magdalene’s mother was saying to her about her prospects; the second to Magdalene’s feelings about (presumably) her sexual frustration caused by the temporary separation from her fiance; while the third I’m not really sure about. The fact that there are characters on the pages in the places where they are does make sense… it’s just that the words as formed by those characters don’t.
The bigger problem is that they don’t make any obvious cryptological sense. The NSA Historian published at least some of them on the NSA’s internal e-message history page asking for people to try to crack them (for inclusion in David Miller’s book), but as yet nobody has succeeded.
So… what do you think? Lovesick random scrawls or calculated encipherment? Reading beyond the short extracts above, I think it’s fair to say that Lady Magdalene De Lancey’s thoughts quickly range from hot flushes to cool calculation: she thinks in a very multimodal (dare I say heteroscedastic?) way, making it hard to pin a single interpretational tail on her historical donkey.
Personally, I’m kind of stuck in the middle here. Even though I’m sure that Sally Smith has transcribed these accurately as possible as characters, I’d much rather see the cipher-like sections for myself before forming an opinion. There must be a thousand ways of steganographically hiding short texts in plain sight (upside-down, left-right mirrored, different inks, embellishments, marks, dots, strokes, pinholes, etc), and my nagging suspicion is that Lady Magdalene may well have employed one such trick to highlight letters within the long random-looking sequences to form her (much shorter) secret message. Not sophisticated, but clever enough to get the point across to her beloved. Hopefully, we shall see… 🙂
Something about the look of it reminds me of childrens’ “codes” or some minor English dialects.
Not as simple as pig Latin; more like the ‘alibi’ talk in the schoolyard, or (more complex)the dialect I heard once on an old record called “Ogden’s Nut Brown Flake”. No idea what that’s called. I’d think this uncomplicated enough, and enough of a novelty for people of Lady M’s class that it could serve as both joke and code – unlikely to be recognised by family or peers.
(Isn’t Geo:St) a key to the first?
Interesting. The capitals in the middle of each sequence suggest some sort of anagramming to me. There’s about one capital to the number of lower case letter you’d expect in a sentence. You’d expect more capital i’s, but it’s not impossible.
Do the letter frequencies match regular English? At first sight it looks like it.
Steganography doesn’t really make sense to me. If you’re hiding things in plain sight, it’s best to keep the sight actually plain. Ie. why give any indication that there is a cipher at all. The whole point is to make the plaintext look innocuous.
I would still like to see the original letter, because it should indicate the complexity of the enciphering scheme. Was the whole thing written in one go, fluid and joined up? Or was it written letter by letter. I think most enciphering schemes require a working copy. It’s not easy to write text this random in long hand.
As for the expression of frustration theory. I note that there’s an ampersand in the last sequence. Somehow, I don’t think she’d use an ampersand if she were really just creating some sort of onomatopoeic representation of her anger.
Yes, where can we find the original letter?
Here’s the letter distribution across all three cipher samples (not using the alternative transcription of the first one):
49 t (14.4%)
38 s (11.2%)
33 e (9.7%)
30 d (8.8%)
28 i (8.2%)
27 r (7.9%)
27 l (7.9%)
23 h (6.8%)
17 g (5.0%)
15 o (4.4%)
9 b (2.6%)
8 j (2.4%)
8 f (2.4%)
6 y (1.8%)
5 c (1.5%)
5 a (1.5%)
4 p (1.2%)
3 n (0.9%)
2 k (0.6%)
1 x (0.3%)
1 w (0.3%)
1 v (0.3%)
Expected frequencies for plaintext English:
e (12.702%)
t (9.056%)
a (8.167%)
o (7.507%)
i (6.966%)
n (6.749%)
s (6.327%)
h (6.094%)
r (5.987%)
d (4.253%)
l (4.025%)
c (2.782%)
u (2.758%)
m (2.406%)
w (2.360%)
f (2.228%)
g (2.015%)
y (1.974%)
p (1.929%)
b (1.492%)
v (0.978%)
k (0.772%)
j (0.153%)
x (0.150%)
q (0.095%)
z (0.074%)
The cipher letters seem to share some of the normal plaintext distribution (for example, the common letters ‘e’, ‘t’, ‘i’, ‘s’, ‘h’, ‘r’, ‘d’ appear most frequently in the cipher text, and the less common letters such as ‘v’, ‘x’, and ‘k’ appear least frequently.)
m, q, u, and z do not appear at all (but m appears once in the alternative transcription).
Pointless side note: When I removed all the spaces and punctuation to look at letter frequencies, I had 340 letters of cipher text. 🙂
It does kinda smell like an anagramming or transposition scheme, without substitutions. Or maybe it’s some kind of overlay (take two messages and interleave them based on some set of rules). One thing I would try is to permute a set of letter removals to see if any words appear. For example, test all possible ways to remove a single letter. Then test all possible ways to remove two letters. Then three. And so on, until the number of possibilities is too large to test (you could keep going by using a heuristic search of some kind). Perhaps in some removal combinations, more words will appear in the remaining text. Surely the NSA has tried all this already.
David: I only posted excerpts from the letters – they do tend to go on and on, and it seemed fairly unlikely to me that they had any overtly cryptographic content. To be honest, I’m reasonably sure that what we’re looking at here is no more than a simple steganographic trick, maybe-just-maybe with some simple monoalphabetic ciphering in there too (though I don’t really believe so).
There are some interesting patterns in there: for example, clusters of i, t, and l characters. So there may alternatively be some kind of verbose cipher at play here.
I notice that the letters whose distribution differs most from that expected include all those in “Magdalene”, the others being justwo – just off to see what I can find on the net about codes using a name as a starting point.
is there a reward for solving this code and who do i contact ive only got the first five words of 3 5 3 7 2 2 letters in each word but i only found it two days ago i should have more by xmas as im busy writing my book STARSHIP STOWAWAY with the Beale Code keys shown in it.
and where can i find more of lady Magdalenes code Stan Clayton
YOUR JOKING THE DE MAGDALENE CYPHER IS SO SIMPLE,JUST NUMBER THE SHTESIR LINE 1—–23.THEN PUT NUMBERS 1243856 17 7 18 10 9 11 15 12 X 14 13 16 19 20 22 21 23 UNDER THEM
READ THESE IN ORDER AND GET, SHE TIRES THE TYTLTRS TO GOD
DO SAME TO PDGRE LINE NUMBER
1-4-5-X 3-8-9-11-10-X-2-X-X-6 X 7-X 12 THIS GIVES PLEDGED BRIDE.
THERE ARE MORE SHOWS UP LIKE LETTERS,BUT IM BUSY WRITING MY BOOK I WOULD LIKE TO SEE THE DECODED VERSION THANKYOU,HOPE IM NOT A NUISANCE STAN CLAYTON
Huh. Haven’t tried my hand at ciphers before but as a dyslexic in letter number 6, Im reading: “She desire that I go to ??? of Sr. ??? of late. She, ????, hardly about ??? wiser than this.”
Again, I’m probably way in left field but my take on letter number 9, is as follows: “That you should speak of ‘seed tilling’ (aka plowing, sowing wild oats, having sex) or sordid ?ing ???? Solely I ???? It just gets ???? pedigreed self???? thinks of little else.
My last pathetic guess is for letter number 12: “Rid the judge/badge ??? the set is gains of??? She said to Dr. ??? about ??? ‘So here, Dr., that there isn’t ???, I bet there’s ??? church ??? the ??? propose its ??? & sure/surely just ??? expertly let its ???”
Carol: tricky, aren’t they? 😉
Judging from the context of the letters, I think it is noteworthy to suggest that the code she used hides sexual or promiscuous interests. Considering the time when these letters were exchanged, it would have been especially scandalous to be exposed for such intrigues, especially by a lady in even the smallest public regard. If the letters had fallen into the wrong hands, he/she could be easily blackmailed, and considering his military affiliations it could have give leverage to interested parties (whatever that could mean.)
I would also like to ask: there weren’t any additional elements to the letters that were successfully ciphered already, were there? No abbreviations or character clusters?
Lastly, I concur that the letter “m” is of no consequence considering it’s absent everywhere else in the code. Even if it represented a “q” it would appear more than that, given how long the string of letters run…
Not exactly ciphers, I think.
In the wiki article on Glossolalia, near the end, see
1817– In Germany…
and
19th century – the words of a Scottish minster which, if we suppose them reported in a newspaper as was not uncommon in those times, might have amused them both.
geo:St means George Street, where the family had its town residence (I’ve worked on Magdalene’s father and on his diaries and correspondance in the National Archives of Scotland, for a thesis. So this is not the key…) Good luck!
To me this looks like the mnemonic writing systems we all develop when trying to take notes from the professor who spoke way too fast. A common feature is omission of short vowels, which would explain the reduced vowel frequency.
Considering “wt” is used to clearly represent the word “with” in the first letter, it seems likely that this was the case.
So I was playing around trying to figure out the sequence in the first letter, and I noticed that prior to the code the author used two 2 letter abbreviation: st and wt(for street and with). I tried to use these words as a key to solve the code.
I noticed that the code was divided into two sections so I decided to try to apply one key word to each section. The missing letters of “with”(i and h) were deleted from the first section, and the missing letters of “street”(r, e and t) were deleted from the second section.
This yielded English words with junk letters between them. The words were “rest to gd(god? good?) of lash hid wish his”. Using a reversed word order and adding an article forms a much nicer sentence: “His wish hid the lash of gd(god or good) to rest”.
Now this doesn’t make much sense so I don’t think I have solved it, but the formation of words using this fairly simple method makes me think that the message was encoded in a similar way.