Louvois' Ciphers in the Archives (1688-1713)

About dozen ciphers from Louis XIV's time gathered from various sources have been listed in "French Ciphers during the Reign of Louis XIV". I added another dozen to the list from BnF fr. 6204, of which I was made aware of by Norbert Biermann's paper (2020). They appear to be Louvois' ciphers (at least some are annotated as such), characterized in that not only syllables and words/names but also letters of the alphabet are mixed in random arrangement.
Now that the list has more than twenty ciphers, I introduced designation of ciphers such as "DE=34/42/97." (The cipher of this example enciphers "de" into 34 or 42 or 97.) This nomenclature will facilitate identifying a new cipher with those already on the list. I have employed similar schemes for American ciphers (e.g., here). (I think I've seen other people use this scheme, but I do not remember specific instances now.)

Two-Part Code between Le Tellier and Colbert (1650)

I added two ciphers used between Le Tellier and Colbert in "Ciphers Early in the Reign of Louis XIV". This is something more than a routine addition, because this shows that the same corresondents used one-part code in November 1650 and two-part code in December 1650. Is this the first two-part code in mainstream French ciphers? Apart from sporadic examples (Duke of Nevers, Edward Stafford), the first two-part code known to me has been Louvois' code (1676). Considering that Louvois was son of Le Tellier, it is probable that he learned of two-part code from his father. It is often said that two-part code was invented by the codebreaker Rossignol. It is wondered whether Le Tellier learned of it from Rossignol or he devised it himself or learned of it from any other source.

Religious Grandmother's Abbreviation Cipher Deciphered on the Web

Back in 2014, a certain JannaK posted what looks like an enciphered message wirtten on an index card by Grandmother (1927-1996), who lived in Minnesota. It begins with:
PDGNHBOBVPNSNHANAOENCNANHPNCPND,NUOCP
NNPNAPNMSMDKBMLPOWP,NAP,]NEENTGBTMLS
HHSSSTMALHFFTMOFPANSTP,NIOOIPNTPNROA
NTRSANTOTTAPOD,PLUADONNOANPBOOLL,PKUA
SASIMAB,PAGA,]PMOMTAVMAJAMMSSSLLAB
....

Thirteen minutes later, harperpitt pointed out that the last AAA on the front side may be "Amen, Amen, Amen." Then, TYAGF appearing earlier may be "Thank you Almighty God for..."
The back side begins with OFWAIHHB, which jessamyn pointed out is used as user names by many people and should be something well-known. Then, harperpitt came up with an idea that OFWAIHHBTN is "Our Father who art in Heaven, hallowed be thy name." This led to revealing the meaning of many phrases. The very frequent PST turned out to be "Please see that ...."
This is not a cipher in the usual sense, but initialism. I didn't know such a system is used for a text of some length until I read Chapter 13 of Elonka Dunin and Klauss Schmeh's book Codebreaking (see another post), where the system is called "abbreviation cipher".

References:
Decoding cancer-addled ramblings
"おばあちゃんが最期に残し20年間も未解決だった謎の暗号がネットの集合知によって爆速で解明へ"

Encrypted Diary in Japanese Solved in 4 Hours on Twitter

I wrote about solution (2014) of an encrypted diary (1995-1996) in Japanese posted on twitter in a separate entry in Japanese. The diary belonged to a deceased uncle of the poster. (Hence, the hash tag #叔父日記暗号 (Uncle's Diary Cipher).) The solution was achieved in four hours. The entries of 1995 turned out to be based on Japanese Morse Code. A character is enciphered by interchange of dots and dashes of the Japanese Morse code. In 1996, a simple Caesar cipher was used, but it was based on old Japanese syllabary I-RO-HA.

「叔父日記暗号」の解読

(Abstract in English)
数年前、「亡くなった叔父の日記」の暗号をツイッターでのやりとりを通じて有志が解読したというニュースを見たのだが、このたび一連のツイートをここで読んだ。

発端は2014年11月19日に投稿された「亡くなった親戚の叔父さんの日記が全然解読できない！ 」というツイート。11月23日になって言語学専攻の院生が関心を示したところ20:34に追加画像が公表され、「#叔父日記暗号」というハッシュタグが付けられて盛んに議論が交わされた。暗号は４時間ほどで解読された。
問題の暗号文は７日分あり、次のとおり。

【1995年のはじめに】
ヌサヌヨ．フ．ヤリミサ
゛ミシ．シホフ．゛ウフ．シホカヨーフ．（イム
ヨーフ）タマンタ．イムヨマヨ．クフツツ．フヌ
ミキ．エソヌア。ヤサチサフ．ヤリミサ．
トツウ．ミサ゛リメマヨ．ツマヨ．リイケヨリ．ヌア。
92872014
【'95 1/5】
ロマネヱリ．トイイケ．チ．ワワ
ムサワ゛トカ．ツーワソ　ヲラヲミヨセヰ．
トタナモヰアヨ（ヤサイキ）ンシマア（イムヨ
ーフ）ロ．タソ．イヨ゜ヨサソヨ．
ノアサヤミ．ケメヨマ゛ソワマイフキヨ
ヲホソミツリ．゛ミシ゛ウキヰ．ケワオ
フカサ．タケワタヌアロミイ．
メサウヱリ．ケテトカヨ．トミイアリケサ．
【'95 1/6】
ミソ．（゛アヨヲヨミケセン）トテロタソ．
エエタマヨシーミソ．ヲヨ(ス)ロフヰケムウ
ーエワミヨ．ミソキヨ．ヲヨ(ス)ツネフアテソー
ウケメヨマ゛ソワウ．アキーンヘソワマ＊イ
ケテケケレマヨ．アミサロ．キヨヰロヲア
キー゛タ．メサウヱリ．゛ミンヨナアリケ
サミイ．
【'95 1/8】
ロミヨヌヨニ゛ウ．レ＊ミヨエテソ．
ケセレフ．ソヰツ゛リ．ヌモアヨ．シマカ
゛リアミイケヨ．イヨセ（ムヨマアミイ）
リアテソンアウ．
タネロネセ．イヨ゜ヨサソヨ．
イヨタノヨケケマイ．
トミイヲ．リッソホニリ．
【'96 1/5】
アラーワス．カオロタヨテヨヨ
ワネヱエンテソヨ．テセフミキサ
ヰヨユメクマンワム．テモワオ
オナーヨン．ロ゛レフスサヨ
ラツエヌテチ．ワツミンキスヱ
ンミヨ．ミヨヰノロン゜アスハ
ロミンユエテム．
【'96 1/6】
フミキサ．イオミン～ミンキイオミ
ン．ノヒソヨカチ．マンスサヨンソ
ヨ．カヘンロンミワユナハカチ
ワツロナソエオヌ．ヱ゛チスヨ゜
ノヒワソヨ．モソユーミカカラ
ヰヨーホスヨン．ワンロンイム．
【'96 1/8】
ヱアミケンチハ．オヒチワナ．セフ
アユワン．ケソヨ．ニスヰノミユ
チエ゜ヘン．クヤヘハ．ヨホン
ヒヰワーハーオ．
セヒム．ノノテタ．ヲワヰロネ．テ
マニンヰ．ケヨチ．タスロ゛ーミヨ．
フスクワナ．アラオネ＊ヌ．
テセヰテアロ．ハヘンオナー?

1995年の分と1996年の分では異なる暗号が使われていた。
1995年のほうは和文モールス信号のトンとツーを入れ替えるモールス反転方式と判明した。たとえば「い」は和文モールスでは・－だが、このトンとツーを入れ替えて－・とすると、これはモールス符号で「た」を表わすことになる。突き止められた暗号表は次のとおり。
アイウエオ とたりおえ
カキクケコ けてそかぬ
サシスセソ んみ？めく
タチツテト いねまき？
ナニヌネノ わろこちふ
ハヒフヘホ を゜のむや
マミムメモ つしへせー
ヤ／ユ／ヨ ほ／ゑ／゛
ラリルレロ れうゑらに
ワヰ／ヱヲ なる／ゆは
ンー゛゜／ さもよひ／
最初の項目の冒頭は次のようになる。
　　ヌ　　　　サ　　　　　ヌ　　　ヨ　．　フ　　　．　ヤ　　　リ　　　　ミ　　　　　サ
・・・・　－・－・－　・・・・　－－　　－－・・　　・－－　－－・　・・－・－　－・－・－
↓
－－－－　・－・－・　－－－－　・・　　・・－－　　－・・　・・－　－－・－・　・－・－・
↓
　　こ　　　　ん　　　　　こ　　　゛　　　の　　　　　ほ　　　う　　　　し　　　　　ん
どのようにしてこのような突飛な暗号方式を見破ったのかの過程に興味があるが、1:29に紗衣さんが「多分解けた．和文モールスでトンとツーを入れ替えてる．ヌサヌヨフヤリミサ→今後の方針 しか確認してないので他は各自復号して下さい．」といって突然解読を宣言しており、途中経過はわからない。想像するとすれば、一連の議論の中でも指摘されていたように、暗号文に「ヰ」「ヱ」が使われており、「゛」「゜」がカナと同様に使われていたことから、電信符号を疑ったのだろう。最も頻度の高い文字「ヨ」が濁点を表わす可能性も言及されていたので（そのツイートを見ていないとしても）両者のモールス符号を見比べてひらめいたということはありうるだろう。

和文モールス反転方式とわかって一件落着だと思いきや、1996年の分はこのやり方では解読できないことがわかった。その後こちらは、いろは歌方式の「シーザー暗号」という、ごく単純な方式と判明した。
イ→ろ
ロ→は
ハ→に
…
ン→゛
゛→゜
゜→ー
のように文字をずらしていくものだ。これはМэй Аренбергさんが解読したのだが、同氏は発想の経緯を簡単に語ってくれている。

「いろは歌」は前半からカンで疑ってた(私はこの筆者が戦前生まれとかなんか思い込んでたんで五十音じゃなくていろは歌だろ～とか勝手に思ってた)んですが、後半の頻度分析で「ん」が最初に来て、日本語の頻度は「い」が最初なので、「ん」を「い」に置き換えて、あれ？「ゑひもせす(ん)」→「いろはにほへと」だからつながってるのでは？と思ってずらしたら文章になったので、いろは歌だっていう結論に達しました。ただ、「ん」は「い」じゃなかったですけどね。

あとから「そう思ってた」というのは簡単なのだが、氏は0:07の時点で「いろは歌でずらすアルゴリズムって作るのむずい？」と問いかけを投げており、このときすでに考えていたことがわかる。同氏が全体の解読文をここに載せてくれている。

巷ではネット民の集合知の成果だなどとして話題になったようだが、結局、解読は二人の個人のひらめきのおかげだったようだ。
とはいえ、解読に至るまでの議論はたしかに面白い。他の暗号を解読するときのヒントになるかもしれないので、その間に出されたコメントをいくつか紹介しておく。

「．」が文節などの境界だろうという当初からの予想は当たった。
「゛」「゜」がそのまま濁点・半濁点だろうという予想は外れた。途中、ヌル（捨て仮名）という仮説も出たが、外れた。
「ヨ」の頻度が高いことは早くから気づかれていたが、結局は濁点（前半）だった。
途中、「ヨ」＝「い」という仮説も出た（日本語の音では「い」が最も高頻度）。こういうとき、その文字が近くに続けて出現するパターンがヒントになることがあるので、「イムいマい」に当てはまるフレーズがいろいろ検討された。「思いたい」という仮説が出てくると、今度はそれらの文字が集中して現われるパターンをさがし、「おいセ（もいたアミお）」というのがみつかった。正解であれば、このあたりから芋づる式に対応が判明していくのだが、今回は袋小路に至った。
「イムヨマヨ」「イヨセ（ムヨマアミイ）」
　こしいたい　　こいつ　しいたけのこ
のような面白い案も出たのだが（www付きで）。

２回連続して出現するツツ、イイ、ワワ、エエ、ケケ、ヨヨは、濁点・半濁点、捨て仮名、棒引き、「ん」ではないとされた。
小さい「マ」が「ヤユヨツ」のどれかだとの予想は当たりで、「ツ」だった（前半）。
「．」を区切りと考え、２度出てくる単語として「イヨ゜ヨサソヨ」「メサウヱリ」が指摘された。
「．」の前にどういう文字がくるかも調べられた。単語の区切りがわかっていれば、単純な頻度分析だけでなく、語頭文字、語末文字の頻度も手がかりになる可能性がある。
今の日本語では普通は使わない「ヰ」「ヱ」がある一方、「コ」と「ル」が出現していないことも注目された。
先鞭をつけた院生のつながりからか言語の専門家が多かったらしく、マレー語とかアイヌ語とか中国語とかアラビア語などの可能性も検討されていたのがすごかった。
書き誤りがヒントになることもある。「ヰ」を「ヌ」と間違えているところがヒントになる可能性も指摘された。
全体が同じ暗号で書かれているという保証はない。項目によって暗号方式が異なり、日記の天気の項が暗号方式を示しているのではという可能性も指摘された。そうだとすると頻度分析などのデータが役立たなくなる。だが天気の違う日に「イヨ゜ヨサソヨ」という共通の文字列が出現していることがわかり、この可能性は排除できた。ただ、その後も天気の記載が「晴」「はれ」「ハれ」と不自然なバリエーションがあることが指摘された。結果的に、前半と後半で２通りの暗号が使われていたので全体を通した頻度分析はノイズの多いデータだったことになるが、幸い「ヨ」の高頻度は埋没せず、解読の手がかりになったと思う（上述）。

Lewis Carroll's Ciphers

I heard of Francine F. Abeles's work on Lewis Carroll's ciphers by a recommendation email of Academia.edu and described them in a new article in Japanese "ルイス・キャロルの暗号".

A New Book on Codebreaking by Elonka Dunin and Klaus Schmeh

I got a new book, Codebreaking, A Practical Guide by Elonka Dunin and Klaus Schmeh, hot from the press.
I have long known Elonka's name by her webpage formerly titled "Elonka's List of Famous Unsolved Codes and Ciphers" without knowing her other achievements in codebreaking, authoring, and public speaking (her bio).
I have known Klaus by his blog Klausis Krypto Kolumne, where many people in the world discuss crypto mysteries and solve hitherto unsolved cryptograms. I'm greatly honored to be credited as one of the proofreaders in the book by these two authors of world fame. (I also contributed a blurb to the official page of the authors.)

This comprehensive book provides separate chapters for just about every major encryption scheme historically used: Caesar ciphers, simple substitution ciphers, homophonic ciphers, codes and nomenclators, polyalphabetic ciphers, complete columnar transposition ciphers, incomplete transposition ciphers, turning grille transposition ciphers, digraph substitution, abbreviation ciphers, dictionary codes and book ciphers, among others. One can see variants of ciphers that may be treated in one chapter in other books are given separate chapters. Each chapter not only provides cracking techniques for each scheme, but also presents "success stories", "challenges", and "unsolved cryptograms" from actual examples.

To me, the chapter "16. Solving ciphers with hill climbing" was the most interesting. Hill climbing is a computer-based technique that has proven very successful in solving historical ciphers. This chapter introduces no less than thirteen "success stories", which I believe is the most comprehensive list up to now of achievements in codebreaking by hill climbing. This chapter may be worth publication as a book on its own, with more details on designing of specific fitness functions and techniques for randomising.

Many actual examples and images, ranging from encrypted postcards to epistles written by famous historical persons, are not the least appeal of this book. (Elgar's Dorabella cipher is well-known, but I didn't know cryptograms of Beatrice Potter and Rudyard Kipling.) Although many of them can also be seen on the authors' websites, it is good to see them in classified chapters in a book.

The DECODE Database

I was invited to collaborate with the DECODE database. It is a part of the interdisciplinary DECRYPT project. The DECRYPT project aims to develop resources and computer-aided tools for decoding of historical documents. The tools facilitate transcription and decryption including (but presumably not limited to) cryptanalysis, among others. The DECODE database is a collection of digitized images of ciphertexts and keys along with metadata. Even nonregistered users can search and view metadata.
During the last week, I made my first uploads about cipher letters of Philip II to Juan de Vargas Mexia (which can be found with the search term "BnF_es132"), which are the ones I want to see deciphered the most (see a previous post). Although the status of the records is "decrypted", it merely means the key has been identified. The plaintext is yet to be recovered by using the key.
I have not yet completed the batch, but after these, I hope to upload many others when time allows.

Euler's "Logogriph"

I heard of breaking of the cryptogram left by the Swiss mathematician Leonhard Euler at Klausis Krypto Kolumne. I compared the solution of the same cryptogram by two people: Pierre Speziali and Hans Rohrbach in a new article in Japanese "数学者オイラーの残した暗号文の解読" and mentioned the cryptogram in "Unsolved Historical Ciphers" (which also covers cryptograms that remained unsolved for some time).
Rohrbach's solution (1973) involves counting symbol frequencies with contact characteristics. So, I also mentioned it in the section "Contact Chart in the Days of Manual Work" in "Creating a Contact Chart with Microsoft Excel".

Independently Enciphered Duplicates

Important dispatches were often sent in duplicates to ensure the delivery of a message. When the message is in cipher, it would be easy to make a copy of an enciphered dispatch. But there are cases in which duplicates are independently enciphered. I added such examples in "Finding the Keys to Philip II's Cipher Letters to Juan de Vargas Mexia".

Practice in Use of Ciphers in the Time of Catherine de Medici

I added a section "References to Cipher in the Correspondence of Catherine de Medici" in my article about French ciphers during the reigns of Charles IX and Henry III. It consists of excerpts of Catherine de Medici's letters mentioning a cipher. They are about instructions to use a cipher, delivery of a cipher, handling a cipher personally, forwarding of letters deciphered or to be deciphered, etc. They may be nothing new, but may allow a glimpse of practice in use of ciphers.

Trevanion's Cipher Still being Sought

Sir John Trevanion was a prisoner at the Colchester Castle during the English Civil War, but managed to escape thanks to a letter which hided a message "PANEL AT EAST END OF CHAPEL SLIDES", which showed him an escape route.
Although this episode is famous as a spectacular feat of steganography, nobody seems to know a primary source of this episode, as I pointed out in 2012 in "Quest for 'Trevanion's Cipher'".
I made some additions to the page (references from 1878 and 1925 and some peripheral information).

Zodiac Killer's Cipher Z340 Broken after 51 Years

Earlier this month (December 2020), Zodiac Killer's Cipher known as Z340 was solved by David Oranchak, Sam Blake, and Jarl Van Eycke. Although I have not paid much attention to this cipher, the puzzle is so well-known that I felt obliged to cover it in a new article in Japanese. (In English, I believe there will be plenty of articles out there.) I also added an entry for this in "Unsolved Historical Ciphers.", which also covers recently solved ciphers. 

 

Ciphers in Catherine de Medici's Correspondence

Some weeks ago, I added ciphers used in the correspondence of Catherine de Medici. (When I said it was the first time I came across ciphers used in her correspondence, I did not remember I already described in the same article a cipher used in a letter from Foix to Catherine de Medici in the reign of Henry III.) Now, I added more ciphers used in letters of the Queen Mother during the reign of Charles IX. As for as the specimens I encountered this time, the Queen Mother used the same cipher as Charles IX. (Not that they did the ciphering themselves. The letters are undersigned by a secretary of state.)
I also made one addition (Tournon?) in "French Ciphers during the Reign of Henry II of France" and another [Hurault?] in "Unsolved Historical Ciphers".

Letters Deciphered by John Wallis Sold at Auction for £29,000

It is well-known that a collectioin of fifty-two letters deciphered by John Wallis was deposited at the Bodleian Library, Oxford. Wallis actually made multiple copies of the volume, and on 7 October 2020, one (possibly Wallis's own) was sold at an auction for £29,000. It even includes a fifty-third letter not included in the copy at the Bodleian Library. I added a reference to this in "John Wallis and Cryptanalysis", and also made a few additions from Philip Beeley's paper (2016) referenced by the auction site.
The auction site has some images.
One is Charles I's letter to his son in February 1647, which I already covered in "King Charles I's Ciphers".
Another is a letter from French agent Graymond to Cardinal Mazarin. The cipher used seems to be the same as Graymond's cipher I recently reconstructed from another source and presented in "Ciphers Early in the Reign of Louis XIV".
Another is a letter from the Earl of Lauderdale to the Countess of Carlisle. I now added this in "King Charles II's Ciphers during Exile".
According to the auction site, nearly half of the collection are "from royalist agents in Breda to Dutch and English merchants and other contacts in London: dating from 1650, these relate to the negotiations by which the exiled Charles II was offered passage to Scotland and a means of recovering the throne in England, in return for accepting the authority of the Scottish kirk and parliament." These may include a key between Charles II and the Duke of Hamilton used in some undeciphered letters. I noted this possibility in "Unsolved Historical Ciphers".

Ciphers of Charles IX and Queen Mother, Catherine de Medici

I added a section "Ciphers of Charles IX and Queen Mother, Catherine de Medici" in "French ciphers during the Reigns of Charles IX and Henry III", now renamed from "French ciphers during the Reign of Henry III of France". (I thought separate treatment of the reigns of these brother kings was not necessary.)
As far as I remember, this is the first time I came across ciphers used in the correspondence of Catherine de Medici. The same ciphers were used in letters to the King and Queen Mother as far as these specimens are concerned.
I have been wondering whether Catherine de Medici brought something new to the French cryptography, but these specimens show no such evidence.

Earliest Known French Ciphers

The earliest known specimen of French ciphertext is from January 1526 (or possibly September 1525). This is a few months earlier than the earliest mentioned in a scholarly paper (2018). This and other French ciphers during the reign of Francis I are described in the new article "French Ciphers during the Reign of Francis I".
It includes some undeciphered ciphers, which are now mentioned in "Unsolved Historical Ciphers".

Two Ciphers Broken by George Lasry

Two of unsolved ciphers presented in "Unsolved Historical Ciphers" have been solved by George Lasry.
One is a ciphertext (Senecey to Archbishop of Lyon) left undeciphered by François Viète (see "Ciphers Broken by François Viète").
The other is an Italian letter in the French archives (see "Undeciphered Letters in BnF fr.4715").
Senecey's cipher employs as many as 79 symbols in a relatively short ciphertext, and was very hard to break even with his sophisticated computer algorithms, which appear to have solved the Italian letter in an instant. I look forward to a paper detailing his method.

Regent Marie de Medici's Cipher

I found undeciphered passages in Marie de Medici's letters (1610) in BnF fr.3789 and included it in "French Ciphers during the Reign of Louis XIII" and "Unsolved Historical Ciphers".
The same source also included Villeroy's letter in cipher. The latter cipher turned out to be the same as one I previously reconstructed from BnF fr.15578 presented in "French Ciphers during the Reign of Henry IV of France".

Cipher between Emperor Charles V and Juan Perez (1527)

I added an unsolved cipher in a letter from Juan Perez, ambassador in Rome, to Charles V and another from Charles V to an unnamed recipient in "Ciphers during the Reign of Emperor Charles V" and added references to these in "Unsolved Historical Ciphers".