At a glanceKnown key applied · complete
- The telegram
- A telegram in Chinese code from Peking, received at Simla by 3 October 1912; the file holds two identical copies, one through Kalimpong. It names neither sender nor addressee.
- Where
- National Archives of India, Foreign Department, Deposit Cypher, 1912, Nos. 86-87, p. 6 (Abhilekh Patal PR_000004118254; duplicate record PR_000006222155). Not on DECODE.
- The code
- The Chinese standard telegraph code, four digits for each character, with no encipherment. The digits were sent in groups of five, so the codes run across the groups.
- How it was decoded
- The copy lays the 25 groups out in four columns. Read by rows, the 125 digits cut into 31 four-digit codes and one filler digit, and the public code table gives the text.
- What it says
- 大總統早經允准復回達賴名號: the President has long since approved restoring the Dalai Lama's title, and allows him to come to the capital.
- Credit
- Read by Brice Dudout with an AI agent (pull request #431). The General Staff at Simla made an approximate English translation on 8 October 1912; the file holds no Chinese text. Lead found in Daniel Bourdeau's Abhilekh Patal sweep; the lead suggested trying the standard telegraph code.
- Still open
- The Chinese reading was made by the contributor’s agent and re-read at integration; no sinologist has re-read it. One code is emended (徇 as copied, read 待). The transcription has not been checked against the image by a second reader. The sender and the addressee are not identified. Chinese printed collections were not searched for the text.
01 The file
The record is titled "Message in Chinese Cypher from Yaoshikinan, Peking which cannot be properly decyphered". The file has 14 pages. Brice Dudout read it after logging in to the archive's online reader and described these pages:
| Page | Content |
|---|---|
| 3 | Viceregal Lodge, Simla, 3 October 1912: a telegram received in an unknown code, "We have been unable to decypher this". |
| 4 | Two identical telegrams, one through Kalimpong; draft to Col. A. H. Bingley of the General Staff. |
| 6 | The groups: 25 groups of five digits, copied in four columns of 7, 6, 6 and 6. |
| 8 | Bingley, 8 October: a translation "so far as we can make sense of it". He suspects a trap, a test of whether the British read the Chinese code. |
| 9 | The General Staff's translation, with two readings (A and B) "according to the placing of the stops". |
| 10–13 | Telegrams of 9 October 1912 to Bengal and to the British Minister at Peking: "Probably it is a trap". |
No image is shown on this page. The pages are on Abhilekh Patal behind a registered login, the site shows a human check to scripts, and the contributor sent transcriptions only, since no reuse licence is shown for the images. The project has not seen the images. What is said here about the file rests on the contributor's description of it.
The telegram has no date code. "Received by 3 October 1912" comes from p. 3 of the file.
02 The code and the reading order
The groups as they stand on the copy, column by column:
col. 1: 71931 24834 09326 00170 11202 25080 60100 col. 2: 88511 84203 67163 33604 97066 69466 col. 3: 29492 36040 51068 51007 01032 39666 col. 4: 04827 21788 25714 92706 71770 56643
Read down the columns, the digits give nonsense in the telegraph code. Read by rows, four groups to a row, they form one stream:
71931 88511 29492 04827 24834 84203 36040 21788 09326 67163 51068 25714 00170 33604 51007 92706 11202 97066 01032 71770 25080 69466 39666 56643 60100
Cut every four digits, this stream is 31 codes of the Chinese standard telegraph code and one last digit, 0, which fills the final group. The code is public: one four-digit number for each character. The table used is the one in the Unicode Unihan database (fields kTaiwanTelegraph and kMainlandTelegraph). The record title says the Foreign Department could not decypher the message properly; the regrouping in fives and the column layout of the copy explain that, and there is no secret key.
The same public table reads two other entries. The 20 plain-code telegrams of Sun Yat-sen's circle copied in Japan in 1913–16 are in the repository folder ctcplain1913, listed on the intercepts page. The telegrams of Chiang Kai-shek and Ma Fuxiang to the Dalai Lama in 1930, from another file of the same archive, have their own page.
03 The reading
電悉。大總統早經允准復回達賴名號，並允到京，格外歡迎優{徇 → 待}。是否通過，速覆。
"Telegram received. The President has long since approved restoring the Dalai [Lama]'s title, and also allows him to come to the capital, where he will be specially welcomed and generously treated. Reply quickly whether [it] has passed or not."
A Chinese telegram has no stops; the punctuation is added. Three points of the reading:
- 優待. Code 1770 gives 徇 as copied, and 優徇 is not a word. 待 is 1769, one digit away, and 優待 "generous treatment" is a set phrase; the General Staff's translation of 1912 also has "generous treatment". This is an emendation of a copy the contributor describes as clear, and the code is not counted as read.
- 復回達賴名號. The General Staff hesitated in 1912 between (A) "the restoration of the Dalai Lama's title" and (B) "the return". A is right: 復回 governs 名號, the title.
- 是否通過. "Whether passed or not" is ambiguous. It can ask whether the telegram got through, whether the matter has passed (been agreed), or whether he will travel through. The General Staff saw the same doubt. The text does not decide, and the translation above leaves it open.
The General Staff's version, as the contributor quotes it from p. 9: "Telegram understood (or for information). The President has already sanctioned the restoration of the Dalai Lama's title (B: the return) and has permitted him (B: also authorised the Dalai Lama) to proceed to the Capital. Extremely gratified at generous treatment. Quickly reply whether or not pass(ed) through." On that wording, "Extremely gratified at generous treatment" mistranslates 格外歡迎優待, which promises a special welcome and generous treatment.
The telegram speaks of the President (大總統) in the third person, so the sender is an official at Peking and not the President, Yuan Shikai, as the lead first had it. "Yaoshikinan" in the record title may be Yao Xiguang (姚錫光); that is a suggestion made when the pull request was verified and it was not checked against any source.
04 Checks
The contributor wrote a protocol before the attack and recorded its SHA-256 hash (frozen/PREREG.md). It fixed the crib words taken from the 1912 translation, the hypotheses and the criteria of success. The order of events rests on the contributor's record: the pull request is a single commit. The plain code read by rows was the first hypothesis and worked at once; the others (a five-digit additive, Campbell's "transparency" relation, the General Staff's column formula, a four-digit additive) were not run.
| Control | Crib score |
|---|---|
| Groups read by rows | 19 |
| Groups read by columns | 0 |
| 1,000 random orders of the 25 groups | at most 6 (mean 1.22) |
| 1,000 random four-digit additives | at most 1 |
| The two halves of the text, separately | 14 and 5 |
Before integration the pull request was verified independently for the project on 12 November 2026. A separate decoder, with a code table built from the Unihan file, gives the same 31 characters. The contributor's control script rerun gives the same output. Other cuts (row offsets of 1 to 3 digits, columns at four offsets, dropping one digit of each group) give gibberish. The hash of the protocol matches, and the crib list in the control script is the list in the protocol. Of the 36 codes one digit away from 1770, only 1769 待 gives a word with 優.
Read share: 30 of 31 codes give the character as written and read as sense, 96.8%. The emended code is a supplied value and is left out of the count.
05 Still open
- The image check. The 25 groups were transcribed from the page only by the contributor's agent, from his screen captures. Nobody else has compared the transcription with the image: the archive shows a human check to scripts, and no way round it was tried. It needs someone logged in.
- Code 1770. Whether the slip is in the transmission, in the Simla copy or in the transcription can only be told on the image or on the original telegraph form.
- Sender and addressee. The telegram names neither. Pages 3–5 and 7 of the file and the Kalimpong duplicate may help.
- Prior art in Chinese. No Chinese text of the telegram was found by web search. The Chinese printed document collections were not searched.
- The date. The contributor notes that historians date Yuan Shikai's order restoring the title to 28 October 1912, while this telegram, received by 3 October, already says 早經允准, "long since approved". (The date of the order is not from the file and was not checked here.)
Full texts
The transcription and the decryption of every document on this page, so that the reading can be used and checked. Short files are shown here; all are in the repository.
Peking telegram, received at Simla by 3 Oct 1912 (NAI Deposit Cypher 1912, Nos. 86-87, p. 6)
Telegram received. The President has long since approved restoring the Dalai Lama's title and allows him to come to the capital, where he will be specially welcomed and generously treated (優待 by emendation of one code). Reply quickly whether it has passed (是否通過, ambiguous). The General Staff's reading A of 1912 (restoration of the title) is confirmed.
Transcription
# peking1912: transcription of the cipher groups # Source: National Archives of India, Foreign Department, Deposit Cypher, 1912, Nos. 86-87, p. 6 (Abhilekh Patal PR_000004118254). # Handwritten copy, 25 groups of five digits, laid out in four columns on the page (7, 6, 6, 6 groups). # Transcribed by a Claude Opus 5.5 agent from screen captures of the online reader (images not reproduced); every digit is clear. # # As laid out on the page (columns left to right): # col.1: 71931 24834 09326 00170 11202 25080 60100 # col.2: 88511 84203 67163 33604 97066 69466 # col.3: 29492 36040 51068 51007 01032 39666 # col.4: 04827 21788 25714 92706 71770 56643 # # Reading order that works: by ROWS (left to right, top to bottom): 71931 88511 29492 04827 24834 84203 36040 21788 09326 67163 51068 25714 00170 33604 51007 92706 11202 97066 01032 71770 25080 69466 39666 56643 60100
Decryption
# peking1912: decryption, code by code
# The 125 digits read by rows form one stream of four-digit Chinese standard telegraph codes (31 codes + 1 filler digit).
# code -> character (standard telegraph code; table: Unicode Unihan kTaiwanTelegraph / kMainlandTelegraph). status: ok, doubtful
1 7193 電 ok
2 1885 悉 ok
3 1129 大 ok
4 4920 總 ok
5 4827 統 ok
6 2483 早 ok
7 4842 經 ok
8 0336 允 ok
9 0402 准 ok
10 1788 復 ok
11 0932 回 ok
12 6671 達 ok
13 6351 賴 ok
14 0682 名 ok
15 5714 號 ok
16 0017 並 ok
17 0336 允 ok
18 0451 到 ok
19 0079 京 ok
20 2706 格 ok
21 1120 外 ok
22 2970 歡 ok
23 6601 迎 ok
24 0327 優 ok
25 1770 徇 doubtful: 1770 = 徇 as written; almost certainly a slip for 1769 = 待 (優待 'generous treatment')
26 2508 是 ok
27 0694 否 ok
28 6639 通 ok
29 6665 過 ok
30 6643 速 ok
31 6010 覆 ok
0 (filler digit, end of the last group)Reading
# peking1912: reading
Chinese text with punctuation (a Chinese telegram has no stops; the punctuation is ours). {x} = doubtful code.
## Peking telegram, received at Simla by 3 October 1912 (NAI, Foreign Dept., Deposit Cypher 1912, Nos. 86-87, p. 6)
電悉。大總統早經允准復回達賴名號，並允到京，格外歡迎優{徇 → 待}。是否通過，速覆。
Code 1770 gives 徇 as written; read 待 (1769), one digit away, which gives the set phrase 優待 "generous treatment"
(the General Staff's translation of 1912 also has "generous treatment").Translation
# peking1912: English translation Telegram received. The President has long since approved restoring the Dalai [Lama]'s title, and also allows him to come to the capital, where he will be specially welcomed and generously treated. Reply quickly whether [it] has passed or not. Notes - 復回達賴名號 "restore (return) the Dalai's title": this settles the General Staff's hesitation of 1912 between reading (A) "the restoration of the Dalai Lama's title" and (B) "the return": A is right; 復回 governs 名號. - 是否通過 "whether passed through or not" is ambiguous, and the text does not decide: whether the telegram got through, whether the matter has passed (been agreed), or whether he will travel through. The General Staff noted in 1912 that it could refer to persons crossing a frontier or to telegrams reaching their destination. (Corrected at integration, 12 Nov 2026: the first version chose "whether [this telegram] has got through".) - 格外歡迎優待: literally "exceptionally welcome, generously treat"; the General Staff rendered it "Extremely gratified at generous treatment", which is a mistranslation. - 優待 is an emendation: the code as copied (1770) gives 徇; 待 is 1769, one digit away.
Key: key.tsv
06 Sources
- National Archives of India, Foreign Department, Deposit Cypher, 1912, Nos. 86-87, on Abhilekh Patal, PR_000004118254 (login needed to see the pages): the groups on p. 6, the General Staff's translation on p. 9.
- Unicode Consortium, Unihan database,
Unihan_OtherMappings.txt, fields kTaiwanTelegraph and kMainlandTelegraph: the code table. - Brice Dudout, pull request #431: transcription, decryption, translation, protocol and controls.
Transcription, decryption, reading, translation, the codes used and the frozen protocol are in targets/peking1912/.
Cite this page
Brice Dudout and Daniel Bourdeau, “The Peking telegram of October 1912 on the Dalai Lama’s title”, Cyphersolver, posted 12 November 2026, https://cyphersolver.com/peking1912.html.