Counting letters to read a hidden message

Choose one short paragraph and one long one, and count how often each letter appears in each. Compare the results and say which paragraph is closer to the general pattern, and why. Decide whether your data involve one variable, and what you would need to make them involve two.

In a table, al-Kindi gives a letter count taken from seven leaves, but the editors' notes emend the numbers and identify three letters omitted from the enumeration ². The record therefore describes both the example and its defects, and does not copy its numbers as a complete modern measurement or as permanent probabilities for all written Arabic ². When the text explains why some letters are so frequent in Arabic, it distinguishes pronounced sounds from short vowels not independently represented in ordinary writing ³. The Damascus edition also compares later treatises when correcting the letter counts ⁴.

The treatise opens with an address to Abu al-Abbas and says it answers a request for an accessible explanation of recovering what concealed books record ⁵. The TDV encyclopaedia lists works by other authors on concealing and recovering messages, and relates official correspondence and expanding administrative departments to the need for concealment and recovery ⁶. Those are the purpose and the setting the record gives: a request from an addressee, and a field of writing tied to official correspondence.

Al-Kindi may be taken for the first person to conceal messages, but the record says that the TDV encyclopaedia calls his the earliest surviving treatise on concealing and recovering messages, not the invention of concealed messages ⁶. The record adds that survival is not absolute priority: an earlier text may be lost, and TDV does report a book attributed to al-Khalil ⁶.

The letter counts from the seven leaves have not reached us intact: the editors emended numbers and noted three omitted letters. Can a data set with known defects still tell you something about the pattern of letters in a language? Which claim would you still accept from it and which would you refuse, and why?

• Al-Kindi's letter-frequency cryptanalysis: the seven-leaf example with its defects, and the difference between surviving and being first.
• Al-Kindi: what the record says about the author of the treatise.