Homophonic cipher

The homophonic cipher is the first serious attempt to break frequency analysis. It appeared in 15th-century diplomatic mail in the Duchy of Mantua, associating each plaintext letter with several cipher symbols (homophones) whose count is proportional to the letter’s natural frequency in the source language.

Principle

In English:

• E (≈ 13 % of letters) gets 10 to 15 different codes.
• T (≈ 9 %) gets 8 to 10 codes.
• Z (< 0.5 %) has only one code.

The sender picks one of the possible codes for each occurrence, often randomly or cyclically. Result: every cipher symbol ends up roughly as frequent as any other (about 1 / total-symbol-count). The distribution becomes flat and plain frequency analysis yields nothing.

Codes are often 2- or 3-digit numbers (000-999), but can also be graphical symbols as in Mary Queen of Scots’ correspondence.

Example

For plaintext CIPHE with a fictional table where E has multiple codes:

C → 42
I → 07
P → 51
H → 23
E → 14   (or 28, or 33, or 09, or 61…)

Result: 42 07 51 23 14.

Weaknesses

The method is still breakable, just not by raw frequency. Attack vectors:

• Digram and trigram frequency analysis: combinations TH, HE, ER in English leave statistical fingerprints even when single-letter freqs are masked.
• Context: short words (THE, AND, OF) stay identifiable by their structure.
• The finite number of codes: with a long enough text, some codes cluster at higher frequency → they map to the same plain letter.

Mary Queen of Scots’ cipher (1586) was broken exactly this way.

Famous variants

• Beale cipher (1820, Virginia) — uses a reference text to generate codes. One of the three Beale papers remains unbroken.
• Mary, Queen of Scots cipher — homophonic enriched with a nomenclator (whole-word codes).
• Arnold cipher — homophonic book cipher used during the American Revolution.

In CipherChronicle

Homophonic is the first step toward statistical security: the player discovers that multiplying codes masks frequencies. Grids built on it can present long texts where digram analysis becomes essential.