The Claim — Two Ciphers, One Play

← Word Cipher main page

A Cryptographic Audit of Sir Francis Bacon’s Word Cipher (1623) · The Keystone

The Claim — Two Ciphers, One Play

What it is — the whole project’s core claim in one picture: Bacon’s two separate ciphers — the biliteral and the word cipher — independently point to the very same Shakespearean play, Troilus & Cressida.
Where it fits — the top of the Word Cipher Outreach Set: the single claim that the nine numbered documents below it are the evidence for. Read this first to see where the whole audit is heading.

The convergence

Two cipher mechanisms, both built by Sir Francis, land on one play

Biliteral cipher · the KEY
FRANCIS ST ALBAN
Bacon’s own post-1621 signature (Viscount St Alban), recovered by the biliteral a/b type-cipher from the Prologue of Troilus & Cressida. The name on the door.

See Experiment One →

Word cipher · the KEYHOLE
orifex
A small opening — a keyhole — occurring exactly once in the whole Folio, in Troilus & Cressida V.ii. The lock the key fits.

See the Eleven-Step Model →

the keythe keyhole

Troilus & Cressida — the one door that both the Word Cipher and the Biliteral Cipher open onto

Neither cipher needs the other to be true. The biliteral could be argued on its own; the word cipher on its own. That they independently select the same play is the striking fact — the kind of convergence that is hard to arrange by accident.

Why he did it

By 1623, twenty years of concealment had produced no reader

The motive, adapted from Margaret Barsi-Greene: by 1623 Sir Francis could see that no apparent progress had been made towards anyone decoding what he had concealed for posterity. Time was running out, and rather than let a twenty-year project disappear altogether, he began an escalating spiral of hints and clews pointing at his two methods — the Biliteral Cipher (Binary Code) and the Word Cipher.

But the two ciphers are already woven through more than nine hundred pages and all thirty-six plays of the First Folio. Splicing fresh clews into that text would have meant re-engineering the ordinary, non-secret writing surrounding every insertion. So the clews went where nothing had to be disturbed: five brief standalone texts added at the very end, of no literary merit, the kind of filler that puzzles scholars who ask why such material is in the book at all.

The longest of the five is the one-page Prologue to Troilus & Cressida — the play that arrived late, out of the Catalogue, at the last moment. That is the answer to why here: a Prologue could be added without touching anything else. The door is where a door could be cut.

This is the project’s reading of Bacon’s motive, adapted from Barsi-Greene — an interpretation of the evidence, not itself a decoded claim. The Folio anomalies set out below are objective facts of the 1623 book.

Why this door

Troilus & Cressida is the Folio’s own anomaly

The choice of play is not arbitrary, and the anomalies are objective facts of the 1623 book, not interpretations: Troilus is the only play missing from the Catalogue (the Folio’s table of contents), and a few copies of the book omit it entirely; it was added late, wedged between the Histories and Tragedies; its pages are almost entirely unnumbered — only two (79 and 80) carry numbers, where every other play is numbered throughout; and it carries a Prologue found only in the Folio, absent from the 1609 Quarto altogether — an addition made at the very moment the plays were gathered and sealed. If a maker wanted to mark an entrance and hide it in the least-expected place, this is the play they would choose — and it is the play both ciphers choose.

And the anomaly was paid for. In the hand-press world of 1623, re-admitting Troilus at the last moment meant real money and real labour: a cancel leaf to hide an abandoned first printing (the play had been set to follow Romeo & Juliet — three pages were already printed), fresh composition and presswork for the Folio-only Prologue, and extra unsigned quires to sew and glue into an essentially finished book. Someone needed this unloved play in the book badly enough to pay for it. The customary explanation — a publisher's rights dispute — rests on two Stationers' Register entries and an imagined middle: no letter, lawsuit, or Stationers' Court record of any dispute has ever been produced. The costly insertion is the documented fact; the dispute is the story told to cover it.

Companion — the full cost ledger
The Price of the Door — what it actually cost the print shop to squeeze this play into an already-finished book. Eleven of the twelve costs are proven by the book itself; the twelfth — the supposed rights dispute — has no evidence behind it at all. Read the ledger →

How to prove it — the honest test

The whole case reduces to one question. Bacon’s biliteral cipher hides a message by setting each letter of the ordinary text in one of two subtly different typefaces — an a-form or a b-form — five of them spelling out each hidden letter. Sorting the Folio’s actual type into those two classes is called the classification. So: is the classification forced by the type on the page, or chosen by the person reading it?

The biliteral signature is 14 letters. Bacon’s cipher spends 5 a/b type-classifications per letter, so the name rides on the first 70 classifications of the Prologue. Whether that name is a miracle or a triviality depends entirely on where those 70 calls come from:

1 in 2 × 10¹⁹
If the a/b calls are FORCED by the type (objective, reader-independent), the odds of the 70 bits spelling this exact 14-letter name by chance are 24⁻¹⁴ ≈ 5 × 10⁻²⁰. Effectively impossible — so the name would be real design.
1 in 1
If the a/b calls are CHOSEN by the analyst, the classification is a free parameter and can be tuned to spell anything. Then the name proves nothing. This was the core of Friedman’s late (1957) critique — though he had spent the start of his career inside Gallup’s own Baconian project, a reversal A. Phoenix documents.
The decisive experiment is therefore not more decoding — it is inter-rater reproducibility: give independent readers the Bodleian First Folio scans of the Troilus Prologue and have each classify the letters as a-form or b-form blind. If they agree at high rates and the agreed classification yields the name, the 1-in-10¹⁹ column wins and the signature stands. If they diverge, the classification is doing the work, not Bacon. This is objective, pre-registrable, and needs no one’s permission.

Note on the word-cipher side: orifex‘s status is already settled — it occurs once in the Folio and in only one other book in the 29-work corpus (Marlowe, 1590). That is a countable fact, not a judgment call. The biliteral key is the half that still turns on the reproducibility test above.

Verdict

A real convergence, with one measurable question left

Strong and objective: the play both ciphers select — Troilus & Cressida — and its Folio anomalies; and the word-cipher keyhole, orifex, whose rarity is a hard count.

Contingent on one test: the biliteral key, FRANCIS ST ALBAN. Its evidential weight is entirely a function of whether the a/b classification is reproducible from the type. That is measurable — and until it is measured, the honest claim is: two ciphers point at one door, the keyhole is confirmed, and the key awaits a blind reproducibility test that would settle it either way.

Send this to someone

Pass this page on by link, by email, or by pointing a phone camera at the code.

References & Sources →

Research by Walter H. Wilkinson · gorhambury.org

← Back to the Word Cipher Outreach Set

Scroll to Top