BNF Fonds Italien 1595 ff. 441r-442r
Translated by Philip Neal
Pavia Monday 4 July 1474
A treatise, without examples, on the extracting of ciphered writings
First be it determined if the writing be in Latin or the vulgar tongue, which shall be known in this manner: consider if the words of the published writing  have but five divers terminations or less or more. If there shall be but five or less, guess that published writing to be in the vulgar tongue; and that because all words common in our idiom are terminated by a vowel, and vowels there be five, a e i o u. Now if the words of the published writing shall have more terminations than five, guess that published writing to be in Latin; and that because words of Latin or of Letters  may be terminated by a vowel, a semi-vowel, which is to say l m n r s x, and a mute, which is to say b c d f g k q p t.
Another rule to determine the same, if the published writing be in the vulgar tongue or else in Latin or Letters. Inspect the published writing, if there be a multiplicity or frequency of words represented by but one cipher; for if it be so, like it is that the published writing is in the vulgar tongue; and that because in the vulgar tongue there be words abounding represented by one cipher only, but that rarely in Latin or in Letters, for in Latin there be no words represented by one only letter or cipher  saving four words, that is to say e a prepositional, o the adverb vocative and i the verb imperative; which monoliteral words seldom are used in Letters saving a prepositional.
Item. Consider if in the published writing there be a multiplicity or frequency of words of two or three letters or ciphers, and if so consider the published writing to be in the vulgar tongue; and that because the like words abound more in the vulgar tongue than in the Latin.
Therefore after it has been opened to you whether the published writing be in the vulgar tongue or else in Latin or Letters, if it be in the vulgar tongue, already you are certain which ciphers represent your vowels, howbeit not determinately but confusedly , and that because ciphers which be at the end of words are ever vowels in the idiom of all Italy; furthermore, when the vowels are known to you in that manner just now said, consider out of the ciphers falling at the end of words which cipher most frequently falls in words monosyllabic and of one letter or of one cipher only, for possible it is and like enough that such a cipher is representative of e; and that because est the verb substantive (and by consequence e) is much used and repeated in vulgar writings and in the same manner so is the copulative conjunction (and by consequence e). 
Item. Words of two ciphers only are much to be considered in vulgar writings, for many of them begin with l; and that because of the articles which are preposed to names appellative, which are lo and la in the singular and li and le in the plural.
Item. Words of three letters only are to be considered, if it be that any such repairs often in the published writing; and that because the word che repairs often in vulgar writings.
But if it has appeared to you that the published writing be in Latin and not in the vulgar tongue, consider then the ciphers which are in the ends of words, which of them repair the most, for belike such be vowels, or s, or m, or else t; and that because the most part of Latin words conclude either in a vowel, or in s, or in m, or in t, and few of them end in a mute unless it be in t, saving ab ad and quod, which abound enough in Letters.
Another rule. Consider the published writing, if there be in it any word represented by one sole cipher, and conjecture that cipher to be a, for in writings which are in Latin but rarely are there found words of one letter only, saving a prepositional as was said before.
Another rule. Consider ciphers which are in the ends of words, which ciphers, as has been said, are frequently representative of one of the vowels, or s and m, or t, and see if you shall find any of them in words of one or two ciphers, for if it be in a word of one cipher then such a cipher is representative of one of the vowels, for no word, nor syllable on that account, can be without a vowel, and that vowel shall be a or e or i or o; but likest it shall be a prepositional as has been said before. Or if it be in a word of two ciphers then let run through your mind all words of two letters only and mostly them which most frequently are found in Letters, such as et ut ad si me te and se. And that it not escape you, I give here all words of two letters only or the most part of them, and they be ab ac ad an and at; da de and do; ea ei eo et ex and es; he hi id ii in ir is and it; me mi na ne and ni; ob os re se and si; tu te ue ui and ut.
Another rule. Consider if in the published writing there be words of three letters only whereof the first be the same one as the third or like to it, such as be ala, ama, ara, ede, eme, ere, ehe, ixi, iui.
Another rule. Consider if in the published writing there be any word or words wherein is a letter triplicate without interval, for such a cipher represents u, as uuula.
Another rule. Consider if in the published writing there be any geminate cipher and that most of all in words of four ciphers, for such a cipher is likely representative of l or s, which are most frequently geminate, as esse and ille.
Another rule and the last, common to vulgar writings and Latin both. Consider if in the published writing there be any cipher which always and everywhere is followed on by one and the same cipher, for such a cipher is representative of q, and the other following is representative of u, for always after q follows u, and the cipher which follows on the cipher representative of u is a vowel always, for always after q follows u and another vowel follows after u.
However these rules which we give can be decluded in many ways, as by ciphering a writing partly in the vulgar and partly in Latin; again by interposition and apposition of ciphers representative of no letters into a writing, and that mostly into words of one or else two or three ciphers or letters; again by writing with ciphers of two utterly diverse alphabets; again by putting one sole letter in place of q and u.
 litterae propositae: the ‘exposed letters’, i.e. the ciphertext. [^]
 litterae: in this context presumably literary or legal Italian. [^]
 littera: letter of plaintext, zifra: letter of ciphertext. [^]
 non terminate tamen sed confuse: i.e. you know the vowels but not which is which. [^]
 i.e. e is very common in Italian because it corresponds to est and et in Latin. [^]