THE UNRAVELMENT OP CIPHERS.
It would be a poor conjuror who would do nothing but set tangled
skeins for his audience, and not fully reveal the means by which they
were deluded, in a book offering to lay bare the art.
We cannot give the reader the patience often necessary to puzzle out
a cipher letter, but we hope he will have the requisite sagacity; and,
besides the secrets already illumined, we add the general rules for the
interpretation of an intercepted message in secret writing.
The difficulties are greatest when one is unaware of the language in
which the despatch is couched: or if it may not be worded in various
tongues: or by words formed from several alphabets; or when the
superfluous words are numerous and artfully interspersed; when the same
letters, words, and syllables are expressed by dissimilar signs ; when
the words are written without breaks, or when they are arbitrarily
separated.
The first step is to make out a list of the signs of the cipher, and
note the frequency of their repetition severally. Then their
combinations are observed, and they are turned and twisted in every
way, until conjectures centre on some certain attribute of this or that
character.
To this end, the majority of the tokens must appear more than once
in the cipher; if the piece is short, and the same letter is
represented variously, the problem becomes most serious. Words composed
of a few syllables, or, better, a few letters, ought to be the first to
be taken in hand. They let the vowels be divined, and once they are
picked out, the discovery of the consonants is facilitated. The
mainstay is the exact knowledge of general orthographic principles.
Gravesand gives an example of unfolding the mystery of a cipher in Latin.
The punctuation marks, and capital letters are not in the original,
but placed by us to assist in our elucidation. This cipher gives 14 f,
14 i, 12 b, 11 e, 10 g, 9 c, 8 h, 8 k, 5 m, 4 a, 3 d, 2 b, 2 n, 2 p, 1
each o, q, r, s, t; in all, nineteen characters, of which only five are
repeated. I see at first glance that hikf is repeated at M after appearance at B; that ikf is apparent but once at F. Lastly, that hekf, at C, and hikf, at
B and M, have some connection between them. It may, therefore, be
concluded that they are possibly the end of words, which I indicate by
the colons. In Latin, it is usual to find the ante-penultimates alone
different in words, or at the four last letters, habitually vowels, as
in amant, legunt, docent, &c, therefore i and e are presumably vowels.
Since fmf (see G) is the commencement of a word, it is not unreasonable to take m or f for a vowel, since there are never three vowels consecutive with two of them alike; and it is more likely the f, forasmuch as that sign appears fourteen times against the five times of m, which latter may be pronounced a consonant.
Thence going to K, we attack gbfbcbg, using the same reasoning to call c a consonant, (f being a vowel) in the bfb; it follows that c must be a vowel, on account of its position in bch.
In L, b is a consonant in gbgrb; and r much more likely so, from there being but one in all the writing. G is also a vowel.
In D, it is impossible to decide on fcgfg being a word or
portion of a word in five vowels. There is no Latin word presenting
such a peculiarity, so that we have been in error in taking fcg for vowels; f must have been wrong, but m is a vowel, and so is b (see K) : f is a consonant. In the place K, the vowel b appears three times separated by only one letter. In Latin, there are several words thus formed ; for instance, edere, legere, munere, si tibi, &c. Now, as the vowel e is the oftenest in use, let ns conclude that b is written for e, and r for i.
Thus working the whole problem out, we translate the entire cipher into—
"Perdita sunt bona; Mindarus interiit: urbs strata humi est," &c.
Contents
|
