Cyclopædia of Political Science, Political Economy, and the Political History of the United States
CIPHER DISPATCHES AND DECIPHERMENT. The art of deciphering secret manuscripts has at all times played an important part in political matters. If we are to believe Comiers d'Embrun, the Hebrews were acquainted with cryptography or the art of using ciphers, therefore with decipherment which is its immediate consequence. The early Christians, according to P. Alex. de Rhodes, used convention signs to keep a knowledge of their affairs from the curiosity of their persecutors.
—Suetonius informs us in his life of the first Cæsars that the emperors wrote to their generals and confidants, transposing the letters of the alphabet. If this be the case, it is quite possible that Julius Cæsar invented the system of ciphers which bears his name. The following is a sketch of the system: A number of conventional signs are made to correspond to the letters of the alphabet, or, better still, with these same letters arranged in a different order, for example:
Thus if we wish to write in cipher, "Put a vase of roses on your balcony; I shall see that it is time to march;" we should write: "bgf 1 hlep aq daepe az jagd mlxnazj; t eslxx epp fslf tf te ftyp fa yldns." The same result will be obtained by conventional signs corresponding to the letters of the alphabet. It is needless to add that in the first case the letters in changed order, in the second the conventional signs, form the secret of the cipher, which should always be known by the person who wishes to use this mode of writing.
—The Japanese and Chinese were instrumental in causing a further development in cryptography without knowing it. The signs used in the writing of these people run vertically from above downward and from below upward; from this originated the idea of the Japanese method, as calculated to puzzle Europeans, whose writing is horizontal.
—Here is as example of this method. "The liberty of man imposes on him as many duties as it gives him rights."
These are simply the letters used in writing the above sentence, placed in order after each other from above downward and from below upward. This cipher becomes, as we see, mere child's play unless it is complicated by some combination.
—In the steganography which he published, Scott considers the method of count Gronsfeld and that of lord Bacon as undecipherable by any one who has not the key to it. Of these two systems we shall take that of lord Bacon, whose originality strikes us, while at the same time we acknowledge its tediousness of execution. The following is its conventional alphabet, each letter of which answers respectively to the ordinary letters of the alphabet:
If we wished to write this sentence: "The enemy is there, be on your guard," we would do it as follows: "abbbb baaba abaaa abaaa bbbab abaaa bbbaa abaab babaa babbb abbbb baaba abaaa babba abaaa, aaaab abaaa bbbba bbbab abaab bbbba aabbb babba baaab aabbb aaaaa babba aabaa."
—This style of ciphering is surely original. Besides, it has this advantage, that it can be carried on under the appearance of a document not having any importance, the letters of which, marked or not by a conventional sign, indicate whether they represent a or b; thence a grouping by fives, the translation of which becomes easy by the aid of the alphabet. This system has a certain resemblance to that attributed to Julius Cæesar, differing from it, however, essentially by the repetition and mingling of the letters a and b; this mingling is well adapted to puzzle the most careful research.
—It is not necessary to say that other signs may be substituted for the letters of the alphabet used in the method we have just mentioned, such for instance, as the Arabic numerals.
—Secret writings being of frequent use in political life, it is often important to be able to find the secret of the cipher which hides the thoughts or projects of its author. We can not here give the reader a complete treatise on cryptography; we must confine ourselves to placing at his disposition certain linguistic observations which become so many means of arriving at the understanding of the greater part of secret writing; the rest is an affair of sagacity and patience.
—Remarks on the construction of words. The German Language. The only letter standing alone is o; monosyllables are very rare; double letters quite frequent at the end of words: the e is often repeated especially in long words; i always in the middle of words of three letters; ck most frequent at the end of words; sch united to l, m, n; r united to e and a to bb in the middle of words; t united to ff; oth, ich, very often at the end; ch frequent; b, l, g, k, p, q, x, z, the rarest of consonants.
—English. The words of one letter are I, a, O; y often appears as final; o doubles and shares this peculiarity with e, from which it will be easily distinguished if attention be paid to the fact that it is always united to f in the word of two letters of; it is often found also with w; e is distinguished in cipher dispatches from double consonants, because it is repeated oftener than any other letter.
—Italian and Spanish. Italian has a strong resemblance to Spanish, but is distinguished from it by the length of certain words and by the frequency of double letters in the middle of words. O is oftener repeated than any other letter; e, i, next, the latter sometimes doubled, the same as o, u. In Spanish, o is very often followed by s; u by e; but the latter principally in the middle of words, the other chiefly at the end; the words of one letter are a, o, y.
—French. French words end most frequently in e, which is often followed by s or nt; ou is met in words of four syllables; the vowels, especially e, are repeated oftener than any other letter; there is no word without a vowel; a word of one letter is always a vowel, or a consonant with an apostrophe; q is always followed by u.
—It is evident that the process of deciphering without the key differs somewhat in each language, varying with its characteristics.
—Let us try to decipher the sentence: "bgf l hlep aq daepe az jagd mlxnazj; t eslxx epp fslf tf te ftyp fa yldns," according to Cæsar's method. We first make a list of the letters in the sentence, noting how often each recurs, as follows: b, once; g, twice; f, 6 times; l, 6 times; h, once; e, 6 times; p, 5 times; a, 5 times; q, once; j, twice; m, once; x, 3 times; n, twice; z, twice; t, 4 times; s, 3 times; y, twice; d, 3 times. Remembering that a and I are the only words (except O which would be rarely used in a cipher dispatch) of one letter in English, we know that l and t must be these two. Suppose l is a, then t is i; t being i, f must be n, s, or t in tf because these are the only letters which can follow i in a word of two letters except f which is the cipher letter. We shall call it t; e in te may be s; a in fa must be o, for only o can follow t in a word of two letters as may be seen by experiment; pp in epp is ee for e and o are the only vowels which can be doubled, and oo could not form a word after s. Now we know that l, t, f, e, a, and p are a, i, t, s, o and e. After making the changes the sentence becomes bgt a hase oq doses oz jogd maxnoxj i ssaxx see tsat it is tiye to yadus. In the word ssaxx it is most probable the xx is ll, for a double vowel could not follow a, and by using the ll and changing the second s to h we have shall instead of ssaxx, and that instead of tsat. In tiye y could in any connection be only d, l, m, n or r; here it is clearly m which gives time: q in oq whatever be the connection would be either f, n, r or x; here we shall use f: and the z in oz is either r or n; call it n. In the first word bgt g must be the vowel u, for we know the others; b is either c or p, let us use c; we have now cut instead of bgt and the word jogd becomes joud. Now d can not be n or t, for we know these; it is r, for this is the only letter left that could follow ou here; j is either p or y; let us take y. We have now your. Knowing or supposing that:
we make the substitutions and the sentence stands: "Cut a hase of roses on your malcony; I shall see that it is time to marnh." It is clear that h in the third word stands for v, that malcony is balcony, and marnh march; n being c, b can not be c, therefore the first word is put, not cut.
—If we wished to send the following dispatch in French; Placez un vase de fleurs sur votre fenêtre; nous saurons qu'il est temps de se mettre en marche, we would write it in this way according to the first cipher alphabet given above: "Bxlnpk gz hlep op qxpgde egd hafdp qpzpfdp zage elgdaze cgtx pof fpybe op ep ypffdp pz yldnsp."
—To decipher this we proceed as follows, remarking only that the reader, by study and experience, might find some other way to decipher it. We make a list of the letters contained in the dispatch, noting the number of times each letter is repeated, thus: b, twice; x, 3 times; l, 4 times; n, twice; p, 16 times; k, once; g, 6 times; z, 5 times; h, twice; e, 9 times; o, twice; q, twice; d, 7 times; a, 3 times; f, 6 times; c, once, t, once; y, 3 times; s, once. We may remark that p occurs 16 times, hence it is very likely a vowel, and probably the letter e. Let us see whether this hypothesis is well founded, and take the shortest words: op, ep, pz. If p is an e, let us suppose, to shorten our demonstration, that pz represents en, since it is one of the French words of two letters beginning with e; then z is an n; op, ep, will therefore be one of the words ce, di, je, le, me, te, se, the only words of two letters in French ending in e. But in the dispatch we see op, immediately preceding ep which has only two letters and de is the only word that can go before a word of two letters: therefore, if p is an e in the word op, o, must be a d.
—Let us now try to find out what e is in the word ep. It is not a vowel. As e is found at the end of many words in the dispatch, we may presume that it is an s; ep would therefore stand for se until proof to the contrary. Gz is another word of two letters, the last letter of which we know to stand for n, whence we conclude that g can be only a vowel, and consequently gz must be one of the following words, an, in, on, un, the only French words of one syllable ending in n. We may reject in which means nothing; in French, an never occurs after a word so long as bxlnpk; therefore, there are left only on and un. G is therefore either o or u; but which? We can not yet decide. Let us take a word in which it occurs, and select, as far as we can, a word which contains the letters presumed to be known. Say the zage. Substituting in this word the letters already known, we have naus; naus has a close resemblance to nous; we may conclude that g is u, and we may also add that a is o.
—Let us take the little word egd: the third letter d is either a cor an r, for only sur or suc can have this form. We shall see this further on. In pef, the first two letters are es; we are led to think that the last is a t, for the word est is the only word of three letters in French beginning with es.
—Let us now take longer words, say hafdp. We know afdp which suggests otceor otre, but only the latter can be taken. Then the word becomes h...otre which has only two analogues in French, notre and votre. Now, as we have seen, n is represented by z, therefore h is necessarily a v. We may therefore add to the foregoing letters, instead of f, d, h, their values t, r, v.
—In the word ypffdp, y is the only unknown letter. Substituting the equivalents already known, we have y...ettre, which can mean only mettreor lettre: after se "lettre" does not mean anything. Everything leads us to believe that y is an m.
—The word we have just deciphered almost translates qpzpfdp, for we find q...enêtre or fenêtre; in like manner qxpgde becomes fxeurs or fleurs; in like manner also fpybe stands for tembs or temps.
—We have deciphered the letters p, y, o, d, f, z, a, e, g, q, b, h, x; with this key let us try to decipher the whole dispatch. By the substitution of known for unknown values we find pllnek un rlse de fleurs sur votre fenêtre, nous slurons cutl est temps de se mettre en marnse. This is not so obscure as it may seem at first. With a little study it becomes apparent that l, c, t, stand for a, q, i. And thus we have: nous saurons qu'il est temps. But pllnek un vase de fleurs stands, a priori, for placez, etc. No great effort of imagination is required to translate "se mettre en marnse," into se mettre en marche.
—What we have just explained will suffice to show the reader the possibility of deciphering. This methods is without doubt purely empirical, but it appears to have been sufficient for the cases presented up to our day.
—To prevent the deciphering of dispatches, combinations of every kind have been imagined, books of signs, even special dictionaries have been published by means of which the secrecy of a dispatch is guaranteed. One of these dictionaries, which seems well planned, has appeared in Paris, published by Berger-Levrault & Co., under the title of Dictionnaire pour la Correspondance télégraphique secrète, by a secretary of legation.*56
Notes for this chapter
End of Notes
Return to top