clue to the labyrinth: Francis Bacon and the decryption of nature, The

Cryptologia, Jul 2000 by Pesic, Peter

These tables resemble the tables used for encipherment and decipherment, though not for natural language but rather "things themselves."30 A cryptanalyst examines the possible correlations between the appearances of certain letters in the cipher text, singly or by pairs or triplets, arranging the results in tabular form. Read negatively, this table also shows which ciphered letters are not correlated with which others. Other tables note the order in which letters are correlated, preceding or following others. Likewise, Bacon's tables marshal parallel data for heat, citing all the known correlations, exclusions, and the degrees thereof (see Figure 4). The Latin word tabula was used extensively for tables of encipherment already in the first printed work on cryptography, Trithemius' Polygraphia (1518). From the earliest sources on, cryptography had relied on such tabular arrays to give the visible key for the encipherment. The word "tabula" or the French "tableau" are already common usage in Porta and Vigenere. By the time of the great Antoine Rossignol (1599-1682) there are distinguished "tables a chiffrer" from "tables a dechiffrer," indicating the greater systematization of deciphering as well as of enciphering. Given Bacon's detailed knowledge, it seems very likely that either he himself tried cryptanalysis, saw work in progress, or heard accounts of it. His posing of a new, more secure cipher shows that he was fully aware of the powers of expert cryptanalysts and, quite likely, of their detailed methods. At the very least, he seems to know of the tables of earlier cryptographers, for he sets out his own biliteral cipher in tabular form (4.445-6; see Figure 3).

The cryptanalyst's tables are a necessary starting point for systematic decryption. They permit certain deeply embedded linguistic features (such as the frequency of the letter "e" in English) to emerge and lead to solution. Bacon's tables proceed by the same logical categories of inclusion and exclusion, of quantity and correlation, that give the cryptanalyst's tables their revelatory power. Far from being undifferentiated lists of instances, his carefully structured tables are intended to be the key to full decryption. In the case of heat, Bacon tries a "First Vintage" or "Commencement of Interpretation." He tries to gather together "all the instances in which the thing itself is to be found" to give a first working hypothesis which he can then test and vex further (4.149).31 He also notes "Striking or Shining Instances," which give paradigmatic examples that are particularly forceful instances of the preliminary hypothesis. Such devices also are tools of the cryptanalyst, for whom the tables are only a beginning to the real work. The tables, after all, must be read, with all the interpretative acuity that word might invoke. Telling passages in the cipher text must be located and probed; hypotheses need to be formed and tested, even if finally discarded, in order that correct order emerge.