## Cryptography

#### The History and Mathematics of Codes and Code Breaking

After the execution of Mary Queen of Scots but prior to the development of more complex ciphers, like the Vigenère cipher, “anybody sending an encrypted message had to accept that an expert enemy codebreaker might intercept and decipher their most precious secrets.” (Singh, p. 45) Because of this, it is safe to assume people writing the encrypted messages would still be very careful with what they were actually saying, and how the things that they wrote could potentially incriminate them. Messages would have been written in vague enough language that even if the text were to be deciphered, the cryptanalysts would either not be able to tell what the exact plan was or, even if they could figure it out, would not be obvious enough to be used in a court of law. (I don’t know how the rules for this worked in that time but today people can claim the way a message is interpreted is completely wrong.)

Going off in a completely different direction now, as I was reading I was thinking of different types of ciphers. When I the name Louis XIV, I fixated on the Roman Numerals. It made me curious about the use of Roman Numerals in cipher alphabets. I think it would add a layer of complexity because the cryptanalysts would have to try to figure out which combinations of letters would represent a number, and then which letter was represented by that number. One weakness, however, would be that it would be very easily identifiable as Roman Numerals because of the small number of letters used in Roman Numerals. I tried to look up Roman Numeral ciphers but nothing came up on a quick Google search.

Previous

Next

## 1 Comment

1. #### Sarah Giordano

I do agree that Roman numerals would add a layer of complexity to a code. Like you said, however, it doesn’t seem feasible, however, that an entire code be created out of 26 different Roman numerals. What if instead of this, Roman numerals were inserted into the plaintext in place of very common words? ‘I’ could become “the”, and “VI” could be “is”, with similar ideas for other selected words and roman numerals. Then, these roman numerals could be encrypted using a monoalphabetic substitution or a homophonic substitution. In the monoalphabetic substitution, any frequency analysis would be thrown off because there would be a good deal more I’s, V’s, and X’s. In addition, it would be harder to find a word to kick of the decryption process; I know that for our second problem set what gave me a starting point was the repetition of the pattern “YGU” which became “the”. With roman numerals, it would be much more difficult to find such a pattern.
Furthermore, if you went an extra step and used a homophonic substitution cipher, it would be just that little bit more secure. In his attempted to break the Great Cipher, Bazeries initially attempted to use frequency analysis on pairs of letters and then used these frequencies in correlation with the most common French digraphs. Even though this didn’t work on the Great Cipher, that could be a way that people would try to break the cipher described here. However, they should be thrown off because the digraphs of I before X or X before I would be more common than other digraphs, leading them to assume it is a more common digraph like “th” or “as”. It would just be much more difficult because nobody thinks of I coming before or after X as something that is typical, so it would be harder to pin down which symbols all were for I.