ATG GAA GAA GAA TAT CGT TAT ATT CCT CCT CCT CAA CAA CAA

This produced fourteen trios, or seven unique varieties: GAA, TAT, CGT, ATT, CCT, and CAA. The charts told him that an English phrase of fourteen letters contained an average of nine different letters. Not far off from the seven he had.

Ando immediately noticed that there was a lot of overlap produced by this system. GAA, CCT, and CAA each occurred three times, and TAT appeared twice. But what really bothered Ando was the fact that GAA, CCT, and CAA each appeared three times in a row. If he assigned each triplet a single letter of the alphabet, there were three separate cases in this short passage of the same letter being repeated three times. He knew enough English to know that double letters were not at all uncommon. But he couldn't think of any English words with triple letters. The only possibility he could think of was situations in which one word ended with a double letter and the next word began with the same letter, e.g., "too old" or "will link".

He picked up an English book he happened to spy nearby and started examining a page at random to see just how often the same letter occurred three times in succession. He'd gone through four or five pages before he found a single instance. The chances of it happening three times in one fourteen-letter sequence were basically nil, he concluded. By contrast, dividing up the forty-two letters into pairs produced just one double letter. As a result, he decided that statistically it made more sense to go with the first option and divide the bases into pairs of letters.

He'd narrowed down the possibilities. From here he could proceed through trial and error.

The AA pair appeared four times, which meant it must correspond to a letter used with great frequency. Consulting another chart, Ando confirmed that the most frequently used letter in English is, of course, E. So he hypothesized that AA stood for the letter E. The second most common pairs in his sequence were TA and TC, occurring three times each. He also noticed that AA was followed by TA once, while TC was followed by AA once. This might be important, since there were also statistics for various combinations of letters. He started trying out various possibilities for TA and TC, constantly referring to his charts.

As far as letters which often follow the letter E and which are also common in and of themselves, the letter A seemed like the best candidate, which meant that TA could stand for A. By the same logic, he thought that TC might correspond to the letter T. Further, by the way it combined with other letters, he guessed that CC might be N. Thus far the statistics seemed to be serving him well. At least, he hadn't run into any problems.

This is what he had:

E_ _EAT_AA_NT_ NTE_ _E

What had once seemed a random jumble of letters now seemed to be taking on the aura of English. Next he tried filling in the blanks based on what he knew of consonant-vowel combinations, always consulting the charts.

Date: 2015-12-24; view: 681