“Etaoin srhldcu” may read like nonsense to most English speakers upon first blush, but as it turns out, the combination is quite significant. It represents, in order, the most used letters in the English language, according to a new survey of 743 billion words conducted by Google’s head of research Peter Norvig.
The survey, which was publicized by Google Research on Monday, was an update to the seminal 1965 survey of some 20,000 words gathered from a variety of printed sources — books, magazines, newspapers — conducted by Mark Mayzner, a former Bell Labs researcher.
Mayzner’s survey involved a lengthy and painstaking process of identifying each word occurrence and transferring it over to Hollerith (IBM) punch cards and running them through a sorter.
Mayzner recently contacted Google’s Norvig via email to see if Norvig was interesting in repeating the experiment using Google’s much more voluminous English language database — the entire Google Books collection of scanned English volumes. Norvig accepted the challenge. Using the Google Books Ngram viewer (which shows word popularity over time), Norvig created a new dataset of some 97,565 unique words, collectively repeated 743.8 billion times, which he noted on his blog is 37 million more occurrences than the 20,000-word sample that Mayzner assembled. Norvig’s sample also included over 3 trillion individual letters.