The document discusses entropy and redundancy in the English language. It provides examples of how rules of grammar, parts of speech, and the inability to make up words make English a redundant language. Shannon calculated the entropy of English using N-grams and by taking the weighted average entropy of words. His calculations estimated the redundancy of English at around 50%. Shannon also had subjects guess letters in phrases to measure redundancy, with subjects able to correctly guess around 69% of letters, suggesting a similar level of redundancy. Statistically measuring redundancy has practical applications like data compression algorithms that take advantage of the non-random nature of English.