Gap analysis and syntax
Victor H. Yngve · IEEE Transactions on Information Theory · 1956
A statistical procedure has been tried as a method of investigating the structure of language with the aid of data processing machines. The frequency of gaps of various lengths between occurrences of two specified words is counted. The results are compared with what would be expected if the occurrences of the two words were statistically independent. Deviations from the expected number give clues to the constraints that operate between words in a language.