[T]he study’s authors compiled scripts from 1,000 movies, and identified, using IMDb, Google and Bing, the lines from those movies that are quoted and remembered the most today. They paired each memorable line with other lines of similar length, voiced by the same character in the same scene, then ran each pair through a computer program to identify linguistic trends.
“Our goal is to understand what general features of language will tell you how memorable a quote will be,” fifth-year Ph.D. student Cristian Danescu-Niculescu-Mizil, one of the study’s authors, told the Huffington Post.
The features that emerged were widespread applicability, straightforward syntax and distinctive diction.
If you think this brings us a step closer to computer-generated movie scripts, you’re on the same wavelength as Danescu-Niculescu-Mizil. Not that it could make them any more formulaic than much of what is currently produced in Hollywood.