Though there’s probably no perfect way, the recent research mined keywords generated by users of the website the Internet Movie Database (IMDb), which contains descriptions of more than 2 million films. When summarizing plots, people on the site are prompted to use keywords that have been used to describe previous movies, yielding tags that characterize particular genres (cult-film), locations (manhattan-new-york), or story elements (tied-to-a-chair). Each keyword was given a score based on its rarity when compared to previous work. If some particular plot point – like, say, beautiful-woman – had appeared in many movies that preceded a particular film, it was given a low novelty value. But a new element – perhaps martial-arts, which appeared infrequently in films before the ’60s – was given a high novelty score when it first showed up. The scores ranged from zero to one, with the least novel being zero. Lining up the scores chronologically showed the evolution of film culture and plots over time. The results appeared Sept. 26 in Nature Scientific Reports.

...Unsurprisingly, the research also suggests that unfamiliar combinations of themes or plots that haven’t been encountered before (something like sci-fi-western) often have the highest novelty scores. “I think this reinforces this idea that novelty is often variations on a theme,” said Sreenivasan. “You use familiar elements broadly, and then combine them in novel ways.”

Sreenivasan’s analysis shows trends within particular genres as well. Action movies are essentially redefined in 1962 with the release of the first James Bond movie. Science-fiction films, on the other hand, show no similar creative uptick during the same period. According to the analysis, novelty in sci-fi has declined essentially since the genre first made it into movies. It’s possible that this has to do with early science-fiction films codifying the major tropes seen in these movies.

Another part of the analysis seem to correspond to theories put forth by social scientists about how much we enjoy novelty in creative works, said Sreenivasan. In general, humans enjoy new things. More specifically, there’s a tendency for people to look at and like things that are new but not too new. “If it’s way out there, it’s hard to palate,” said Sreenivasan. “And if it’s too familiar, then it seems boring.”

A model known as the Wundt-Berlyne curve illustrates this result. The amount of pleasure someone derives from a creative piece goes up as its novelty increases. But at a certain point, there is a maximum of enjoyment. After that, something becomes too unfamiliar to stomach anymore. Using the revenue generated by different films as a measure of its mass appeal, Sreenivasan found that more novel films sold more tickets until they reached a score of about 0.8. Afterwards, they appeared to decline in popularity and revenue.

(From the standard errors & shuffled results, the decline in revenue from 0.8 to 1.0 happens very fast, so one probably wants to undershoot novelty and avoid the catastrophic risk of overshoot.)