Mining Books To Map Emotions Through A Century

Were people happier in the 1950s than they are today? Or were they more frustrated, repressed and sad? 

To find out, you’d have to compare the emotions of one generation to another. British anthropologists think they may have found the answer — embedded in literature. 

Several years ago, more or less on a lark, a group of researchers from England used a computer program to analyze the emotional content of books from every year of the 20th century — close to a billion words in millions of books. 

This effort began simply with lists of “emotion” words: 146 different words that connote anger; 92 words for fear; 224 for joy; 115 for sadness; 30 for disgust; and 41 words for surprise. All were from standardized word lists used in linguistic research. 

But Alex Bentley, an anthropologist at the University of Bristol involved in the research, says no one expected much when they set their computers to search through one hundred years of books that had been digitized by Google. 

“We didn’t really expect to find anything,” he says. “We were just curious. We really expected the use of emotion words to be constant through time.” 

Instead, in the study they published in the journal PLOS ONE, the anthropologists found very distinct peaks and valleys, Bently says. “The clarity of some of the patterns was surprising to all of us, I think.” 

With the graphs spread out in front of him, Bentley says, the patterns are easy to see. “The ’20s were the highest peak of joy-related words that we see,” he says. “They really were roaring.” 

But then came 1941, which, of course, marked the beginning of America’s entry into World War II. It doesn’t take a historian to see that peaks and valleys like these roughly mirror the major economic and social events of the century.

“In 1941, sadness is at its peak,” Bentley says.