Gloomy Vocabulary in Books Reflects Poor Economic Climate (Op-Ed)

Mood expressed in English language books was measured by recording the frequency of words expressing unhappiness across a database of more than eight million digitised books published in the past 100 years.

This “literary misery index” was found to have a strong correlation to the annual US economic misery index, which is the sum of inflation and unemployment rates.

The researchers found that some periods, particularly the years following World War I, the aftermath of the Great Depression and the 1975 energy crisis, were clearly marked by “literary misery”. The results seemed to follow a pattern of Western economic history, shifted forward by a decade.

The team began to look into this field in a paper published last year. Alberto Acerbi, who was involved in the research, explained the first paper “demonstrated a new methodology. It showed that some clear trends can be extracted from the mass-analysis of digitised books. Now, we are starting to try to explain these trends”.

“Our results show that a robust correlation exists between economic mood, and social mood as expressed in books,” he said. “Given the size of the database we used, one can think about it as a sort of barometer of the general mood”.

The research adds a new dimension to the recent flurry of research into how emotive language online can be used to assess and predict wide socio-political and economic trends. Previous studies have investigated how large samples of language on social media and web search engines can be used to statistically predict future consumer activity, the stock market, and voting intention.

However, by taking this premise and applying it to past literature, Alex Bentley and colleagues' paper suggests that a much larger range of writing can be opened up to such analysis.

Following the recent mass digitisation of millions of past books, similar research into public feeling in a variety of subjects could be expanded to include a past before the internet.

This would allow cultural trends and moods to be tracked over a much larger span of time than is currently possible.

“We are in the era where non-computer scientists, for example social scientists, are able to work with large-scale inputs of data and test hypotheses that in the past was infeasible,” Vasileios Lampos, who was also involved in the study said.

However, Josh Cohen, Professor in Literary Theory at Goldsmith’s, remains doubtful as to the efficacy of these claims.

“There is the crudest kind of causal determinism at work here, the kind that’s been eliminated from social and political theory as much as from literary studies,” he said.

“The most dubious implication of all is that the ‘misery words’ somehow signify the same things in the same way across all literary texts. Without even the most rudimentary reference to how words signify, their bald presence is almost completely meaningless.”

“The number of texts must be so vast, the variety of concerns so diverse, I just can’t credit the idea that all these different uses of ‘negative emotional’ vocabulary mean the same thing and can be used to substantiate the same claim.”

Of course, an enormous quantity of factors differentiate literature from the kind of word samples used in online language analysis. This means that similar forays into the past may not be so imminent. However, the study certainly reveals the new directions that are offered by this kind of mass literary analysis.

This article was originally published at The Conversation. Read the original article. The views expressed are those of the author and do not necessarily reflect the views of the publisher. This version of the article was originally published on LiveScience.