Frequent feelings

In my research into emotions I’ve found it useful to look at Google search counts to gauge the relative popularity of terms like happiness, melancholy, and mood. Now Google provide n-grams, which show the occurrences of words (or any string of characters) during any year in the entire corpus of books that Google have scanned for online access — normalized as a percentage of the words in the database (for each year), with a timeline.

So you can see that melancholy gave way to the approximate synonym sad in English language books some time around 1850. There’s nothing special about that date as far as I can see; just the crossing of two trends.

Around 1600 melancholy was tied in with medical conditions. As everyone knows, it means black bile. Google also provides a dateline so you can access books in its database that use that term. First up in the period 1648—1781 is a scan of John Payne Collier’s An Antidote Against Melancholy: Made up in Pills, Compounded of Witty Ballads, Jovial Songs, and Merry Catches published in 1661 — mainly drinking songs as far as I can see.

Around melancholy’s peak there’s the book by Brother Democritus Junior (1871), The Anatomy of Melancholy, What it is, with all the Kinds, Causes, Symptoms, Prognostics and Several Cures of it. By the year 2000 melancholy is treated as a theme in history, or an intellectual theme, e.g. Jonathan Flatley’s Affective Mapping: Melancholia and the Politics of Modernism.

Unlike the word melancholy, sad is a term anyone could drop into their writing: to be sad about missing a train is to state a commonplace, whereas to be melancholic about it indicates a more thoughtful condition. There are lots of books now about sadness, mostly self help books it seems.

Enchanting words

I discovered n-grams when someone decried the dialogue in the costume drama Downton Abbey: “Dodgy! He just said dodgy. People never used that word in the 1920s.” Someone on Twitter had looked it up and linked to the n-gram. Dodgy was in use in the 1920s, but not nearly as common as it is now.

Here are some other words that only rose to prominence in and beyond the 1980s: smilies, pay per view, audience engagement, film noir and n-gram. Cybernetic peaked in about 1975, and is now on the wane. Deconstruction had its peak around 1995.

No doubt there’s room for caution in the use of such bibliometrics, but massive scope for scholars to match up events and word usage, and all kinds of analyses, e.g. there’s a paper by Saif Mohammad called “From once upon a time to happily ever after: tracking emotions in novels and fairy tales,” demonstrating that “fairy tales have a much wider range of emotion word densities than novels.”

So there’s something there about emotion and enchantment. Apropos the last few posts on enchantment, perhaps a magical, fairly-tale world is a world rich in feeling — or at least emotional language.

References

• Br Democritus, Junior. 1871. The Anatomy of Melancholy, What it is, with all the Kinds, Causes, Symptoms, Prognostics and Several Cures of it. New York: W.J. Widdleton
• Collier, John Payne. 1661. An Antodote Against Melancholy: Made up in Pills, Compounded of Witty Ballads, Jovial Songs, and Merry Catches. London: Mer. Melancholicus
• Flatley, Jonathan. 2008. Affective Mapping: Melancholia and the Politics of Modernism. Cambridge, MA: Harvard University Press
• Mohammad, Saif. 2011. From once upon a time to happily ever after: tracking emotions in novels and fairy tales. Proc. LaTeCH ’11 Proceedings of the 5th ACL-HLT Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities: 105-114. Stroudsburg, PA: Association for Computational Linguistics.

Notes

• See Google n-grams site. In the example above I compared sad (an adjective) with melancholy (a noun). Sadness (a noun) is in less common use than sad, but melancholy (a noun) is more common than melancholic (an adjective). That’s an interesting asymmetry.
• The new word selfie/selfy (taking a picture of yourself and friends from a camera at arm’s length) seems to have been around for years, but that’s mainly due to scans from old documents where self (felf) followed by a comma got scanned as selfy. It’s still not a common word, so the n-gram for selfy is really just noise.