How Data Science Pinpointed the Creepiest Word in “Macbeth”

From OneZero:

Macbeth is a creepy play.

Actors have long been superstitious about acting in it. That’s partly because performances have been riddled with accidents and fatalities; indeed, actors consider it bad luck to even utter the name of the play. (They call it “The Scottish Tragedy”.) And it’s partly because the basic substance of the plot is eldritch: You’ve got black magic, witches, a gore-flecked ghost and walking forests.

But fans of Macbeth often say its freaky qualities are deeper than just the plot devices and characters. For centuries, people been unsettled by the very language of the play.

Actors and critics have long remarked that when you read Macbeth out loud, it feels like your voice and mouth and brain are doing something ever so slightly wrong. There’s something subconsciously off about the sound of the play, and it spooks people. It’s as if Shakespeare somehow wove a tiny bit of creepiness into every single line. The literary scholar George Walton Williams described the “continuous sense of menace” and “horror” that pervades even seemingly innocuous scenes.

For centuries, Shakespeare fans and theater folk have wondered about this, but could never quite explain it.

Then a clever bit of data analysis in 2014 uncovered the reason. (The paper is here.)

It turns out that Macbeth uncanny flavor springs from the unusual way that Shakespeare deploys one particular word, over and over again.

That word?


. . . .

Initially, the scholars who did this analysis — Jonathan Hope and Michael Witmore — didn’t think to look at something as mundane as “the”.

They began by reading through scholarly work on the play, looking for various previous hypotheses. They pondered some of the more obvious weird things about how Shakespeare uses language in Macbeth, such as how witches speak in trochaic tetrameter (“Bubb-le | bubb-le, | toil and | trou-ble”). That rhythm is jarringly different from the train-chugging-along iambic pentameter of everyone else (“So fair | and foul | a day have | not seen.”). But the witches don’t appear onstage often enough for their odd prosody to contaminate the feel of the entire play.

Then Hope and Witmore moved on to another point that scholars historically made about Macbeth, which is that the play has a lot of repetition. The witches talk about battles “lost and won”, Duncan uses these precise words when he enters, too. When Lady Macbeth first greets Macbeth, she uses phrasings very similar to when the witches first greet Macbeth.

. . . .

So that made them wonder: Maybe they should do a word-frequency analysis of Macbeth. Perhaps it’d show the recurrence of certain words that would help identify the source of floating menace.

. . . .

So they did an analysis of the “log-likehood” of words in the play. “Log likelihood” is a metric of whether a word is used more or less often than normal. So they compared word-usage in Macbeth to Shakespeare’s overall writing. What words did he use in Macbeth more frequently than in his other plays?

The results are in the chart below. The higher the “Log likelihood” number, the more frequently Shakespeare used it in Macbeth …

Sure enough, you can see several words used with unusual frequency! This includes several we could justifiably call creepy — like “knock”, “cauldron”, “tryant”, “weird”, “trouble”, “dagger”, “fear”, and “horror”.

Cool. But the thing is, that still can’t really explain what’s going. Sure, Shakespeare used these words more often than normal — but they don’t occur so frequently that they’d change the entire mouthfeel of the play’s language. Nor is their overoccurrence terribly surprising. Of course “cauldron” would appear more often in a play with witches, and “thane” in a play about, well, a thane. “This result,” as Hope and Witmore noted drily, “is not very interesting … We hardly need computers and advanced statistics to tell us this.”

Ah, but then Hope and Witmore looked at the list again. And realized there was one word that was pretty odd to see …

. . . .

How exactly did Shakespeare wind up using “the” so frequently? To figure out, Hope and Witmore began combing through the play, looking for uses of the word “the”.

They began to notice a pattern. Consider this example below; it’s Lady Macbeth speaking. The Macbeths are getting all jittery and nervous, and they’re startled by some noises in the night. Lady Macbeth explains the noises thusly …

Now, that’s a weird way to talk about that owl. Imagine you and I were walking through the woods and we suddenly heard a hoot. I’d probably say, “oh — it’s an owl!” An owl. Not the owl. If you say “the owl,” you’re referring to a specific owl that you, and everyone around, you is already familiar with.

By saying “it was the owl that shriek’d”, Lady Macbeth is — in a quite deliciously creepy way — implying that everyone already knows what owl she’s talking about.

It is a collusively strange way for a character to talk. And it makes us, the readers, feel slightly alienated from our own sense of ourselves, and our own knowledge of the world. (Man, maybe I do know that owl? What the hell is going on???) It’s very subtle effect, but it sends a little shiver down your spine.

Hope and Witmore have another, different way of looking at it: By saying “the owl”, Lady Macbeth makes the bird seem like “a generalised, mythical or proverbial owl … The owl becomes an idea, rather than a thing.”

This curious use of “the” is all over the play. Shakespeare just kept on doing it. Here’s Lady Macbeth again, when she’s counseling Macbeth on how to lie

Same thing here with “serpent”! Normally you’d say, “be a serpent” — but “be the serpent” sounds so much more specific and freaky.

Here’s one last example, from when Macbeth is steeling himself to stab Duncan to death …

Again, interestingly awkward word choice! As Hope and Witmore note, you’d expect Macbeth to refer to “my hand” and “my eye”. By writing it as “the hand” and “the eye”, Shakespeare neatly evokes the way Macbeth is beginning to be tormented by his own decisions; he disassociates from his own body. In a few acts he’ll be a totally unravelled mess.

Link to the rest at OneZero and thanks to HG for the tip.

8 thoughts on “How Data Science Pinpointed the Creepiest Word in “Macbeth””

  1. They didn’t put the th’ stuff in red.

    Also… to be fair, a ton of languages refer to “my” hand or “his” hand as “the” hand. But interestingly, both Gaelic and English don’t.

    So yeah, that is a bit weird… and I think they’re right about this being deliberate.

  2. The scholars have the computer, producing the results that they exhibit to the world as the explanation for the creepiness of The Scottish Tragedy.

    Well, harmless at least, unlike some other “scholars” that have only superficial knowledge of mathematics, how to relate it to real world matters, or that computers have no more insight than their users.

  3. I wonder if those “scholars” are familiar with the concept and uses of archetypes.
    “Be like the serpent” makes perfect sense in a play of eldritch concepts. Lady MacBeth seems to me to be referring to the archetype of all snakes, not a specific breed; a collective rather than a singular.
    The culture and language has changed since the Bard’s day; one might wonder if the target audience would have found the constructions to be weird.

    Yet again I am reminded of Asimov’s THE IMMORTAL BARD.

  4. Read this aloud to housemate who promptly pointed out a famously creepy pieces of poetry “the Raven” does not do this. We’ve been discussing it, anyway, how Poe made it so very effective, which recital available on YouTube is best, that sort of thing.

    Thinks the Raven’s effect is in the whiplash from calm to frenzy. But is now going to poke at MacBeth to make up own mind. Also remarked that as someone who also write poetry (with strict rythmn and rhyme) sometimes you need a single syllable word to make the line fit.

  5. I still think that the creepiest aspect of Macbeth (I’m not an actor, I’m a literary scholar!) is that the ultimate “winner” — not quite a deus ex machina himself — exists only because of a pre-play deus ex machina: He survived a c-section… in eleventh-century Edinburgh, which even then had a reputation as among the foulest, least-sanitary cities in all of Europe. (Either that, or he was really a Faerie changeling, which would be a different kind of creepiness indeed given the recent-at-the-time-of-first-performance publication of the first part of Spenser’s The Faerie Queen…) He didn’t just “survive,” but he thrived to become a leading claimant for the throne.

    We’ll leave aside the possibility that it wasn’t the Birnham Wood itself that was the victorious army, but dead trees (paperwork) wielded by lawyers at the fictional Magic Circle firm Burnham Wood… (Culture note: To this day, there’s a set of major English law firms known as the Magic Circle.) Now that is a frightening thought indeed!

  6. Now you’re talkin’.

    “The”, of course it’s “The”.

    Remember, The play’s the thing.

    Magic is useful. Sometimes people stumble across it, or actually know what they are doing. The basic rule of using magic is, don’t tell people that you are using magic.


    This clip has the same flavor.

    “The Good Man.” “The Doctor.” Yes, as above, referring to “archetype”.

    Eighth Doctor Regenerates into War Doctor | Paul McGann to John Hurt | Doctor Who | BBC

    BTW, As an example of “magic”, I’m reading The King in Orange by John Michael Greer, about the magic used in the past election cycles. I see the book as a supplement to Evil Geniuses by Kurt Andersen.

    Most illuminating.

  7. Tough to take “scholarship” seriously when they can’t get the Shakespeare right:

    (“Bubb-le | bubb-le, | toil and | trou-ble”). –From the article. Was that Disney? Or did we also misremember as kids? I found a bunch of references but no Disney. Shows how relying on memory invites failure. But I knew that wasn’t Shakespeare.

    This is Shakespeare: From Macbeth, Act IV, Scene 1

    Round about the cauldron go:
    In the poisoned entrails throw.
    Toad, that under cold stone
    Days and nights has thirty-one
    Sweated venom sleeping got,
    Boil thou first i’ the charmed pot.

    Double, double toil and trouble;
    Fire burn and cauldron bubble.

    Fillet of a fenny snake,
    In the cauldron boil and bake;
    Eye of newt and toe of frog,
    Wool of bat and tongue of dog,
    Adder’s fork and blind-worm’s sting,
    Lizard’s leg and owlet’s wing.
    For a charm of powerful trouble,
    Like a hell-broth boil and bubble.

    Double, double toil and trouble;
    Fire burn and cauldron bubble.

    Scale of dragon, tooth of wolf,
    Witch’s mummy, maw and gulf
    Of the ravin’d salt-sea shark,
    Root of hemlock digg’d i’ the dark,
    Liver of blaspheming Jew;
    Gall of goat; and slips of yew
    Sliver’d in the moon’s eclipse;
    Nose of Turk, and Tartar’s lips;
    Finger of birth-strangled babe
    Ditch-deliver’d by a drab,
    Make the gruel thick and slab:
    Add thereto a tiger’s chaudron,
    For the ingredients of our cauldron.

    Double, double toil and trouble,
    Fire burn and cauldron bubble.

    Cool it with a baboon’s blood,
    Then the charm is firm and good.

    (public domain)

Comments are closed.