Friday, July 1, 2011

Tragedy, farce, and other, rarer modes: a post on Kafka

In almost any interesting narrative, everything happens at least twice. The first time to set the rules of this here kind of fictional event, so that the second time its stakes and its moves are intelligible to the audience. That intelligibility may be necessary to the protagonist too -- the first time may occur as practice, in the Smersh training circuit, on the flight simulator, in the dojo, in the school yard (no less in The Wire or in Buffy than in the opening of Abel Gance's Napoleon). Or it may be only there for us, as when Chigurh flips his coin to determine his suppliants' fate in No Country for Old Men.

It's really hard to think of a story with an interesting plot that doesn't rhyme incidents in this way: Madame Bovary, Proust, Austen, Waiting for Godot, any Shakespeare play, Dickens, Fielding, Tolstoy (well, maybe not Tolstoy: there may be something of Kafka in him). It's the principle of revenge stories as of love stories. It is to be found in every fairy tale. Hansel drops stones and the children find their way home. The second time he drops bread-crumbs and they don't. The wolf blows down two of the pigs' houses, and then he fails.

There's one obvious reason, to which I plan to return in a later post, for this repetition (sometimes multiple repetition as in the story of the three little pigs as well as the story of Goldilocks). Events in narrative had better be doing at least double duty. They are interesting in themselves, but they also convey information crucial to the set-up of later and still more interesting events, events that make the narrative particularly interesting, events that particularize the interest of this narrative. Narratives have to teach you the rules of the game they're playing (or in genre fiction the rules of the variant of the game that they're playing), for you to appreciate the moves made according to those rules. A Swedish client of my father's visited the U.S. in 1956 for the first time, and his host took him to see Game 5 of that year's World Series. The client was bored out of his mind. Nothing happened, that he could see: every Dodger batter who went to the plate returned to the dugout.  Don Larsen retired every single one, pitching the only perfect game in World Series history. But the Swedish client didn't know the rules, and he missed the unbelievably exciting experience that he was present at.

If a narrative doesn't set up its own rules early, and then follow the constraints imposed by those rules, you get only chronicles, what E. M. Foster calls plot as opposed  story. Such constraints don't only include whether people can fly in this fictional world or not; whether animals can talk or not; they include subtler things like establishing the balance between consistency and freedom in a character, between a character's capacity to act surprisingly and gratuitously and our sense that a character is acting plausibly, consistently with what we know about her and what she has become at some point in a narrative. Such rules will also include psychological rules about how much time or experience can change characters: is their psychology Jamesian? Dickensian? Woolfian? Proustian? Is remorse or repentance or degeneration or despair possible in their fictional world?

An interesting vignette about such a change, early on, will sound the note for the possibilities of such changes over the course of the narrative. As with so much of Eliot, the opening two of Daniel Deronda makes the point explicitly:
Men can do nothing without the make-believe of a beginning. Even science, the strict measurer, is obliged to start with a make-believe unit, and must fix on a point in the stars' unceasing journey when his sidereal clock shall pretend that time is at Nought. His less accurate grandmother Poetry has always been understood to start in the middle; but on reflection it appears that her proceeding is not very different from his; since Science, too, reckons backward as well as forward, divides his unit into billions, and with his clock-finger at Nought really sets off in medias res. No retrospect will take us to the true beginning; and whether our prologue be in heaven or on earth, it is but a fraction of that all-presupposing fact with which our story sets out.

Was she beautiful or not beautiful?
The first paragraph tells us how to read the question that opens the second: we are at an arbitrary moment that nevertheless tells us that there are four different directions, towards four quadrants in the narrative field, that the curve of the story might trace. We know both that a character might be difficult to pin down, and, what's more, that he might be difficult to pin down because he has difficulty pinning another character down.  Daniel's difficulty says something crucial both about Gwendolyn and about himself, and the possibilities of their respective adventures are given by the narrative anecdote with which the book opens, and by the narrative vignette with which that anecdote opens.

There's a chaotic dialectic between randomness and determinism in good narrative, a dialectic itself stabilized by the practice of repetition, often as with Eliot in the mode of self-similarity, where repetitions occur across different orders of magnitude.  A good narrative approximates a Koch Curve:

From Wikimedia: as we zoom in on the curve you can see its every part is made up of similar parts
To take some obvious narrative examples, in a trust narrative, a character must learn either trust or suspicion.  She'll learn them on the basis of an early narrative experience, and the mode may either be once burned, twice shy or first antagonistic, later life-saving.  The second schema is to be found in any number of heart-warming narratives, from Huck Finn to Lisbeth Salander's relation to Mikael Blomqvist.  Of course you can get various combinations of the two, meta-versions where modes of repetition are themselves repeated with a difference.

Indeed the second story probably requires the first: "Once burned, twice shy" is the reason a character may start out antagonistic; if she becomes life-saving afterwards, what's happened is that once burned, twice shy has been repeated with a difference.  If the once burned, twice shy part of the story is itself well told, so that trust leads to distrust, the return to trust will require all that much more narrative invention, force, resourcefulness, and management of the relation between the audience's desire and its knowledge.  These are things that Dickens and Homer do particularly well.  The rightly suspicious Achilles trusts Priam in the end, and Priam trusts him; the rightly suspicious Odysseus trusts Penelope, and she trusts him.

If we put this very schematically, we can describe Frye's four basic narrative arcs in terms of such repetitions.  The replacement of a sad event with its happy repetition is comedy.  The replacement of a happy event with its sad repetition is tragedy.  Happy to happy is farce.  Sad to sad is melodrama.  Again, these things can occur on meta-levels, some more self-similar, some less, again with different emotional outcomes.

I was thinking about this because I was thinking about Kafka's fairy tale openings.  They're so hard to write -- no one trusts themselves any more with this kind of unexplained directness, which seems a feature of oral memory, not of meditation and invention. In the twentieth century maybe only Kafka ever wrote them convincingly.  Gregor Samsa wakes up as a giant insect: so do fairy tales begin.  The Emperor has a message for you.

All the fairy tales that open in this way would repeat their incidents in a way that advanced towards some conclusion with a difference.  You wake up as a bug and later on, when you wake up as a tree your experiences as a bug will help you negotiate your new, arboreal circumstances.  The Emperor's message comes, or doesn't, and later on when you hear a more important message has been sent, you'll know how to insure that you receive it in time this time.  Leopards break into the temple, and later on they'll become a part of the ceremony.

But, in Kafka, what almost always happens -- this is the rarer mode -- is that things only happen once.  You wake up as a bug, and that's it.  You stay a bug and nothing rhymes with that experience.  You don't get the message.  You don't get to the next village. The first part of the book will be about getting to the Castle.  K. never gets there.  Or it will be about waiting for the trial.  But the climax never comes. Once you pass the first stage and enter the doorway of the Law, that's when things will get complicated.  But you never enter the doorway.  God calls unto Abraham to sacrifice his son!  Abraham assumes some other Abraham was meant: he can't be important enough for this test.  Fairy tales without the repetition that offers narrative satisfaction: that's the amazing thing in Kafka.


  1. The ultimate practitioner of the Koch curve being James Joyce. Has anyone ever practiced it more completely or more frequently? The mistake I've made in my own fiction, I think, was to try to pull this off thematically rather than plotwise, raising things to too great a level of abstraction. Oh well.

    This also reminds me of Kurt Vonnegut's short little piece on fairy tales in Palm Sunday. I hadn't thought of it in years.

  2. For example, the Reverend letter, which begins: "Reverend" and also "riverrun" and ends "ALP" and also "a last a long a loved the."

    Merrill does it really well too, everywhere (of course) but in Ephraim especially.

    I still have to read your novel. Once the never-settling dust finally settles. It's getting there. I do think that's the mistake that Jonathan Franzen makes, though. Among many others. At least in The Corrections. The idea that every paragraph or so has to have some example of a correction being made or failing to be made: spelling errors, markets, criminals, manners.

  3. And a subscribing friend comments in an email: "When you say 'its own', it gets me to thinking: does narrative, even if it merely obeys convention, need to cast itself nonetheless as constructing them?"

    Nice -- exactly. It does need to cast itself as constructing the conventions. That's part of the necessary and central convention of narrative (of authorial) omniscience.