Understanding Kishoutenketsu

Kishoutenketsu (起承転結) is the name for a certain Chinese and Japanese style of composition, particularly interesting in that there are no good descriptions of it anywhere on the English internet (do not hope for one here), despite the fact that everyone will tell you it’s nothing like Western composition. What makes it even more interesting is that there are actually two usages of the style — argumentative and narrative/poetic writing — which bear the same “philosophy” but are different in practice.

I want to discuss the narrative/poetic use of kishoutenketsu, because it’s not clear on the surface how it differs at all from the Western structure, but you’ll hear a lot of empty words on how it creates narrative without conflict or the like. The argumentative use has been well documented in scholarly work on cross-linguistic editing and ESL learning in Japan, because it contradicts the thesis-evidence-conclusion structure of Western argumentative writing.

Anyways, the structure of kishoutenketsu is as follows:

  • 起: the introduction
  • 承: the development
  • 転: the twist and climax
  • 結: the conclusion

Okay, let’s try comparing this to the Western story structure:

  • exposition
  • rising action
  • climax
  • falling action/resolution

Wait! They’re literally the same!

When understood in this loose sense, it’s easy to apply kishoutenketsu far too widely, and lose the uniqueness that it claims to carry. If you read about kishoutenketsu online, you’re likely to run into a generalized explanation which could just as well be replaced with a copy of The Hero with a Thousand Faces.

This Gamasutra interview on Mario has a short section which raises kishoutenketsu as a design principle:

It’s very similar to a narrative structure that you find in four-panel comics. Something that’s talked a lot about in Japanese manga, for example, is a phrase, kishoutenketsu, where you introduce a concept, and then in the next panel you develop the idea a little bit more; in the third panel there’s something of a change-up, and then in the fourth panel you have your conclusion.

So that’s sort of what we try to do with the way people relate to gameplay concepts in a single level. We provide that concept, let them develop their skills, and then the third step is something of a doozy that throws them for a loop, and makes them think of using it in a way they haven’t really before. And this is something that ends up giving the player a kind of narrative structure that they can relate to within a single level about how they’re using a game mechanic.

We’ll come back to this interview towards the end. For now, I want you to try asking: how is this any different from exposition/rising action/climax/resolution? If your impression is that they seem to be the same, you might learn something from this article.

The difference is ultimately a classic problem of “lost in translation”. Specifically, the connotations of ten and twist or climax are not comparable, and the simple translation fails to be accurate. In order to understand what makes kishoutenketsu unique, we must examine the ten.

What’s in a “Twist”?

Consider Romeo and Juliet. There are two major events which could be dubbed climaxes: the duel between Romeo and Tybalt, as well as the final suicide scene. Both are plagued with uncertainty. We have no reason to predict that Tybalt will kill Mercutio, nor that Romeo will kill Tybalt in response, but both events are reasonable given prior events and characterization. In the suicide scene, it is pure coincidence that Romeo arrives moments before Juliet wakes up, but after this random occurrence, the rest of the scene plays out as expected. Both of these events can be considered as twists and climaxes. Note, though, that twists always follow from what precedes them: they do not break the logic of the world.

If I were to offer a one-word translation for ten in kishoutenketsu, it would be: recontextualization. The purpose of ten is not to advance a narrative, but to modify or subvert the reader’s understanding of the writing — to break the logic of the world. (Subversion is notably common in comics.) Because the reader’s knowledge of the writing is always partial and based on assumptions brought from real-life experience, it is always subject to attack. ten is precisely that attack.

The example of kishoutenketsu I find easiest to explain is this four-panel comic from Extra Fabulous Comics:

Image for post
Image for post
Read left to right, top to bottom.

Starting from the top left, we have the introduction, which lays out the setting within a mostly familiar framework (“I’m not ready to die” is the starting point for any number of myths). Moving to the top right, this framework is developed in an unsurprising fashion (“No man is truly ready [to die]” is a common truism). Then, in the bottom left, this framework (which the Reaper is still playing out) is interrupted by recontextualizing the boy’s utterances within a different framework (that of being dressed for the occasion). In this case, the recontextualization is violent, emphasized by the forefronting of the boy and his dialog box over the Reaper (but it need not be). Finally, the consequences of this recontextualization are shown.

Again, the major point is recontextualization: whereas we begin reading this comic in terms of running from death, the third panel forces us to rethink it in terms of welcoming death in style. No such rethinking occurs in Romeo and Juliet: it is a romantic tragedy through and through.

While I use the term “framework” loosely, my notion is equivalent to the concept of genre in frame-based cognitive linguistics. A genre guides interpretation and constrains organization of a piece of writing; in essence, it is a collection of tropes and rules for understanding tropes. While we often think of genre at the level of “science fiction”, they can be much more low-level; one may talk about local genres like 1880’s California literature on race relations, or genres that follow a specific line of thought, such as poetry that focuses on the loneliness of autumn evenings. I’ll use “genre” only when discussing high-level groupings.

What specifically is a trope? In classical Japanese poetry, certain words always signal the season in the poem (these are called kigo). For example, if you say “frog”, then the poem takes place in spring; if you say “Milky Way”, then it’s autumn. Other words always signal the emotion of the poem. “Mountain temple” invokes the loneliness of autumn evenings. Any such pattern of form and meaning can be considered a trope, although when talking about genres we should want to focus on the unique tropes they offer.

Image for post
Image for post
This will be important later

Comics are technically narrative, but as mentioned, there is a similar poetic use of kishoutenketsu. The analysis is basically the same, because what matters is the framework the reader uses to understand the writing, not the narrative events. Here’s a poem by this guy which was rendered in Japanese on the kishoutenketsu wikipedia page and which I loosely translate here:

Against the blue river waters I see birds ever more white

Against the mountain green I see flowers bloom ever more red

Once again does this spring flee from me as I gaze upon it

How long will it be until the day I may return to my homeland?

The first two lines function within the basic poetic framework of “recounting the beauty of nature (in spring)”. The third line recontextualizes this by placing it within the framework of “growing old”. Note that in this example, the frameworks merge, instead of dominating each other as in the comic, and we have a haunting picture of “growing old by seeing spring pass by”. The power lies in the tension between the two frameworks; spring is associated with youth and beauty, age with winter. It would not have been much of a ten if the first two lines were recounting the bleak landscapes of winter. Finally, a consequence develops out of the merged frameworks: a nostalgia for homeland (associated with youth and therefore spring) which the narrator worries will only be fulfilled with age, if ever.

Ketsu is in general the consequences that develop from the collision of the two frameworks. It may present even more new information, like in this quatrain. Emphasis on ketsu can differ greatly, and in the case of humor, ketsu may be removed or minimized.

Brave New World, the least dystopian dystopian novel ever written, is an example of how you might merge both narrative event and recontextualization to construct a doubly powerful climax. After Linda’s death (a “twist”), John resolves to “free” the people from the state’s ideology. To this end, he interrupts the distribution of soma to factory workers and throws all the drugs out the window. Obviously, the workers get pissed and attack him. Of Helmholtz and Marx, the other two misfits, Helmholtz helps John and Marx is indecisive on the sidelines. This is a Western-style climax in that it brings the three misfits into open conflict with the ideology of their society and solidifies the position of these three who were previously unsure of whether or not they should rebel.

The ten in Brave New World, and the reason why it is far less dystopian than commonly claimed, is the chapter that follows, wherein Mustapha Mond recontextualizes everything “wrong” with their society — social structure, drugs, religion, art, science — within the framework of human happiness and societal stability. Inspired by 1984 or by tales of Nazi Germany, a reader might initially consider the society of BNW as serving corrupt government and authoritarian power, and therefore read it in terms of the political framework of corrupt dystopian government. This chapter forces the reader to instead think about stable societal organization. It is this reversal and subversion of perspective, wherein the reader’s initial framework is no longer sufficient, that characterizes ten. Like our comic, this subversion is violent; there is no merger of perspective.

“If it were really like Othello nobody could understand it, however new it might be… because our world is not the same as Othello’s world.

The world’s stable now. People are happy; they get what they want, and they never want what they can’t get. They’re well off; they’re safe; they’re never ill; they’re not afraid of death; they’re blissfully ignorant of passion and old age…

Actual happiness always looks pretty squalid in comparison with the over-compensations for misery. And, of course, stability isn’t nearly so spectacular as instability. And being contented has none of the glamour of a good fight against misfortune, none of the picturesqueness of a struggle with temptation, or a fatal overthrow by passion or doubt. Happiness is never grand.”

“God isn’t compatible with machinery and scientific medicine and universal happiness. You must make your choice. Our civilization has chosen machinery and medicine and happiness. That’s why I have to keep these books locked up in the safe. They’re smut.”

As a final comment before we move on, let’s dicuss the difference between recontextualization and “reveals” (I am your father, etc). Recontextualization forces a revision of how the reader understands the writing, whereas “reveals” are, fundamentally, a subclass of unexpected narrative events (much like the Western “twist”). Star Wars is a space opera before and after this reveal; there is a complication in character relationships and motivation, but it doesn’t trigger a change of framework. A “reveal” preserves the framework; in other words, it preserves the methods by which you interpret the writing. On the other hand, BNW initially suggests a framework of “this is about authoritarian power in a new era of technology” — but it is really about human happiness and security in a new era of technology, and you must rethink your initial conclusions about its “dystopian” social structure. (Note that people often forget this ten in BNW and thus tend to make erroneous comparisons between BNW and modern politics, where human happiness and security are explicitly deprioritized in favor of capitalist power. Why this “misreading” occurs is a topic for another day.)

Reveals can be (but are usually not) also recontextualizing. Consider The Murder of Roger Ackroyd by Agatha Christie. The book presents itself as a mostly traditional detective crime novel, with the retired detective Poirot stepping up to solve a murder investigation. As in Sherlock Holmes — a key contributor to the genre/framework of detective novels — the book is narrated by Poirot’s assistant, specifically the local doctor Sheppard. When it is revealed at the end that Sheppard was the murderer, the genre falls apart. The genre’s basic assumption that the assistant is smart and innocent breaks; the genre’s basic assumption that the narration is direct and truthful fails. (A few early critics actually complained that it was “cheating” to make the narrator-assistant the murderer!) Because the genre can no longer account for the exceptional reveal, the reader must recontextualize the entire book. Sheppard actually does this in the last chapter, when he reveals that the book, rather than being a disembodied narration, was intended to be his record of Poirot’s failure to solve this final mystery, and that he took great pleasure in writing about his experiences while omitting everything that picked him out as the criminal.

I am rather pleased with myself as a writer. What could be neater, for instance, than the following:

‘The letters were brought in at twenty minutes to nine. It was just on ten minutes to nine when I left him, the letter still unread. I hesitated with my hand on the door handle, looking back and wondering if there was anything I had left undone.’

All true, you see. But suppose I had put a row of stars after the first sentence! Would somebody then have wondered what exactly happened in that blank ten minutes?

When I have finished writing, I shall enclose this whole manuscript in an envelope and address it to Poirot.

Today, though, we have a lot more stomach for these kinds of reversals, and we’re also a bit more distant from the Holmes model of detective fiction. It’s thus not obvious that the recontextualization would be as strong today, or if the modern genre of detective fiction even makes the assumptions which Roger Ackroyd defies.

Tomo-chan to Tomo ni Kishoutenketsu

Image for post
Image for post

1: Ki. Carol approaches Misuzu and initiates a discussion.

2: Shou. The discussion continues in expected fashion.

3: Shou. The discussion continues. The ten is forecast by way of the emphasized height difference and the line “but first…”.

4: Tenketsu. The discussion is recontextualized by zooming out and revealing that it took place in a strange setting. The framework of “normal discussion” no longer applies given Carol’s strange positioning. The consequences (ketsu) of Misuzu’s response and people looking on worriedly are minor and merged in the same frame.

Image for post
Image for post

1: Ki. The setting is: Akechi is at the culture festival for her daughter’s school.

2: Shou. Logically developing the setting, Akechi comes across Misaki, who is there for the same reason. As it’s been established that they know each other, it’s logical that Akechi performs a greeting. We read these two panels in terms of the framework “old acquaintances’ chance meeting”.

3: Ten. Misaki ignores Akechi. The “old acquaintances’ chance meeting” framework is subverted and we must reinterpret the previous panels in terms of some kind of grudge that would lead to this response.

4: Ketsu. The consequences of the clash between “old acquaintances’ chance meeting” and “old acquaintances’ grudge” are displayed.

Image for post
Image for post

1: Ki. The setting is: Misuzu and Carol are wandering around at the culture festival. They see Jun and Tomo.

2: Shou. Jun and Tomo are embracing, which is slightly out-of-character and justifies Carol’s callout in the first panel.

3: Shou. Carol continues her callout. She explicitly states the framework by which we should understand the comic, which is “lovey dovey”. Misuzu forecasts the ten with her comment.

4: Ten. The embrace is recontextualized as: Jun is restraining Tomo from going on a rampage. In this case, there are no consequences; the humor is primarily in the violent subversion of the “lovey dovey” framework.

Image for post
Image for post

This one is a bit more playful with the structure; it’s actually a second-degree instance of subversion.

1: Ki. Instead of running the race, Carol and Misuzu are looking at wildlife. (Previous chapters show the race starting, so there’s no surprise here.)

2: Ten?. Carol and Misuzu are juxtaposed with Jun and Tomo, who are taking the race way too seriously. There is a contradiction between the two different perspectives on how to run a race. Thus, this appears to be a ten.

3: Ten?. Carol and Misuzu clearly see the contradiction between the perspectives. It follows that, in the next panel, they should comment on the contradiction (therefore ending with a ketsu).

4: Ten. Carol and Misuzu do not comment on the contradiction. In doing so, they subvert the framework the reader was using to understand the first three panels (something like “characters should comment on overt contradictions between their opinions and others’”).

As it is well-known that fair use only allows posting four pages from a manga in a single article, the last example will be from XKCD. I’ve seen a lot of Western comics that employ the philosophy of kishoutenketsu, and SMBC’s one-panel captioned comics might offer a particularly unique structure for it, but I don’t have space for that.

Image for post
Image for post

1: Ki. Playing boring video games to escape from a boring life is perhaps not the image that people had of video games 20 years ago, but it’s a standard joke (and therefore a meaningful trope!) today.

2: Shou. The other interlocutor agrees with the framework.

3: Tenketsu. The recontextualization is that the character on the right is playing video games to escape from playing video games. The standard framework is subverted and replaced with an absurdist reimagining. The consequence is the passing angel, and a possible pointer towards a resolution (in the psychology of motivation).

Responsibility in Narrative

English as a practiced language is writer-responsible. The writer must state the thesis clearly at the very beginning, signpost everything, and then state the thesis again at the end. This is especially apparent in technical writing. As my research professor told me when I asked how to structure a research paper: “First say what you’re going to say, then say what you have to say, finally say what you said.” Hence the emphasis on a strong thesis statement, rigid organization, and the pruning of unimportant information which might be distracting: it is entirely the writer’s responsibility to ensure that the reader understands.

Japanese, on the other hand, is largely reader-responsible. In the kishoutenketsu argumentative style, the thesis may not be stated until the ten or ketsu sections, until which it is entirely up to the reader to organize and make sense of the partial information conveyed in the first sections. The structure of the writing is associationist rather than hierarchical (one must wonder whether this is linked to the associationist and drop-heavy syntax of Japanese). Likewise, unimportant information can be appreciated for how it illuminates the author’s wide knowledge, and it is up to the reader to incorporate it or throw it away.

We can extend this logic to narrative/poetic writing as well. Subverting or shifting perspective puts a large amount of stress on the reader. Therefore, we can expect it to occur much more often in reader-responsible languages. Kishoutenketsu and all its philosophies are thus a part of language practice.

You may have an impression that Western modern art (especially the era around WW1, with the Imagists and Surrealists and the like) also pushes more responsibility on the reader, whether that be by removing structures of clear interpretation or rebelling against standard artistic frameworks. The question follows: was the philosophy of kishoutenketsu adopted by Western artists around this time?

Yes and no. It’s not clear that Western artists at this time studied kishoutenketsu. However, it is well known that they studied haiku, which, while a different poetic form from kishoutenketsu, might be said to bear some of the same philosophy. One of the key tenets that Western artists drew from haiku was juxtaposition, which often appeared as a structural system of two “setting” lines and one “cutting” line. Here are two famous haiku, the first from Moritake and the second from Bashou (both translated mediocrely by me), where the third line cuts across the process of the first two lines.

I look again at the

fallen blossom returning to its branch —

a butterfly

old pond

frog jumps in

sound of water

And here is a famous “basically a haiku” from Pound, one of the early proponents of haiku in Anglo-American poetry, which has the same structure:

In a Station of the Metro

the apparition of these faces in a crowd;

Petals on a wet, black bough.

Kishoutenketsu’s recontextualization and haiku’s juxtaposition are clearly similar concepts, and examining their genealogy would likely bring up some interesting points of intersection. I may talk more about their similarities and differences in a later post on haiku.

Wrapping Up

So that’s sort of what we try to do with the way people relate to gameplay concepts in a single level. We provide that concept, let them develop their skills, and then the third step is something of a doozy that throws them for a loop, and makes them think of using it in a way they haven’t really before. And this is something that ends up giving the player a kind of narrative structure that they can relate to within a single level about how they’re using a game mechanic.

We can now clarify what makes this different from the Western notion of exposition/rising action/climax/resolution. While the exposition and rising action are mostly the same, the key difference is ten: where, instead of progressing the narrative further, the writer recontextualizes the narrative. What’s key in this interview is “mak[ing] them think of using it in a way they haven’t really before”: this is a recontextualizing ten.

Do Mario levels actually recontextualize the mechanics they introduce? I don’t play Mario so I can’t answer that. I can say, though, that a few of the shrines in Breath of the Wild recontextualize your basic Sheikah Slate mechanics. The bomb and stasis mechanics are particularly fun for the breadth of ways you can use them, and some of the shrines do explore this.

Most games don’t recontextualize their mechanics. In some cases, this may be simply because the mechanics are too “straightforward”. I don’t see how it would be possible to recontextualize Dark Souls or Persona gameplay, because gameplay is based on firm and consistent rules rather than openly interacting systems (BoTW) or new and unexplored mechanics (Mario, presumably). This isn’t a knock on them; using mechanics that can be recontextualized is a game design tradeoff.

That’s all for this piece. I’m inspired to write on haiku, so expect something on that within a few months. Thanks for reading.

Image for post
Image for post
This is only here for the cover image.

Other Reading

  • On responsibility in kishoutenketsu versus the Western argumentative system.

Johnson, Jeffrey. 2011. Haiku Poetics in Twentieth-Century Avant-Garde Poetry.

  • On the philosophy of haiku as adopted and transformed by continental, Anglo-American, and Latin American poetics.

Liddicoat, Anthony J. 2009. “Communication as Culturally Contexted Practice: A View from Intercultural Communication”. Australian Journal of Linguistics 29:1, 115–133.

  • On differences in cultural structuring of discourse (ie. the tropes of discourse). The section on digressions is especially interesting.

Nutt, Christian. 2012. “The Structure of Fun: Learning from Super Mario 3D Land’s Director”. Gamasutra.

  • The interview I brought up twice.

Software engineer, epic gamer, and Touhou developer. All my writing is licensed under CC BY-SA 4.0 unless stated otherwise.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store