Sethian Sethian Sethian

Spoilers for Sethian are many and unmarked. There’s one in the second sentence.

Imagine talking to a chatbot, with a writing system you barely recognise, and with all the linguistic capability of a three-year-old. Now imagine that the conversation is on religion and philosophy, and you’ve got the central conceit of Sethian. A wildly ambitious game, made by developer Grant Kuning after a successful Kickstarter campaign, Sethian revolves around trying to learn an alien language, Sethian, in order to find out what happened to a colony, Sethian, and its people, the Sethians.

Naming hilarity aside, the game itself consists of asking a computer simple questions, looking at its output, relying on your magical journal-dictionary-combo to translate them for you, and using these tools to figure out more about the language and the narrative of what happened. On the Kickstarter page and on his own blog, Kuning talks about his inspirations behind Sethian-the-language, (Mandarin) Chinese and American Sign Language. This makes it, refreshingly, a constructed language, or conlang, not based on an Romance language, nor one that uses an alphabetical writing system, and with a fairly different grammatical structure to English. In a blog post, Kuning says that one of the things Sethian can do is explore the culture of a language’s speakers, and that’s one of the things you will have to contend with in order to figure out what happened to them all.

But there is one major issue standing in the way of Sethian-the-language, and it’s the game. It’d be easy to blame Sethian’s limited vocabulary, but that’s not the problem — the problem is with Sethian’s grammatical structure. While it’s perfectly fine to design a conlang that is meant to be easy to learn, Sethian takes it a step further by stripping out grammatical features. The language presented to you in the game only has noun-verbs (most Sethian words can either be verbs or nouns, depending how they’ve been marked using punctuation), and punctuation. Sometimes this works well — “thought” and “to think” are represented by the same character, and that’s fine. But in order to fit the strict noun-verbs only restriction, the word for “because” gets classed as a verb (it’s not a verb), and the word “for” (which isn’t a noun) is stuck in the subject, or do-er, position of a sentence. Of the five pages of 4-by-5 grids presented to you, only a handful get used at all. Most remain undefined even in the in-game dictionary. The six punctuation symbols, therefore, need to convey a lot of information — far too much.

This choice, to keep only noun-verbs and punctuation, obviously stems from an ease-of-learning perspective. It keeps down the number of new things you have to learn, and means that every symbol you see will be unique. But the noun-verb system itself is already hard to learn, especially for a native English speaker, because it’s not something English does. The fact that punctuation has to carry so much information means that they’re used in irregular ways, as demonstrated by the solid horizontal dash symbol “—”.  Initially you might be forgiven for thinking it’s analogous to English’s “for”, or “that”, in a sentence like this:

But then you get a sentence like this.

Look at the bit between “high go” and “all person”. That combination of dots, dashes and solid lines reads, in order, “to”, “is”, “to”, making the sentence “[this person] said, ‘A person who ascends that is for completing (a) person.’”

The problem arises because the sentence is fairly complex, and can be bracketed like this.

[(Person) said,
    [For a person to ascend
        [is (for the purpose of)
            [completing that person.]
        ]
    ]
]

There’s punctuation between each set of square brackets, but it’s not all the same punctuation. It changes, from the left-aligned dot to the long dash, to the end-of-sentence space. To make matters worse, the breaks in each “chunk” of the sentence aren’t all breaking for the same reasons, which means the two long dashes are technically marking different things.

But none of this would be as big an issue of the game itself didn’t get in the way of the language, as well. The game itself is very light on setting and exposition; the only things you’re told about yourself is that stumbling on deserted Sethian would be “good for your career”. Presumably it’s in the future, since your journal mentions stasis pods, and (spoilers!) the “bad ending” involves you getting plugged into the Sethian Matrix. And the game very quickly swings towards the philosophical and the transhuman; two of the most important terms in the game are “ascension”, and “the Great Ascension”, and unlocking the “good ending” involves going off-script and trying to work out what some of the more obscure keywords are, keywords which your magic journal will not translate.

This is compounded by the fact that you’re technically not translating a language, but a written language. A language is never quite written the same way it’s spoken, and more to the point, Sethian’s writing system is badly designed for a game and as a writing system. Here’s an example of what I mean:

Those symbols in the grid, of which the game uses only three (circled in white), aren’t good for writing. Humans are lazy writers. Through repeated copying, and then copying the copies, we simplify the shapes of these glyphs (the shapes of each character, letter, or what have you) and sometimes end up with glyphs that are only minimally distinctive from each other. Consider, for example, the difference between “b”, “d”, “p”, and “q”, or the difference between “Q” and “O”. Now look at the glyph circled in red.

In one of his blog posts, Kuning writes that inspiration for some of the characters comes from proofs from Euclid’s Elements. While a novel approach to glyph design, it’s not particularly friendly for humans. The glyph layout is grouped by vague similarity of shape, except when similarly-shaped glyphs are spread out over multiple pages. They aren’t arranged by meaning, given how many of them remain untranslated. It’s reasonable to assume that this arrangement is meant to hint at many more unseen and unused glyphs, especially since the ones on later pages look more complex.

However, the layout also obscures this design goal, as the complex glyphs feel arbitrarily intricate. What we functionally end with are four columns, each with five characters, each with a vague theme in shape, as if the Sethians said, “Hey, for column theme, let’s just auto-generate five glyphs and call it a day!” That regularity is itself unnatural. One of Sethian’s real-life inspirations, Chinese characters, come in groups that hugely vary in size: in the Kangxi Dictionary (think OED but for Chinese), there are 44 characters that incorporate “八”. There are 794 that incorporate “人”.

Even claiming that Sethian might be a computer language or an alien language doesn’t make sense. From the Kickstarter page, the Sethian colony is explicitly descended from the player’s culture, and since the player is human (or close enough), it follows that Sethian was made by the same beings. And this computer didn’t come up with the language, it just speaks it. In any case, this is a language constructed by a “human”, based on other “human” languages.

From a design perspective, these symbols are very confusing. They’re complex and busy, which makes them hard to distinguish from each other. I’ve played through the game three times and still regularly confuse them. For a game whose developer has been very careful to lay out how the language was designed to be easy to learn, the writing system curiously sacrifices function to form.

A desire to keep the language simple and easy to learn means that Sethian has a very basic grammar, with which it now has to try and explain very complicated philosophical concepts that might have no direct analogues to those that players might widely understand. It’s kind of like handing someone a black crayon and telling them to draw a rainbow: not impossible, but you’ll get a more complex and colourful result if you just expand your toolkit. Worse, the more complex “good” ending  requires demonstrating to the computer that you understand, and are already capable of, some of these complicated concepts, therefore you don’t need to be plugged into the Matrix and would just very much like to go now, thanks.

Trying to discuss these concepts in English is hard enough, let alone a constructed language that we’ve had all of an hour or two to learn. One such central concept is a compound word, made up of the word for “me” or “I” and the word for “freedom” or “to be free”. One of the ways of triggering the good ending is stating that “I do me-freedom”, or “I am me-freedom”, but there’s no real way of knowing what the Sethian concept of me-freedom is, nor what Kuning means by me-freedom, because it’s never translated. He even acknowledges it in a Kickstarter update, saying, “I’m starting to realize the problem is that it’s just very difficult to talk about abstract subjects in a language you don’t understand well, particularly things which develop over long periods of time”. And given the vastness of human culture, let alone sci-fi cultures, there’s very little ground for us as human, English-speaking players to stand on when we try to figure out what these concepts are. I triggered the good ending, saw its triumphant final lines, sped away on my spaceship through the stars, and I have absolutely no idea what crock of lies I fed to the chatbot to get it to let me go.

Even far-future intergalactic travellers need to talk about mundane things like arms, legs, plants, animals and rocks; why not base a game around more concrete things? In linguistics, there are lists of words called Swadesh lists that are used as a diagnostic tool in historical-comparative linguistics. The words are chosen because they crop up in the vast majority of human languages that we know of. It’s no surprise that they’re mostly about things like human body parts, landscape features like “mountain” or “sea”, or simple verbs such as “walk”, “sleep”, “give” and “die”.

Ultimately, it is Sethian’s insistence on being a game that hobbles it. In a dev update, Kuning noted, “I’m starting to feel like my gameplay is defining my politics – like I have to promote Great Man Theory because it’s just what works best for the game, even though I don’t really believe in it.” There is noticeable and unwelcome tension between Sethian as a language, and Sethian as a game. The ambition of the Sethian language as something that can and does discuss complicated philosophy, paired with a non-English grammar and very different writing system, clashes with the relative conservatism of Sethian’s game design. Languages don’t have power curves, always-meaningful progression, or a sense of fun, but Sethian has been designed to include them, and it suffers for it.

Thanks to Z.L. for his invaluable contributions, knowledge, and advice on this piece, not to mention willingness to waffle about conlang trivia for three solid hours.

One thought on “Sethian Sethian Sethian

  1. Oooh! The explanations are pretty good!

    “I triggered the good ending, saw its triumphant final lines, sped away on my spaceship through the stars, and I have absolutely no idea what crock of lies I fed to the chatbot to get it to let me go.”

    That summarizes so much of what is… bad about the game.

Leave a Reply

Your email address will not be published. Required fields are marked *