New Worlds: Writing Systems

(This post is part of my Patreon-supported New Worlds series.)

Given that the New Worlds essays are all about worldbuilding for stories, it’s a little surprising that I haven’t discussed writing before now. So let’s remedy that! (Full disclosure: I am deeply indebted to NativLang’s Thoth’s Pill series of videos for helping me sort this into manageable order and providing examples.)

Writing is a huge watershed in human society, right up there with things like agriculture. And, like all major innovations, not everybody approved of it: we have ancient documents that record elders decrying the kids these days, letting their memories atrophy by leaning on the crutch of written records, losing the precious face-to-face interaction of spoken communication, making it so their words can be read (and probably misunderstood) by total strangers years from now. All those complaints about kids these days and the radio/TV/internet/etc? Not remotely new.

It’s also been invented multiple times — at least three times (in Mesopotamia, China, and Mesoamerica), with other writing systems developed from or least inspired by those source scripts. It isn’t impossible to run a complex state without writing, but it definitely helps, because of the need for record-keeping. In time I’ll dig into the many social, political, economic, religious, and aesthetic facets of writing, but let’s start with the basics: different types of writing systems.

Mythological origins aside, they don’t generally spring fully-formed from nowhere. Pictographs are “proto-writing” — literal images, albeit sometimes stylized, of the things they’re meant to represent. They’re not good at dealing with abstract concepts, though, so from pictographs we move to ideographs, which represent ideas. For an example, consider the no symbol, which communicates the idea that something is prohibited. This leads in turn to logographs: signs that represent specific words. Once you have logographs, you can have a full writing system, capable of representing your entire language.

The rebus principle then allows you to develop phonetic readings off your logographs: for example, sometimes the sign for “bee” (as in the insect) might get used to represent the verb “to be,” or anything else containing that syllable. For clarity, you can add a determinative sign, which provides semantic information to distinguish various homophones and near-homophones from one another. This shows up all over the place in Chinese writing, where the determinatives get called radicals (but some radicals are phonetic instead of semantic, because this stuff gets complicated).

Rebus-based phonetic writing brings us around to a syllabary, a writing system where each sign represents a syllable. What constitutes a syllable varies from language to language; in most cases it will incorporate a vowel, but there are exceptions, such as Japanese “syllabic n.” (Not actually a syllable, if you want to get super-technical, but for the purposes of discussing writing, that term will suffice.) Syllabaries may cope with things like final consonants either by underspelling — leaving off the final consonant entirely — or by using an echo vowel, writing Mayan balam as ba-la-ma or Minoan Knossos as Ko-no-so (underspelling and echo vowel both!)

You can solve that particular conundrum with an alphabet, where each sign represents a single phoneme. (In theory. English spelling assigns multiple phonemes to some letters, especially vowels.) In an abugida or alphasyllabary, like Devanagari, the consonants form the core of each sign, and the vowel is added onto that as a kind of ligature. An abjad like Arabic simplifies even further, writing only the consonants and leaving the vowels to be filled in from context; an “impure” abjad includes some vowels or diacritics for marking vowels. Hebrew, Arabic, and Syriac also use matres lectionis: specific consonants that can also be used to represent vowels.

And finally, no rundown of writing systems would be complete without a shout-out to Korean and its featural script, Hangul. In this approach, the symbols are systematically designed to tell you things about how they should be pronounced: details like the place of articulation (e.g. labial vs. glottal) and the manner of articulation (e.g. sibilant vs. plosive). It’s fascinating to imagine how Hangul might adapt to unfamiliar types of sounds — is there a way to represent click consonants? What if we meet aliens who make sounds the human body is incapable of?

Such speculations aside, we still have a dizzying array of writing options. Reading through it, you might be forgiven for thinking that it’s an evolutionary path: that the alphabet/abugida/abjad end of things is the “best” writing system, and things like logographs are archaic holdovers. But in truth, each system has its benefits and downsides.

Take logographs. You need a lot of these to represent your language; the usual estimate is that you have to know three to four thousand characters to be considered fully literate in Chinese. That seems complicated — and in a way, it is. But the flip side is that logographic writing can communicate (albeit imperfectly) across languages: you can recognize the characters, even if you have no idea what phonetic readings they’re supposed to have. It’s possible to be partially literate in a logographic script, in a way that doesn’t really happen with an alphabet (very few people stop at knowing nine of English’s twenty-six letters).

An alphabet, by contrast, has many fewer signs: dozens at most, instead of thousands. It’s easy to learn all the signs in an alphabet, and easy to adapt to different languages, assigning different phonetic readings to existing letters or adding on one or two more where needed (and dropping the ones you don’t need). But it won’t help you understand across language boundaries; you can write English, Hawaiian, and Hungarian with slight variations on the Latin alphabet, but if you don’t speak all three of those languages, you won’t understand what the other two say. And god help you if the language includes a lot of homophones! My favorite online Japanese dictionary demonstrates the problem with the (admittedly artificial) sentence niwano niwaniwa niwa niwatoriwa niwakani wanio tabeta. Written out in a pure syllabary, it isn’t much more comprehensible. But add in the Chinese characters, and it’s easy to see that the sentence means “Two chickens in Mr. Niwa’s garden suddenly ate an alligator.”

But if you look at the sentence behind that link, you’ll see it isn’t written purely with logographs, because they lack a certain kind of flexibility. Mandarin Chinese is an analytic language, and as such works very well with this system, because it communicates things like the past tense by adding on more words. Japanese, however, uses more inflection, so that a verb like kaku (to write) may have the forms kaita, kakimasu, kakeba, and so forth. When Japan borrowed China’s writing system, they had a problem: how could they fit their inflections into a system not designed for such changes? They found ways — several of them, in fact — but the one that won out was to create a syllabary — two of them! — which are used in combination with the Chinese characters.

Because despite what the descriptions above may imply, languages often intermix these types of script. Mayan writing was partly logographic, partly syllabic, and sometimes glued syllabic signs onto logographic ones as phonetic complements. Egyptian combined an abjad with logographic determinatives. Heck, look at texting today: the proliferation of emoji means that pictographic proto-writing is coming back into vogue, right alongside our alphabet, which sometimes simplifies into an abjad, because writing is srs bsns.

The relevance of all of this for worldbuilding is that it affects how easily people become literate, how readily their writing system adapts to other languages, how intelligible the writing is between languages, and so forth. And that’s before you get to all the other aspects mentioned before — the social, political, economic, religious, and aesthetic concerns — but this is already half again as long as a normal essay, so those will wait for a later day!

The Patreon logo and the text "This post is brought to you by my imaginative backers at Patreon. To join their ranks, click here!"



About Marie Brennan

Marie Brennan is a former anthropologist and folklorist who shamelessly pillages her academic fields for inspiration. She recently misapplied her professors' hard work to the short novel Driftwood and Turning Darkness Into Light, a sequel to the Hugo Award-nominated Victorian adventure series The Memoirs of Lady Trent. She is the author of several other series, over sixty short stories, and the New Worlds series of worldbuilding guides; as half of M.A. Carrick, she has written The Mask of Mirrors, first in the Rook and Rose trilogy. For more information, visit, Twitter @swan_tower, or her Patreon.


New Worlds: Writing Systems — 39 Comments

  1. I hope at some point you address the “secret” language found in Chinese harems. Possibly a proto-Mandarin. The women were not allowed to read and write the newer or “real” language but wrote beautiful poetry, literature, and even political commentary in their own secret form of communication, among themselves and across to other households.

    Our own Alma Alexander wrote her beloved and award winning novel “The Secrets of Jin-shei” based on this.

    • Well, I just added codes, cryptography, and special languages to the list of topics, which already included gendered language, so . . . eventually! 🙂

      • Have you read Don’t Sleep, There Are Snakes?

        Fascinating book by a linguist missionary about his experiences in the Amazon with a tribe whose language is unique.

        Anything else would be spoilers.


        • No, I haven’t!

          The pedant in me wants to point out that all languages are unique in the strict sense of the word. 😛 But it’s true that some languages have very unusual features not seen in many (or any) other places.

          • Understood. But describing how the particular language and culture the book is about would be pretty spoiler-ish, and also I didn’t want to write a dissertation or a review. Anyway I found the book fascinating.

            The language has a men’s dialect and a women’s dialect. There’s one situation in which the men of the tribe have one reaction and the women another. But though the author’s wife and young daughters (the latter at least probably knew the language?) were with him, he doesn’t report talking to the women about the situation, either himself (the dialects are mutually comprehensible) or via his wife or daughters. I found that puzzling.


            • I’ve heard interesting things about the Pirahã language, yeah.

              he doesn’t report talking to the women about the situation, either himself (the dialects are mutually comprehensible) or via his wife or daughters. I found that puzzling.

              Puzzling indeed — and off the cuff it sounds like a failure to follow through on his part, though maybe there were contextual reasons why he didn’t.

            • I probably missed something in the book; I’d thought the dialectal difference was because women have an extra letter in their alphabet.

              The Piraha take nationalism to mind-boggling extremes (other people build boats – we don’t wanna learn from them!)

              • Maybe I was unclear about what I was saying. Even if the women’s and men’s dialects were mutually incomprehensible (which I *don’t* believe is the case), he could have asked his wife and daughters what the women thought about the incident in which the men’s and women’s reactions were entirely different. But if he did, that information is missing from the book. It’s information I would have liked to know.

                A book that *does* include women’s perspective (including re-visiting 19th century anthropological reports) is Women and Power in Native North America,, which I also found fascinating.


                • he could have asked his wife and daughters what the women thought about the incident in which the men’s and women’s reactions were entirely different

                  This is a non-trivial problem in anthropology, because even in societies that aren’t obviously gender-segregated, there’s often a difference in how women will behave around men and vice versa. It’s why you get some husband-and-wife teams of anthropologists out in the field: they’re able to coordinate and compare notes for a much more well-rounded view of any given situation.

                • *nods* I doubt there are much in the way of differences between the two – beyond something like “we say CAT, they say CAD” given how little separation seems to exist between any group in the Piraha community…men and women, young and old, etc.

                  My apologies for misunderstanding you and possibly being unclear in my reply.

                  One can hope information like that (about what differences, if any, between men and women in the Piraha language, given that Everett claims the Piraha pretty much have no pragmatic speech) would be in related texts about the Piraha, like …I forget the title, but it was part of a book series & the individual book was focused on the languages of the Amazon.

                  Thank you for the recc.

  2. Pingback: New Worlds: Writing Systems - Swan Tower

  3. I confess to being spoiled by Korean being my mother tongue. Every time I make a conlang, the script ends up looking a little like Korean because it is so efficient and logical that I don’t want to do anything else. (The high language alphabet is influenced both by Hangeul and Japanese kana.)

    By the way, I am blackly amused that one of the chief complaints by the yangban (literati) when King Sejong attempted to introduce the Hangeul alphabet was that “it’s so easy even a woman could learn it.”

    I use determinatives in my note-taking and have since high school, mainly as a sort of shorthand so I don’t have to write so many things out. And then there are just plain abbreviations; when I was taking US Government, I decided that “dp” was “due process” because it came up ALL THE TIME.

    • There seems to have been a pattern.

      ‘When it was first developed, hiragana was not accepted by everyone. The educated or elites preferred to use only the kanji system… Hence hiragana first gained popularity among women, who were generally not allowed access to the same levels of education as men. And thus hiragana was first widely used among court women in the writing of personal communications and literature.[8] From this comes the alternative name of onnade (??) “women’s writing”‘

      Hiragana being the dominant syllabary used in Japan. And Nushu, one of the Chinese women’s (though not harem) systems that Phyllis mentions, was syllabic.

      Regarding evolution and ease of learning, I think I’ve read that Sequoyah basically recapitulated writing evolution in trying to come up with a Cherokee writing system, inspired by simply knowing that white people wrote down speech somehow. “Pictographs? Not complete enough. Logographs? Too hard. Syllabary? Just right.” And once completed, learnable in a few days by previously illiterate Cherokee.

      • A writing system that gets consciously designed for the language it’s going to be used to write, rather than being an ad hoc evolution from some not-quite-appropriate model, often winds up being very easy to learn. (At least if the person designing it knows what they’re doing.)

        But I also have to remind myself that I’m pretty sure children in, say, Spain, have an easier time learning to read than Anglophone children do. The alphabet works a lot better when you have a more consistent match between symbol and sound.

        • Oh yes. Speaking as a career teacher of small children, the fact that English makes no sense whatsoever to kids used to cause much lack of joy.

          • I pretty much assume that in Spanish-speaking parts of the world, or German, or other places where the orthography makes actual sense, there is no such thing as years of spelling tests during early education. If you know how to say it, you know how to spell it.

            • Estonian writing is based on the German system but we still have problems spelling words as hard as ‘eat’ or ‘sell’. Never underestimate the idiots.

    • I suspect I’ll wind up talking about that “ease of learning” thing when I post about literacy — because we’re accustomed to thinking of widespread literacy as a good thing, but not every society has agreed . . .

      I’m curious what kinds of determinatives you employ! I’ve been known to use ? as my notation for “person” or “people,” because it’s one of the instances where a Chinese character is just flat-out faster than the English equivalent. And when I took archaeology classes, “kya” (thousand years ago) and “mya” (million years ago) were very handy abbreviations.

  4. *nods* and, like your link points out, the determinatives aren’t always supposed to be pronounced…one of the reasons I like Mandarin: fewer exceptions to the rules (that i learned, anyway) than English or Akkadian.

    And sometimes the very complexity of a writing system is seen as a feature rather than a bug — not just English nowadays, but the partial bilingualness of Sumerian-Akkadian writing, or remembering where to bury your knives in Ancient Egyptian (always in the demons)

      • Gladly. When I was learning about Ancient Egyptian in school, one of the little factoids of the book was that, when scribes wrote “Apep” (Apophis, in Greek and scifi tv) either spelled out or depicted as a serpent, they had to add daggers to stab into Apep — because Apep was dangerous and you don’t want to give him even that little bit of strength accidentally. (also, i imagine, loyalty – the scribes were all in the employ of people who claimed association one way or another with Apep’s enemies…Ra and Osiris)

  5. As to homophones, Estonian is extremely rich. Working as a translator for the EU I ran into a boring Regulation about ‘concentrated butter or butter or butter equivalent’. Simple, right ? Except that the word for ‘butter’ is ‘või’. And the word for ‘or’ is ‘või’. Dunno if you can see the o with the squiggle on top but we had that ‘või’ five times in a row. Mister Niwa can go home to feed his chickens.

    (We ended up doing something one should never do when translating legal documents. We debated, chewed our keyboards, banged our heads, shouted at each other, howled to the moon, debated some more, and replaced ‘or’ with ‘and’.)

    • Oh, yikes! That’s an incredibly unfortunate confluence of words — every bit as bad as Mr. Niwa and his chickens, but without any need for so contrived of a sentence. You could easily wind up with Mr. Niwa’s garden — niwa no niwa — or Mr. Niwa’s chickens — niwa no niwatori — or chickens in the garden — niwa ni niwatori — or even just two chickens — niwa niwatori — but nothing quiiiiiite as bad as või või või või või.

      It makes me think of the (again, incredibly contrived) English sentence “Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo.” But that relies on some really arcane tricks to make it work, while again, yours was pretty normal.

      Not that I have any kind of authority to say this, but I think you guys made the right choice with your translation. Which would lead to greater confusion: saying “and” instead or “or,” or the same word five times in a row with two different alternating meanings?

      • I can see the problem, but, as a (recovering) lawyer, I will point out that “and” and “or” have very different legal meanings in English. You probably took that into consideration, because it depends on the context whether it matters or not. But note the difference between “A and B are required” and “A or B are required.”

        Language is complicated. Translation is an art and a science. I greatly admire anyone who can do it at all, and worship those who do it well.

      • No one ever complained so we probably got it right. (And õ sounds something between a beer belch and the o in ‘no’ but we honestly do have this sound as an integral part of our language. Even the Finnish laugh at õ.)

        Estonian has ‘tee’, too. It can mean ‘road’, ‘tea’ or ‘go make’. Go figure. Go make tea, tee tee.

        I once had the dishonor to doctor a translation by someone who always used the very first word from an abridged dictionary. English can be very confusing, too. A Soviet anecdote went, how do you manage to artificially inseminate all the cows on your collective farm? Oh, we just use the bull afterwards! Well, I felt I was the bull.

        And I am literate in Cyrillic, too. So brave missionaries Cyrillos and Methodios set out from Constantinople to enlighten the Russians of Kiev but during their arduous journey they forgot whatever literacy they might have had and ended up inflicting the ugliest, most unreadable and difficult to write alphabet on a nation of millions. The Russians themselves complain their writing ‘thwarts the eye and assaults the aesthetic sense’. Never mind the Soviet attempts to force Cyrillic on nations already having written languages in Latin or Arabic alphabets as that was an epic fail.

        • I once had the dishonor to doctor a translation by someone who always used the very first word from an abridged dictionary.

          Oh dear.

          The Russians themselves complain their writing ‘thwarts the eye and assaults the aesthetic sense’.

          Oh dear oh dear. <lol> That’s . . . I’ve never thought Cyrillic was especially ugly, but I’ve also never really dealt with it to any notable extent.

          • While Latin letters go up and down a lot, thus having some pleasant variety and giving the reader tiny visual pointers, Cyrillic letters tend to be all the same size, and the shapes just never look as distinct.

            However poor Cyrillic has its merits, too. They only need a single letter for ‘schtsch’. And another one for ‘tsch’.

            It gets really funny when Russians transliterate Latin names. See, they go by the sound. You can probably guess who Uiljam Shekspir was but what about Dzhek Vens?

            Hmmm. There might be some good ideas for stories where cultures clash just because their writing systems mismatch.

            • Jack Vance? If so, it took me a while, but I did get there eventually. (Then again, I’m used to dealing with Japanese transliteration, which can sometimes be impenetrable. Basukettobooru is relatively easy to parse as “basketball,” but makudonarudo threw us for a loop for quite a while — it’s MacDonald’s.)