Wednesday, February 4, 2009

The appearance of text and script

This post was inspired by the internet discussion that has prompted my latest surge of visitors - welcome, all of you who've never visited before! It is also quite relevant to my language design workshop (fortunately!).

I'm talking about letters. That is, how a language is written down. In previous posts about designing languages (see the label "designing languages"), I've talked about how different writing systems correspond to different sounds, so today I'm going to come at it from a different angle - what those characters look like, and how to discuss the appearance of text, letters and script in a story.

When I was designing my Varin world, I actually designed the language after I drafted the novels for the first time, and I had to go back and look for every instance in which I mentioned the appearance of the text. Lucky for me, this didn't involve instances of naming letters so much as places where people were looking at books or handwritten notes. So I sat down and designed a single-sound-based alphabetic character system with a print and a cursive form.

The trick with designing alphabets is that you want them to be simple enough to duplicate, and yet distinct enough for them to be easily recognized. The alphabets I created as a kid were almost always too complicated to duplicate with any reasonable degree of speed, as are some of the character systems I've seen used in published fiction. I've also seen many character systems in published fiction in which the characters were not easily distinguished from one another. Tolkien's Elvish/Mordor script is of course wonderful - you see it and you immediately think you're looking at a foreign language, but that it's a language.

Another factor to consider is what tools your people use to write. If they just use pens, then you're pretty flexible. If they use a reed stylus and clay, then the form of the characters will be influenced by that.

Back to my Varin alphabet design story. The Varin alphabet (which is written with pens) has the basic elements of dots, vertical lines, horizontal lines, and diagonal lines. No curves or circles, unlike our writing system. Instead of orienting itself at mid and high distances from a single line at the bottom of the text, the Varin text orients itself along a central axis. The curves come in when people start writing in cursive. So when I went about describing it in the context of the story, I used these elements. Examples: "He stared at the note until it became just meaningless dots and lines sprayed across the paper." "He had gorgeous handwriting, with axis serifs at the end of each word."

That will take you some distance toward solving the problem - but suppose you want to deal with particular characters, as when a child learns their ABC's? I have several recommendations here:

1. Unless your people actually speak English or another language that uses the Latin alphabet, don't use the appearance or name of the letters we use.

2. Don't try to list out the entire alphabet and the names for all the letters. Keep the description general or appearance-based and restrict yourself to naming two or three, maybe four of the letters.

3. Make sure that the names of the letters are short, easy to pronounce and remember.

4. Make sure that the sounds of your letter-names match the phonological system of the names of your characters. A made-up word for a letter will always stand out, but it may stand out more than you want if your character's name is Aramia and the first letter is called grixbat. Not that this is impossible linguistically, of course - but the contrast will be noticed. So take a look at the sound characteristics of the words you've already created, and match them with one another and with the names of the letters (I have already made posts on sounds; see articulatory phonetics).

I hope you find this helpful. Please feel free to comment or ask questions.


  1. I don't know how I stumbled across your blog but it's like I've found a candy tree. I'm really enjoying your posts and just wanted to feed back. :)

  2. your character's name is Aramia and the first letter is called grixbat

    Quite clearly we have a language in which living things are named fluidly and abstractions and non-living things have angular names. I bet a horse is a "winnywu" and an automobile is a "klankaxaon."

  3. Thanks, Sunil!


    You got me again :-)! That was why I kind of hedged about the topic and said it could be done but might be distracting. I love your idea, though. The example in the forum discussion had very English-sounding names like "Grayson" and was concerned about giving a random name to the letters. With good reason. The two should match under some principle - and I think your idea is an interesting one!

  4. There is precedent. Irish Gaelic uses two sets of numbers: one for counting people, one for counting things.

    It also occurs to me that the Gaelic letters were called by "tree names." Each letter was called by a tree whose name began with that letter. Anciently: beit (burch), luis (rowan), nuin (ash)... (B, L, N, etc.). Modern Irish follows Roman alpha-beta order.

  5. I'd like to hear your thoughts on pictograms (like in Japanese/ Chinese languages) in this context. I don't really have a specific question, except that frequently the name of the pictogram won't match the sound. Or it has multiple sounds.

  6. Mike,

    Japanese doesn't have different sets of numbers, exactly, but its numbers require a suffix which categorizes the noun. Humans are -ri, while swords are -furi, small four-legged animals are -hiki, long cylindrical things are -hon etc.

  7. To my anonymous poster:

    If you're interested in Chinese and Japanese script, I do have an earlier post on the topic (look under the label "writing systems"). The different pronunciations for the same character has to do with the fact that the Chinese script (which the Japanese borrowed) represents meanings rather than sounds. Because the Japanese borrowed words from the Chinese more than once over their shared history, you can have multiple pronunciations for the same character:
    1. the Japanese native word for that object
    2. the Japanese phonological interpretation for how a Chinese person said his word for that object when it was initially borrowed
    3. the Japanese phonological interpretation for how a Chinese person said his word for that object two hundred years later when it was borrowed again


    This is one reason why simplification of the writing system has historically been a contentious issue in Japan.

  8. Mike, how splendid!

    Fit this idea together with Byron's idea about all flying things, be they birds, bats or rocketships having similar root based names, and you begin to see a language taking shadowy shape. As with Byron's suggestion, I'll squirl yours away.

    Also, though I am of Irish descent and though I have studied the Irish language, I didn't know about the letters. That is quite lovely. Think that's another tidbit to add to my squirl hoard.

  9. Hmmm. I have not yet considered how or even if my marine species has any form of writing.

    Perhaps that's why the guilds guard their secret knowledge - it was lost when the land flooded and all those books drowned :)

  10. Sounds reasonable, David. So they were thrown back, in effect, to the conditions of a preliterate society.

  11. That's the idea, Catreona. I'm planning on leaving clues througout the story, showing that the arcati once had a much higher tech level.

    For example, they mention a helix or double helix in relation to, well, relations. They know that biological inheritance involves a particular shape, but how do they know that? They don't have the technology to look at DNA.

    Of course, what they do have is pretty impressive. Even if they can't write it down.