On writing systems

Most conlangers work with text: text files, wordlists, and the like. It’s very much a visual process, quite the opposite of “real” languages. Yes, we think about the sound of a language while we’re making it, but the bulk of the creation is concerned with the written word. It’s just easier to work with, especially on a computer.

Writing, of course, has a long history in the real world, and many cultures have invented their own ways of recording the spoken word. For a conlang, however, the usual form of writing is a transcription into our own alphabet. Few go to the trouble of creating their own system of writing, their own script. Tolkien did, to great effect, but he was certainly an outlier. That makes sense. After all, creating a language is hard enough. Giving it its own script is much more effort for comparatively little payoff.

But some are willing to try. For those who are, let’s see what it takes to create writing. Specifically, we’ll look at the different kinds of scripts out there in this post.

Alphabet

The alphabet is probably the simplest form of script, from the point of view of making one. You don’t really need an example of an alphabet—unless this post was translated into Chinese while I wasn’t looking, you’re reading one! Still, our familiar letters aren’t the only possibility. There’s the Greek alphabet, for example, as well as Cyrillic and a few others.

Alphabets generally have a small inventory of symbols, each used (more or less) for a single phoneme. Obviously, English is far from perfect on that front, but that’s okay. It doesn’t have to be perfect. The principle stands, even if it’s stretched a bit. None of our 26 letters stands for a full syllable, right?

That’s why alphabets are so easy to make, and why they’re (probably) the most common form of writing for conlangs. You only need a few symbols—and there’s nothing saying you can’t borrow a few—and you’re all but done. Writing in the script you make can be as simple as exchanging letters for glyphs.

Abjad and abugida

These two foreign terms name two related variations on the alphabet. The abjad is a script where only consonants are directly written; vowels are represented by diacritics, if at all. That’s the basic system used by Arabic and many of its cousins, as in “ةباتك” (kitāba). Note that Arabic isn’t a “pure” abjad, though. The third letter (reading right-to-left) stands for the long a, while the final a has its own letter. As with English, that’s fine. Nobody’s perfect.

The abugida is similar to the abjad, but it does mark vowels. Unlike an alphabet, this is usually with some form of diacritic or as an “inherent” vowel, but it’s always there. Many of the various languages of India use this type of script, such as the Devanagari used by Hindi: लेखन (lekhan). This particular word has three “letters”, roughly standing for l, kh, and n. The vowel a (actually a schwa) is implicit, and it’s omitted at the end of words in Hindi, so only the first letter needs a diacritic to change its vowel. Once more, the scheme isn’t perfect, but it works for a few hundred million people, so there you go.

Syllabary

Alphabets, abjads, and abugidas all have one thing in common: they work on the level of phonemes. That makes intuitive sense, particularly in languages with complex phonotactics. When there are hundreds of thousands of possible syllables, but only a few dozen individual phonemes, the choice is clear. (That hasn’t stopped some crazy people from trying to make a syllabary for English, but I digress.)

The syllabary, by contrast, gives each syllable its own symbol. Realistically, to use a “pure” syllabary, a language almost has to have a very simple syllabic structure. It works best with the CV or CVC languages common to Asia and Oceania, and that’s probably why the most well-known syllabary comes from that region, the Japanese kana: てがき (tegaki).

A syllabary will always have more symbols than an alphabet (about 50 for Hiragana, plus diacritics for voicing), but not an overwhelming number of them. Syllabaries made for more complicated structures usually have to make a few sacrifices; look at the contortions required in Japanese to convert foreign words into Katakana. But with the right language, they can be a compact way of representing speech.

Featural

A featural alphabet is another possibility, sitting somewhere between an alphabet and a syllabary. In this type of script, the letter forms are phonemic, but they are constructed to illustrate their phonetic qualities. Korean is the typical example of a featural script: 필적 (piljeog). As you can see (hopefully; I don’t seem to have the right font installed on this computer), each character does encode a syllable, but it’s obviously made up of parts that represent the portions of that syllable.

Featural alphabets might be overrepresented in conlanging, because they appeal to our natural rationality. Like agglutinative languages, they’re almost mechanical in their elegance. They only require the creation of an alphabet’s worth of symbols, but they give the “look” of a more complex script. If you like them, go for it, but they’re probably rare in the world for a reason.

Logographic

Finally, we come to the logographic script. In this system, each glyph stands for a morpheme or word, with the usual caveat that no real-world system is perfectly pure. Chinese is far and away the most popular logographic script these days: 写作 (xiězuò). Chinese characters have also been borrowed into Korean, Japanese, and other neighboring languages, but they aren’t the only logograms around. Cuneiform, hieroglyphs (Egyptian, Mayan, or whatever), and a few other ancient scripts are logographic in nature.

It should be blatantly obvious what the pros and cons are. The biggest downside to logograms is the sheer number of them you need. About half of Unicode’s Basic Multilingual Plane is composed of Chinese characters, and that’s still not enough. Everything about them is harder, whether writing, inputting, or even learning them. In exchange, you get the most compressed, most unambiguous script possible. But the task might be too daunting for a conlanger.

The mix

In truth, no language falls neatly into one of the above categories. English is written in an alphabet, yes, but we also have quite a few logograms, such as those symbols on the top row of your keyboard. And with the advent of emoji, the logographic repertoire has grown exponentially. Similarly, Arabic has alphabetic properties, Japanese uses Chinese logograms and Latin letters in addition to its syllabic kana, and the phonetic diacritics used by languages such as German are essentially featural.

For your conlang, the style you choose is just that: a style. It’s an artistic choice. Alphabets (including abjads and abugidas) are far easier. Syllabaries can work if you have the right language, or are willing to play around. Logograms require an enormous effort, but they’re so rare that they might be interesting in their own right. And featural systems have the same “logical” appeal as conlangs like Lojban. Which you choose is up to you, but a natural script won’t be limited to one of them. It will borrow parts from the others.

Creating a script for a conlang can be a rewarding task. It’s not the type of thing to undertake lightly, however. It’s a lot of work, and it takes a bit of artistic vision. But you wouldn’t be making a language if you weren’t something of an artist, right?

Leave a Reply

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