Vox Audita Perit, Littera Scripta Manet

Roman proverb, "the heard voice perishes, the written letter remains." I was thinking about alien alphabets. It took me forever to find an angle to approach the process from. Then I found an article on the sadly defunct "Lyzrd's Stomp" website, about making alien alphabets for model-building, and I've created two alphabets based on its principles. I thought I'd share what I learned from that site here, but add in some thoughts about alphabets from a linguistic standpoint. I am, after all, a science fiction writer.

1. Classification
   
   "Alphabet" is actually, often, an inaccurate term. Hebrew and Arabic are actually "abjads", and Devanagari and Thai are actually "abugidas". Japanese (kana, anyway) is a "syllabary", and Chinese is a set of "logograms".
   
   See, an alphabet is only an alphabet if it writes consonants and vowels with the same type of character. An abjad, like Arabic, writes the vowels as diacritics, if at all (though it uses some of its consonants for long vowels—e.g., y is "î", w is "û", and ' is "â", in Arabic). And an abugida assumes that every consonant has an intrinsic vowel; in most Indian abugidas the intrinsic vowel is "a". Other vowels are written with diacritics, as is the "vowel canceler" (at least in abugidas for languages with biliterals).


2. Development
   
   As far as we know, all four of the world's alphabetical systems began as logogram sets. And yes, there are only four—cuneiform, hieroglyphics, hanzi, and Meso-American glyphs—and cuneiform and Meso-American are extinct. Other than Chinese, Japanese, and Korean, all writing systems in the world are, directly or indirectly, derived from hieroglyphics. Actually, Linear A, which might be related to the Luwian glyphs and possibly the Indus Valley script, represents a fifth, but it was pretty much extinct by the Golden Age of Greece, and we don't actually know that it wasn't derived from Egyptian (or possibly cuneiform) somewhere along the way.
   
   What's interesting is, everyone starts with logograms. Then they eventually start writing in syllabaries or abugidas, sometimes in combination with the logograms. Then come abjads, and finally alphabets. Usually, when people try to create alphabets for their languages (if they don't just adapt another alphabet, e.g. Greek becomes Etruscan becomes Roman), they create syllabaries; Persian derived its from cuneiform, Sequoia derived his from Roman, and Kobo Daishi derived his from Siddham Sanskrit (and the man'yogana system for writing Japanese phonetically in hanzi).


3. Can You Write It With a Ballpoint Pen?
   
   Right. So. My brilliant teacher from Lyzrd's Stomp started with, quite sensibly, "look at our alphabet". Notice that the characters are made up of different kinds of stroke. For example, in Arial and other "Helvetica" scripts,
   
   A E F H I K L M N T V W X Y Z
   
   are all straight lines, while
   
   C O S
   
   are all curves, and
   
   B D G J P Q R U
   
   are a mix of curves and straights.
   
   Now, I do something a little different from him, since my purpose wasn't merely decorative. Take a good long look at, say, the Romulan alphabet from Star Trek: what the hell tool do you write that with? I mean, I get that that's a typeset font, and all, but what does the handwritten form look like? It's really hard to tell.
   
   So what I do is, I keep my scripts' basic form simple (because my script has to stand up to the "would anyone actually be able to write that?" test). I make a bunch of glyphs—more than I need, so I can choose my favorites—from those elements, curves, straights, and mixes. Then I assign them to phonemes. It doesn't matter if some of your letters look like some of ours, though you might wanna avoid having the doubles stand for the same sounds. Actually having some false cognates is realistic, look at Cyrillic, with its H=N, C=S, and P=R (and cursive M=cursive T, which I have never understood).
   
   At this step, it helps to think about your script's history. For instance, my felinoids' script derived from (a tiny portion of) one of their civilizations' logograms, that one of its successors used it as the basis of an abugida. But then a third civilization changed it to an alphabet. As a result of its history, though, the vowels—which were originally the abugida's vowel-diacritics—are written half the height of the consonants.
   
   Speaking of, though it's quite likely a script will have both a script and print form, it's extremely unlikely that it'll use the case-distinction we use (the only other alphabets that have capital and lowercase copied them intentionally from Latin languages). The only exception is Greek (which acquired its two cases in a parallel development with Latin-script languages), and though it has capitalized letters in proper names, it doesn't put them at the beginning of sentences (come to think of it, German uses our alphabet but capitalizes every noun). There might be different situations where another alphabet switches between script and print forms (remember how spell names are always italicized in D&D?), and, like in my felinoids' alphabet, there can be other reasons for a size distinction.
   


4. Linguists Don't Write Fruit Receipts
   
   I know, I know, you love hangeul to death, and the "turn it 90 degrees" thing in Aboriginal Syllabics is kinda neat. But those scripts could not possibly be, or look, more artificial, not even if every glyph had "© Dow Chemical" written real tiny at the bottom. Unless the culture the writing is for is a totalitarian dystopia dominated by an academic ideology (like Joseon Korea was), or was taught by missionaries from somewhere else (like the various Indians who use Syllabics), they'd never have a writing style like that. Scripts don't develop like that. Both Sinai script and kana derived from using logograms phonetically, not from any linguistic analysis.
   
   I urge you not only to resist the urge to nerd out while creating your script, but actually build in little historical oddities. The order of the Roman alphabet comes from casting off, then re-acquiring, Greek letters; the Cyrillic "i" looks like a backwards N because it's based on a cursive eta (apparently already pronounced "i"). Not only do my felinoids write their vowels small because of an abugida's vowel-diacritics, they also write the glottal stop (which is a consonant) as small as a vowel, because originally it was the abugida's vowel-canceller. They write the two "purred" vowels the same size as the consonants, though, because those two characters derive from the abugida's vowel-initial syllables (I'm not sure why, maybe the purred-vowel diacritics were too similar to the others, or ugly, or something).


5. Numbers
   
   Take care when writing your numbers. Above all, no dots. Sure it's easy, and there have been real-world systems that did it (most Meso-American peoples denoted numbers with dots, and it's how Romans wrote fractions), but it's also lazy and unattractive. Plus, think about writing: how easy are dots to blur together? Pretty damn easy.
   
   Personally, I like acrophonic numerals. Our own number system (the Hindu one) is mostly acrophonic, except that 1, 2, and 3 are cursive tallies (2 and 3 are sideways), while all the others are the first letters of their Sanskrit name, written in Brahmi. Cool, huh? Plus you don't have to come up with more letters, which is nice (maybe strikethrough or underline them when they're being used as letters), and you can have people make jokes by reading numbers phonetically, the way people do in Japanese (like how Shinigami's phone number in Soul Eater is 42-42-564, or "dying-dying-killing").