The Art of Language Invention

“Careful now. I must say this right. Upe, I have killed your husband. There’s a gold hairpin in his chest. If you need more soulprice, ask me. Upe, bing keng … No, that doesn’t start right. Wait, I should say, ‘Upe, bing wisyeye keng birikyisde… And then say sorry. What’s sorry in this stupid language? Don’t know. Bing biyititba. I had to… She can have the gold hairpin, and the other one, that should be enough. I hope she didn’t really love him.”

This is a tiny fragment from the novel I was writing when I started hacking on Krita… I finished the last chapter last year, and added a new last chapter this year. The context? Yidenir, one the protagonists, an apprentice sorcerer, is left alone by her master in a Barushlani camp, where she lives among the women, in the inner courtyard. When she learns she has been abandoned, she goes to the men’s side of the tent, argues with the warlord and to make sure he understand she’s a sorcerer, kills his right-hand man, by ramming one of her hairpins in his chest. Then she goes back, and tries to figure out how to tell that henchman’s wife that she has killed her husband. A couple of weeks isn’t long enough to learn Den Barush, as Barushlani is called in Denden (where ‘barush’ is form of the word for ‘mountain’).

Together with the novel, I wrote parts of a grammar of Barushlani. I had written a special application to collect language data, called Kura, and a system that used docbook, fop and python to combine language data and descriptive text into a single grammar. I was a serious conlanger. Heck, I was a serious linguist, having had an article published in Linguistics of
the Tibeto-Burman Area

But conlanging is how I started. I hadn’t read Tolkien (much, the local library only had Volume II of Lord of the Rings, in a Dutch translation), I didn’t know it was possible to invent a language. But around 1981 I started learning French, English and German, and with French came a grammar. A book that put down the rules of language in an orderly way, very attractively, too, I thought. And my mind was fizzing with this invented world, full of semi-hemi-demi-somewhat humans that I was sculpting in wax. And drawing. And trying to figure out the music of. My people needed a language!

So I started working on Denden. It’s no coincidence that Denden has pretty much no logical phonology. Over the years, I found I had gotten sentimentally attached to words I invented early on, so while grammar was easy to rewrite and make more interesting, the words had to stay. More
or less.

Then I started studying Chinese, found some like-minded people, like Irina,
founded the Society for Linguafiction (conlang wasn’t a word back then), got into a row with Leyden Esperantist Marc van Oostendorp who felt that languages should only be invented from idealistic motives, not aesthetic. I got into a memorable discussion in a second-hand bookshop when a philosopher told me smugly that I might have imagined I had invented a language, but that I was wrong because a) you cannot invent a language and b) an invented language is not a language.

I got into the community centered around the CONLANG mailing list. I did a couple of relays, a couple of translations, and then I started getting ambitious about my world: I started working on the first two novels. And then, of course, I got side-tracked a little, first by the rec.arts.sf.composition usenet group, where people could discuss their writing, and later on by Krita.

These days, when we need words and names for our long-running RPG campaign, we use Nepali for Aumen Sith, Persian for Iss-Peran. Only Valdyas and Velihas have proper native language. The shame!!

And apart from RPG and now and then writing a bit of fiction, I had more or less forgotten about my conlanging. The source code for Kura seems to be lost, I need to check some old CDR’s, but I’m not very hopeful. The setup I used to build the grammars is pretty much unreconstructable, and the wordprocessor documents that have my oldest data don’t load correctly anymore. (I did some very weird hacks, back then, including using a hex editor to make a Denden translation of WordPerfect 4.2.)

Until today, when young whipper-snapper David J. Peterson’s book arrived, entitled “The art of language invention”. Everything came back… The attempt to make sense of Yaguello’s Les Fous du Langage (crap, but there
wasn’t much else..) Trying to convince other people that no, I wasn’t crazy, trying to explain to auxlangers that, yes, doing this for fun was a valid use of my time. The Tolkienian sensation of having sixteen drafts of a dictionary and no longer knowing which version is correct. What’s not in David’s book, but… Telling your lover in her or your own language that you love her, and writing erotic poetry in that language, too. Marrying at the town hall wearing t-shirts printed with indecent texts in different conlangs, each white front with black letters shouting defiance at the frock-coated marriage registrar. (I don’t believe in civil marriage.)

Reading the book made me realize that, of course, internet has changed what it means to be a conlanger. We started out with literally stenciled fanzines, swapping fanzine for fanzine, moving on to actual copiers. Quietly not telling my Nepali/Hayu/Dumi/Limbu/comparative linguistics teacher what I actually was assembling the library of Cambridge books on Language (the red and green series!) for.

Linguistically, David’s book doesn’t have much to offer me, of course. I adapted Mark Rosenfelder’s Perl scripts to create a diachronically logical system of sound changes so I could generate the Barushlani vocabulary. I know, or maybe, knew, about phonology, morphology, syntax and semantics. I made my first fonts with Corel Draw in the early nineties. I had to hack around to get IPA into Word 2. But it was a fun read, and brought back some good memories.

And also some pet peeves… Dothraki! I’m not a Games of Thrones fan, I long for a nice, fun, cosy fantasy series where not everyone wants to kill, rape and enslave everyone else. I found the books unreadable and the television series unwatchable. And… Dothraki. David explains how he uses the words and names the author had sprinkled around the text to base the language on. Good job on his side. But those words! Martin’s concept of “exotic language” basically boils down to “India is pretty exotic!” It reads like the random gleanings from the Linguistic Survey of India, or rather, those stories from the Boy’s Own Library that deal with Hindoostan. Which is, no doubt, where the ‘double’ vowels come from. Kaheera’s ee is the same ee as in Victorian spellings of baksheesh and so on. Harumph.

BUT if the connection with television series helps sell this book and get more people having fun conlanging, then it’s all worth it! I’m going to see if I can revive that perl script, and maybe do some nice language for the people living in the lowlands west of the mountain range that shelters Broi, the capital of Emperor Rordal, or maybe finally do something about Vustlani, the language of his wife, Chazalla.

Let’s go back to Yidenir, doing the laundry with poor disfigured Tsoy… Tsoy wants to sing!

Yidenir, ngaimyibge?” Another fierce scowl.

“What did you say? — do I sing? Er…” Yidenir was silent for a moment. Was this girl making fun of her? Or was she just trying to be friendly?

Sadrabam aimyibgyi ingyot. Aimyibgyi ruysing ho,” Tsoy explained patiently.

“Er, singing, is good, er allowed? when doing laundry? Oh, yes, I can sing… Denden only, is that all right? Er, aimyipkyi denden?


“All right, then… Teach you a bit of Denden, too? Ngsahe Denden bingyop?” Yidenir offered.

Een wereldtaal, De geschiedenis van het Esperanto

By Marc van Oostendorp

Not so long ago I hacked languages instead of painting applications, and I cannot, in fact, promise that I’ll never hack languages again. And not programming languages, but human languages. I’ve invented quite a few languages for my invented world, the setting of two novels that I’m trying
to sell. I’ve had the languages bug since I first discovered that our school grammar of French wasn’t all that well laid out and could be improved upon. Later I learned about Tolkien, about Roland Tweehuysen (but not Mark Okrand — I never was a trekkie). I joined a club of people interested in designing imaginary countries, worlds and languages.

And then, one day, I found myself studying Sinology in Leyden, and still hacking away on Denden. I was outed (hacking languages was decidedly uncool in those days) and one day a journo from the Leyden University rag, Mare, asked me whether I wanted to tell something about this weirdo hobby, inventing languages. I pliantly complied, something I’d never do nowadays.

I said I invented languages for fun, because it was a creative endeavor, and also because playing with languages helped me understand how languages work, which came in handy, since I wanted to go on and study comparative linguistics of the Sino-Tibetan languages. And then the dreaded Esperanto questions was posed.

“Why invent another invented language if there was already Esperanto?”

I answered I invented languages for aesthetic enjoyment, not to improve the world and bring brotherhood to all mankind. Besides, Esperanto, I said, is boring, staid and ugly. I’d learned it, but not enjoyed it.

I shouldn’t have said that. There was one Esperantist left in Leyden, and he tried to start a polemic. Now I wasn’t, in those pre-Usenet, pre-blogosphere days, as experienced in the noble art of flame-warfare, so I answered back, and when he answered back again, I shrugged and
considered that in order to be an Esperantist you needed to have a really powerful reality-distortion field and ignored him from that moment on.

And I may be very much mistaken, I won’t say I’m not, but in the haze recesses of my memory the name “Marc van Oostendorp” rings a bell. Was he my opponent in Mare? The author of the book under consideration — a birthday present given me by Irina — is about the right age, and he’s an Esperantist. He’s now a researcher with the P.J. Meertens Instituut, learned Esperanto as a child and was the first Dutch professor in Interlinguistics (whatever that may be) and Esperanto at the University of Amsterdam. It may be him, and, if so, he’s done better out of linguistics than me.

Anyway, this book is perhaps written in a style that most closely resembles a high school essay, something a lot of semi-popular science books in Dutch suffer from, but it’s full of diverting anecdote and little bits of knowledge that I didn’t know previously. I’m still not tempted to drop Greek and start boning up on Esperanto, though. Esperanto is still a boring language.

PS: I even went as far as writing a complete linguistic description  application that kind of fused my need for research in Sino-Tibetan languages from Nepal and my fictional linguistics. It’s now maintained by Peter Bouda, and available at Institut für Allgemeine und Typologische Sprachwissenschaft in München. [ETA: link gone. Bummer.]

Dutch doesn’t exist

Well, really, I mean. It’s not just George van Driem who maintains, with some justification, that Dutch is a Low-German dialect. (We do have a navy, after all — even though I don’t know what it’s good for. The North Sea is a Mare Nostrum seen from the POV of the NATO anyway, and we had better get rid of the Dutch Antilles, the populace of which insists on electing certified corrupt separatist politicians.) But take a look at Dutch in past two or three centuries…

In the heady days of the Bataafse Republiek, written Dutch was a calque of French, in the sense that sentence construction and vocabulary were modelled on French. The influence is evident in authors like Rheinvis Feith. Before that, circa Hooft, Dutch was quite original, and Hooft is a lot of fun to read — but it smacks of Flemish. (The best Frenchified Dutch is found in erotica, as collected by Leonard de Vries.)

Later, after Rheinvis Feith, we arrive at the German era of Dutch. Authors like Querido or even Bomans or Belcampo wrote a Dutch that was to all purposes and semblances modelled on German. It had a kind of reinforced concrete-ness to it that one still finds in the Frankfuerter Zeitung. In the less gifted authors, for instance of non-fiction books about history or the great achievements of Dutch genius, like the Afsluitdijk, the tendency is even more apparent.

It would made a good story to say that that ended with the Second World War, but that isn’t quite true. There was a period of slow transition that started around 1900 to the modern Dutch, which, even from the best of literary authors, reads like it has been — badly — translated from English.

So then, what’s Dutch? Literary, written Dutch? I’m sure I don’t know — I’ve quit writing in Dutch since the latest spelling reform, which was too silly for words. I’ll keep to English, which hasn’t — in its written form — changed so much that texts a century old are all but unintelligible. I mean — today, I translated a little snippet of 1930’s vintage Wodehouse for Naomi and shell fell from the sofa laughing. Absolutely fell from the sofa. But the same text, translated into 1950’s Dutch leaves her cold, because it has grown too alien. She said, when I know enough English, that’s a book I’m going to really enjoy — and I bet she will. But she won’t enjoy Belcampo as much. Wanna make a bet?

Of course — I haven’t done any real research. One should establish a contemporary corpus (not forgetting vectors of influence — who read  what), choose some kind of tagging standard, compare sentence structures, do a statistical analysis, perhaps, and then write a thorough paper on the findings. But that wouldn’t be a blog entry, it would be something fit for Acta Linguistica Twincastriana and take a lot of time. So all you get are my impressionistic impressions.

Kura runs on OS X!

Thanks to some tips from Michael Dunn, I’ve finally managed to get Kura
working on OS X. Actually, it wasn’t so much a matter of getting Kura to work, as it was a matter of untangling various previous attempts at getting PyQt and sip to work.

Here’s the setup I am using:

If you have managed to compile everything (and you need some Terminal skills for that, this is not drag and drop in your /Applications installation), you will need to use a special a shell script to start Kura, because the default shell script doesn’t work. You will also be running Kura from out of the kura-2.0 directory, you should not try to install it.

Qt needs the full path to the program; not a relative path like ./kuraclient/, but /Users/boud/prj/kura-2.0/kuraclient/ And you need to
use pythonw, not python, otherwise OS X won’t know you’ve started a GUI application, and you will never be able to activate the pretty window. So take a peek at the included script, and then run it.

Note that you will be greeted by the message importing db api failed. This is not an error. It merely means that you will be using a big file for your data, instead of a MySQl databse. Ignore it, open a test database (it ends with .dbobj, and read the fine manual.

This is a development snapshot of Kura; you might have equal success with a regular release, but I don’t think you should install it, instead run it from the package directory. Also don’t be afraid to mention any bugs, or fix them, if easy to do so…