A Farewell to Romlangs

Many conlangers make up romlangs, i. e. fictional Romance languages. The most famous of these, and the one to start this fashion, is Andrew Smith's Brithenig. Brithenig is an attempt at fabulating out what a Romance language of Britain could look like. Brithenig prompted many other conlangers to come up with their own Romance conlangs.

I was one of these epigones of Brithenig, and started my own Romance conlang: Germanech - a language that would be to German what Brithenig is to Welsh: I based it, more or less, on the application of the sound changes of German to Vulgar Latin. Alas, that language, now known as Roman Germanech, soon became something I lost interest in, so I "fixed it up" and left it at that.

Why Germanech failed

So why did Germanech fall short of satisfying me? What was it that rendered me unable to achieve a satisfactory result? There are several reasons.

1. Grafting sound changes onto another language does not work well

The basic idea behind Germanech was to apply the sound changes of one language - German - to a different language - Vulgar Latin. This did not work well. In fact, it could not: Vulgar Latin and Common West Germanic have very different phonologies, so I had to bend a lot to make things work out, and they did not work out well. The result did have some nice points, such as the Vulgar Latin open-mid vowels leapfrogging the close-mid vowels and becoming high vowels, but the whole thing was not much of a success.

2. I could not come up with a plausible alternative timeline

The original idea behind Germanech was that in an alternative timeline, the Romans managed to conquer Germany after Varus won the Battle of Teutoburg Forest. Sure, the area we now call Germany would probably have become Latin-speaking in such a world (though we cannot be sure: Latin failed to take foot in Britain, and that could also have happened in Germany). But it would have thrown history as we know it too far off the path, completely altering the patterns of the Völkerwanderung and perhaps even securing the survival of the Roman Empire. Thus, I decided to drop the alternative timeline and pursued the idea of having Germanech as a lostlang.

3. Germanech does not really work as a lostlang

But even that was a failure. For a Germano-Romance lostlang, I had to find a spot somewhere in Germany where a Romance language had a marginal chance to survive. I chose the vicinity of Trier on the Mosel river, which had been part of the Roman Empire, and where a Romance language is known to have survived until about 1100. But such a Romance language would almost certainly not have developed in the direction of Germanech. Apart from the fact that it is historically implausible that any language mimics the sound changes of a larger language it is in contact with, it was in the wrong area. The sound changes of Germanech were modelled on those of Standard German - which is based on a dialect spoken in Thuringia, about 300 kilometres further east, where no Vulgar Latin was ever spoken. It was hard to justify my phonology. At least, the Odenwald region where I finally put it was a workable compromise.

4. Germanech bored the shit out of me

The setting of Germanech meant that I could not pursue my own taste in conlangs in that language. To be plausible, it could not have been anything else than a Standard Average European language, and such languages are overdone. I simply could not motivate myself to carry on with this uninteresting language.

What Germanech taught me about romlangs in general

The failure of the Germanech project taught me a lesson about romlangs in general. I now feel that there simply is no way to do them well.

1. Romlangs are overdone

There are now so many Romance conlangs that anyone who works on his own has to ask himself, "Do we really need another Romance conlang?" He should only carry on if he has a brilliant idea of his own, and as we shall see below, there is not much room for that in romlangs.

2. All the good ones are taken

There aren't really many plausible extensions to the Romance language family tree. Most of the areas where Latin was ever spoken by the majority of the local population around 400 AD still speak Romance languages. Some marginal possibilities exist, such as Britain; and these marginal possibilities have already been explored.

3. The more interesting, the less plausible

The real-world Romance languages are all quite similar in many ways. Romanian has developed some peculiarities in its isolation in southeastern Europe; French has undergone some sweeping sound changes that considerably altered its pronunciation. But overall, all of them are quite much alike. There is thus little reason to assume that any additional Romance language would be different in interesting ways. Brithenig goes as far as one could get from the standard Romance model with its Welsh-style initial mutations, and even that is not without problems. (Consider that English never developed anything even remotely similar to initial mutations!) Some conlangers have invented very fanciful stories to come up with oddball romlangs such as "Sino-Romance". Alas, there usually is no plausible way how such languages could have come about.

Conclusion

There are many interesting and edifying things a conlanger can do; Romance languages aren't really among them.

There is something to add...

Meanwhile, Ray Brown has started a new Romance conlang, Britainese, like Brithenig a Romance language of Britain, but with a very different approach. Instead of apply Welsh sound changes to Proto-Romance, Ray extrapolates from the existing western Romance dialect continuum to arrive at a language that could have plausibly spoken north of France. This probably will not result in a language with many features untypical of Romance languages, though some nice and interesting peculiarities (such as the letter u having the value [i]) are emerging. So it is perhaps possible to make an interesting and plausible Romance language. Indeed, I considered redoing Germanech in a similar way (what about something that forms a brdige between eastern French dialects and Rhaeto-Romance?), but I currently know too little about the relevant Romance dialects, and have eough other projects already.