Do you ever wish that people who don’t speak English could still understand your writing? Would you like your software’s UI to have a reasonably reliable automatic translation?

Sadly, English is a challenging language to translate from. There are many ways to produce ambiguity, and even if your text is fairly unambiguous, it may still be ambiguous to AIs that lack the life experience we have all had.

I’ve started developing a new language called Ungglish: a language similar to English, but much easier to learn and much less ambiguous. Like most of my projects, Ungglish is intended for multiple purposes simultaneously:

  1. Primary purpose: Ungglish is a teaching language. It helps non-English speakers learn a simplified, phonetic, regular, and heavily tweaked version of English quickly and easily, with roughly 1/8 as much study/practice time as real English requires. Students can then go on to learn about real English by taking courses written in Ungglish.
  2. Secondary purpose: to help create multi-lingual documents and web sites. The idea is that first, you learn Ungglish, and then you use it to write documents suitable for automatic translation to other languages. Ungglish will almost always reliably translate to English, and if your Ungglish text is written clearly, it will translate well to other languages too, with much less chance of mistranslation compared with a direct translation from English. Since auto-translation isn’t the primary purpose, the language does not try to prevent ambiguity by construction as, for instance, Lojban does. Instead, the translation tool should be designed to detect potential mistranslations and ask you for clarifications (which, I assume, could be encoded permanently in the text somehow so that the translator asks fewer questions when you edit and re-translate).
  3. Tertiary purpose: international auxiliary language. Because Ungglish is so much easier to learn than English, it’s would make sense if someday everyone on Earth learns it (by contrast, less than one third of humans speak English). If one single country decides to teach Ungglish to its children, this may provide a kernel from which Ungglish will spread across the world, allowing everyone to eventually be able to communicate with everyone else.

Compared with a human translator, automatic translation is not only cheaper - it also allows instant retranslations when the original text changes. What software developer wouldn’t love that?

In order to make Ungglish easier to use for native English speakers, by the way, an Ungglish editor should help you write text in Ungglish in much the same way as an IDE helps developers. For example, you write “naval” and the editor auto-replaces this with “naevyic” (meaning “of or relating to the navy”); you write “light” and the editor proposes “liet” (photons) or “deshevi” (opposite of heavy); or you write an idiom like “get in touch with” and it suggests “contacta”.

Building this language and its tools will require lots of volunteers to provide manpower, deliberation, and knowledge of foreign language grammars. I cannot do this alone, so if you’re interested in helping out, visit ungglish.loyc.net and then contact me.