book

Creating an inclusive auxlang

Raye Chell Mahela

Published by

Raye Chell Mahela

Raye Chell Mahela

Programmer, artist, musician, and language lover.

Read in EnglishLegu Esperante

An international auxiliary language (sometimes abbreviated as IAL or auxlang) or interlanguage is a language meant for communication between people from different nations who do not share a common first language. An auxiliary language is primarily a second language.

from Wikipedia

Not all conlangs are meant to be auxlangs, but some are – like Ido and Esperanto. However, can a language invented by one man or one small committee be inclusive?

Becca’s post, “How universal can a language be?“, mentions a few things:

For queer people, learning any language can be a very invalidating experience.

[…]

Learning a constructed language can be even more invalidating. Constructed languages have been made with a particular goal in mind, and queer people soon discover that this goal did not involve them.

[…]

When thinking about the possibility of a queer language, it is hard to imagine constructing such a thing without invalidating someone. Any constructed language is very likely to push, consciously or unconsciously, the particular biases of the author.

Which got me thinking about how could we achieve a language that includes as many people as possible. What are some of the challenges that would arise?

1. Be created by many

A single person cannot reasonably create a language that includes everybody and excludes no one. Again, from Becca’s article:

For example, if a transmedicalist were to construct a language designed to be inclusive to trans people, the author would probably make sex equivalent to gender, erase the concept of being cis or trans altogether and strictly assert the gender binary. A person who does not believe in gender, on the other hand, may choose to erase any concept of gender from their language altogether. Yet to many trans people, either of these would be less inclusive and less validating than a Romance language

I think that the only way to reasonably come to rules for the language that most people can agree to, and most people can feel represented by, is to have a group of people create the language. Not just Europeans, not just straight people, not just one gender or another. As many people as possible need to be able to give their input – whether or not they are linguists.

2. Be fluid

With certain conlangs, such as Esperanto, the community in general is very resistant to change, thinking that it might end up killing the language and defeat Esperanto’s goal of being spread to everybody as a second language.

But, in order for a language to be inclusive, it has to be open to changing – after all, even with a committee of people from various backgrounds working on a language together, somebody is bound to be left out. Therefore, the language would need to be open to change when people voice their concerns.

3. Be versioned

Of course conlangs go through various drafts, but I think that it is important to not just stop at v1.0. Each version needs to have people using it and refining it, with new features added for new versions.

That might sound like programming language development – even ol’ C++ has major differences between version 1998 and version 2011. 1998 is a solid language, and many people still use 1998 exclusively, but 2011 adds a lot of modern features that people have come to expect from modern languages.

Perhaps spoken languages should be similar.

If you look at Ithkuil, it is versioned as well — each “version” is marked by a year: 2004, 2007, 2011. I have not learned it myself, but if anyone out there has input on how the Ithkuil community deals with this, please let me know! :)

4. Be modular and extendible

It might seem daunting to build such a fluid language! What if some people want aspects of Láadan’s evidence markers, but want Ido’s reversibility when it comes to word building? You have to choose one focus!

No you don’t! Why not include everything?

The language shouldn’t be written into a corner so that it has to follow one paradigm, but should be built in such a way that it can be expanded upon with minimal pain for the core language itself.

Again, if you’re a programmer, think of libraries of code. Libraries for C++ are built with C++’s rules, but extend the functionality of the language – so, for example, your programs don’t have to just be console-based, white text on a black screen. (Though it’d be interesting to have namespaces in branches of the language, hmm…)


So how do we achieve this?

How could we possibly collaborate on an auxlang, bringing in many voices and allowing for evolution over time? How could we allow people to work on off-shoots of the language, and once refined, asked to be made part of the core language? How do we keep track of all of the changes made to the language over time?

Revision Control.

Ho, ve. That’s a little programmery, isn’t it? But a lot of conlangers I know are programmers.  That isn’t to say that the language should be built by a diverse group of programmers (everyone knows the field of CS has its diversity problems…), but revision control can be a really great tool for this sort of project, and non-programmers can learn to use it, too.

I would love to see a conlang develop on GitHub, or Bitbucket, or Sourceforge, or on its own server with its own website, and see tools develop to aid in teaching and using that language. It would take a lot of effort and a lot of time, but perhaps it’s an experiment that should happen at some point.

But Rachel, how do we get people to learn such an auxlang?

Honestly, if you want people to learn a language, there has to be stuff to do in that language. This can be chatting with others, but there is more to that. Perhaps if we are able to create films and animations and video games and news websites and everything else in such a language, we build value for the language.

It’s hard to learn a language specifically on ideals, and it’s very hard to learn a language that has virtually no resources out there but a few language lessons.

But creating content is something that is required for the language itself to grow and evolve. We would need to use it for our entertainment or daily lives, find out what is lacking, and build onto it.

I do not think that having an evolving language would hinder this too much. There is still entertainment from older versions of English that get adapted and are still shared today, and with revision control history (and, hopefully, branches for each new ‘version’), all of the historical data would be there to enable somebody to adapt their work to newer versions, or other works from older versions.

What do you think?

  • Do you know any conlangs built by a group, with the intent of being inclusive?
  • Do you know any conlangs that are being built on GitHub or with other open source methodologies?
  • Would you be interested in taking part in such a project, either by building out the core, testing the language by using it, or creating resources otherwise?
  • What downfalls do you foresee?

(One problem I foresee is that I’m writing this in English, and to get people from around the world contributing, we’d need resources in each language – at least to learn the core language, then communicate with that for language building.) :)

6 thoughts on “Creating an inclusive auxlang”

  1. I *love* the idea of modularity. Back in the early days of auxlangs there was a flawed concept that single languages “devolve into dialects” over time. Nowadays it is understood that every language is a dialect and vice versa, and standardized or “prestige” languages are created out of dialects. In other words, the dialects came first; they are not the result of decay.

    Modularity in an auxlang would allow it to acknowledge the reality that sometimes queer people have our own ways of speaking, sometimes women have our own ways of speaking, etc., and that this is wonderful and beautiful and ought to be celebrated.

    Of course, I personally am not sure I would have the time for this.

    1. And time aside, I think that an American trying to solve the language problem would probably suffer from the same problems as a straight person trying to find the answer to heterosexism. But I wish anyone attempting this best of luck. Certainly you would at least need to involve interpreters to do this fairly, if it is even possible.

      1. I am not clever enough to begin to come up with an original grammar or vocabulary for such a language, though I would definitely be willing to offer support via programming, providing resources or building resources, and such. Perhaps this is something that should be brought up in a community of conlangers; preferably one that isn’t an English forum.

  2. In my conlanging career, I’ve had first-hand experience with two group projects. The first was “creation for the sake of creation”, the second was tied to a popular french fantasy series of book for teens (where mention was made to the language of an elf-like people).
    Both projects died slowly within weeks, as the interest vaned, with several sparks of resurgence (driven by bad conscience) but that was it.
    I mention it to show that keeping the involved minds interested is much more complicated than getting a lone conlanger out of bed. A possible solution, discussed on other boards, would be to delimit clearly the responsibilities and duties of each of the participants.
    Interestingly, there was a group project that fared well some years ago, without any conscious planning. More info here : http://dedalvs.com/smileys/2006.html

    Back to the specifics of your proposal, I know of two modular conlangs, that is, allowing parts of the grammar to be added or withdrawn when willing to be more precise/to facilitate learning.
    — Lojban, the logical language, is designed to allow every feature of natural languages to be added in order not to push a particular worldview upon the minds of the speaker (while being based on formal logics) : alienable/inalienable possession for instance (the distinction “my spoon/my arm” may be grammaticalized), and even features from conlangs, as the system of evidentials of Láadan.
    — Elko, “language of the gods”, crafted by French Ziecken Azuris on theonymic roots, has 4 levels of complexity, all allowing expression, but with more nuances the higher it gets. There is also a possibility to use the language under an isolating or agglutinating grammar.

    Well, Lojban is flexible in its cultural imprint, Elko in its grammar ; none of them has a specific say on sex/gender (worth noting that the facultative gender prefixes are exclusively binary in Elko, though).

    I’d advocate a Lojban approach to this, having a “rougher” basis that lets room for personal innovation, but then it’s something we can already do with natlangs and existing conlangs (that are really spoken) : if you, as a group, want to say things from another angle than the mainstream, and that you’re nevertheless understood by all, (grudgingly or not), isn’t it enough ? You can well decide to use -ičo in Esperanto, or suppress -ino in your daily speech, you’d be understood ; the only ones having a problem with that are the self-proclaimed advocates of the “purity of the language” (another name for conservatism…).
    There is much argument in France (but to my knowledge, not in Québec) over the feminization of professional nouns. The body charged with “regulating” the language, the Académie Française, is against it, saying “there’s a closed list of feminine nouns passed upon us by our ancestors, period”, but many add the feminine marker to male nouns without a second thought, and what they write is not agrammatical.
    In the end, it’s what people accept as well-formed sentences that defines the language, not director boards.

    That being said, I can hardly resist a call to conlanging 😉
    A link to the French conlanging forum L’Atelier : http://aphil.forumn.org/forum

  3. Kelkaj homoj de la komunismaj subreddit-oj klopodis krei novan internacian helplingvon sur la bazo de komunisma moralo, sed la klopodo ne daŭris longtempe. Post du semajnoj la kreintojn ne interesis la lingvo.

    Sed mi samopinias. Unu homo ne povas krei vere inkluzivan lingvon, ni devus krei ĝin kune.

  4. I still think that Esperanto offers the best base to create an inclusive language. Within (and around) Esperanto I’ve seen a lot of people willing (or even eager) to change / reform the language. Certainly a greater number than I could ever imagine coming together to create a new language from scratch! (Well, I’d be happy to be proven wrong… only for the time being my personal alliances lie with Esperanto.)

    Hmm… I’m trying to grasp now, why would you not want to use Esperanto as your foundation? Like, even if you’d decide to not work with any Esperantists (be they too conservative or whatever..), it would be much easier to adapt an existing language than to make a completely new one, wouldn’t it? Or is there so much more wrong with Esperanto than I can see?

    Well, my main point was to say that I do see possibilities for reforming Esperanto. There has been talk about it :
    * http://jadeo.net/index.php/esperanto/genro-kaj-egaleco/
    * http://lingvakritiko.com/2014/10/16/esperanto-kaj-sekso/
    * http://lingvakritiko.com/2015/01/31/seksa-egaligo-en-la-lingvo-laufundamente/
    … and some people are actually using “ri” and “-iĉ” – most notably the popular band La Perdita Generacio.

    The main “problem” that I foresee is, that there will be one group of people who care about inclusion, who will use a reformed / an evolved version of Esperanto (or maybe a variety of different versions?!), and another group who care less about inclusion, who will continue to use old Esperanto. I fear this cannot be helped.

Leave a Reply

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