Is a Collaborative Conlang Even Possible?
Autore: Gary J. Shannon
Data della SM: 07-20-2012
Data FL: 09-01-2012
Numero FL: Florida-00000C-00
Citazione: Shannon, Gary J. 2012. “Is a Collaborative
Conlang Even Possible?” Florida-00000C-00, Fiat
Lingua,
September 2012.
Diritto d'autore: This work is licensed under a Creative
Commons Attribution-NonCommercial-NoDerivs 3.0
Licenza non trasportata.
http://creativecommons.org/licenses/by-nc-nd/3.0/
Fiat Lingua è prodotto e gestito dalla Language Creation Society (LCS). Per maggiori informazioni
riguardo alla LCS, visita http://www.conlang.org/
Is a Collaborative Conlang Even Possible?
Gary J. Shannon
The hobby of constructing artificial languages is the admittedly odd pursuit of a
small handful of individuals who, aside from this strange avocation, could probably pass
for normal in the society of their fellow humans. But as satisfying as this activity might
be when done in secret, it seems it should be all the more interesting when shared with
others. Infatti, the first constructed languages were meant to be shared with the whole
world. The idea of one universal language for all of mankind goes back to the Biblical
account of the Tower of Babel, but one of the first practical attempts to construct such a
language also became one of the earliest collaborative conlang projects of which we have
any knowledge.
In the mid 1880’s a constructed language called Volapük, the creation of a
German priest, Johann Martin Schleyer, became all the rage in France. Both the
intellectuals and the working class people were caught up in the enthusiasm for “A new
Golden Age of brotherhood, unencumbered by the chains of Babel”[1]. One could
frequently overhear Volapük being spoken in the streets of Paris. In 1888 alone 182
Volapük textbooks were published. By 1889 there were 25 regularly published magazines
in or about Volapük, and at the third International Volapük congress in Paris in May 1889
everyone, down to the waiters in the dining hall, spoke Volapük.
But beneath this appearance of strength and vitality was a foundation that was
falling apart. Factions were coalescing around mutually incompatible goals and intentions
for the language. Some wanted a “rich and perfect” language suitable for the highest of
literary forms, while others favored a simpler more practical Volapük for science and
commerce. The Dutch polymath Auguste Kerckhoffs, as secretary of the French
Association for the Propagation of Volapük, was asked to submit all questions about the
language to the International Academy of Volapük where each issue would be deliberated
and decided upon. Invece, he submitted a complete published grammar, or rather his own
vision of what the grammar of Volapük should be. Meanwhile, Schleyer, as inventor of
the language insisted that his decisions must always be final regardless of the preferences
of the Academy.
Tensions mounted and, in spite of the support and enthusiasm of the hundreds of
thousands of correspondents involved with the language, the whole project collapsed
almost overnight. In 1889 there were approximately 200,000 active Volapük
correspondents and 283 clubs around the world, but by 1891 the number of active clubs
and associations dropped to only four and by 1902 the list of registered correspondents
had fallen to 159.
Since the beginning of the Internet, we have seen the same life cycle play out time
and again, and apparently for the same reasons. The short lifespan of collaborative
projects can be seen in the posting activity the groups and forums formed around various
collaborative projects from the past. The activity on these groups follows a nearly
identical pattern in all of the prior projects for which data is available. [2] The graph
1
below shows the number of posts per month for each of the first 24 calendar months of
the life of seven representative projects. [3]
The monthly post numbers were tallied as a percentage of the total number of
posts for each forum so that the averages would not be skewed by the couple of
languages that had an overall higher number of participants throughout their lifespan.
Regardless of the number of participants in each project, Tuttavia, half of the total posts
ever made for that project were made in the first two months, followed by a sharp drop
off in the third month. Beyond the third month participation continued to drift downward
until the one or two posts per month were often nothing more than spam not related to the
conlang project.
Although the time scale is much compressed due to the high speed of modern
communication, the general shape of the curve matches that described for Volapük over a
century ago.
What are the reasons for the failure of such collaborative projects to have any
longevity? There seem to be a number of significant reasons, and any given conlang
project might stumble at a different point in its life cycle depending on a variety of
factors.
The earliest failures are those that don’t attract enough initial participants to keep
things rolling. Real life has a habit of intruding on one’s hobbies, and there are times
when any given group member might not have the time to participate. For a project with
2
a large enough membership the occasional absence of one or two members would not
adversely effect the group’s level of activity, but if there are only three or four people in
the whole group and two of them have real-life duties that take them away from the group
for a week or two, the project will almost certainly die for lack of participation.
Assuming there are enough interested people to sustain activity, the next hazard is
disagreements about features of the language. Conlanging, being a normally solitary
pursuit, has made conlangers accustomed to calling all the shots themselves. But if some
participants are pushing the project in one direction while other members are pulling it in
a different direction, something’s got to give. If there are enough participants it may
happen that the two factions split apart and launch two different dialects of the conlang.
Examples of this failure mode include the Esperanto/Ido split and even the rift between
the relatively obscure language Lingua Sistemfrater (Frater) and Bartlett’s revised
Frater2. With a smaller number of participants the two factions may not split. Invece
members may simply abandon the project. I’m reminded of one post made by a member
of the Kalusa project: “Ack! I go away for three weeks, and return to find Kalusa
defeated! Oh well. It had a good run. …[addressing certain contributors]… Looks like
you’ve successfully driven everyone off, including me.” Some members suggested
(although I was not able to verify this) that one of the most active “factions” was, in fact,
a single person posting under multiple user names.
If the project survives disagreements the next hazard is the loss of frontiers.
Conlangers are, by definition, interested in creating languages, but it does not follow that
they are interested in using a language that is mature enough that it no longer presents
any significant problems to be solved. When a language reaches a certain level of
usability the design process begins to wind down. Once a painting is finished the artist
sets it aside, and once a conlang is complete it is only natural that a conlanger would set it
aside. Of course a conlang can never be complete in the sense that a painting is complete,
but it can reach a point where only the boring rote tasks remain, such as filling out the
lexicon with names for common tools, food items, plants, animals, colors, weather
phenomena and so on.
A case in point is Larry Sulky’s Ilomi which went public in November of 2005.
One month into the project Larry posted a proposal that the members divide up the
Universal Language Dictionary [4] and each tackle the job of coining words to fill in a
particular lexical category. Through the remainder of that month and halfway into the
next there were a few discussions on grammatical matters, but it was clear that the focus
had moved on to vocabulary building. The language was not complete, by any measure,
but it was clearly ready to be released “into the wild”. It was time for the engineers to
stand down and turn the project over to the user community. Only there was no user
community, so when the builders and tinkerers stood down a month later there was
nobody left to carry on.
What would it take, Poi, for a collaborative conlang project to achieve vitality
and longevity? Clearly each of the hazards mentioned above would have to be avoided.
Primo, the number of people involved in the project would have to be reasonably large.
3
Secondo, the structure of the project should be such that it is resilient in the face of
disagreements, and can accommodate diverse points of view without fracturing. And
Finalmente, there would have to be a user group; people who were more interested in using
the language than in tinkering with it.
Rather than tackling each issue separately, we should consider how the three
problems are, in fact, three facets of the same single issue. That issue is nothing more nor
less than the very reason for the existence of language in the first place. Each of the
hazards discussed above can be addressed from the perspective of this single question:
Why language?
Languages do not prosper because people are interested in language, they prosper
in the service of some greater, or at least more urgent or more desirable end. A language
wins over users for reasons having nothing to do with phonology or syntax, but rather to
fulfill some immediate social need, whether in the market, among friends and family, A
gain status or power, or in service of political, religious, or artistic pursuits. IL
discussions among typical users of the language are not “Should verbs inflect for number
of arguments?” but the more mundane: “Help! The goat has fallen into the well.”
It follows that for a conlang to prosper it must occur in a context that has a larger
purpose. The use of the language would need to be a small part of a larger system of
experiences and rewards. As an example, within the community of Star Trek fans, social
status accrues to those who can speak Klingon. In the real world knowing the local
language means being able to buy and sell chickens and cabbages, or plan a picnic, o
help build a bridge. In an online virtual world like Everquest, World of Warcraft, o
Second Life, rewards such as status or power might accrue to those who learn and use the
“native language” of that world. In each case there is some desirable reward for mastering
the language in question.
Although creating and organizing a large physical community in the real world is
out of the question, online communities are a viable alternative. Virtual online worlds
typically have tens of thousands or even hundreds of thousands of citizens. The online
world Second Life has a population of some 18 million. No matter how interesting or
innovative a collaborative conlang project might appear to linguists, it’s hard to imagine
such a project ever attracting even 18 participants, let alone 18 million.
For a conlang to maintain a large community the language itself must remain
intatto. When a person first learns to speak German it would never occur to them to
suggest ways of “improving” or changing the vocabulary or grammar of the German
language. The language is a given; stable, and immutable. Yet who can approach a
conlang for the first time without soon discovering ways in which the language can be (o
even should be) modified and “improved”. My own experience with Esperanto is a case
in point. Within the first few weeks of beginning my study of Esperanto I had discovered
a dozen or more things that were clearly “wrong” with the language, and that urgently
needed to be “fixed”. When the language itself is seen as something artificial and
arbitrary, this urge to tinker with the language is quite natural, especially to conlangers.
4
But that way lies fragmentation of the community as factions split off to champion one
variant over another as was seen in 1889 with Volapük.
Yet if the language is seen as cast in stone then it is no longer a conlang project,
but a dead language. In order for a language to survive intact through its early formative
stages it must be capable of absorbing change without fracturing into splinter groups.
This is not so much a function of the language itself, but of the community of users. That
community must tolerate diversity, which means that some other “glue” beyond the new
language must hold that community together. The perceived rewards for maintaining
community solidarity must outweigh the perceived rewards for disrupting the community
merely for the sake of promoting a variant language. Continuity also means that there
cannot be any single person or group of people who are seen as the ultimate authority on
proper usage. Such authority, by resisting or prohibiting innovation, invites rebellion.
While there will always be those who would wish to set themselves up as the final
authority, IL “common folk” must outnumber, and thus overwhelm the criticisms of the
self-appointed “elite guardians of linguistic correctness”. Linguistic anarchy must prevail.
Ain’t no other way to keep the language alive.
Any changes to the language should come about not as the result of discussions,
negotiations, or planning by a central governing body, but as the result of spontaneous
innovation and experimentation by the “common folk” while using the language to
discuss ordinary daily matters quite unrelated to linguistics. These innovations might
extend what is already considered “proper” or they might violate the established “rules”.
Either way the survival of the innovation depends solely upon whether that innovation is
picked up and used by other speakers of the language, not upon that feature’s
“correctness” or lack thereof. Language features are, in other words, not engineered, Ma
used into existence.
In addition to such grass-roots random mutations, there might also be certain
respected “literary figures” who would be in a position to introduce innovations or
refinements in their writing. Di nuovo, they would not write about linguistics in general or
the language in particular, but would use their innovations in their writing on other
subjects.
Having addressed the three major hazards that bring about failure of collaborative
projects, Tuttavia, we seem to have arrived at a point where we are no longer discussing
a constructed language at all. There is no designer or builder, and nothing about the
language is “constructed”. Noi siamo, in other words, describing an online contact language
or pidgin. In that case the “conlang project” would not be to design a language, but to
engineer an environment in which a language could be born and evolve in the same way
that all languages have been born and evolved. This would not be a “constructed
language”, affatto, but a “created language”.
Natural languages are never invented out of whole cloth, but evolve and descend
from pre-existing parent languages. A language might descend from a single parent or it
might be result of the blending of two or more parent languages as in the case of pidgins
5
and other contact languages such as Tok Pisin or Chinuk Wawa. For a created language
to evolve in a naturalistic manner within a large community of users there would have to
be a parent language from which it could descend. This language would serve as the
starting point from which alterations and modification could be explored in daily use. Se,
Tuttavia, the parent language were to be a well-established and completely functional
natural language there would be no incentive for innovation since the language would
already be usable as it stands. In that case the language would evolve no faster than
natural languages do, and we might as well just observe natural language evolution “In
the wild.”
D'altra parte, if the parent language were to be well defined, but far too
simplistic in its initial state to support normal social and commercial intercourse, there
would be good reason for the user community to “push the envelope” and introduce
innovation in vocabulary and grammar. A good candidate for such a proto-language
would be Sonja Elen Kisa’s Toki Pona with its vocabulary of only 123 words. [5] IL
culture of Toki Pona, Tuttavia, is quite different from the purpose we are exploring here.
Toki Pona is, by declaration, a simple language that is meant to remain simple and
uncontaminated. That is the fundamental guiding principle of its design, and to violate
that principle is considered heresy by the user community. The language has over
600,000 Google hits, and a user group far more active than any collaborative created
language project we’ve discussed so far [6]. D'altra parte, having a grammar and
vocabulary cast in stone for all time is a distinguishing feature of a dead language, Quale
also runs contrary to our purposes.
In its defense, the Toki Pona forums web site does have a forum for “Tinkerers
Anonymous”. [7] To quote the forum header, “Some people can’t help making changes to
‘fix’ Toki Pona. This is a playground for their ideas.” Questo, Tuttavia, is still a roped-off
and quarantined arena for discussing proposed changes, not using them. Using
innovations outside this sandbox is still frowned upon. Inoltre, Toki Pona is governed
by copyright and Creative Commons License, which also limits its release into the wild
as a proto-language. It’s takes little imagination to speculate on what might happen to,
Dire, French if someone where to successfully copyright the language itself.
While Toki Pona itself would not be suitable for the reasons mentioned above,
something along those general lines would be a suitable proto-language for our purposes.
So where does that leave us? We would need, first and foremost, to discover an
overarching higher goal or greater purpose to serve as the justification for the existence of
the user community. That purpose would also need to support some justification for
needing a new language. One could imagine a group of “super villains” intent on taking
over the world wanting a secret language to use among themselves, but that scenario is
neither realistic nor wholesome and productive. Another higher purpose might revolve
around SETI and attempts to communicate with extra-terrestrials using a special
language, but it seems that’s the kind of language that should be engineered. Fans of the
occult might fancy themselves inventing a language of magic, or “intuitively
rediscovering” the lost language of Atlantis. Science fiction fans might become
6
enthusiastic about a fictional alien language that was entirely in the public domain, Piuttosto
than being the tightly controlled property of some movie studio or book publisher. Come
fantasies aside, Tuttavia, discovering an overarching context that meets the necessary
criteria could well be the most difficult part of the entire process.
Once a suitable pretext is found the project structure would need to be designed. It
seems unlikely that a collaborative created language project could ever prosper using the
established techniques of the individual conlanger. In order for the project to succeed the
founders of the project would need to be willing to relinquish all control over the project,
and allow anarchy to reign. Such a project could only happen in the public domain, free
of all procedural and institutional fetters.
That very requirement, Tuttavia, goes contrary to the inclinations of the average
conlanger who wants to build a language of his or her own, to his or her own
specifications. A project of this sort seems to fall more readily into the category of
experimental linguistics, where the founders of the project are reduced to passive
observers and documenters of the process. Sono, in other words, forced from being
prescriptivists into being descriptivists, the very antithesis of conlanging. This would be
especially odious to the conlanger if his creation took on features he strongly disliked.
The ultimate irony of all this is that, whatever the design of the project, it’s unlikely that
this is the kind of undertaking that a “true conlanger” would be interested in.
In summary then, given all the requirements we’ve touched on, what would a
successful collaborative language creation project look like. Here is one possible set of
design specifications. Note that these are not goals for the design of the language, Ma
goals for the design of an environment in which a language can be created by non-
linguist participants.
1. Some overarching purpose must exist which is served by the language. This could be
social, commercial, artistic, religious, or whatever. The only requirement is that it be
something which is inherently interesting and engaging to sufficient number of people to
support a vibrant and growing community of participating users. Users should come from
a variety of different language backgrounds so that the new language does not tend
toward a relexification of some single language such as English.
2. The culture of the overarching purpose must be one that encourages loyalty and
tolerates linguistic diversity. A participant should never feel the need to leave the
community or to form a splinter group because of linguistic differences of opinion. IL
goal is for such linguistic diversity to eventually even out as a “standard” emerges of its
own accord. Until that occurs there is no right and wrong way to say anything in the
language.
3. Participants may join and leave the group any time they feel like it. There are no
commitments to make.
7
4. The web sites and forums used by the user community should be accessible to people
of any linguistic background. Newcomers should be gently introduced to the proto-
language in their own native language, but beyond that, participation in the group should
be a total immersion experience. No languages other than the conlang itself are allowed
within the group. This forces use, Quale, in turn, forces growth and evolution.
5. The seed language, or initial proto-language from which the conlang is to evolve
should be simple enough that it can be learned quickly, and easily used and expanded
upon by anyone who has even the most basic familiarity with it.
6. The proto-language should be fun to use and play with so that people will enjoy
writing things in the language. This is necessary because it is usage, not authority that
will drive the evolution of the language.
7. There should be a minimum set of initial grammar rules that describe the proto-
language, but these rules are to be considered as nothing more than preliminary
suggestions. Once the language is minimally usable, those rules and guidelines fade away
to be replaced entirely by whatever becomes common usage.
8. If a new rule of grammar is proposed, it can be discussed, among participants, but the
final arbiter is common usage. If people begin to use this new grammatical principle then
it will, by definition, be a part of the language. There is nobody with the authority to veto
or sanction any new grammatical idea. It is either used by the community or not. IL
proper way to advocate for a given grammatical principle is to showcase it in a piece of
prose or poetry that others will want to read for its own sake.
9. No grammar rules will be written down by any “official” body. Grammar is taught and
learned by example only. If anyone wishes to write down the rules of grammar as they
exist in common usage it will be understood that these are descriptive only, and will not
constitute a prescriptive grammar. The structure of the overarching organization may be
fixed and governed by rules and regulations, but linguistically, anarchy must be allowed
to prevail.
10. An automated dynamic dictionary and concordance should be maintained online with
entries drawn from the corpus of all contributed writings. This dictionary will not
prescribe words and definitions, but will only describe how the words are being used in
the corpus. Any participant may submit a word to the automated dictionary were it is
added, without the need for review or approval by any higher authority.
11. No idea proposed, or word coined, or principle put into actual use by a particular
writer is inherently superior or inferior. Posterity will judge it to be more useful or less
useful, and no amount of arguing, bickering, or flaming will change the judgment of
posterity. It is not for us to decide what is best for the language. It is for us only to put it
to practical use and let posterity pass judgment by what they keep and what they change.
8
12. While individual writers may certainly copyright their own original works in the
language, no word or grammatical feature of the language may be copyrighted or claimed
as the exclusive property of any person.
In conclusion, Poi, for a community of users to create a new language requires
that the creation of this novel language be incidental to some higher shared purpose. Così
strictly speaking a collaborative conlang project cannot be successful. If a novel language
is to be created, in service of some higher purpose, then it must be created by a
community of non-conlanger language users who have some practical, social, o
ideological reasons for wanting to use that language.
How, Poi, could a conlanger participate in such a project? Perhaps the only practical
contribution a conlanger could make would be to design some minimal proto-language
along the lines of David Peterson’s Wasabi [3], and then find a way to fire up the
enthusiasm of a large number of potential users for participating in the “birth of a New
World Language”, and turn this community of users loose to do its own thing, with their
only mission being to become evangelists “to help the whole world understand each
other”. But this is not conlanging, or even auxlanging. I’m not even sure what to call
something like that.
FOOTNOTES
———
[1] Kahn, Davide, The Codebreakers, Scribner 1967.
[2] The projects for which data was collected are: Universal Language (Oct. 2000),
KuJoma aka KuJomu (Mar. 2002), Madjal (Jun. 2004), Ilomi (Nov. 2005), Kalusa (May
2006), Pandumia (Nov. 2007), and Tak aka Lese, aka Minujaninya (Oct. 2008). Data for
Madjal, Kalusa, and Tak is no longer online, but as I am the originator of those projects, IO
have the data offline.
[3] Numerous other collaborative conlang projects have been brought to my attention
including David Peterson’s Wasabi project (http://dedalvs.com/wasabi.html), e
Kenakoliku (http://colla.conlang.org/). The first was a one-time classroom based project
and the second, an online bulletin board based group effort whose recent posting activity
appears to be following the classical pattern with these most recent 2010 posting numbers
to an ongoing word creation thread: Apr: 99 posts by 9 people; May: 12 posts by 4
people; Jun: 4 posts by 2 people, July: 1 post by 1 person.
[4] http://www.uld3.org/uld27/index.html and
http://tech.groups.yahoo.com/group/elomi/message/374
[5] http://en.tokipona.org/wiki/What_is_Toki_Pona%3F
[6] http://groups.yahoo.com/group/tokipona/ maintained an active membership with
number of posts averaging between one or two per day to 10 or 12 posts per day. In late
2009 the traffic moved to a dedicated forum at http://forums.tokipona.org/ where 87
members have made nearly 7,000 posts.
[7] The follow message is posted at the top of the forum by the creator of Toki Pona:
9
Tinkerers Anonymous
As the inventor of the Toki Pona language and author of its official website, I have a very
specific vision for my creative work. While I am open to ideas and suggestions, IO
unfortunately don’t have the extra time to discuss and give attention to everyone’s ideas of
how to “improve” Toki Pona according to their differing scopes and creative visions.
There will always be people who will want to adapt Toki Pona in one way or another to
better suit their personal goals. Of course you are welcome to invent such derivative
works if you follow the Toki Pona website’s licence. I have created this forum jan nasa li
wile ante e toki pona as a playground for these people’s ideas.
Sonja Elen Kisa
10