How Not To Conlang?

[alice]

It's easy to give examples of things you shouldn't do when creating other forms of Art, such as "don't use every synonym of 'said' you can think of" and "don't switch viewpoints within one sentence", or "keep your eyelines consistent". Beyond obvious things like "don't spell /e/ as <k>" and "don't slavishly copy the syntax of your native language", what such advice would you apply to conlanging?

[mèþru]

If you are going to be non-naturalistic, you better have a good reason.

[mèþru]

Also, don't have one-to-one lexical correspondences between the conlang and (a) natlang(s) or even between conlangs.

[TurkeySloth]

Also, don't have one-to-one lexical correspondences between the conlang and (a) natlang(s) or even between conlangs.

So, is my ease-of-translation use of one script for two languages bad, then?

[chris_notts]

This thread reminds me of the following book, which I found pretty hilarious:

How Not to Write a Novel: 200 Classic Mistakes and How to Avoid Them--A Misstep-by-Misstep Guide

My guideline, which I don't always follow myself, would be that in many cases "less is more". It's easy to see stuff in other languages and decide that your language needs them. Maybe I decide to have 20 noun cases, agreement with up to 4 arguments of the clause and rigid word order and completely ignore the fact that all of these three things are partly serving the same purpose of role disambiguation. While a language like this might exist, most languages will have at most two of these three things.

Similarly, within most categories in most languages there are only a small number of highly grammaticalised options. While languages with 20 different moods or aspects or tenses do exist, most make do perfectly well with 2 - 3 core options and lexical marking when speakers need to be more specific.

Oh, and don't rigidly impose boundaries just because that's the way linguists usually think about things. Markers in real languages often serve multiple functions depending on the context, e.g. directional/aspectual (preverbs in some languages), aspectual/modal/temporal, ergative/genitive, ..., or distinctions are neutralised in some contexts. A language which has V-A-T-M with a unique slot for each category and with no blurring of boundaries or interaction between those categories seems possible but a bit unnatural.

So to summarise: less is more, and blur boundaries.

[HourouMusuko]

Don't get bogged down in complex phonology and fail to progress any further into your language. Phonology can be a roadblock. But it's not the be-all end-all.

Unfortunately I see that sometimes.

[alice]

This thread reminds me of the following book, which I found pretty hilarious:

How Not to Write a Novel: 200 Classic Mistakes and How to Avoid Them--A Misstep-by-Misstep Guide

To paraphrase Roger Lass, nobody with an interest in writing novels should be allowed to not have read this book.

I'm glad you mentioned it, because I meant to say that it seems a lot harder to come up with similar rules for conlanging than for other forms of Art, and it's not entirely clear why.

[mèþru]

Because there isn't a whole professional academic field that makes it their business as well as a multi-million dollar industry for teaching language creation?

[chris_notts]

Because there isn't a whole professional academic field that makes it their business as well as a multi-million dollar industry for teaching language creation?

Maybe it's just me, but I think that creating a good artificial language is harder than writing a good novel, or at least the knowledge is a lot rarer. You don't need to be an expert in any academic discipline to write a good novel or indeed to have much specific experience of anything else except reading and writing. The following is a list of successful books published by authors with fairly limited life experience:

https://en.wikipedia.org/wiki/List_of_b ... _teenagers

In a novel, you can rely on your audience to fill in a lot of the blanks, so while a novel needs to be logically consistent it doesn't need to be complete. Further, the only limits in a novel are what you can imagine and how far your audience are willing to follow you.

A constructed language, on the other hand, needs to not only be a consistent system but a complete system and a functional sytem. It's much more like architecture than novel writing: an architect might want to create a pretty building, but they probably also want to design one that's possible to build, safely, at an affordable price, and given the laws of physics. Lots of people would probably like a floating castle, and think floating castles might be a good idea, but none have been built for very good reasons.

Of course, making a bad conlang is a lot easier, and making an incomplete but (possibly) good conlang is also a popular option. I think most of us think of conlanging as more of a journey than a destination...

[xxx]

this kind of rules that can make see a representation of thing as a real thing, to be appreciate as art, are not relevant to conlangs...
A conlang is not a fake thing that seems real, it has to be able to replace any real language..
in this way conlang is not art...

[Salmoneus]

It's easy to give examples of things you shouldn't do when creating other forms of Art

No it isn't.
Indeed, that's almost the definition of art. What your examples are are examples of things you shouldn't do when creating commercial products in the hope of maximising profitability. You shouldn't switch viewpoints within one sentence because people well get confused and won't want to buy your book. Dan Brown mostly doesn't switch viewpoints within one sentence - but James Joyce often does, because James Joyce doesn't give a shit whether or not you buy his book. Because he's writing for the sake of art, whereas Dan Brown is writing for his bank account.

So it's easy to create rules of how not to conlang if you're doing it for commercial reasons. If you want actors to be able to pronounce it, or readers not to laugh. But otherwise, there are no rules. There is nobody who has the authority to decree rules for everybody else to follow.

If you are going to be non-naturalistic, you better have a good reason.

Why? Sez who? Who appointed you the Pope of Conlanging? Why should I have to justify my personal tastes, whatever they might be, by giving you "reasons" that you consider "good"? You've given us the "you better!" clause, but I fail to see the "or else!" - what threat can you possibly levy that would compel me to follow these rules of yours?

What "reason" could ever be "good"er than "because it's my artwork, and because I fucking feel like it"?

[Salmoneus]

Because there isn't a whole professional academic field that makes it their business as well as a multi-million dollar industry for teaching language creation?

Maybe it's just me, but I think that creating a good artificial language is harder than writing a good novel

Define 'good'. I have some sense of what it might mean in a novel, but I have no idea what it might mean in a language.

, or at least the knowledge is a lot rarer. You don't need to be an expert in any academic discipline to write a good novel

That's very true. But you also don't need to be an expert in any academic discipline to be able to score 50 goals a season in La Liga. It doesn't mean it's easy - tens of thousands of people have tried, or have tried to reach a position where they could try, but only one man has ever accomplished it. It's really, really hard. How academic something is has nothing to do with how hard something is.

Writing is indeed not an academic subject; but it is, nonetheless, an extremely rare talent - millions try every year, and only a handful are truly artistically successful. [only a fraction even successfully write a novel, let alone a good one]. Indeed, it's hard because it's not a matter of knowing things, but a matter of taste, and talent, and finesse, and inspiration.

Certainly, speaking for myself, I know I could, if I applied enough effort, create a conlang that would at least satisfy my own tastes, if not necessarily yours - but I doubt I would ever be able to write a satisfying novel.

or indeed to have much specific experience of anything else except reading and writing. The following is a list of successful books published by authors with fairly limited life experience:

https://en.wikipedia.org/wiki/List_of_b ... _teenagers

And your point is? Why is the metric of difficulty or worth "how much life experience is required"? In that case, the pinnacle of art is a good set of tree-rings.
Also created by teenagers? Mozart's violin concertos and Exsultate jubilate. Der Erlkoenig. Jodi Foster's performance in Taxi Driver, and Nadia Comeneci's performance at the Olympics. As teenagers, Saul Kripke was invited to become a lecturer at Harvard and John Stuart Mill had invented an entire calculus of formal logic.
There's certainly no reason a teenager couldn't be a good conlanger. Sure, some knowledge of language is helpful - but Mill was reading the classics of ancient Greek, in the original, when he was 3. Champollion learnt Latin, Greek, Hebrew, Arabic, Syriac and Chaldean before presenting his first major linguistic paper at 16. Maria Agnesi conversed fluently on learned subjects at 11, and was conducting philosophical salons at 15 (having taken time out to research ballistics in between).

In a novel, you can rely on your audience to fill in a lot of the blanks, so while a novel needs to be logically consistent it doesn't need to be complete. Further, the only limits in a novel are what you can imagine and how far your audience are willing to follow you.

All of which completely true of a conlang also.

A constructed language, on the other hand, needs to not only be a consistent system but a complete system and a functional sytem.

Sez who?
Of course it must 'function' in some sense, but then a novel must also function. A novel must be complete!
[I don't even know what logically could be meant by a language being 'complete'...]
More importantly: again, I don't remember the papal election. Who are you, with all due respect, to tell me what my conlang 'needs to be'?

It's much more like architecture than novel writing: an architect might want to create a pretty building, but they probably also want to design one that's possible to build, safely, at an affordable price, and given the laws of physics.

You're thinking of structural engineers...

Lots of people would probably like a floating castle, and think floating castles might be a good idea, but none have been built for very good reasons.

Sure, economics. But economics is not a relevant constraint in the case of art - as opposed to the construction industry, which is a commercial enterprise.

[zompist]

Indeed, that's almost the definition of art. What your examples are are examples of things you shouldn't do when creating commercial products in the hope of maximising profitability. You shouldn't switch viewpoints within one sentence because people well get confused and won't want to buy your book. Dan Brown mostly doesn't switch viewpoints within one sentence - but James Joyce often does, because James Joyce doesn't give a shit whether or not you buy his book. Because he's writing for the sake of art, whereas Dan Brown is writing for his bank account.

So it's easy to create rules of how not to conlang if you're doing it for commercial reasons. If you want actors to be able to pronounce it, or readers not to laugh. But otherwise, there are no rules. There is nobody who has the authority to decree rules for everybody else to follow.

Wow, you've just proved that criticism doesn't exist. "There is nobody who has the authority to decree rules for everybody else to follow" in music, in novels, in painting, in poetry, in comics. The artist can always say "fuck 'em all!" And so what? How does that remove people's ability to judge the work? When was the papal election that gave you the authority to ban all artistic criticism?

Profitability is a red herring. What's the formula for making a profitable conlang? There is no set of features shared by Quenya, Na'vi, Klingon, and Dothraki but no other conlangs, nor are these clearly "Dan Brown" sellouts unlike the lofty "James Joyce" conlangs. Making your words pronounceable is not something you do to make money; it's something you do to avoid hassle. But that may or may not be a goal for a commercial project. Surely one of the most "profitable" conlangs ever is Klingon, which is not at all easy to pronounce... indeed it was purposely made to be alien for English speakers.

It seems pretty strange to forget that this board has some pretty high demands on conlangs-- mostly coming down to linguistic knowledge, with a side order of not-being-SAE. I seem to remember you bringing up in another thread that you'd created a test for how SAE a phonology is. Consistency, completeness, and naturalness are all pluses here.

Another place might have quite different standards-- especially places where auxlangs or loglangs are favored. Criticism is not a monolith.

And sure, like any other artist, a conlanger can do whatever they want. What's the downside? Maybe a contentious forum thread. Or maybe quiet awe as everyone reassesses their criteria, who knows. But in any case, people are free to like or dislike the conlanger's work, and their opinion might even be worth listening to.

[Raphael]

It seems pretty strange to forget that this board has some pretty high demands on conlangs-- mostly coming down to linguistic knowledge, with a side order of not-being-SAE. I seem to remember you bringing up in another thread that you'd created a test for how SAE a phonology is. Consistency, completeness, and naturalness are all pluses here.

Sorry for the dumb question, but what do you mean by "SAE" in this context?

[zompist]

See here: viewtopic.php?f=3&t=177

(It's a dumb term-- "Standard Average European". One of the links in that thread suggest the more intriguing "Charlemagne Sprachbund". It mostly seems to come down to French + German and their varying influence on neighboring languages.)

[Alargule]

Raising the bar too high for yourself, thinking that one day (asap, preferably) your conlang will have grown into a fully fledged language that could match any natural language (if that was your aim, at least), considering decisions you made earlier on as set in stone, even if they don't make sense anymore, keep tinkering with certain aspects of your conlang until it has reached perfection (losing sight of the bigger picture all the while), measuring your work against the work of other conlangers...

I've stepped in all of these traps - and still do, occasionally.

[Xwtek]

Be prepared that a lot of grammatical features affect usage of other grammatical features. For example: if your language have a hard time relativizing obliques, there will be a lot of valence modifying suffix.

[Raphael]

See here: viewtopic.php?f=3&t=177

(It's a dumb term-- "Standard Average European". One of the links in that thread suggest the more intriguing "Charlemagne Sprachbund". It mostly seems to come down to French + German and their varying influence on neighboring languages.)

Thank you!

[Salmoneus]

Wow, you've just proved that criticism doesn't exist. "There is nobody who has the authority to decree rules for everybody else to follow" in music, in novels, in painting, in poetry, in comics. The artist can always say "fuck 'em all!" And so what? How does that remove people's ability to judge the work? When was the papal election that gave you the authority to ban all artistic criticism?

It would be nice to imagine that, after a decade and a half, I'd earned some shred of generosity or benefit of the doubt, even if never any genuine respect. But it's long been clear that that's never going to happen.

First, regarding the difference between conlangs and novels, I already specified: I think I have some idea what a good novel is. Specifically, criticism of novels is valuable only to the extent that novels represent a shared human project, with goals broadly agreed upon through consensus. When that consensus does not obtain, criticism does indeed become unconstructive - imagine the most postmodern, static, hallucinatory, middle-aged-literature-professor-considers-having-and-affair novel you can think of, and imagine it being reviewed by someone who only reads mills and boon. Or imagine mills and boon being reviewed by someone who only reads Literature. Each reviewer can point out things the other book fails to do - but this is not particularly helpful, since the author of the book was not trying to do those things, and their audience did not want them to. In the case of conlangs, when we're talking about the broad sweep of possible conlangs and creating rules that say that one type of conlang is inherently better than another - naturalistic over non-naturalistic, "complete" over suggestive, fewer cases over more cases, no lexical correspondence with English, etc - then I don't accept that there IS a consensus project, against whose purposes conlangs can be judged. The person who invents a pan-Germanic auxlang, the person who makes a language for alien slug-creatures, the person who tries to create a simulacrum of an obscure amazonian language, the person who creates an attractive-sound near-cypher that players need to learn to complete a computer game, these are all doing fundamentally different things, and their languages therefore have to be assessed against entirely different standards. They're not like two Dickens novels and you can say which one best achieves Dickens' aims - they're like a catechism and porno story. Both the catechism and the porno are forms of literature, but it makes little sense to say that one is better than the other, or that one is 'how not to' write literature.

Second, you're right: criticism, as you're defining it, doesn't exist - or at any rate shouldn't, as is generally recognised by any critic worth paying attention to. "Criticism" in the artistic sense is a term we use to describe two very different enterprises: reviewing, and critical analysis. Reviewing is where someone describes the impact a work of art has had on them, and that they believe it may have on others, and attempts to explain why this may be. Critical analysis is where someone attempts to tease out the implications of the work - either as seen by the audience or as intended by the author - and how those implications (intentionally or unintentionally) arise. Neither of these enterprises involves comparing the artwork to the sort of fixed set of "rules for writing" that alice describes, and people who attempt to concoct such rules are generally regarded critically by reviewers and critical analysts. Perhaps some people thought they should do that, in Victorian times, but we have progressed. These days, such "rules" are generally limited to commercial ambitions ("how to write a bestseller" and the like), where they can be judged, at least in theory, by the objective success of their rules by the standard of commercial utility.

So if people propose such rules for novels, but "it seems a lot harder to come up with similar rules for conlanging", the message we should take from that isn't that we should work harder to invent such rules for conlangs, but that the proposed (and invariably inconsistent, and out of touch with either actual acknowledged classics or mass popularity) rules for novels should be regarded more skeptically.

And third, regarding your "I know you are but what am I?" argument: I never said I wanted to ban artistic criticism. I never even said I wanted to ban bad artistic criticism. But criticism is itself liable to be criticised, and when people criticise badly, their criticism should be criticised. Criticism that holds art up to preformed "rules" and "standards" is bad criticism. Bad as criticism, that is - it may, of course, still be excellent art.

Profitability is a red herring. What's the formula for making a profitable conlang? There is no set of features shared by Quenya, Na'vi, Klingon, and Dothraki but no other conlangs, nor are these clearly "Dan Brown" sellouts unlike the lofty "James Joyce" conlangs. Making your words pronounceable is not something you do to make money; it's something you do to avoid hassle. But that may or may not be a goal for a commercial project. Surely one of the most "profitable" conlangs ever is Klingon, which is not at all easy to pronounce... indeed it was purposely made to be alien for English speakers.

My throwaway remarks were not intended to be a 'formula', but a suggestion of the sort of guidelines that could be appropriate for commercial conlanging. Whether they are actually the best guidelines is another question, and not one I'm interested in. My point was that when you have a commercial project, you have an inherent standard of success - a commercial one. But when you have not committed to a commercial project, or some other project with goals define by others (work on commission, for example), then there are no objective standards that can meaningfully be imposed.

It seems pretty strange to forget

Then maybe, MAYBE, in the spirit of decency and charity, you could try imagining that I have not "forgotten" anything, and that if your interpretation of my words relies on an assumption of my unaccountable and 'strange' forgetfulness, perhaps your interpretation is flawed.

that this board has some pretty high demands on conlangs

I agree that some people on the board have indeed been unkind and judgemental in their criticism of other conlangers. I don't agree that this is objectively a good thing, or that this unkindness has been in any way objectively justified. To the extent that people have been unkind to one another, they have been wrong to be so.

But there is an important distinction to be drawn here between our opinions of a conlang qua art, and our opinions of a conlanger and their processes qua craft. When you talk about the "demand" of linguistic knowledge - that's not a demand made of a language, it's a demand made of a linguist.

When we see a conlang that appears non-naturalistic, or that appears a cipher of English, it is perfectly legitimate to draw attention to this, and ask whether this is what the conlanger intended. If it seems that it is NOT what was intended, then we can criticise the construction of the language, the craft of the linguist, by its own standards - it has done something the author did not intend. The author can then accept that her language is not what they intended, either change her intent to match her work, or change her work to match her intent. Pointing out that there is a conflict between intent and effect is entirely reasonable. There is also a natural, but sometimes perhaps over-hasty, tendency to assume, as a default, that the intent of a stranger is similar to our own intent - so, because most of us here intend to create naturalistic, non-European languages, there is a default assumption that others will do too, and provided that that is expressed with appropriate respectfulness, there is nothing wrong with that. Conversely, on a forum that skewed auxlang, there would be other assumptions about intent.

But what should we do when that assumption is denied? When somebody says that they are not intending to create a naturalistic conlang? Should we, with methru, say "oi, you'd better have a good reason for that!" Should we say that according to the "rules" and "standards" and "demands" of conlanging, what they have done is "how not to conlang"? No, I don't think we should. Because then we've moved from criticising the craft, criticising the work according to the standards laid out for it by its author, to criticising the standards set by the author. The former is the domain of reason, but the later is the domain of the soul, and of personal taste. It's meaningless to say that Klingon, or Esperanto, or Lojban, or Ithkuil, or indeed Quenya, are bad conlangs, examples of how not to conlang, that they have broken the "rules" or fallen short of the "standards" - we can only say they may not be to our taste. Because they were never meant to be. They each had their own objectives. Similarly, an auxlanger is equally unjustified in dismissing a naturalist conlang - because it never intended to be an auxlang, and so the standards that might be applied to auxlangs are irrelevant.

That is, I think the appropriate form of "criticism" of conlangs is not "you shouldn't do that, it's against the rules", but "did you mean to do that?"

I seem to remember you bringing up in another thread that you'd created a test for how SAE a phonology is.

Well, it's mostly other people bringing it up, unfortunately, and I always feel a bit bad, because it's not a great test. But the point is, I never suggested that having a less SAE phonology made a conlang better - indeed, most of my conlangs have been fairly SAE in their phonology. The purpose of such a test, beyond curiosity, is in helping to bring unchallenged assumptions to the surface - so that an author can realise how specifically European their phonology is, or indeed how remarkably weird their phonology is, and decide for themselves how they feel about that. So that the nature of their work can be more informed by conscious choice, and hence more in line with their own determined preferences, and less forced upon them by unexamined assumptions that may be in conflict with their own ambitions.

It's true that some people seem to have regarded that test in a more moralistic way, treating a 'less SAE' score as in some way objectively superior. But to the extent that people have done that, I don't endorse it, or agree with it.

And sure, like any other artist, a conlanger can do whatever they want. What's the downside? Maybe a contentious forum thread. Or maybe quiet awe as everyone reassesses their criteria, who knows. But in any case, people are free to like or dislike the conlanger's work, and their opinion might even be worth listening to.

But if their opinion is simply that I have done conlanging "wrong", then it's not something I should be cowed into accepting.

[Pabappa]

We didn't invent Standard Average European .... we just started applying it to languages that were wholly unlike all others and realized it could use a few new criteria.