Conlanging for books/comics/etc

[linguistcat]

What are some things that you would find acceptable in a conlang in a book or other fiction, but not in a standalone conlang? This could be too much similarity to English (or whatever language the conlanger is writing in), in story things that might affect the language ("This sub group actively changed the language so it would be hard to understand by the main group that speaks a related language" or the like) or just other details.

If you don't judge them differently, do you feel that a "conlang" shouldn't be included unless it's done well, or does it just factor in slightly to how much you enjoy the work as a whole? Would it ruin a book that is otherwise ok but not great? Would it sour a book that you'd otherwise think is fantastic? Would you still enjoy a book that is ok, but complain to fellow language enthusiasts about the poor conlang work?

For myself, if the language was based on a real language, I would understand if someone just took sounds from the real language and tweaked them slightly to make the "related" language, as long as the tweaks were consistent; Whereas, if someone were deriving a standalone conlang that was supposed to be related to a real language, I'd want to see how they derived things from a shared proto-language. I'd also accept if things were a little too similar to English (or again, the conlanger's first language) especially when it came to phonology. I personally have a love of θ/ð even though I know they're rare phonemes cross-linguistically, so it would be unfair if I judged fellow conlangers who are also writers for similar things.

[mèþru]

I hold conlangs in books to the same standards as standalone. If it's a relex, code or cipher it isn't a conlang.

Would you still enjoy a book that is ok, but complain to fellow language enthusiasts about the poor conlang work?

Yup, that's me.

[Curlyjimsam]

I think what we see of the conlang in the book (or whatever) should be held up to more-or-less the same standards as any other conlang. Although I also think that in general what we see of the conlang probably shouldn't be a great deal, so there might not be all that much to judge. For example, in many cases a book might have some names from a conlang and nothing else. So there's not going to be that much anyone can say about whether or not the syntax is any good, and so the conlanger doesn't need to bother working out much or any of the syntax if they don't want to, and that's fine. And if your conlang-in-the-book did have enough information for somebody to give feedback on loads of complex syntactic ideas or whatever, you've probably put too much of it in, because the ordinary reader is going to be put off by all that.

It's not going to massively dent my enjoyment of an otherwise good book if the conlang is poor, though.

If a conlang is being used for names and such, it's reasonable in general for it to be constructed such that the typical reader has some hope of pronouncing them more-or-less correctly (I don't think we should expect perfection here). So a conlang-in-a-book's phonemes should typically be more-or-less a subset of English phonemes, its syllables should mostly be allowable English syllables, etc. But it's still possible to create languages that don't look or sound much like English within those parameters. For example, you can create a good approximation of Japanese using only English phonemes and English-permitted syllables, but typical Japanese words are quite different from typical English ones.

I think general guidance is always going to have exceptions. I have a (fairly minor) character in a book called Xwungvirq. This is pronounced [xʷuŋwiʁq], something I don't expect the average reader to guess. But the point is it's supposed to be a bizarre alien-looking name, and mangled pronunciations are something I can cope with. Even the characters remark on how hard it is to pronounce properly. Actually, even I end up pronouncing it in my head as [zʌŋgviɜːk] or something similarly totally off most of the time.

I wrote an essay on this sort of thing once which can be found here, though five years on I'm not sure I still agree with all my points.

[mèþru]

I disagree about it having to conform to the audience's phonology - for the most part, the audience is not going to care to even try pronouncing it correctly, so you may as well ignore the audience.

[Curlyjimsam]

I disagree about it having to conform to the audience's phonology - for the most part, the audience is not going to care to even try pronouncing it correctly, so you may as well ignore the audience.

That might be reasonable. I think it's probably helpful if the audience at least feel they can make a good attempt though. An orthography too far removed from their ordinary experience is liable to be off-putting and get in the way of the story. Also, in the hypothetical case that you were a Big Name Author, your fans might find it handy to be pronouncing things more-or-less the same way.

[xxx]

the genius of an author, is to make us believe real, what he describes in a few words...
a good author does not need to include a real conlang in his book to make us believe such ...
on the contrary, it would even be an admission of weakness...

[alynnidalar]

Honestly, I don't really care how good a conlang in a book/comic/movie/etc. is. As long as it isn't obviously terrible, I'm not likely to spend any time analyzing it to see if it's good or bad anyway. Even if it is terrible, if the rest of the media is entertaining, it probably won't interfere with my overall enjoyment at all. (e.g. all Star Wars movies and games have ghastly alien "languages", but I still like them plenty)

The only thing that would truly bother me is if a conlang is obviously bad and the writer thinks it's good. e.g. if the language is terrible and the author writes up all kinds of appendices and blog posts and whatever on it, as if it isn't just a relex of English/bad Latin/etc. Or if the author has Bad and Wrong ideas about language in general--e.g. that "primitive" people have "primitive" language, or severe Sapir-Whorfism, or "they don't understand love because their language has no word for it!", or whatever. I may well privately laugh at a bad conlang, but I'll probably still read the book, you know?

The thing is, most authors are not likely to be interested in linguistics. You and I love them, naturally; otherwise we wouldn't be on a language invention forum. But I understand that most authors don't and I'm willing to put up with this for the sake of an otherwise good book/movie/game. It's like geology. I don't need an author to draw up a detailed tectonics map to explain the geological history of all the mountain ranges; unless they're obviously terrible, I'm not going to care, and even if they are obviously terrible, I probably will continue reading the book anyway. It's just not something I value in a book enough to ruin it.

[Pabappa]

all Star Wars movies and games have ghastly alien "languages", but I still like them plenty)

Im a purist, and I think this might be part of the reason why. I remember working in a toy store and seeing Star Wars action figure names like "Malakili" with translations on, and thinking it all was real. I was disappointed when I looked deeper and saw that a lot of them were just distortions of English words. I think by this time, LOTR had spoiled me and led me to think that all authors were as thorough as Tolkien had been.

The only thing that would truly bother me is if a conlang is obviously bad and the writer thinks it's good. e.g. if the language is terrible and the author writes up all kinds of appendices and blog posts and whatever on it, as if it isn't just a relex of English/bad Latin/etc.

I mostly agree with this, because it's an insult to the rest of us that that guy got published and thinks he's the best when he's the worst. I agree also most authors arent interested in lingusitics and even though I myself am a purist with regards to conlangs in fiction, I dont consider authors who use "cheap" conlangs to be bad writers unless they put themselves on a pedestal and claim theyre the next Tolkien.

If a conlang is being used for names and such, it's reasonable in general for it to be constructed such that the typical reader has some hope of pronouncing them more-or-less correctly (I don't think we should expect perfection here). So a conlang-in-a-book's phonemes should typically be more-or-less a subset of English phonemes, its syllables should mostly be allowable English syllables, etc. But it's still possible to create languages that don't look or sound much like English within those parameters. For example, you can create a good approximation of Japanese using only English phonemes and English-permitted syllables, but typical Japanese words are quite different from typical English ones.

I think general guidance is always going to have exceptions. I have a (fairly minor) character in a book called Xwungvirq. This is pronounced [xʷuŋwiʁq], something I don't expect the average reader to guess. But the point is it's supposed to be a bizarre alien-looking name, and mangled pronunciations are something I can cope with. Even the characters remark on how hard it is to pronounce properly.

I handle this by running everything through an intermediary language that doesnt have tones or exotic consonants, and which occasionally mishears the words being transliterated. So for example the hideous-looking Gʷidiʕìləs can become "Vidīlas" and ʕʷĕle becomes "Wele".

[Moose-tache]

Insisting that a conlang featured in a work of fiction about something else must fulfill the same standards as a conlang in a reference grammar about that conlang makes no sense to me. I'm sure somewhere on a geology forum people are complaining that the hills in a movie are from a different geological province than the purported setting of the film, or that the orogeny of the area wasn't well thought out by the director before filming. Let's not be so myopic as that. In any work of fiction, the settings and characters only need to serve the story. A one-dimensional character may work great as a secondary villain in a comic book movie, less so as the subject of a three-volume biography. Planets with impossible climates may be perfectly acceptable in a tongue-in-cheek science fantasy, less so in an academic paper about habitable exoplanets. Purpose and context are key.

To give an example of a conlang, it would be ridiculous for someone to present a descendant of English in the year 3000 AD as a stand-alone conlang with words randomly changed from current English. Realistically, the changes would need to be regular. Plus, some words probably wouldn't change much at all while some would become unrecognizable after just a few centuries. But in your 3000 AD cyber punk novel, the words need to be uniformly different from English: unchanged words will seem awkward and confusing, and unrecognizable words will be missed entirely. Your cyber punks will say things like "Hum Sannich" not "A~ Ha~diss" or "Ham Thndkh" or "Xean Sandwich."

Does your conlang clearly have Celtic words in it when the story takes place in a fictional world and there are no Celts? No problem!

Does your warrior race use more fricatives than a Semitic blender? No problem!

Does your fictional rabbit language combine the same four or five roots over and over in different ways so the reader can keep up? No problem!

The goal in any work of fiction is to create good fiction. Everything in the story only really needs to serve that goal.

I disagree about it having to conform to the audience's phonology - for the most part, the audience is not going to care to even try pronouncing it correctly, so you may as well ignore the audience.

That might be reasonable. I think it's probably helpful if the audience at least feel they can make a good attempt though. An orthography too far removed from their ordinary experience is liable to be off-putting and get in the way of the story. Also, in the hypothetical case that you were a Big Name Author, your fans might find it handy to be pronouncing things more-or-less the same way.

I think orthography in this case is far more important than phonology. If you have a castle named "Tmarak," nobody cares if they can say it properly. It's intended sound is pretty obvious, so you can probably hear it in your head even if you struggle to make that cluster. But a castle named "Xou" or "Qarkh'a" is basically gibberish. The worst offenses here are probably decorative apostrophes, and deliberately using letters or sequences with ambiguous pronunciations, like X, Q, or GH.

[Curlyjimsam]

I handle this by running everything through an intermediary language that doesnt have tones or exotic consonants, and which occasionally mishears the words being transliterated. So for example the hideous-looking Gʷidiʕìləs can become "Vidīlas" and ʕʷĕle becomes "Wele".

This is an interesting idea. I sometimes do something similar where I partially anglicise things as an English speaker in contact with the culture might. Fo example the Viksen city name Wésósje becomes "Wisusia" (in this case the difference in pronunciation is minimal). Or with Viksen place names ending in the common place name suffix -ze I use -ia instead.

I think orthography in this case is far more important than phonology. If you have a castle named "Tmarak," nobody cares if they can say it properly. It's intended sound is pretty obvious, so you can probably hear it in your head even if you struggle to make that cluster. But a castle named "Xou" or "Qarkh'a" is basically gibberish. The worst offenses here are probably decorative apostrophes, and deliberately using letters or sequences with ambiguous pronunciations, like X, Q, or GH.

For all apostrophes can be derided by conlangers, I don't actually think they're a particularly bad idea in the fictional context. Admittedly I wouldn't want them to be "decorative" - they should indicate something about the pronunciation - but the ordinary reader isn't going to be able to tell the difference. They'll probably just ignore the apostrophe most of the time when trying to say the words, and sometimes it may even guide them toward a more correct pronunciation - e.g. they'll have a better stab at something like /kaʔi/ if it's written Ka'i rather than Kai.

q isn't all that ambigious; I think most of the time people will just pronounce it /k/. gh is not much more ambiguous than g is - how should something Ges be pronounced? But the main point is a valid one - there are letters/sequences where English speakers are likely to be torn between different pronunciations, and maybe even these should be avoided.

(This can be true even of not particularly outlandish words and names. It's not clear from the text if G.R.R. Martin's Arya is supposed to be /ɑrjə/ or /ɑ:riə/ or what.)

[Moose-tache]

They'll probably just ignore the apostrophe most of the time when trying to say the words

They won't ignore it as such (they're not reading it out loud after all). Their eyes will just gloss over the word at full speed, treating it like a hieroglyphic. "This character is named [random sequence of letters]." These cases usually derive from the author's desire to create a fun sequence of letters rather than a unified aesthetic. As I said, I don't think a conlang in a work of fiction necessarily has to have a well reasoned phonology, but an orthography the reader can make sense of is the difference between having a naming language in your book and having a blank space that literally represents nothing in the reader's mind.

Let's say you want to create a group of desert merchants who ride camels and like oil. Any name you write with <kh> will be interpreted as [x]. But maybe you really like the letter X, and you think digraphs are ugly anyway, so in this language [x] is represented by <x>. Congratulations, your reader has now lost the aesthetic impact of the Semitic hints you were dropping, and is now reading the characters' names as "X guy 1" and "X guy 2." This is why a good orthography is important, and why it's not the same as a linguistically pleasing orthography.

[Pabappa]

I handle this by running everything through an intermediary language that doesnt have tones or exotic consonants, and which occasionally mishears the words being transliterated. So for example the hideous-looking Gʷidiʕìləs can become "Vidīlas" and ʕʷĕle becomes "Wele".

This is an interesting idea. I sometimes do something similar where I partially anglicise things as an English speaker in contact with the culture might. Fo example the Viksen city name Wésósje becomes "Wisusia" (in this case the difference in pronunciation is minimal). Or with Viksen place names ending in the common place name suffix -ze I use -ia instead.

Thank you. Sometimes I add -ia, sometimes I dont ... Im not very consistent. But it's not really much different than what we have in real life where names like India get suffixes but names like China and Japan dont. And then, sometimes it's -land instead (Finland, Greenland, Thailand).

Another thing I do a lot of is just use English names that arent tied to conlangs at all. For example, Dreamland is a major country and historic empire on planet Teppala. I'm just using this name because there's no one historical name that stands out as the most proper one, as the country survived for thousands of years. (Also, I keep changing the languages, so they'd be wrong anyway.) This idea probably wouldn't fit well in a professional work but I've been doing it for quite a while.

For all apostrophes can be derided by conlangers, I don't actually think they're a particularly bad idea in the fictional context. Admittedly I wouldn't want them to be "decorative" - they should indicate something about the pronunciation - but the ordinary reader isn't going to be able to tell the difference. They'll probably just ignore the apostrophe most of the time when trying to say the words, and sometimes it may even guide them toward a more correct pronunciation - e.g. they'll have a better stab at something like /kaʔi/ if it's written Ka'i rather than Kai.

Yeah everyone has their own opinion here. I'm on the "anti" side, but I'll make an exception for someone who has a creative use for apostrophes and can show that it wouldnt be better served by either a full size letter or a diacritic. I used apostrophes in one variety of Moonshine once to show silent yer-like things that affect the voicing of the previous consonant, but that was mostly for my own benefit to help remember etymologies .... if I were to show that conlang publicly I would probably spell the words phonetically, using t for /t/ instead of using both t and d' depending on etymology.

[Nila_MadhaVa]

Hey everyone, forum newbie here with my first post.

So, I’m hoping for some advice re romanization. I’ve recently decided to basically redo my main conlang from my conworld (the first version was made ages ago and it was my first conlang; it really was a terrible Latin clone with bits and pieces from other langs). I did a one to one phoneme to letter romanization that required using diacritics and non-standard letters which I was quite happy with, but after some feedback from family/friends that I use as sounding boards, I decided it might be better to make a more intuitive romanization with digraphs instead. So I did.

The problem I’ve run into is that due to the phonotactics of version 2.0, the digraphs are colliding with clusters and I’m fretting over the ambiguity. An example is the digraph <ng> /ŋ/ and the cluster <ng> /ng/. The only solutions I’ve come up with are living with the ambiguity, or marking whether it’s a cluster or not (which brings me back to using diacritics), e.g. <ṇg> /ng/. Just FYI, I’ve got 14 digraphs and the same number of ambiguous clusters (yeah, it’s a large consonant inventory). Another thing is that I personally find this second romanization rather cumbersome to the eye, particularly when digraphs form clusters, but that’s something I’m willing to live with if the overall scheme is more intuitive to the average Joe (or Jane).

I’ve been going back and forth on what to do for a while, so any advice or suggestions would be greatly appreciated.

[Pedant]

Hey everyone, forum newbie here with my first post.

So, I’m hoping for some advice re romanization. I’ve recently decided to basically redo my main conlang from my conworld (the first version was made ages ago and it was my first conlang; it really was a terrible Latin clone with bits and pieces from other langs). I did a one to one phoneme to letter romanization that required using diacritics and non-standard letters which I was quite happy with, but after some feedback from family/friends that I use as sounding boards, I decided it might be better to make a more intuitive romanization with digraphs instead. So I did.

The problem I’ve run into is that due to the phonotactics of version 2.0, the digraphs are colliding with clusters and I’m fretting over the ambiguity. An example is the digraph <ng> /ŋ/ and the cluster <ng> /ng/. The only solutions I’ve come up with are living with the ambiguity, or marking whether it’s a cluster or not (which brings me back to using diacritics), e.g. <ṇg> /ng/. Just FYI, I’ve got 14 digraphs and the same number of ambiguous clusters (yeah, it’s a large consonant inventory). Another thing is that I personally find this second romanization rather cumbersome to the eye, particularly when digraphs form clusters, but that’s something I’m willing to live with if the overall scheme is more intuitive to the average Joe (or Jane).

I’ve been going back and forth on what to do for a while, so any advice or suggestions would be greatly appreciated.

Might I suggest using <ng> for /ŋ/, <nng> for /ng/, and <ngg> for /ŋg/?

[Nila_MadhaVa]

Might I suggest using <ng> for /ŋ/, <nng> for /ng/, and <ngg> for /ŋg/?

That is such a simple and elegant solution! It would mean that I can't disambiguate <nng> /nːg/ clusters, but I'm actually pretty ok with that since I doubt most people would intuitively make that distinction, even intervocalically. Besides, those geminates are only allophonic anyways, e.g. <nn> realized as /nː/, so the ambiguity would only exist in a handful of edge cases. Or I could just extend my prohibition against <ngg> /ngː/ type clusters to their <nng> /nːg/ cousins (breaking them up with an epenthetic /ə/ like this <nəgg> /nəgː/, <nnəg> /nːəg/). Hmm, I still have some thinking to do, but at least I feel like I've taken a big step forward, so thank you very much for that.

[Pedant]

You’re very welcome!

[bradrn]

Might I suggest using <ng> for /ŋ/, <nng> for /ng/, and <ngg> for /ŋg/?

That's lovely - I've never thought of that! Another solution that I quite like is having <ng> for /ŋ/, <n'g> for /nɡ/, and <ngg> for /ŋɡ/. The advantange of this solution is that you can scale it up to other digraphs, so things like <th>/<t'h>/<tth> or <gh>/<g'h>/<ggh>, say.

[Pedant]

Might I suggest using <ng> for /ŋ/, <nng> for /ng/, and <ngg> for /ŋg/?

That's lovely - I've never thought of that! Another solution that I quite like is having <ng> for /ŋ/, <n'g> for /nɡ/, and <ngg> for /ŋɡ/. The advantange of this solution is that you can scale it up to other digraphs, so things like <th>/<t'h>/<tth> or <gh>/<g'h>/<ggh>, say.

An excellent addition, if I may say so!

[Nila_MadhaVa]

That's lovely - I've never thought of that! Another solution that I quite like is having <ng> for /ŋ/, <n'g> for /nɡ/, and <ngg> for /ŋɡ/. The advantange of this solution is that you can scale it up to other digraphs, so things like <th>/<t'h>/<tth> or <gh>/<g'h>/<ggh>, say.

Another great suggestion - I think I'll file that away for a daughter that has lost the ejective stops for which I'm using C'.

[mèþru]

Personally, I'd just go with the 1:1 correspondence. <ŋ ng ŋg nng ngg> /ŋ nɡ ŋɡ nːɡ nɡː/