How Not To Conlang?

[alice]

... and you may say to yourself, "My God! What have I done!" I think Sal's saying that it's not possible to have objective standards for judging conlangs. Well, if I may put my head on the chopping-block and let it be kicked around for a bit, as it were: here's what I remember of the verbal system of my first conlang.

- Infinitives had the form CVVCVC, for example teozik "to have", pronounced /so.zaik/ (don't ask).
- The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.
- Past and future were each expressed by adding particles to the present: ek teo za "I had", ek teo fu "I will have". The forms of these particles were not always obviously related to the infinitive (i.e. I made them up at random).
- There was a "second conjugation" with seven-letter infinitives and different letter-patterns for the finite forms.

Now, is anybody willing to try to defend this as even barely acceptable, never mind any good, conlanging?

[Xwtek]

... and you may say to yourself, "My God! What have I done!" I think Sal's saying that it's not possible to have objective standards for judging conlangs. Well, if I may put my head on the chopping-block and let it be kicked around for a bit, as it were: here's what I remember of the verbal system of my first conlang.

- Infinitives had the form CVVCVC, for example teozik "to have", pronounced /so.zaik/ (don't ask).
- The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.
- Past and future were each expressed by adding particles to the present: ek teo za "I had", ek teo fu "I will have". The forms of these particles were not always obviously related to the infinitive (i.e. I made them up at random).
- There was a "second conjugation" with seven-letter infinitives and different letter-patterns for the finite forms.

Now, is anybody willing to try to defend this as even barely acceptable, never mind any good, conlanging?

It's not realistic indeed, but it sounds interesting to me.

[Xwtek]

... and you may say to yourself, "My God! What have I done!" I think Sal's saying that it's not possible to have objective standards for judging conlangs. Well, if I may put my head on the chopping-block and let it be kicked around for a bit, as it were: here's what I remember of the verbal system of my first conlang.

- Infinitives had the form CVVCVC, for example teozik "to have", pronounced /so.zaik/ (don't ask).
- The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.
- Past and future were each expressed by adding particles to the present: ek teo za "I had", ek teo fu "I will have". The forms of these particles were not always obviously related to the infinitive (i.e. I made them up at random).
- There was a "second conjugation" with seven-letter infinitives and different letter-patterns for the finite forms.

Now, is anybody willing to try to defend this as even barely acceptable, never mind any good, conlanging?

Not realistic indeed, but it looks interesting to me.

[HourouMusuko]

I thought SAE stood for “Standard American English”. I’ll be.

Anyway, I agree with Salmoneus. Though after reading this thread I am afraid to post my IE-based conlang here.

[Salmoneus]

... and you may say to yourself, "My God! What have I done!" I think Sal's saying that it's not possible to have objective standards for judging conlangs. Well, if I may put my head on the chopping-block and let it be kicked around for a bit, as it were: here's what I remember of the verbal system of my first conlang.

- Infinitives had the form CVVCVC, for example teozik "to have", pronounced /so.zaik/ (don't ask).
- The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.
- Past and future were each expressed by adding particles to the present: ek teo za "I had", ek teo fu "I will have". The forms of these particles were not always obviously related to the infinitive (i.e. I made them up at random).
- There was a "second conjugation" with seven-letter infinitives and different letter-patterns for the finite forms.

Now, is anybody willing to try to defend this as even barely acceptable, never mind any good, conlanging?

...well obviously. Look, here's how to defend this as acceptable conlanging:

"this is acceptable conlanging".

There, defended beyond reproach!

Now, to rebut my defence, you'll have to show that it is not "acceptable". And since I have accepted it, that seems to be impossible. More generally, you'd have to explain why each thing supposedly cannot, or should not, be accepted, by reference to some rule, itself not in debate, that prohibits it. I don't agree that there is such a rule, and I don't see what such a rule could be.

Now, you can absolutely argue "these things are not in favour with the Naturalistic School". Note that they do not fail to be 'realistic', in that each of your points is perfectly plausible and functional in a human language - opaque orthography, complex templatic morphology and wildly irregular inflection to the point of suppletion are all perfectly natural - but merely fail to be 'naturalistic' as one particular ideology defines it. They could be made considerably more 'naturalistic' by explaining each point away through individually plausible diachronic processes, because the Naturalistic School places as much emphasis on process as it does on content, although inherent improbability would be a remaining notch against it (Naturalism favours slightly (and hence interestingly) improbable features, but discourages excessive improbability - a whole bunch of natlangs would probably not measure up to its standards.

This board happens to be largely and historically (but not exclusively) dominated by naturalists - I'm broadly naturalist most of the time myself. But there's nothing to say that the naturalist school inherently defines the purposes of conlanging. You can conlang however you want. Indeed, I'd love to see more non-naturalistic conlanging, since it's often much more interesting. Your language you describe here is at least not boring! Huge swathes of conlangers explicitly eschew naturalism - naturalism is almost certainly a minority position. Esperanto did not aim to be naturalist; Klingon did not aim to be naturalist; Lojban did not aim to be naturalist. Quenya did not aim to be naturalist, although it's more 'realistic' than those others - but look at Tolkien's etymologies and diachronics, filled with ad hoc decisions by "the loremasters" on the basis of what they thought was prettiest or most fun, to see the extent to which he saw his languages as, as it were, "cultivated" languages maintained by scholarly tongue-gardeners in a way that is completely at odds with actual human language processes. But these are not "unacceptable" languages. Indeed, they're four of the most widely accepted conlangs in the world.

[Salmoneus]

Though after reading this thread I am afraid to post my IE-based conlang here.

Post away! Many of us have indeed dabbled in the forbidden arts of Indo-European - though fewer than in the past, when board's collective aesthetic was less severely defined. There have even been IE languages that have received acclaim here, although I think you have to work harder at it than if you present something less familiar.

[Pabappa]

As an a priori purist, I appreciate the new polished look of every conlang I make. I think doing an a posteriori language is actually harder in general, and appreciate the hard work people put into them. Definitely more difficult than loading up a word generator and sifting through a few thousand roots.

However if your IE lang is so close to a particular natlang that it could be seen as merely a dialect of it, I agree that it might be seen as less artistically advanced than one that branches from far back on the tree.

[HourouMusuko]

Essentially, my conlangs have used IE as a grammatical framework to which I add my own a priori vocabulary. To create the grammar, I've sourced mainly from PIE reconstructions (such as Sihler and Ringe), but also taken some inspiration from Sanskrit, Greek, Hittite, and Latin. A true IE a posteriori lang would be more difficult than what I've done, but it's no cipher; I've given myself actual headaches over it. And believe me, I hate romlangs as much as the next conlanger

Anyway, I don't know if that is quite up to snuff for this site; I've posted that material elsewhere. I may use this site as a chance to work on something entirely different. I've never attempted a pure a priori lang, probably because studying ancient IE languages is my primary linguistic interest, so it follows that my conlanging is related to that as well. But I think I would like the experience of creating a truly pure a priori language. Making it interesting would be another task...

[zompist]

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.

It's a bit unconvincing to both say "how dare you mischaracterize my words!" and also "you have entirely correctly characterized my words!"

But whatever this lapse of charity was, how about applying it to alice, mèþru, and chris_notts? They were obviously expressing opinions. And rather than countering their opinions, you pretend that they are setting themselves up as popes. After 15 years, I like your posts but this need to dial up the rhetoric up to 11 has never gotten endearing.

On criticism, I think we have too many disagreements to easily proceed.
* I don't believe we have a consensus idea on what a good novel is
* I don't think criticism relies on this consensus that doesn't exist
* I don't think "rules for writing" are inherently bad: writing is hard and rather than let everyone stumble, it's fine to offer some guidance
* On the other hand, I don't think we can demand that criticism be "constructive". Criticism is itself done for all sorts of reasons, and helping the author or even the reader is not always a priority
* I don't think "commercial" is a despective
* I don't think anyone here but you takes "rules" as some sort of legal regulation with legal punishments, or any at all.

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.

I didn't say a word about kindness, much less against it. (Nor for that matter did I say that I agreed with the board's standards. I have consistently said that conlanging is art, and you can do as you like.)

But what should we do when that assumption is denied? When somebody says that they are not intending to create a naturalistic conlang?

Gosh, why do you think I said "Another place might have quite different standards-- especially places where auxlangs or loglangs are favored. Criticism is not a monolith."? Perhaps because I meant that there can be quite different standards, and that criticism is not a monolith.

The actual discussion, before you attempted to shout it down, didn't really come up with anything horribly inappropriate to other types of conlangs. The closest is mèþru's statement on non-naturalism. I wouldn't agree with it as stated, but it's not insane or something, even if you're creating an auxlang or loglang or something wilder. At least if we charitably accept what a "good reason" is. E.g. Toki Pona has an unnaturalistically small lexicon. But it has a good reason for that, namely the quickness of acquiring it, and the particular fun of making do with strong constraints.

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.

I agree that more knowledge is good! That's what I write books about, in fact. Whether you're writing a novel or doing a conlang, it's good to be aware of more directions you can go.

But I think it's a little disingenuous not to recognize that behind this idea is a standard or a judgment— that a more informed choice is a better choice— and it's no better supported than any of the "rules" you find obviously wrong. You could certainly have a contrary standard that praised idiosyncrasy or enthusiasm or purity of intent instead.

Art is tricky, because you can't just say something is "wrong"— virtually any beginner's mistake is also some virtuoso's great art. And yet it's also possible to teach art. You shouldn't do this by stomping on the beginners, of course. A lot of the teacher's job is to provide education, just as in conlanging: "here are some things you can do that you probably didn't think of." And a lot is training in skills, in areas where skills are important. If you want to learn perspective, there really are rules and methods to learn— it doesn't mean the artist can't say "fuck perspective!", it means that if you do use perspective, you have to learn how to do it.

[zompist]

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.

The Wikipedia article doesn't do the concept any favors. Whorf and Gode weren't Indo-Europeanists or experts on sprachbunds and don't seem to have characterized the idea on more than an informal level.

The only other major name is Haspelmath, and his paper is certainly interesting, but he seems to be an outlier in accepting the idea and the term. (It's from his paper that I got the term "Charlemagne Sprachbund".)

His Map 107.13 (p. 1505) shows, I think, the strengths and weaknesses of the idea. Assuming his analysis is correct, it doesn't at all point to something "European"; it's basically Western Romance plus Dutch and German. The concept is useless for Celtic and misleading for Slavic.

(It would be interesting to have a detailed comparison of this proposed sprachbund with the Balkan one. Also to have Haspelmath's map over time— what did it look like in AD 800?)

I mostly dislike the term, but I also think we have to be careful about taking something as "what linguists say" when it's really the hobbyhorse of a small number of people.

[mèþru]

I think there is something to it, because Czech seems more SAE than other Slavic languages and North Germanic also seems more SAE than the Baltic. Certainly many of the features in SAE were not inherited from PIE skipping Romanian and Russian.

[Bob]

I'm not too picky but I do have my own preferences. I'm really into logographic writing systems but really don't see much of them. Actually, I generally don't like translations of mundane texts. I like my conlangs to refer to epic things like myths or gripping historical texts.

There's a tendency among linguists today to use sample sentences for their grammars which are mundane. I don't really get into that. This is part of why I prefer to study ancient languages, because they're different and exciting. I had some people - most people - complain about my conlang post that included the Ancient Greek myth which it was translating. Ancient Greek myths are very compelling to me. But I look around a bit and see that other people are not translating Ancient Greek myths into their conlangs. I hardly see any sizable texts being translated aside from the one thread where people talk about mundane things.

Me, I think it's fine to have preferences but that it's bad form among conlangers to present one's preferences in a certain way. Few people are interested in conlangs and the possibilities for them are quite broad.

[Pabappa]

- Infinitives had the form CVVCVC, for example teozik "to have", pronounced /so.zaik/ (don't ask).
- The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.
- Past and future were each expressed by adding particles to the present: ek teo za "I had", ek teo fu "I will have". The forms of these particles were not always obviously related to the infinitive (i.e. I made them up at random).
- There was a "second conjugation" with seven-letter infinitives and different letter-patterns for the finite forms.

I think I remember this language. You posted about it way back when and it was already old even at that time. I agree it's not plausible, except for infinitives being CVVCVC, which seems like something that might arise in a langauge that is highly formulaic and uses infixes.

[zompist]

The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.

Well, this is a challenge. Assuming these are 1s 2s 3s 1p 2p 3p, and that the patterns (not the letters) apply to all verbs of this type...

The basic form of the root is *zok.
The plural is formed by vowel change: *zik.

There's a suppletive root *to which is associated with 1st/second person-- especially 1s 2s. Why every verb has a suppletive root is hard to explain. Perhaps the underlying form is **tozok, the accent pattern varied by person, and then the unstressed syllable was lost.

At some point the non-3 forms were *to *tozik, with *zik used as a pluralizer.

An infixed *e changed *to to *teo for the 1st person. This narrowed *to a marker of 2nd person.

We would expect *teozik for the 1p. Instead we get the root *zok plus the infixed 1st person *e = zoek. Perhaps *to was thought of as singular, meaning that the root *zok had always been used for 1p; when *e developed it was applied to all 1st person forms.

Now *to *tozik were 2s 2p rather than non-3. As happens frequently, the 2p was used for politeness' sake and eliminated the 2s form, giving us tozik for both.

This leaves only 3p teozik to account for. I suspect that 1s teo is replacing some earlier morpheme. It may be telling that this is also the infinitive. The infinitive might have developed from the maximal forms of the two suppletive roots, and then been applied to a similar-sounding 3p form.

Edit: OK, I think this is even better:

Original forms: 1s tózok, 1p tozókin, 3s tozók, 3p tózokin. (There might be 2s/2p forms; they're lost.)

The reversed accent pattern in the plural is, of course, based on polarity changes in the Somali definite article.

The plural ending could be anything, but it triggers a vowel change: 1s tózok, 1p tozóikin, 3s tozók, 3p tóizikin.

Now, lose unaccented syllables unless there's one further on to lose: 1s to, 1p zoik, 3s zok, 3p toizik.

CV words get a broken diphthong, so to > teo. This becomes interpreted as a signal of 1st person, so zoik > zoek. And later on toizik is change to teozik to reduce the number of anomalous forms.

That leaves only the 2s/2p form. Presumably 3p teozik was used as a polite form, and simplified to tozik.

[Xwtek]

- Infinitives had the form CVVCVC, for example teozik "to have", pronounced /so.zaik/ (don't ask).
- The present tense was formed with the personal pronoun followed by some or all of the letters (not phonemes; I didn't have a full linguistic education at that age) of the infinitive taken in regular but essentially arbitrary orders, of which I remember teo tozik zok zoek tozik teozik; "I have" was ek teo.
- Past and future were each expressed by adding particles to the present: ek teo za "I had", ek teo fu "I will have". The forms of these particles were not always obviously related to the infinitive (i.e. I made them up at random).
- There was a "second conjugation" with seven-letter infinitives and different letter-patterns for the finite forms.

I think I remember this language. You posted about it way back when and it was already old even at that time. I agree it's not plausible, except for infinitives being CVVCVC, which seems like something that might arise in a langauge that is highly formulaic and uses infixes.

Where is it?

[chris_notts]

The Wikipedia article doesn't do the concept any favors. Whorf and Gode weren't Indo-Europeanists or experts on sprachbunds and don't seem to have characterized the idea on more than an informal level.

The only other major name is Haspelmath, and his paper is certainly interesting, but he seems to be an outlier in accepting the idea and the term. (It's from his paper that I got the term "Charlemagne Sprachbund".)

Not sure anyone else has mentioned it, but Heine and Kuteva also wrote an entire book on the diachronic convergence of European languages. I have a copy, but it's been a long time since I read it. I guess the interesting thing is that they view the emergence of a typical European language as a process rather than a timeless truth.

[Birdlang]

Using too many diacritics on one letter like
For example


s̞̝̖̃̆̃́̅̄̕̚

That’s an eye mess right there.
I made that mistake once and the language I did it in now uses digraphs because of that.
Or using English spelling for your conlang. That’s a little too noobish.

[chris_notts]

Using too many diacritics on one letter like
For example


s̞̝̖̃̆̃́̅̄̕̚

That’s an eye mess right there.
I made that mistake once and the language I did it in now uses digraphs because of that.
Or using English spelling for your conlang. That’s a little too noobish.

Some software and fonts don't seem to render lots of stacked diacritics very well either. I had some pain with Linux Libertine, XeTeX, and a stacked nasal tilde + acute accent a while ago.

[Birdlang]

Using too many diacritics on one letter like
For example


s̞̝̖̃̆̃́̅̄̕̚

That’s an eye mess right there.
I made that mistake once and the language I did it in now uses digraphs because of that.
Or using English spelling for your conlang. That’s a little too noobish.

Some software and fonts don't seem to render lots of stacked diacritics very well either. I had some pain with Linux Libertine, XeTeX, and a stacked nasal tilde + acute accent a while ago.

I had that problem in Windows 10 with some fonts.

[Raholeun]

A cypher is one type of work that seems to be universally unappreciated by the various conlang communities. Posting something along the lines of "KLRIONG QWE TYVYKKIRE ZXWRERG means 'cyphers are generally NEG-liked'" will inspire no one that is used to reading grammars or thinking about the problems of language ambiguity and/or global communication. Reading a language you know, but with a different transcription is a waste of everybody's time.

Also:

Edit: OK, I think this is even better:

Original forms: 1s tózok, 1p tozókin, 3s tozók, 3p tózokin. (There might be 2s/2p forms; they're lost.)

The reversed accent pattern in the plural is, of course, based on polarity changes in the Somali definite article.

The plural ending could be anything, but it triggers a vowel change: 1s tózok, 1p tozóikin, 3s tozók, 3p tóizikin.

Now, lose unaccented syllables unless there's one further on to lose: 1s to, 1p zoik, 3s zok, 3p toizik.

CV words get a broken diphthong, so to > teo. This becomes interpreted as a signal of 1st person, so zoik > zoek. And later on toizik is change to teozik to reduce the number of anomalous forms.

That leaves only the 2s/2p form. Presumably 3p teozik was used as a polite form, and simplified to tozik.

IMO, explaining one or two incongruous forms by positing a situation of language contact in the past would only add to the artistic value of the conlang. Let's say there was a related language or dialect from which the verb form was introduced, or maybe a completely unrelated language and you'd comment on how the exotic verb was integrated into your language by adapting its sounds, changing the syllable structure to fit your language's phonotactics, etc.